A UNITEO

NBS TECHNICAL NOTE

STATES

DEPARTMENT OF

COMMERCE

640

PUBLICATION

Considerations
for the Precise

Measurement

of Amplifier Noise

U.S.

APARTMENT
OF

COMMERCE
National

Bureau

QC
^

loo
JL57

53

,f

NATIONAL BUREAU OF STANDARDS

The National Bureau of Standards was established by an act of Congress March 3, 1901.
The Bureau's overall goal is to strengthen and advance the Nation's science and technology
1

and

facilitate

To

their effective application for public benefit.

this

end, the Bureau conducts

research and provides: (1) a basis for the Nation's physical measurement system, (2) scientific

and technological services for industry and government, (3) a technical basis for equity in trade,
and (4) technical services to promote public safety. The Bureau consists of the Institute for
Basic Standards, the Institute for Materials Research, the Institute for Applied Technology,
the Institute for Computer Sciences and Technology, and the Office for Information Programs.

THE INSTITUTE FOR BASIC STANDARDS

provides the central basis within the United

and consistent system of physical measurement; coordinates that system
with measurement systems of other nations; and furnishes essential services leading to accurate
and uniform physical measurements throughout the Nation's scientific community, industry,
and commerce. The Institute consists of a Center for Radiation Research, an Office of Measurement Services and the following divisions:
States of a complete

Mechanics Heat Optical Physics Nuclear

Applied Radiation Quantum Electronics Electromagnetics Time
and Frequency Laboratory Astrophysics Cryogenics
Applied Mathematics
Sciences

Electricity

"

".

THE INSTITUTE FOR MATERIALS RESEARCH

conducts materials research leading to

improved methods of measurement, standards, and data on the properties of well-characterized

materials needed by industry, commerce, educational institutions, and Government; provides
advisory and research services to other Government agencies; and develops, produces, and
standard

distributes

reference

materials.

The

Institute

consists

of

the

Office

of Standard

Reference Materials and the following divisions:

Analytical

Radiation

Chemistry Polymers
Physical Chemistry.

Metallurgy

THE INSTITUTE FOR APPLIED TECHNOLOGY

Inorganic Materials

Reactor

provides technical services to promote

in industry and
Government; cooperates with public and private organizations leading to the development of
technological standards (including mandatory safety standards), codes and methods of test;
and provides technical advice and services to Government agencies upon request. The Institute
consists of a Center for Building Technology and the following divisions and offices:

the

use of available

technology and to

Engineering and Product Standards

facilitate

technological innovation

Invention and InnovaWeights and Measures

Technical Analysis
Electronic Technology
Measurement Engineering
Structures, Materials, and Life Safety '
Building
Environment *
Fire Technology.
Technical Evaluation and Application '
tion

Product Evaluation Technology

THE INSTITUTE FOR COMPUTER SCIENCES AND TECHNOLOGY

conducts research
and provides technical services designed to aid Government agencies in improving cost effectiveness in the conduct of their programs through the selection, acquisition, and effective
utilization of automatic data processing equipment; and serves as the principal focus within
the executive branch for the development of Federal standards for automatic data processing
equipment, techniques, and computer languages. The Center consists of the following offices
and divisions:

Computer Information
Information Processing Technology.

Information Processing Standards

Systems Development

Computer

Services

THE OFFICE FOR INFORMATION PROGRAMS

promotes optimum dissemination and

and other agencies of the Federal
Government; promotes the development of the National Standard Reference Data System and
a system of information analysis centers dealing with the broader aspects of the National
Measurement System; provides appropriate services to ensure that the NBS staff has optimum
accessibility of scientific information generated within

accessibility

to the

scientific

NBS

information of the world. The Office consists of the following

organizational units:
Office of Standard Reference

Library

Data

Office of Technical Information

and Publications

Office of International Relations.

1
Headquarters and Laboratories at Gaithersburg.
Washington, D.C. 20234.
2 Part
of the Center for Radiation Research.
< Located at Boulder. Colorado 80302.
* Part of the Center for Building Technology.

Maryland,

unless

otherwise

noted:

mailing

address

wo
(LI

OO
53

Considerations
for the

Precise Measurement

of Amplifier Noise

David

F.

Wait

Electromagnetics Division
Standards
National Bureau of Standards
Institute for Basic

*#

Boulder, Colorado 80302

<*

-*&

U.S.

DEPARTMENT OF COMMERCE, Frederick

NATIONAL BUREAU OF STANDARDS, Richard W.

Issued August 1973

B.

Dent, Secretary

Roberts, Director

National Bureau of Standards Technical Note 640

Nat Bur.

Stand. (VS.), Tech. Note 640, 129 pages (August 1973)

CODEN: NBTNAE

For

sale by the Superintendent of

Documents,

U.S.

Government Printing

(Order by SD Catalogue No. 03.46:640). 81.25

Office,

Washington, D. C. 20402

CONTENTS
Page
1.

2.

3.

Introduction

1.1

The Basic Measurement

1.2

Accuracy Considerations

Hot and Cold Standards

Mismatch Ambiguity

4
5

5.

Mismatch Error
The Bandwidth Problem--

6.

The State-of-the-Art

10

7.

An Example

17

4.

7.1

Measurement of the Post Amplifier Noise

20

8.

Conclusions

22

9.

References

24

Appendix A.

Measurement Errors that Depend on

T
and T
hot
cold

Appendix

Mismatch

B.

26

49

B.l

Mismatch Uncertainty

49

B.2

Mismatch Error

50

B.3

Derivations

52

Appendix C.

Effect of Front End Loss on Noise Figure

and Effective Input Noise Temperature

83

Appendix

Computer Programs

86

D.

D.l

Appendix E.

Figures

Through

87

D.2

Figures Bl Through B7

95

D.3

Table

99

D.4

Tables AI-AXXI

101

D.5

Tables BIII-BXXII

105

Reprints of Useful Papers

[1]

[3]

[8]

109

Wait, "The Measurement of Amplifier

Noise"
110
Engen, "A New Method of Characterizing Amplifier Noise
Performance"
115
Wait, "Thermal Noise from a Passive
Linear Multiport"
121

in

List of Figures

Figur e

Page
The state-of-the-art errors in measuring
noise figure

11-14

B1-B7

Mismatch uncertainty

55-61

CI

Alternate input reference planes for an

amplifier

1-4

84

List of Tables

Table

Page

Thermal noise sources

Errors for various

Translation between effective input noise

temperature T (K)
and noise figure, F, R

T,

hot'

cold

combinations-

AA- XXI

Me asurement errors that depend on

T,

ot

16

and

cold
Key to mismatch uncertainty figures

50

B- II

Key to mismatch error tables

62

B- III
B- XXII

Maximum mismatch error

D- I

Computer program for figures 1-4

89

D- -II

Print out of D-I program

93

D- III

Computer program for figures B1-B7

96
98

B-

D-

IV

Print out of D-III program

D-

Computer program for table

63-82

100

D- VI

Computer program for tables A-I through A-XXI 102

D- -VII

Computer program for tables B-III through

B-XXII

IV

106

CONSIDERATIONS FOR THE PRECISE MEASUREMENT

OF AMPLIFIER NOISE

For the best accuracy in measuring noise figure,

attention needs to be given to the choice of the hot
and cold noise standards and to mismatch problems.
Tables and graphs are presented to aid in choosing
the proper measurement conditions, and an example is
given to demonstrate their use.
This paper essentially
supplements a previous paper (included in an appendix)
treating in more detail topics that become important
when state-of-the-art measurements are required.
Key words:
Amplifier noise; effective input noise
temperature; mismatch error; mismatch uncertainty;
noise figure.

1.

Introduction

The problem of measuring amplifier noise is extensively

documented in the literature [1,2].

For the reader's con-

venience, a general overview of this problem is reprinted in

Appendix

(along with other related papers)

This technical

note provides additional detailed information concerning the

accuracy of measuring amplifier noise with the various noise

This additional information

standards presently available.

should be helpful in evaluating or designing measurements of

amplifier noise as accurate as

1%

in effective

0.3 dB in noise figure or

input noise temperature.

With the present state-of-the-art, the least inaccuracy

to which amplifier noise figures can' be measured is approximately

0.1 dB in noise figure or 2% in effective input noise temp-

This paper is especially addressed to those who in-

erature.

tend to make such a state-of-the-art noise measurements.

The Basic Measurement

The most accurate measurements of amplifier noise use

some version of the so-called Y-factor method.

In this method,

two noise standards with noise temperatures of

and T

T,

hot

cold

are sequentially connected to the input of the unknown ampli-

fier*, and the ratio of the noise powers, Y, out of the unknown

amplifier is measured.
of the Effective

amplifier, T

Input Noise Temperature** of the unknown

follows:

as

Because

T,

hot

This parameter Y may be written in terms

and T

hot

V^coid

CD

and Y is measured, then T

are known,

coldj

V-

'

'

is known:

=
e

(T

hot

Also in common use as

noise figure, F, R

Y T

"

cold

)/(Y

1).

(2)

measure of amplifier noise is

To avoid the ambiguities and difficulties

of the IEEE definition of F, R discussed in [1]

of F, R used in [1]

the definition

used here, namely

is

dR

10

log[l

T /290].
e

(3)

*In this paper, the term "unknown amplifier" is used to refer

to the amplifier whose Effective Input Noise Temperature, or
Noise Figure we want to know.

**The definition of T

which is
and F Jr> are discussed in [1]
L
J
dB

reprinted in Appendix E.
2

In crude

0.1 dB require
n

Accuracy Considerations

terms, measurements of T
T,

hot

>> T

>> T

cold

within 2% or F, R within

and require
that the ren

flection coefficients of the noise standards closely match the

reflection coefficient of the "antenna"* in both magnitude

and phase angle.

There are at least seven important sources of error in a

typical measurement of T
T,

hot',
.

(2)

uncertainty in the value of

(1)

uncertainty' in the value of T

in the measurement of Y,

(4)

..

uncertainty7

(3)*

cold'

amplifier -gain instability during

the time required to measure the ratio Y,

(5)

inequality of

the reflection coefficients of the "antenna" and the two stan-

dard noise sources,

(6)

uncertainties in input connector

losses with the "antenna" and each of the two standard noise

sources connected to the unknown amplifier, and

errors in

(7)

correcting for the noise originating in the measuring system

(cascade error).

As we shall see, the magnitude of these

various errors depend

on the choice of
r

T,

hot

and T

,.

cold

To minimize the measurement error requires two steps.

First the hot and cold standards are selected, then the other

measurement conditions are investigated.

Some major factors

that influence measurement accuracy are discussed first, then

the state-of-the-art is discussed, and last a measurement

example is given to clarify the use of this technical note.

*The term "antenna" in this paper refers to all of the components that will be attached to the input of the unknown amplifier when it is being used in its intended application.
(If the end use is not known, then it is usual to assume that
However, the concept of
the "antenna" is reflect ionless
an unknown "antenna" is not in the spirit of a precision amplifier noise measurement.)
.

lo avoid

breaking up the text, the long series of tables or

graphs are collected in the appendices.

Hot and Cold Standards

Some of the choices for the hot and cold standard noise

temperatures that are available are shown in Table

Table

1.

1.

Thermal Noise Sources

Typical Effective Temperature

and state-of-the-art uncertainty

Source*

9.

Neon gas -discharge

Argon gas -discharge
WR15, WR62, WR90-NBS Standards
WR284--NBS'-Standard
Commercial coaxial hot
standard
14 mm coaxial --NBS
Standard
Unregulated ambient
temperature load
Regulated ambient
temperature load
Commercial LN load

10

14 mm coaxial,

1.
2.
3.
4.
5.
6.
7.

18,000
11,000

270 K (1.50%)
165
(1.50%)

1250
692

0.9

(0.24%)
(0.13%)

373

0.5

(0.13%)

373

0.15

(0.04%)

300

(0.33%)

300

0.1
1

(0.03%)
(1.25%)
(0.25%)

80

LN

WR90-

NBS

Standards

11

Commercial LH

12

WR90--NBS LH

10

0.2

0.5

0.1

the
or F JT,, namely
J
dB

standard

standard

Four of the error contributions to T

uncertaintyJ in the values of

T,

on the choice of
bility,
depend
7
r
'

hot
T,

(12.5%)

(2.50%)

'

instaT ,,, Y, and gain

to
cold'
.

hot

and T

,.

cold

A
Appendix
rr

*Frequency coverage for the NBS 14 mm coaxial source is d.c

to 1.2 GHz, for WR284 is 2.60 to 3.95 GHz, for WR90 is 8.2
to 12.4 GHz, for WR62 is 11.9 to 18.0 GHz, and for WR15 is
50 to 75 GHz.

contains tables that list these dependent measurement error

contributions for various values of T

of

T,

hot

and T

selected from Table

cold

or F JT) for combinations

dB

Table

1.

some of the results given in Appendix A.

under F,_ in Table

F,
is

Thus by F,

is

summarizes

The error listed

the magnitude of the uncertainty of

0.1 dB, we mean the true value of F,

between 7.9 and 8.1 dB.

The uncertainty in Y, and varia-

tions in amplifier gain used in Table

mate the present state-of-the-art.

in the parameters used,

and Appendix A approxi-

For different uncertainties

one needs to refer back to Appendix A

and make the modifications explained there.

Mismatch Ambiguity

Mismatch ambiguity (or uncertainty as it is called in [1])

is the ambiguity in effective

input noise temperature because

of an ambiguity in specifying the reflection coefficient of

the "antenna,"
'

ambiguity (see

[1]

...

ant

It

is

best to measure the mismatch

in Appendix E)

but to give an idea of

the magnitude of the mismatch ambiguity,

it

is

examined below.

The simplest assumption for amplifier noise is that the

amplifier is linear (that is, its output voltage y is related

to

its

input voltage x in the form y

are constants).

ax

where a and

If in addition to being linear, we assume

isolation such that an impedance change at the amplifier's

p
03

U
H
T3

E-h

C
H

H
03

03

H
n3

p
CJ

(A)

03

a
o

<

rH
42

TJ

Sh

a
Ph
<

HH HH
1 hH l-H
1

>

HH
1

ti

>>>>

1
ti h1 r- h-

>
>>

HH i
1 1 i

f>S

?S

r^% l>s

l-H

XX

Ps ^s ?s ^s PS ?s ?s

l^s

<<<<<<<<<<<<<<<;<<<;<;<<

4H

""

o3

V)

>

o\

Sh

11

Ph

TJ

Sh
Sh

P
c

t/>

a
o

T3

Sh

O
a
P

P
a

0)

-C

P
O

+1

+1

(3

bO

CD

03

erf

>N

Hh

m
o

Jh
Sh

CD

P
H

'

fH

Sh

3
Crt

rH

O 43
P
O

O
u
E-h

\e

!>.

+1 +1 +1

'

CM
vD

t~~

VO

+1

^'v

+1

+1

r-\

+1

+1

i\

'

LO LO **

LO

LO

>

CM rH <* CM CM

CO \o LO cn vD vO <*

,^
o\o

CM

r-^ /
o\ o\=
V

CM

<^J

^-^

to

o\

o\o

o\

o\=

r-\

r-\

CM

^-t

r-i

(XI

+1 +1

+1

+1

+1

LO to cn

+1 +1

+1

+1 +1

o\=

o\=

o\= o\ o\=

LO rH OJ rH LO

to

+1

+1

+1 +1

+1

+1

r-i

t-i

^-i

r-i

r-A

+1

<-{

r-H

r--

r-t

^-1

C--

o\o

CXI

o
o
o
o
t-i

to

t-^ t--

o\=>

CM oa co CM CM

r-i

,_^ ,_,

o\=>

CM

r-H

O
oo
oo
rH LO

OO
LO CM

r-<

CM

r-l

rH

+1

+1

OO
O CO

LO
r-A

r-<

+1 +1

+1

"*

tO

oo
o CO

r-i

CM rH

!-H

+1

"*

+1

+1

Oo
o CO

+1 +1

^t

to

tn

LO iH CM

CM rH

ooooo OO

rH

CM

T~\

OOOO

r-H

+1 +1

+1

+1

Oo
O CO

+1

+1

+1

+1

oo
o CO

tf

"*

CO

+1

Oo

r-\

+1

+1

"* **

CO CO

to

to

bO
~

Ha
Sh

rH T3

+->

O
U a
H

V)

03

T3

bO

Sh

43

a
03

CD

LO to CO
i CM

LO

13

T3

^
H

LT)

CD

0)
Sh

r-~^

P
3

ooo
ooo
ooooOOOoOoooo Oo
ooooooooooooo
o LO LO o o LO o o o o o LO o o o o
LO H
CM
CM 1
1^ CM
rH to CM
rH
o o o o o o o o o CD O o o LO O O o o o
o o LO o o LO
o LO ro O LO to o o LO cm LO
LO
o rH CM
to rH
CM

Ph

03

rH

rH rH a> CO

^-t

vO to CO tO
o O CM o o o

CM *

T3

42

03

^~>

o
+1

+1 +1
*' v *

LO

o\= o\ o\ o\=

r-<

o3

03

f-- r~-

+1

CO

U
p
p

Sh

oo
oo
oo
oo
t~~

rH

CXI

rf

Cx]

P
P
H
P
03
a
P CD
Sh
CD

P
03

V)
fH

+->

T3

+1

tO CM CM to to to

w
w

3
o

oo

Sh

fc^

<D

Sh

o\=>

CXI

CD

42
t3

o\

43

>

r\r-\

CM

&

Sh
03

cr>

rH

(J

<* >*

Sh

03

t3

+1 +1
V

Sh

(/)

cn

Sh

V)

43

+1

00 <* LO CO

E
P

+1 +1
v

03

>

0> CT>
rH rH

ID

Oh

T3

O
U

+1
/*

r-

P
oi

42

nd

>H

bo

Ph

e
o

+1

>

TJ

a3

H
rQ

'

Ph

<tf

1-

+1

03

oO

HP W

03

CO LO
rH rH rH rH rH I CD

CO

PH
O

H
P

I-~-

t>- l>-

LO LO LO
lh LO LO
yD vO ^o

cm cm

CsJ r-i

r-\

r-t

+l

+1 +1

+1

+1

+l

en en cn LO LO

to to ro
+1

+1

CnJ

s
03

00 00 CO
rH r-\ rH

r-t r-\

r-A

rH

r~i

<-l

rH rH

t-A

^-{

r-i

oOoooOoO

+1 +1

ooooooooo
o o o o o o LO LO LO
o o o o o o CM CM

"4-1

43 H
rH
Ph

E-i

ooo

+1

+1 +1

+1 +1 +1

+1

T-i

r-i

r-i

r-^

+1

+1

+1

+1

OoOO
OcOO

CM CM CM to to to to to
a. cn cn F- r-- t-~ t~- I>.
^o vO \0 to to to to to to to to to

'

'

output -termination does not alter T

dependence of T

with input reflection coefficient is

t
T

,,
r
(ant)

1+b

r
'

!nt- g

[3]

2
)

(4

ant

'

T
where the parameters
r

be terminal

then the most general

b,

|r

'

'

an t

-3

'

and

|r

'

'

ant

are chosen to
1

invariant (i.e., their value does not depend on

the location of the input or output terminals provided the

choice is limited to lossless regions)

The parameter T

the amplifier's characteristic noise temperature, bT

is

is

the

magnitude of reverse radiation (i.e., the noise temperature

of the radiation from the amplifier as seen by the "antenna"),
3

can be thought of as a measure of the correlation of the

reverse radiation with the internal noise or alternately it

may be thought of as

measure of the difference in conditions

for maximum power transfer and minimum noise figure,

and

r'

ant

"

ant

=
1

amp
C5J

ant amp

where the asterisk implies the complex conjugate of the

t

"amplifier's" reflection coefficient. Note that

for

maximum power transfer, and has magnitude unity when

|r
1

it

ant
fc

is

1.

For the case that

and

ant

amp

are small, then

convenient to use the approximate form of equation

r

In

Appendix

mismatch ambiguity.

an t

(5)

an t

amp

cfi~i
D>
*-

are graphs which can be used to estimate

Frankly,
of b, and
is

it

is

a bit of a

problem to decide what value

to use for any given amplifier because so little

known about what are typical values of

has measured

|@|

0.13, b

0.03, b

0.35, T

vacuum tube amplifier,

0.65, T

Engen

3.

[3]

496 K for an X-band

For an X-band tunnel-diode amplifier

crystal mixer amplifier.

|3|

and

|3|

825 K was measured.

For a 30 MHz

0.22, b

0.59, T

161 K was

measured.
*

For any
case, and a typical
value of
7
J *

'

'

ant

-r

amp

the mismatch ambiguity is less than about 2% of T

F,,,)

dB'

0.02.

if the magnitude
of the uncertainty
of
&
'

ant

0.1,
'

'

(0.1 dB of
is less than

This means that precise measurements of amplifier noise

of
require
fairly
accurate knowledge
}
n
&

ant

Mismatch Error

Mismatch error is the error resulting when the "antenna,"

the hot standard, and the cold standard do not have identical

reflection coefficients.

The maximum mismatch error is listed

in tables B-III to B-XXII

for ideal linear amplifiers.

As pointed out in

[1]

(see Appendix E)

match errors using the tables in Appendix

estimates of misare hazardous.

When it is practical, the measurements suggested in

preferable

[1]

are

The Bandwidth Problem

One problem that complicates the measurement of ampli-

fier noise figure is caused by the dependence of the ampli-

fier's noise on the variations in the "antenna" reflection

coefficient versus frequency.

For example,

if local oscil-

lator power in the superheterodyne amplifier leaks out the

signal input port, then the mixer diode bias depends in part
on the amplitude and phase of the reflection coefficient of

the "antenna."

This causes the noise figure and other ampli-

fier parameters to vary depending on the phase of the "antenna"

reflection coefficient at the local oscillator frequency

In

addition, the noise figure depends on the phase of the "antenna"

reflection coefficient at the signal frequency

The resulting

combination effect modulates the amplifier noise figure and

other parameters with a period characteristic of the IF amplifier frequency.

In addition,

the noise figure in theory can

depend on the reflection coefficients of any pair of fre-

quencies within the passband of the amplifier so that to

check out reflection coefficient dependent effects, one needs
to consider an unwieldly range of possible

dences.

frequency depen-

These difficulties can be eliminated or reduced

if the reflection coefficient of the standard noise sources

can be adjusted to equal that of the "antenna" at all frequencies.

There is evidence in the literature that significant

measurement errors can occur by not using the correct frequency

dependent reflection coefficient

[9]

The State-of-the-Art

6.

Several factors which affect the accuracy of measuring

amplifier noise have been mentioned, and details of their

effects examined in the appendicies.

In figures 1-4,

resulting& error caused byJ the uncertainties of

Y,

T\

hot'

the
T

,,,

cold'

gain, connector loss, and mismatch are shown for four com-

binations of hot and cold noise standards.

For these figures

it is assumed that Y is measured within 0.01 dB,

gain is

stable within 0.1% over the measurement time, and the worst
case uncertainty in loss in the connectors joining the hot
or cold standards

to the amplifier during the measurement

is

0.01 dB (see Appendix C for the effect of connector loss on

measurement accuracy).
to

|r

ant

-r

.,

0.01 and

<

s t u.

properties,

The mismatch error selected corresponds

0,

|r

ant

0.05.

<

For the amplifier

were selected not because these are

typical, but because the error of this assumption is within a

factor of

of the probable situation for the various ampli-

fiers represented in the range of amplifier noise indicated.

In figures

1-2,

the measurement error is expressed in decibels.

As for Table 2, a 0.1 dB error say at

noise is between 7.9 and 8.1 dB.

error is a sizable fraction of F, R
a

dB means the amplifier

For values of F, R where the

,

the figures no longer have

simple interpretation other than "this is no way to be

measuring amplifiers with such small noise figures."

10

0.20

n
HOT

10,000 +150K

~i

COLD

= 300 0,5K

0.18

QUADRATURE SUM
MISMATCH

0.16

CONNECTOR LOSS
GAIN

0.14

Y-FACT0R
CO
T3

0.12

LINEAR SUM
0.10

Ufflnmm

COLD

iZ^MW^

0.08

0.06

0.04

0.02

Figure

1.

dB

The state-of-the-art errors in measuring noise

figure, F,y,, using an argon gas discharge noise

source (about
10,000
k
>

K)j

as the hot

standard,

>

T,

^
n ot'

and a room temperature resistive termination (about

The uncer300 K)' for the cold standard, T ,,.
cold
tainty assumed for Y-factor, gain, connector loss,
and mismatch are 0.01 dB, 0.11, 0.01 dB, and
< 0.05
(for mismatch)'
r
-r ^,\ < 0.1 and
v
ant
std
ant
as discussed in the text.
1

'

11

0.20

0.18

0.16

0.14

0.12

0.10

0.08

0.06

0.04

CO
T3

o
or
en
LU

0.02

0.00

dB

Figure

2.

The state-of-the-art errors in measuring noise

figure using the National Bureau of Standards' WR15,
WR62 or WR90 "black-body" standard as the hot standard, and a room temperature resistive termination
as the cold standard.
Other parameters as noted in
figure 1.

12

5.0

T
-

4.5

H0T

= 373 0.5K

C0LD

= 80 1K

QUADRATURE SUM
MISMATCH

4.0
i

CONNECTOR LOSS
GAIN

3.5

Y-FACTOR
T,

3.0
en

o
2.5

2.0

.5

.0

0.5

Figure

(K)

The state-of-the-art errors in measuring the effective input noise temperature, T e using a commercial
hot standard near boiling water temperature and a
cold standard near liquid nitrogen boiling temperaOther parameters as noted in figure 1.
ture.
,

13

5.0

4.0

3.0

o
en
en

UJ
en
IS)

2.o m

.0

Figure

4.

(K)

The state-of-the-art errors in measuring T e using

a room temperature resistive termination as the hot
standard, and a liquid helium cooled resistive termi
The other parameters
nation as the cold standard.
as noted in figure 1.
14

The inaccuracy of hot and cold standards used is indi-

cated in each figure.

In general,

the measurement situation

depicted is typical of the best measurement conditions, and

at

some frequencies

(e.g., where no calibration service for

noise sources exists) somewhat better than the best conditions

The four combinations of hot and cold noise standards

selected are 10,000 K and 300 K, 1270 K and 300 K, 373 K and
80

K,

and 300 K and

K.

The combination 1270 K and 300 K

differs from the others in that it assumes the use of the NBS

primary standards, or if you will

the best NBS can measure

The results from these figures can be converted

noise figure.
to

--

either noise figure or effective input noise temperature

using table

3.

In figures

1-4,

the dashed line is the quad-

rature sum (root mean square addition) of the errors.

see,

As we

the quadrature sum is typically half of the linear addi-

tion of errors.

The quadrature sum is the appropriate sum

if the errors meet three criteria:

(1)

the error sources are

the errors are equally likely to be positive

independent,

(2)

or negative,

and (3) the errors are more likely to be small

rather than large.

In

other words, quadrature addition is

appropriate if the error distributions are something like a

gaussian distribution about the value used in the measurement
calculation.

Unfortunately, three corrections, namely clipping, connector loss, and mismatch correction, seldom satisfy the

15

"

Table

Translation between effective input noise temperature

3.

(K)
-

and noise figure

The asymmetry in the

dB

error statements, because of the logarithmic non-

linearity of F,

dB

at

10
15
20

30
50
70

100
150
200

300
500
700

+
+

+
+

+
+

i%
1%
1%

-L

s-

9-

9-

1000
1500
2000

9-

9-

3000
5000
7000

1%
1%
1%

10000
15000
20000

30000
5 0000
70000

li

1%
1%

11

avoided by using the slope of

15

.22
29

43

.69
.94

29
.81

28

.08
35

4
5

33

.48
90

7
8

97

10 .55
12 61
14 00

50
22
18 45
15
17

20
22

23

19
39
84

dB

1%
1%

is

the corresponding T

:k;

+
+

+
+

+
+

dB

+
+
+

.1 dB
.1 dB
.1 dB

.0041 dB
.0064 dB
.0084 dB

.1
.1
.1

.0111 dB
.0143 dB
.0177 dB

.0221 dB
.0275 dB
.0307 dB

10
11
12

.0337 dB
.0364 dB
.0379 dB

13
14

15

.0396 dB
.0410 dB
.0417 dB

16
17

18

.0014 dB
.0021 dB
.0028 dB

1
2

.0422 dB
.0426 dB
.0428 dB

.0430 dB
.0432 dB
.0433 dB

16

+
+

19

21

22

23
24

+
+

(K)
v
J

75

11

169
238

dB
dB =
dB

433
627

864

.1
.1
.1

dB
dB =
dB

1163
1539
2013

.1
.1
.1

dB
dB =
dB

2609
3360
4306

.56%
.50%

46%

.1
.1
.1

dB
dB
dB

5496
6994
8880

42%

.40%
33%

dB

11255

.1
.1

dB = 14244
dB
18007

.1
.1
.1

dB
dB
dB

.1
.1
.1

dB = 45671
dB
57572
dB
72554

22745
28709
36218

22%
24%
62%
831
371
.07%

83%
.74%
.63%

36%
35%
34%
.33%
33%
32%

32%
.31%
.31%

dB

quadrature conditions.
it is more practical

'

In principle

this can be arranged but

to keep the level of these errors

low,

and then add their contributions linearly to the quadratic

sum of the remaining errors

Frequently the errors due to the uncertainty in the hot

noise standard, the cold noise standard, the measurement of
Y,

and the amplifier gain instability add in quadrature.

In

this case, the contributions from these four sources will be

about half of their linear accumulation of error.

An Example

To illustrate a measurement problem we begin with the

following measurement specification provided by a "buyer".

WR15 mixer-preamplifier specifications:
Local Oscillator:
I.F.

10-110 MHz

Bandpass:

Maximum Noise Figure:

dB (1500 K)
< o.i

<

Ir

J
ant

Input

GHz

60

<

.1

<

.1

Gain:

>

10 dB

Gain Instability:

<

0.005 dB/min

Measurement Accuracy Goal:

0.1 dB

Output

First

reference

r|

[1]

r|

for

comment on the specified accuracy goal.

(see Appendix E)

17

From

we note that a calibration from

'

NBS exists so figures 1-4 can reasonably represent the state-

of-the-art.

Using figure

we see that 0.1 dB error is near

1,

the state-of-the-art using an argon gas discharge noise standard and a room temperature standard providing some of the

error contributions can be placed in quadrature.

figure

2,

Further, from

appears NBS should be able to provide

it

ment verification if needed.

However,

it

measure-

clear that the

is

specified accuracy goal is going to be difficult to obtain.

To compare the specified accuracy goal with the estimated

mismatch ambiguity,
bT

300 K so that b

would use figure B4 (it is my guess that

=

0.2, and

0.2 to be consistent with

Engen's
measurement
&
|r
1

^.j
std

ant

about 1.1%.

0.1.

[1])
i '
L

with

r
'

ant

amp

'

0.2, and
'

The mismatch ambiguity

indicated is
J
&

This is several times the accuracy goal specified.

At this point one should stop and reevaluate whether it

is

economically more effective to lower the amplifier noise

figure to allow for the mismatch ambiguity, or to specify the
If the specified accuracy

antenna impedance more precisely.

goal is maintained, a difficult measurement is being undertaken

with a real risk that the "buyer" cannot make use of it.
a

situation like this there must be

In

very clear mutual under-

standing with the "manufacturer" of the conditions under which

the amplifier will be utilized.

unintentionally become invalid.

18

Otherwise the measurement can

As a final consideration of the accuracy goal, we note

from [10]

or

from table

in

[1]

(see Appendix E)

that

connector losses as low as 0.01 dB have been measured in WR15.

from experience at NBS, a 0.1 dB loss can occur if con-

But

nectors from different manufacturers are mated without care-

inspection to see that the flanges close properly.

ful

Thus,

flanges must be examined and handled with "laboratory" care,

right through the time the amplifier is finally installed.
The gain of the WR15 mixer-preamplifier in this example
is

not great enough so that the cascade noise contribution

due to a post amplifier can be ignored.

This contribution is

approximately equal to T (post) divided by the power gain of

the WR15 mixer-nreamplif ier

For a F Jr> (r>ost)

of

dB (627 K)

the post amplifier contribution of 63 K represents a 0.1 dB

increase of F, R (spec
T

(post)

post)

over F, R (spec).

We need to know

and the gain of the specified amplifier within about

10%.

At this point, we can list a set of conditions which will

make it probable that

1.

measurement of

F_,

is

within

.1

dB.

A quadrature error situation using an argon-gas discharge

noise standard and room temperature standard.

a.

For example

An argon-standard adequately calibrated by a

standards laboratory.
b.

A Y- factor

measurement system accurate to within

0.01 dB (Including resolution limitations).

l'J

A thermometer accurate to within about 0.1 K

c.

measure the room temperature standard.

to

Mismatch ambiguity measured for

2.

|r

0.1

(with the

most unfavorable phase) preferably to better than 1%.

to meet the specification,

(Note:

the measured

must be lower than that specified by this mismatch

ambiguity.
3.

Connector losses should be less than about 0.02 dB.

4.

Cascade correction

amplifier gain measured within 10%.

a.

60 GHz

b.

Post amplifier noise temperature measured within 10%

Measurement of the Post Amplifier Noise

The example of measuring the WR15 mixer -preamplifier is

But the problem of measuring the post amplifier

completed.

noise temperature within 10%

is

0.3 dB at

dB noise figure)

sufficiently different problem that it merits further

To be specific,

discussion.

assume the manufacturers speci-

fications of the amplifier chosen for this task are:

Bandpass 10-110 MHz

Noise Figure
Input

Output

Gain

>

70

<

.1

dB

Gain instability
1

dB (627 K)

0.05

<

|r|

<

<

0.005 dB/min

dB gain compression at 100 mw or greater.

20

The gain selected permits one to operate a bolometer power

bridge with 0.1 milliwatt output when a room temperature noise

source is connected to the WR15 mixer-preamplifier.

The gain

compression level selected is such that less than 0.1% of the

noise power will be clipped when the hot noise source

is

connected to the WR15 mixer-preamplifier.*

To estimate the mismatch uncertainty for the post amplifier

(due to the low frequency noise standard's reflection

coefficient being different than the output reflection coefficient of the mixer-preamplifier)

would use figure B7 (no

hard information for this choice) with

r
l
1

iant

"

omJ
amp

~
l
'

r a n1ant

"

r^

std
.

0.1.

The result is off the

graph but extrapolates to be near 131 (or

A 101 overall accuracy

the standards

is

.4

dB via table 3).

our goal so we would like to tune

(which decreases their accuracy) to equal the

output impedance of the mixer-preamplifier.

But to match the

output impedance from 10-110 MHz is probably impractical so

like it or not, we may have to accept the 131.
For the post amplifier noise measurement, a solid-state

noise source with

T,

hot

near 10,000 K, and a room temperature

cold standard are reasonable choices.

Scaling the errors from

*From Cohn [4] average amplitude/saturation amplitude is approximately equal to 0.07 for 0.11 clipping in a square law
Bolometer-power bridges work accurately up to 10 mw
detector.
This hard
of power so this suggests > 147 mw saturation level.
limiting model is not easy to interpret for real amplifiers.
From our experience, 1 dB compression at 100 mw is adequate to
For other discuskeep clipping correction to less than 0.1%.
sions of the effects of clipping see Deutsch and Hance [5],
Bell [6], Van Vleck and Middleton [7].
21

table A-IV for T

T

700 K,

10 K contributes

Qld

=
e

T,

not

1.6%, Y

3% contributes

4.401,

0.1 dB contributes 3.6%,

and gain stability o 1% over the measurement time contributes

1.6%, connector loss of 0.05 dB contributes

.05 dB to F Jr>

etc

(see appendix C)

both have
bT /T
a

|r
1

or 1.69% via table

<

0.02, then from table B-IX,

|r
1

ant

.3,

(this is just
a guess
but in line with the measJ
b

urements in [1]) and using F(dB)

for

If the noise standards

3.

>

.1

'

(because

not listed),

mismatch error of 0.068 dB (2.1% using&

table 3) needs to be added to the 13% mismatch uncertainty

for a total mismatch error of about 15%.

measurement error expected

the hot source
(3.6%)

is

27.9%.

If the errors due to

(4.42%), the cold source

and the gain instability (1.6%)

The grand total

(1.6%),
are

the Y-factor

in quadrature,

then the grand total error reduces to 22.8%, or still uncom-

fortably large compared with the 10% goal.

8.

Conclusions

Under favorable measuring situations, noise figure meas-

urements within 0.1 dB or effective input noise temperature

within 2.5%

is

about the best that can be done.

greatest problems to accurate measurements

is

One of the

mismatch error.

This problem is compounded by the general ignorance of the

magnitude and variation of the pertinent amplifier parameters.

22

The accuracy achieved in a particular measurement of

amplifier noise depends not only on whether

matic noise figure meter

is

commercial auto-

being used, or whether

refined

Y-factor measurement scheme is utilized, but also on whether

a

national reference noise source exists at the frequency of

interest.

also depends on the stability of the amplifier

It

properties during the measurement.

But in addition,

the

accuracy depends on the experience and skill of the metrologist

This experience and skill needs to be learned, preserved, and

shared if accurate amplifier noise measurements are to become

a

reality.

of the noise

This paper provides information on the accuracy

figure measurement in terms of the uncertainties

of various measurement parameters without stating how to

estimate the uncertainties of these measurement parameters.

For these important estimates we anticipate that the average

metrologist will utilize the information provided by the

manufacturers of the test equipment he chooses to use.
the metrologist hopes to have

If

individual error contributions

combined in quadrature, then an even greater skill and understanding

is

required.

We hope that

for some of the amplifiers

whose noise figure he has measured, he will have the National

Bureau of Standards verify the measurement.
an opportunity

to

This provides

assure the metrologist that his methods and

his equipment are adequate,

and to give him the confidence to

share his ability with others.

2 3

[1]* D.

References

Wait, "The Measurement of Amplifier Noise," Microwave

F.

Journal, pp. 25-29, Jan. 1973.

[2]

M.G. Arthur, "A Guide to the Measurement of Noise Per-

formance Factors," NBS Monograph

[3]* a.

_,

1973.

Glenn F. Engen, "A New Method of Characterizing

Amplifier Noise Performance," IEEE Trans. Instr.

Meas., Vol.
b.

Glenn

IM-19, No.

4,

pp.

344-349, Nov. 1970.

Engen, "Mismatch Considerations in Evaluating

F.

Amplifier Noise Performance," IEEE Trans. Instr.

Meas., Vol.
[4]

IM-22, Sept.

1973.

Charles Erwin Cohn, "Errors in Noise Measurements Due to

the Finite Amplitude Range of the Measuring Instrument,"
Rev.

[5]

Sci.

Instr., Vol. 35, pp.

Ralph Deutsch

Harold V. Hance

701-703, June 1964.

,

"A Note on Receivers for

Use in Studies of Signal Statistics," IRE Convention

Record, Part
[6]

R.

Bell,

L.

I,

Information Theory, Vol.

1,

pp.

7-13,

1953.

"Linearity Range of Noise-Measuring Amplifiers,"

Wireless Engineer, pp. 119-122, April 1947.

[7]

J.

Van Vleck

H.

David Middleton, "The Spectrum of

Clipped Noise," Proc.

Jan.

IEEE, Vol.

1,

pp.

2-19,

1966.

[8]* D.F. Wait,

"Thermal Noise from a Passive Linear Multiport,"

IEEE Trans. Microwave Theory

No.

54, No.

9,

pp.

Techniques, Vol. MTT-16,

687-691, Sept. 1968.

*Copies of these papers are in Appendix E

24

9]

C.K. Mayer,

"Improved Noise Measurements Using Ferrites,"

IEEE Trans. Microwave Theory

pp.
10]

B.C.

24-28, Jan.

at

60

Techniques, Vol. MTT-4,

1956.

Yates and C.J. Counas

Evaluation

"Summary of WR15 Flange

GHz," NBS Technical Note 642, Oct. 1973.

25

Appendix A.

Measurement Errors that Depend on T

hot and T cold

The tables in this appendix are summarized in part in Table

2

on

page 6.

program noted

These tables are computer print outs using the

in

Appendix

meanings (see equations

D.

(1)

The symbols have the following

(2)

Symbol

and

(3)

in

text)

Meaning

TC

DY

(A-l)

hot

.01

dB

(A-2)

cQld
Y-factor measurement inaccuracy of

0.01 dB

(A-3)
(A-4)

TE(K)

Amplifier gain instability of 0.11

T
expressed in degrees Kelvin

F(DB)

(A-6)

Y(DB)

Y expressed in decibels,
= 10 log Y
Y

DG

.101

(A-5)

dB
i.e.

(A-7)

dB

ETH

Error to T

ETC

Error to T

EY

Error to T

EG

Error to T

e
e

due to uncertainty
in
'

(A-8)J

T,

due to uncertaintyJ in T

hot^

(A91
v
*

,,

cold

due to uncertainty in Y
e

(A-10)

due to uncertainty
in
'

amplifier gain
+

(A-ll)
(A-12)

The error listed next to value of T

under TE(K)

sum of the errors listed under ETH, ETC, EY and EG.

this error is converted

Then

into the corresponding error in

and listed next to the value of the corresponding F, R

the

is

F,,,

As an example of using Table A-l, note that an ampli-

fier with an effective input noise temperature near 7000 K

(noise figure of 14.0 dB)

can be measured to an inaccuracy of

26

2.1%

18000

(+

0.087 dB in noise figure)

270 K, the cold standard is 300

(near 5.35 dB)

is

caused by the

standard, 0.02%

the Y-factor

K,

is

measured with an inaccuracy of 0.01 dB,

is

and the gain is unstable within

1.59%

standard

if the hot

is

270

Of the 2.1% inaccuracy,

0.1%.

uncertainty in the hot noise

caused by the

uncertainty in the

cold noise standard, 0.34% is caused by the 0.01 dB in meas-

uring the Y-factor, and the remaining

gain instability of 0.1%.

0.15% is caused by the

Mismatch error, connector loss

error, and cascade error are neglected in these tables.

instead of the

270 K, 1 K,

0.01 dB

racies assumed in Table A-l for

T,

respectively we had

K,

270K,

hot

,
'

and

cold

Y(dB)
>-'

0.005 dB

and gain
&

then the corresponding error contributions to T

inaccu-

0.1%

and

If

0.5%,

in our

example would expand or contract in proportion to 1.59%,

0.04%, 0.17%, and 0.75% for

total error of

2.55%.

Simi-

larly for any other measurement uncertainty situation the

tables can be modified to obtain the appropriate error

contributions to T

27

Table A-

TH=18000 + 270.00
TC=
300 +
1.00

F(DB)

Y(DB)

(1.50%)

(0.337.)

DG=.10%

DY=.01 DB

TE(K)

ERROR CONTRIBUTIONS TO TE
ETH
ETC
EG
EY

+67.9%
+45.9%
20 +34.9%

0.15 + .093
0.22 + .098
0.29 + .098

17.64
17.57
17.51

47.29%
32.33%
24.41%

10.13%

30 +23.9%
50 +15.1%
70 + 11 .3%

0.43 + .097
0.69 + .096
3.94 + .095

17.37
17.12
16.39

100 + 8.5%
150 + 6.3%

16.56
16.36

200 + 5.2%

1.29 + .094
1.31 + .093
2. 23 + .392

+ 4.1%
530 + 3.2%
700 + 2.3%

3.08 + .090
4.35 + .033
5.33 + .037

1000 + 2.6%
1530 + 2.3%

2330 + 2.2%

6.43 + .036
7.9Z + .085
3.97 + .335

3000 + 2.1%
50 00 + 2.1%
70 00 + 2.1%

5.39%

7.26%
4.92%
3.75%

3.15%
2.14%
1.63%

16.73%
10.63%
8.06%

3.40%
2.04%
1.46%

2.53%
1.64%
1.24%

1.12%
3.71%
3.54%

15.61

6.10%
4.58%
3.31%

1.02%
0.68%
3.51%

0.94%
3.71%
0.59%

0.41%
0.31%
3.26%

14.84
13.64
12.72

3.05%
2.44%
2.13%

0.34%
0.21%
3.15%

0.43%
0.39%
3.35%

3.21%
0.17%
3.15%

.65

10.35
9.39

.93%
1.33%
1.75%

3.11%
2.37%
0.06%

0.32%
3.30%
0.30%

0.14%
3.13%
3.13%

10.55 + .085
12.61 + .036
14.30 + .08 7

3.34
6.37
5.35

1.68%
1.62%
1.59%

0.04%
0.02%

0.30%
3.32%
0.34%

3.13%
0.14%
0.15%

10030 + 2.1%
15000 + 2.2%
23330 + 2.3%

15.53 + .09 3
17.22 + .394
13.45 + .093

4.34
3.34
2.72

1.57%
.56%
1.55%
1

0.02%
0.01%
0.01%

0.33%
3.44%
3.50%

0.16%
0.19%
0.22%

33033 + 2.5%
51'? 30 + 2.3%
70330 + 3.2%

20

.
9 + .136
22.39 + .122
23.34 + .133

2.33

.54%
1.53%
1.53%

0.01%
0.01%
0.31%

3.39%
1.15%

13
15

32)0

.31

3.93

79 %

03

S3

7,

3.27%
0.39%
0.53%

Table A-II

TH=13339 + 273.00

(1.50%)

TC=

(1 .257.)

DY

F(DE)

TE(K)

30 +
.31

.30

DG=.10%

DB

Y(DB)

ERROR CONTRIBUTIONS TO TE
ETH
ETC
EG
EY

10
15

+26.6%
+18.3%
23 +14.2%

3.15 +.039
3.22 +.039
0.29 +.04 3

23.31
22.78
22.56

13.56%
9.54%
7.53%

10.35%
6.73%
5.33%

2.08%
.47%
1.16%

0.93%
3.64%
0.53%

33 +10.1%
53 + 6.8%
73 + 5.4%

3.43 +.341
0.69 +.343
3.94 +.045

22.15
21.43

5.52%
3.92%
3.23%

3.35%
2.31%

3.85%
3.63%
3.53%

0.37%
3.26%
0.22%

133 + 4.3%
153 + 3.5%
230 + 3.1%

.29 +.043
1.81 +.052

20.32
13.97

2.28 +.055

13.13

330 + 2.7%
533 + 2.3%
703 + 2.2%

3.08 +.359
4.35 +.065
5.33 +.368

16.33
15.34
13.83

1303 + 2.1%
1530 + 2.0%
2000 + 2.3%

6.48 +.071
7.90 +.074
3.97 +.376

12.45
13.91
9.83

3303 + 2.3%
5033 + 2.3%
7030 +2.3%

13.55 +.379
12.61 +.382
14.33 +.084

3.34
6.56
5.43

13303 + 2.1%
15303 + 2.1%
20333 + 2.2%

15.53 +.337
17.22 +.091
13.45 +.395

4.44
3.43
2.77

30303 + 2.4%
53333 + 2.3%
70303 + 3.1%

0.19 +.134
22.39 +.120
23.34 +.136

2.03
1.33
3.99

20.31

29

.44%

.01%

2.71%
2.31%
2.11%

0.63%
0.51%

3.42%
0.36%
0.33%

0.18%
0.16%
0.14%

.91%
.75%
1.68%

3.34%
3.21%
3.15%

0.30%
3.28%
3.27%

0.13%
3.12%
3.12%

.63%
.59%
.57%

0.11%
3.36%

3.26%
0.26%
3.27%

3.11%
3.11%
3.12%

1.55%
.53%
1.52%

0.34%
3.33%
0.32%

3.28%
3.33%
3.32%

3.12%
3.13%
3.14%

.52%
1.51%
1.51%

0.32%
0.01%
3.31%

3.36%
3.43%
0.49%

0.16%
3.19%
3.21%

1.51%
1.51%
1.51%

3.31%
0.01%
0.31%

3.62%
0.33%
1.13%

3.27%
3.38%
3.49%

1
1

1
1

3.0 7%

Table A-III

TH=1S300 + 270.23

TE(K)

F(DB)

TC=

DY = .01

DB

2.50

K
K

(1.53%)
(12.50%)

DG=.10%

Y(DB)

ERROR CONTRIBUTIONS TO TE
ETH
ETC
EY
EG
2.10%
1 .90%
.80%

5.33%
3.34%
2.50%

3.32%
3.29%
3.28%

.70%
.62%
.59%

.67%
.00%
3.72%

0.26%
3.25%
3.24%

3.11%

.56%
.54%
.53%

3.53%
3.34%
3.25%

3.24%
3.24%
3.24%

3.10%
3.10%
3.13%

.52%

1.5 1%

3.17%
3.10%

1.51%

2.3 7%

3.24%
2.24%
3.24%

3.13%
3.12%
3.13%

.51%
.50%
.50%

3.05%
0.04%
3.03%

0.24%
3.25%
0.25%

3.11%
3.11%
3.11%

.5 3%

3.27%
3.29%
0.32%

3.12%
0.13%
0.14%

3.0 1%

7 ,S%
+ 5 .7%
20 + 4 .7%

3.15 + .311
3.22 + .312
3.29 + .013

31.39
29.77
23.76

30 + 3 .7%
50 + 3 ,3%

0.43 + .015
0.S9 + .019
3.94 + .022

27.25
25.24
23.33

22.41

2 3.71

230 + 2 1%

1.29 + .027
1.81 + .033
2.23 + .338

300 + 2 ,0%
503 + 2 ,3%
700 +
.9%

3.38 + .045
4.35 + .354
5.33 + .059

17.30
15.65
14.24

.9%
.9%
,9%

6.48 + .0 64
7.92 + .069
8.97 + .3 72

12.77
11.13
9.99

.9%
.9%
2 ,0%

10.55 + .275
12.61 + .330
14.00 + .232

3.45
6.52
5.53

.50%

.5 3%

0.32%
3.31%
0.01%

10003 + 2 0%
+ 2 1%
22300 + 2 9 %

15.53 + .335
17.22 + .393
13.45 + .394

4.47
3.42
2.79

.53%
1.50%
1 .50%

0.31%
0.31%

2.3S%
0.42%
3.49%

0.16%
0.13%
0.21%

30003 + 2 ,4%
50003 + 2 .8%
70330 + 3 .17.

23.19 + .133
22.39 + .1 19
23.34 + .135

2.34

.34

0.99

0.33%
3.00%
3.33%

0.51%
0.37%
1.13%

3.27%
3.33%
0.49%

10
15

70 + 2 ,7%

130 + 2 .4%
153 + ;> ,2%
.

300 +
1500 +
2000 +
1

3000 +
5330 +
7000 +

1
1

15 300

19.53

30

1
1

1
1

.5 3%

.53%
.50%

1
1

3.14%
0.13%
3.12%
1

1%
1%

Table A-IV

TH=11330 + 165.33
TC=
333 +
.30
1

DY=.01

(1.53%)
(3.33%)

DG=.13%

DB

ERROR CONTRIBUTIONS TO TE

13 +63.67.
15 +45.4%

23 +35.3%

30 +24.1%
53 +15.2%
70

+11.4%

Y(DB)

F(DB)

TE(K)

ETH

EG

6.86%
5.15%

7.34%
4.93%
3.79%

3.19%
2.16%
.55%

2.61%
1.66%

15.53
15.44
15.37

4 7.33%
32.33%
24.67%

0.43 +.093
3.69 +.097
0.9 4 +.39 7

15.24
14.99
14.76

16.96%
12.79%
3.15%

3.44%
2.37%

.29 +.096
1.81 +.094

6.17%
4.63%
3.86%

2.23 +.093

14.43
13.94
13.50

303 + 4.1%
530 + 3.3%
700 + 2.9%

3.33 +.39
4.3 5 +.389
5.33 +.338

12.75
11.53
13.63

3.03%

+2.6%

2.4%
2030 + 2.3%

6.48 +.337
7.^3 +.03 7
8.97 +.037

9.65
8.42
7.52

2.30%
1.85%

3333 + 2.2%
5033 + 2.2%
70 03 + 2.2%

13.55 +.333
12.61 +.393
14.33 +.092

6.23
4.80
3.92

1.73%

10339 + 2.3%
15003 + 2.4%
23033 + 2.6%

15.53 +.096
17.22 +.133
13.45 +.109

3.39
2.33

303 03 + 2.8%
53300 + 3.5%
733 30 + 4.1 %

23.1.9

15 03 +

EY

3.15 +.399
0.22 +.399
3.29 +.099

100 + S.6%
150 + 6.4%
200 + 5.2%

1300

ETC

+.123
22.39 +.149
23.34 +.176

.34

1.31

3.34
0.62

31

2.4 7%

2.23%

.77%

13.29

.24%
3.69%
0.52%

3.96%
3.72%
0.63%

3.41%
3.31%
0.26%

0.35%
0.21%
0.15%

3.49%
2.40%
^^7
'0 /o

2.21%
3.17%
3.15%

0.11%
3.03%
3.36%

3.34%
3.32%
0.32%

0.15%
0.14%
3.14%

3.33%
3.36%
0.43%

2.14%
3.16%
3.11%

3.23%
3.25%
3.29%

.43%

'A
J

. \J

.61%

3.34%
0.03%
0.02%

.59%
1.57%
1.57%

0.02%
0.02%
2.01%

3.4 7%

1.56%
.55%
1.55%

0.01%
0.31%

2.39%

.13%

.26%

0.72%
0.55%

63 %

.31%

3.57%
3.6*%

1
1

.32%
.75%

0.39%
3.57%
0.76%

Table A-V

TH =

TC=
DY

TE(K)

F(D3)

1000 + 155.30 K
80 +
1.03 K

.01

(1

.50%)
.25%)

DG=.10%

DB

Y(DB)

(1

ERROR CONTRIBUTIONS TO TE
ETH
ETC
EC
EY

10
15

+26.7%
+18.4%
23 +14.3%

0.15 +.039
0.22 +.339
0.29 +.043

20.38
23.64
20.42

13.60%
9.57%
7.55%

13.03%
6.72%
5.05%

2.39%
1

.47%
1.16%

3.91%
0.64%
0.50%

30 +10.1%
53 + 6.3%
73 + 5.47.

0.43 +.041
0.69 +.044
0.94 +.346

20.01
19.29
13.68

5.54%
3.93%
3.24%

3.5 7%
2.32%
.4 5%
1

0.35%
0.61%
3.50%

3.37%
3.26%
0.22%

103 + 4.3%
150 + 3.5%
203 + 3.!%

1.29 +.3 43

17.93
S.36

2.23 +.355

2.72%
2.32%
2.12%

.32%
0.63%
3.51%

0.42%
0.36%
3.33%

0.!3%
0.15%
0.14%

333 + 2.7%
530 + 2.4%
703 + 2.2%

3.33 +.059
4.35 +.065
5.33 +.069

14.73
12.97
11.76

.91%
.75%
.58%

0.3 4%

0.21%
0.15%

0.30%
3.23%
3.27%

3.13%
0.12%
3.12%

+ 2.1%
1503 + 2.1%
2003 + 2.3%

5.43 +.072
7.90 +.075
3.97 +.077

13.46
8.98
7.96

.63%
.59%
1.57%

3.11%
0.03%
0.36%

3.27%
0.23%
3.29%

3.12%
0.12%
0.12%

3333 + 2.0%
5030 + 2. !%
7400 + 2.1%

13.55 +.033
12.61 +.334
14.03 +.383

6.58
4.98
4.35

.55%
.54%
.53%

0.34%
0.03%
0.02%

3.33%
0.34%
3.38%

3.13%
0.15%
0.17%

10303 + 2.2%
15333 + 2.3%
20033 + 2.5%

15.53 +.092
17.22 +.099
13.45 +.106

3.19
2.37

.52%
.52%
.52%

0.32%
0.02%
0.01%

0.45%
3.55%
3.66%

0.19%
0.24%
0.29%

30333 + 2.8%
50303 + 3.4%
70002 + 4.0%

23.19 +.1 19
22.39 +.145
23.34 +.172

.52%
1.51%
.51%

3.01%

0.37%

.29%
1.71%

0.33%
3.56%
3.74%

103

1.31 +.052

5.32

1
1

1
1
1

1
1

.89

1.35

0.36
3.63

32

3.01%

Table A-VI

TH=11090 + 165.03
TC=
4 +
0.50
DY

TE(K)

F(D3)

.31

(1.50%)
(12.53%)

DG:.10%

DB

Y(DE)

ERROR CONTRIBUTIONS TO TE
ETH
EG
ETC
EY

13 + 7.67.
15 + 5.7%
20 + 4.77.

3.15 + .011
3.22 + .012
0.29 + .013

28.96
27.63
26.62

2.13%

30 + 3.7%
50 + 3.0%
73 + 2.7%

0.43 + .015
3.69 + .319
0.94 + .022

25.11
23.11
21.75

100 + 2.4%
150 + 2.2%

1.29 + .027
1.81 + .333

200 + 2.1%

2.28 + .038

23.28
18.63
17.40

330 + 2.0%
500 + 2.3%
700 + 1.9%

3.33 + .045
4.35 + .054
5.33 + .060

15.73
13.58
12.21

1.52%
1.51%

1000 + 1.9%
1500 + 1.9%
2003 + 1.9%

6.48 + .365
7.90 + .070
8.97 + .073

3300 + 1.9%
5030 + 2.3%
7003 + 2.1%

.83%

5.31%
3.34%
2.51%

0.32%
0.29%
0.28%

3.14%
3.13%
0.12%

1.70%
.52%
1.59%

1.67%
1.33%
0.72%

3.26%
0.25%
0.25%

0.1 1%

.5 6%

0.53%
0.34%
3.25%

0.24%
0.24%
0.24%

0.13%
3.13%

.51%

0.17%
0.13%
0.33%

0.24%
3.24%
0.25%

0.13%
0.11%
3.11%

10.77
9.20
8.12

.51%
.50%
1.50%

0.35%
0.04%
0.03%

0.25%
0.26%
0.27%

0.11%
0.11%
0.12%

10.55 + .077
12.61 + .082
14.00 + .336

6.58
5.05
4.10

1.53%
.53%
1.53%
1

0.02%
0.01%
0.31%

3.29%
3.34%
3.33%

0.13%
0.15%
0.16%

10033 + 2.1%
15333 + 2.3%
20033 + 2.4%

15.53 + .090
17.22 + .393
18.45 + .134

3.22
2.39
1.93

1.53%
1.53%
.53%
1

0.31%
0.01%
0.01%

0.44%
0.54%
3.65%

0.24%
0.23%

30333 + 2.7%
50033 + 3.3%
70303 + 3.9%

20.19 + .113
22.59 + .144
23.84 + .170

1.36
0.86
0.63

1.53%
.50%
1.53%

0.31%
0.01%
0.31%

0.36%
1.23%
1.73%

33

.9 3%

.54%
1.53%
1

3.11%
0.11%

0.137.

3.197.

0.37%
3.55%
0.747.

Table A-VII

TH= 1253 +
TC =
303 +

DY=.31 DB

3.30
0.13

K
K

(3.24 7.)
(3.93%)

DG=.12%

ERROR CONTRIBUTIONS TO TE
F(DB)

TE(K)

+24.7%
+15.8%

Y(DB)

ETH

ETC

EY

EC

1.33%
0.39%

9.47%
6.44%
4.93%

2.337.

0.15 +.33 6
0.22 +.03 6
0.29 +.336

6.39
6.04
5.99

9. 79%
7.

0.677.

5.89
5.73
5.52

3.477.
2.217.

0.457.
0.277.

70 + 4.3%

0.43 +.335
3.69 +.336
0.94 +.036

1.67%

100 + 3.3%
150 + 2.5%
200 + 2.17.

+.037
+.037
2.28 +.038

5.23
4.93
4.62

300 +
500 +
700 +

.5%
.4%

3.33 +.339
4.35 +.042
5.33 +.044

1000 + 1 .5%
1500 + 1.57.
2300 + 1.7%

6.48 +.049
7.9 3 +.05 6
S.97 +.364

3030 + 2.0%
5000 + 2.77.
7000 + 3.37.

10.55 +.079
12.61 +.109
14.00 +.139

10300 + 4.4%
15030 + 6.1%
20033 + 7.8%

33300 +11 .3%

53003 +18.57.
70000 +25.8%

13
15

20 +12.87.

30 + 8.3%
50 + 5.67.

1
1

.8%

6.63%
5.35

4.11%

2.14%

0.23%

3.41%
2.21%
1.69%

0.96%
3.73%

1.26%
0.95%
0.79%

3.14%
0.13%
3.38%

1.31%
1.02%
3.33%

0.57%
3.44%
0.33%

4.12
3.40
2.93

0.63%
0.51%
0.45%

0.05%
0.04%
0.33%

0.757.

0.63%
0.68%

0.33%
0.29%
0.29%

2.33

0.417.

0.38%
0.36%

3.02%
3.02%
0.32%

0.71%
0.30%
0.91%

3.31%

1.34
1 .53
.13

3.35%
3.33%
3.33%

0.317.

0.72
0.53

1.13%
.61%
2.09%

0.49%
0.73%
0.91%

15.50 +.184
17.22 +.260
18.45 +.336

3.33
3.25
0.20

0.33%
0.32%
3.32%

3.31%

2.81%
4.02%
5.24%

.22%
1.75%
2.27%

20.19 +.488
22.39 +.797
23.34 +1.1 15

0.13
0.03
3.06

3.32%
0.32%
0.32%

0.31%
3.01%
3.01%

7.69%

3.33%
5.44%
7.57%

1
1

.29
.31

34

0.01%

0.317.

3.017.
3.317.

69 %

17.38%

.48%

0.357.

3.39%

Table A-VIII

TH=

TC=

1250 +
321 +

DY=.01

TE(K)

(3.24 7.)
(0.25%)

DG=.10%

DB

Y(DB)

F(DB)

3.03
0.20

ERROR CONTRIBUTIONS TO TE
ETH
ETC
EG
EY

7.7%
5.3%
23 + 4.2%

3.15 +.01
3.22 +.011
0.29 +.012

11.46
11.24
11.34

2.31%
1.62%

33 + 3.3%
53 + 2.1%
70 + 1 .7%

0.43 +.012
0.69 +.013
3.94 +.314

10.66
10.33
9.44

3.94%
0.67%
0.55%

.29 +.315

13 +
15 +

133 +

.23%

2.15%
1.44%
1.09%

2.23%
1.53%
1.25%

0.97%
3.63%
0.54%

73 %

0.44%
0.32%

3.92%
0.67%
0.56%

0.40%
3.29%
0.24%

.4%
.2%
1.1%

+.017
2.28 +.019

8.75
7.34
7.14

3.46%
3.39%
3.36%

3.23%
3.16%
0.12%

3.43%
3.42%
0.43%

0.21%
0.13%
0.17%

300 + 1.3%
533 + 0.9%
700 + 0.9%

3.08 +.321
4.35 +.026
5.33 +.329

6.11

4.33
3.98

3.32%
3.30%
0.29%

0.39%
0.06%
3.05%

3.39%
3.40%
3.43%

0.17%
3.17%
0.19%

1333 +
.0%
1500 + 1.1%
.'2%
2303 +

6.48 +.334
7.93 +.341
3.97 +.047

3.19

0.04%
3.33%

.94

3.28%
0.27%
3.27%

3.43%
0.57%
3.67%

3.21%
3.25%
3.29%

3333 +
.5%
5030 + 2.1%
7033 + 2.6%

13.55 +.063
12.61 +.385
14.00 +.110

1.40
0.93
0.66

3.26%
0.26%
0.26%

0.32%
3.02%
0.32%

0.36%

3.37%
3.54%
3.71%

13033 +3.5%
15330 + 4.9%
23333 + 6.3%

15.53 +.147
17.22 +.238
18.45 +.270

0.43
0.32
0.25

0.26%
0.26%
0.26%

3.32%
0.32%
0.02%

2.23%
3.22%
4.21%

3.97%

33300 + 9.1 %
53003 +14.9%
73330 +20.7%

23.19 +.393
22.39 +.64 2
23.34 +.89 6

0.17
0.13
0.37

3.26%
0.26%
3.26%

0.02%
3.32%
0.02%

6.19%
10.21%
14.32%

2 . 63 %

150+1
203 +

1
1

.31

2.41
1

.3

33 %

.25%
1.64%
1

.43%
1.82%

4.43%
6.12%

Table A-IX
TH= 1250 +
4 +
TC=
DY

F(DB)

TE(K)

10 +
15 +

.01

DB

(2.50%)

DG=.10%

ERROR CONTRIBUTIONS TO TE
EG
ETH
ETC
EY

Y(D3)

0.347

.017.

0.33%

0.14%

0.307.

0.63%

0.307.

0.137.

0.29%

0.517.

0.23%

0.12%

0.43 + .004
0.69 + .005
0.94 + .007

15.76
13.82
12.51

0.277.

0.347.
0.217.
0.157.

0.277.

0.26%
0.26%

0.127.
0.117.
0.117.

1.29 + .008
1.81 + .010

11.13
9.59

0.26%

0.1 1%

,77.

0.277.
0.277.

3.12%
0.122
3.137.

.4

23 +

,27.

30 +

.07.

7.

.87.

2 .87.

100 +
150 +
200 +

3 ,77.
.77.

2.28 + .012

3.52

300 +
500 +
730 +

,17.

3.08 + .015
4.35 + .023
5.33 + .024

1000 +
1503 +
2000 +

0, ,9%

,11
.37.

,07.

,17.

6.43 + .029
7.90 + .036
8.97 + .042

030 + 1 .47.
5303 +
.9%
7030 + 2 .47.

10.55 + .055
12.61 + .078
14.03 + .102

3 .27.
4 ,67.
5 ,9 7.

33033 + 3
50330 + 13
70030 + 19

(0.247.)

19.54
13.23
17.24

.37.

10000 +
15300 +
23300 +

K
K

0.15 + .003
0.22 + .003
0.29 + .003

50 +
70 +

3.00
0.10

.67.

.9

7.

.4

7.

15.50 + .136
17.22 + .194
13.45 + .252

20.19 + .368
22.39 + .601
23.34 + .333

0.26%
3.25

7.

0.257.
0.257.
0.257.

0.117.
0.077.

7.07

0.24

7.

0.047.

2.29%

5.41

0.24%

0.33%

0.337.

0.14%

4.42

0.247.

0.027.

0.3 67.

0.167.

3.50
2.62
2.10

0.247.

0.027.
0.017.

0.427.

0.187.

0.24%
0.24%

0.51%
3.60%

0.267.

3.97

2.24%
0.24%

3.71

3.247.

3.51

0.35
3.26

3.24
0.24%
0.24

0.18
3.1

0.03

3.24%

1.51

0.01%
2.017.
0.017.
0.017.

0.79%

3.01%
3.017.

2.28%
3.01%

3.31%

3.937.

0.24%

2.01%

0.247.

0.017.
3.317.

5.30%
9.56%
13.42%

7.

7.

36

0.36%

1.167.
1

.537.

0.22%

2.347.
0.507.
0.667.

3.9 3%
1

.337.

.71%

2.517.
4.127.
5.747.

Table A-X

TH =
TC =

692 +
33 +

3.9.3

2.10

DG=.10

DB

DY=.31

(3.13?.)
(3.037.)
7.

ERROR CONTRIBUTIONS TO TE
TE(K)

F(F)B)

Y(DB)

ETH

ETC

1.797.
1.237.
3.917.

12.73%
8.72%
6.69%

5.55%
3.79%
2.91%

3.61%
0.38%
3.28%

4.67%
3.35%
2.37%

2.33%

.86%

0.31%

.4S7.
.317.

3.647.
0.577.

1.17%
1.12%
.17%

3.51%
0.49%

3.517.

.29%
1.55%
1.32%

3.5 5%
3.67%
3.79%

EY

EG

13 +2 7 ,2%
15 + 3. .5%
23 + 14 ,2%

3.15 + .039
0.22 + .343
3.29 + .043

3.55

7.127.

3.51

4.32%

3.47

3.677.

33 + 9 .3%
53 + S .4%
73 + 4 ,9%

3.43 + .343
0.69 + .041
0.94 + .041

3.40
3.26
3.14

2.537.

3, ,8%
3 ,3%
2 ,6%

1.29 + .34 2
.31 + .344
2.23 + .045

2.97
2.72
2.51

3.927.
0.697.

+ 2 ,2%
533 + 2 ,0%
733 +
,17.

3.33 + .349
4.35 + .056
5.33 + .363

2.13

0.4
0.3

1000 + 2 ,2%
1503 + 2 .5%
20 3 2 + 2 ,9%

6.4 3 + .374
7.93 + .392
3.97 + .1 10

1.14

3.3

37.

3.04%

0.36

0.047.

63

0.23%
3.26%

10.55 + .147
12.61 + .223
14.03 + .293

3.49
0.31
0.23

0.25%
3.24
3.24%

0.03%
0.03%
3.03

2.39%
3.55%
4.727.

2.35%

15.53 + .404
17.22 + .539
13.45 + .777

0.16

3.24%
3.23%

0.037.

3.11

6.49%
9.43%

3.03

3.237.

2.31%
4.09%
5.37%

133 +

+
203 +
15?)

3 '3 3

r>

3033 +
530 3 +
7303 +

3, ,7%
5 ,4%

10003 +
15333 +

9 ,6%

23000 +

18, ,27.

7 ,37.

3 ,37.

33033 + 27

,07.

53 330 + 46
70333 + 69

.4'.

,4

7.

23.19 +1.163
22.39 +2.331
23.84 +3.00

1.617.
1.217.

0.5 77.

1.73
1.44

67.

3.33%

77.

3.0 6%

3.33%

0.35%

7.

0.0S

3.33

0.23%
0.23%

0.32

37

3.23%
3.14%
0.11%

23 7.

3.33%

7.

3.33%
3.33

7.

. 03 %
0.037.
. 33 7.

1
1

12.52%
13.327.

32.3 77.
50.397.

1
1

.337.
.037.

.04%

1.547.

7.96%
13.23%
13.71%

Table A-XI
TH=
TC:
DY

692 +
80 +
.01

3.90
0.20

(3.137.)
(3.257.)

DG=.10Z

DB

ERROR CONTRIBUTIONS TO TE
TE(K)

+
+
20 +

Y(DB)

FCDB)

ETH

ETC

EY

2.29%

2.33%

,3%
,9%
3 ,3%

3.15 + .013
3.22 + .013
3.29 + .311

3.92
8.72
3.52

0.937.
7,

.54%
1.15%

30 + 2 ,37.
53 +
.9
73 + 1 ,6%

3.43 + .01
3.69 + .012
0.94 + .013

3.17
7.56
7.36

0.547,
3.387.
0.3 2%

3.79
0.43%
3.36%

1.29 + .314
1.31 + .316
2.28 + .318

6.43
5.64
5.03

0.26%

3.25%
3.13%
8.15%

4.17
3.13
2.52

3.197.
0.177.

10
15

100 +
150 +
20 3 +

300 +
5:53 +
733 +
1000 +
1500 +
2000 +

3000 +
5 303 +
7300 +

7.

,37.

1
1

.1%
,0%

3.69%
0.87%

0.3 3%
3.38%
3.4 6%

13.55 + .33"
12.61 + .136
14.00 + .183

0.79
0.49
0.36

3.15%
0.15%
0.15%

0.04%
0.04%
0.34%

15.53 + .253
17.22 + .371
18.45 + .49^

3.26
3.17
0.13

0.15%
3.15%
0.15%

0.03%
3.03%
0.03%

20.19 + .729
22.39 +1.224
23.34 + .754

0.39
3.35
0.34

2.15%
0.15%
3.15%

0.03%

.77.

2 ,27.
3 .3%
4 .4%

7.

.57.

3.4 7%

0.3 6%
3.0 5%
0.34%

7.90 + .353
8.97 + .3 65

.4

3.23%
0.21%
0.20%

3.16%
3.15%
0.15%

6. 48 + .043

,4%

.07.

3.54%
0.49%

1.95
1.42
1.12

,27.

33303 + 17
50003 +28
73303 + 40

3.437.

3.32%
3.27%

3.257.

.47.

.33%

0.73%
0.61%

3.58%

7.

7.

.33%

0.16%

1%

,77.

3.21

1.347.

3.73%
3.58%

0.21%
3.23%

,3

23 %

.63%

3.47%

10000 + s
15030 + q
20302 + 11

0.5 27.

3.38 + .821
4.3 5 + .327
5.33 + .353

7.

3.74

3.33%
3.3 6%

,3
,3

7.

.32%

3.117.

EG

38

33 %

0.33%

.057.

.43%

2.18%
2.937.

4.067.

5.9S%
7.36%
1

.737.

19.82%
23.65%

3.62%
3.95%
1

.27%

.76%

2.53%
3.43%
5.34%
3.35%
11.72%

Table A-XII

692 +

TH=
TC =

3.90
0.10

DYr.31 DB

(0.13%)

(.2.57)7.)

DG=.13%

P.RROR

F(DB)

TE(K)

10 +
15 +

20 +

CONTRIBUTIONS TO TE

Y(OR)

ETH

ETC

FY

0.33%
0.30%
3.29%

3.14%
3.13%
3.12%
3.127.

,77
.3%
,1%

3.15 + .00 2
3.22 + .033
0.29 + .303

17.00
15.71
14.72

0.13%
0.167

1.02%
3.69%
3.52

9%

3.43 + .004
0.69 + .005
0.94 + .006

13.27

0.157,

0.35%

.30

3.147
0.14%

0.227.

3.27%
0.27%

0.16%

0.277.

.29 + .00 7
1.01 + .039
+ .311

3.82
7.33

3.147.

0. 127.

3.13
0.13 %

0.08%
0.36%

0.29%

3.00 + .015
4.3 5 + .020
5.3 3 + .026

5.14
3.74
2.96

0.13/1

3. 05 %

3.137

.33

0.15%
0.17%

3.03

0.34%
3.437
0.47%

6.43 + .033
7.90 + .344
3.9 7 + .05 5

2.27

3.57%
0.74%
3.90%

25%
o!32%
0.39%

0.90

70 30 + 3 ,0%

+ .375
. 5 5
12.61 + .110
14.00 + .160

10030 + 5 .3%
15003 + 7 ,7%
20003 + 10

15.53 + .223
17.22 + .320
13.45 + .453

0.29

3000? + 15 3%
50300 +25 .3%
70000 +35 .6%

20.19 + .64 5
02.39 + .300
23.04 +1.539

33 +
50 +
70 +

p,

.7%
.7%

100 + g ,6%
150 + 3 ,67.
2 30

300 +
530 +
700 +
10-20

1530 +
?030 +

,6%
.7

7.

r/

.7%
Q ,3%
l

.0%

.2/5

.4

7.

3300 + l ,9?
5003 + o .9%

7.

11
1

. 1

6.41

1
1

.64
.25

0.177.

7.

3.13
7
. 1

37

0.137
3.13%

7.

0. 3?%
0. 32%

3.02%

20 7
3*

7.

3.12%

0.127
0.13%
3.13%

?.

.13%

3.327.

l.47

3.13%
3.137

0.02
0.22%

1.917.

3.54%
0.03%

2.53%

.12%

0.32%

3.53%
5.27 7
6.967

.5-7

0.15

0.13 %
3.13%
3.13%

2.23%
3.317

Z. 13

0.13

0.31

3.5
3.4

5
1

<3

0.0 6
0.04

39

.13 7
0.137.

7,

? %

3.327

10.33%

7.
1 7.

3.31

17.4 7%

25.36%

4.47%
7.41%
13.33%

Table A-XIII

TH=
TC=
DY

+
30 +

3 73

0.50
1

(3.13%)
(1.25%)

DGr.13%

DB

.01

.00 K

ERROR CONTRIBUTIONS TO TE
F(P3)

TE(K)

Y(DB)

ETK

ETC

10
15

+13.5%
+12.7%
20 + 9.8%

3.15 +.027
0.22 +.027
3.29 +.327

6.29
6.11
5.94

.54%
1.33%
3.35%

13.37%
3.33%
6.71%

33 + 6.9%
53 + 4.6%
73 + 3.6%

0.43 +.023
0.69 +.329
0.94 +.333

5.64
5.12
4.73

0.63%
0.44%

133 + 2.9%
153 + 2.4%

1.29 +.332

230 + 2.1%

EG

EY

2.71

1.33%
1.54%

1.13%
3.34%
0.67%

0.3 7%

4.53%
2.39%
2.16%

1.16%
0.86%
0.75%

0.53%
0.33%
0.32%

2.23 +.033

4.23
3.57
3.11

0.31%
0.26%
3.24%

1.61%
1.19%
3.98%

0.67%
0.63%
0.63%

3.29%
0.27%
3.27%

333 +
.9%
530 + 1.9%
730 + 2.1%

3.33 +.343
4.35 +.053
5.33 +.053

2.48
1.78
1.39

0.22%
0.20%
0.19%

3.77%
3.60%
0.52%

3.67%
0.33%
0.94%

3.29%
a. 35%
3.41%

1330 + 2.3%
1533 + 2.8%
2303 + 3.4%

6.43 +.373
7.90 +.103
8.97 +.12S

1.04
0.74
3.57

0.13%
0.13%
3.13%

0.47%
3.43%
3.43%

1.17%
1.55%
1.94%

0.51%
3.67%
3.34%

3303 + 4.5%
5033 + 6.7%
7303 + 9.3%

10.55 +.177
12.51 +.275
14.33 +.3 74

3.39
3.24
0.13

0.13%
3.17%
0.17%

0.33%
3.37%
3.36%

2.72%
4.30%
5.83%

1.13%
1.36%
2.55%

20033 +24.0%

15.53 +.522
7.22 +.773
18.45 +1 .329

3.12
0.08
3.36

3.17%
3.17%
3.17%

0.35%
3.27%
12.33%
0.35%
3.35%
16.53%

33033 +36.4%
53300 +55.3%
73000 +135.3%

23.19 +1 .566
22.39 +2.318
23.14 +4.575

1.81 +.335

13
15

303 +12.4%
300 +13.1%

0.04
3.03
0.32

()

0.17%
3.17%
3.17%

3.3 5%

3.34%
3.34%

3.57%
5.29%
7.31%

13.53%
25.33%
17.74%
7.32%
79.76% 25.53%

Table A-XIV
373 +

TH=
TC=

DY=.01 DB

F(DB)

TE(K)

0.53
0.53

K
K

ERROR CONTRIBUTIONS TO TE
ETH
ETC
EG
EY

YCDB)

0.15 + .333
0.22 + .009
0.29 + .339

14.37
13.13

30 + 2.4 7.
53 + 1.7%
73 + 1 .4%

0.43 + .313
3.69 + .011
0.94 + .012

133 + 1.2%
153 + 1.17.
200 + 1.07.

1.29 + .314
1.31 + .316
2.28 + .319

330 + 1.17.
500 + 1.27.
700 + 1.37.

3.33 + .323
4.35 + .332
5.33 + .043

3.45
2.39

1533 + 2.3%
20 03 + 2.4

1.36
0.95
0.73

3.147.

7.

6.43 + .352
7.90 + .372
8.97 + .392

3033 + 3.3%
5033 + 5.1%
7330 + 6.9%

13.55 + .131
12.61 + .209
14.00 + .2S7

10003 + 9.67.
15033 + 14.17.
20303 +13.7%

15.53 + .435
17.22 + .532
13.45 + .302

30333 +:23.2%
53033 + 49.17.
70333 + 74 . 5 %

23.19 +1 .214
22.39 +2.113
23.34 +3.221

10 33

.57.

(12.507.)

DG=.137.

+ 5.9%
+ 4.1%
20 + 3.2%
13
15

(0.13%)

0.19%
0.17%
3.16%

5.19%
3.50%
2.66%

0.31%

0.15%
0.13%

0.297.

0.137.

13.74
8.94
7.77

0.157.

1.82%
1.15%
0.36%

0.29%
0.29%
0.29%

3.127.

3.15%
0.14%

6.53

0.14%
0.14%
0.14%

3.64%
0.47%
0.39%

3.31%
3.34%
0.36%

0.13%
0.15%
3.15%

0.147.

0.30%
3.24%
0.21%

0.43%
0.55
0.67%

3.13%
0.247.
3.297.

0.367.

0.37%

1.17%
1.48%

0.517.

0.14%

0.19%
0.17%
0.16%

0.53

0.147.

0.157.

3.31

0.22

0.14%
3.14%

0.15%
0.14%

2.11%
3.36%
4.61%

0.92%
1.46%
2.30%

0.1S

0.147.

0.147.

0.1

0.14%
3.14%

0.14%
0.14%

6.53%
9.63%
12.92%

2.31%
4.17%

5.31

4.49

0.14%
0.14%

.33

0.03
0.05
0.03
0.32

41

3.14%

0.14%
0.14%
0.147.

0.14%
0.14%
0.14%

0.337.

7.

19.66%
34.33%
54.41%

0.12%
0.15%

0.64%

5.547.
8.297.

13.91%
19.73%

Table A-XV

TH=
TC=

373 +
303 +

0.15
0.10

DY=.01 OB

K
K

(0.04%)
(0.03%)

DG: .10%

ERROR CONTRIBUTIONS TO TE
FCDB)

TEC K)

YCDB)

ETH

ETC

6.37%
4.32%
3.29%

5.25%
3.54%
2.69%

37.46%
25.70%
19.34%

16.26%
11.16%
3.61%

EY

EG

+65.3%
+44.7%
20 +34.4%

0.15 +.095
0.22 + .096
3.29 +.096

0.92

30 +24.2%
50 +16.0%
70 +12.5%

0.4 3 +.098
0.69 +. 102

2.26%

0.94 +.136

0.37
0.82
0.78

.44%
.09%

1.34%
1.16%
3.37%

13.93%
9.34%
7.39%

6.07%
4.06%
3.21%

100 +13.0%
150 + 3.2%
200 + 7.4%

.29 +.112
1.31 +.121
2. 23 +.131

0.73
3.65
0.59

0.32%
0.62%
0.51%

0.65%
0.43%
3.39%

5.97%
4.95%
4.52%

2.59%
2.15%
1.96%

300 + 6.8%
503 + 6.9%
730 + 7.4%

3.03 +.153
4.35 +.139
5.33 +.229

3.53
3.33
0.31

3.41%
0.33%
3.29%

0.31%
3.24%
0.21%

4.25%
4.41%
4.84%

1.34%
1.91%
2.13%

1003 +3.5%
1503 +10.6%

0.24
0.17
3.14

0.27%
3.25%
0.24%

0.19%
0.17%
3.16%

5.64%
7.11%
3.66%

2030 +12.3%

6.48 +.238
7.95 +.336
3.97 +.435

3003 +17.3%
5030 +26.7%
7303 +36.6%

10.55 +.686
12.61 +1 .096
14.03 +1 .523

3.10
0.06
0.04

3.23%
3.22%
3.21%

3.15% 11.34%
0.15%
13.50%
3.14% 25.64%

171

15

13000 +53.0%
15 030 +8 7.5%
20333 +142.2%

15.53 +2.233
17.22 +3.727
18.45 +6.335

0.91

0.89

0.03

3.32
0.02

42

1
1

3.21%
3.21%
3.21%

2.45%
.

33 %

3.74%
5.39%
7.34%
10.64%

14.9-3%
3.14% 37.74%
0.14% 64.66% 22.47%
3.14% 111.11% 30.72

Table A-XVI
TH =
TC =
DY

F(DB)

TEC K)

23 + 3.3%

0.15 +.013
3.22 +.313
3.29 +.011

33 + 2.8%
53 + 2.3%
73 + 1.6%
133 + 1.4%
15-3 + 1.2%
230+ 1.2%

10 + 7.0%
15 + 4.9%

+
80 +

3 73

.01

D8

Y(D3)

0.15
0.20

(3.34%)
(0.25%)

DGr.13

7.

ERROR CONTRIBUTIONS TO TE
ETH
EG
ETC
EY

6.29

0.4 6%

S.ll

5.94

0.32%
3.26%

0.43 +.311
3.69 +.012
3.94 +.314

5.64
5.12
4.70

0.19%
0.13%
3.11%

1.29 +.315
1.31 +.313

2.23 +.321

4.20
3.57
3.11

33 3 + 1 .2%
530 + 1.3%
733 +
.5%

3.33 +.326
4.35 +.336
5.33 +.04 6

1333 +
.3%
1533 + 2.4%
2333 + 2.9%

2.61%

2.71%

.13%

.93%
.54%

0.34%
3.67%

3.92%
3.53%
0.43%

1.16%
0.36%
0.75%

3.53%
3.33%
3.32%

3.39%
3.33%
0.07%

0.32%
3.24%
3.20%

0.67%
0.63%
0.63%

3.29%
3.27%
3.27%

2.43
1.73
1.39

3.36%
0.36%
3.36%

3.15%
0.12%
3.13%

3.67%
0.33%
3.94%

3.29%
0.35%
3.41%

6.43 +.361
7.90 +.336
^.97 +.1 11

1.34
0.74
3.57

3.0 6%

0.05%
0.05%

3.09%
3.39%
3.33%

1.17%
1.55%
1.94%

3.51%
3.67%
3.34%

3333 + 4.3%
5000 + 6.3%
7300 + 3.5%

13.55 +.160
12.61 +.253
14.00 +.357

3.39
3.24
3.13

3.35%
3.35%
3.35%

3.03%
0.37%
0.37%

2.72%
4.30%
5.83%

1.13%
1.36%
2.55%

10333 +12.3%
15333 +17.7%
2333 3 +2 3.6%

15.53 +.505
17.22 +.756

3.12
3.03
0.36

3.05%
0.35%
3.35%

3.37%
3.27%
0.07%
12.33%
0.37%
16.53%

33333 +36.3%
53333 +64.9%
73333 +135.4%

23.19 +1 .549
22.39 +2.3>J1
23. S4 +4.553

13.45+1.012

3.34
3.33
0.02

43

3.05%
3.35%
3.35%

.77%
1.3 4%
1

1
1

3.57%
5.29%
7.31%

3.37% 25.33%
10.53%
3.37% 47.32%
17.74%
3.37% 79.76% 25.50%

Table A- XVI

TH =
TC=

3.15
3.10

373
4

DY=.01

(3.04%)
(2.50%)

DG=.10%

D3

ERROR CONTRIBUTIONS TO TE
F(DB)

TECK)

Y(D3)

ETH

ETC

EY

EG

1.34%

3.33%
3.31%
29 %

3.15%
3.13%
3.13%

+ 1.6%
.2%
23 + 1.0%

3.15 +.002
3.22 +.333
3.29 +.033

14.37
13.13
12.14

3.36%
3.35%
3.35%

0.73%
3.53%

33 + 2.87.
53 + 3.7%
70 + 3.6%

0.43 +.333
3.69 +.334
3.94 +.335

10.74
3.94
7.77

3.35%
3.34%
3.04%

3.36%
3.23%
3.17%

3.29%
3.29%
3.29%

3.12%
0.12%
3.13%

.29 +.30 7

.31

6.53
5.31
4.49

3.34%
3.34%
34 %

0.13 %
3.39%
3.03%

3.31%
3.34%
3.36%

3.13%
3.15%
3.16%

3.36%
0.05%
3.34%

3.43%
3.55%
3.67%

3.13%
S.24%
3.29%

3.36%
1.17%
1.43%

3.37%
3.51%
3.92 %

10

15+1

130 + 3.6%
153 + 3.6%

+.339

230 + 3.6%

2.23 +.31

333 + 3.7%
533 + 3.9%
703 + 1.3%

3.33 +.31
4.35 +.324
5.33 +.332

3.45
2.3^
1.33

0.34%
3.34%
3.04%

1030 + 1 .3%
15 33 +
.*%
2333 + 2.2%

6.43 +.344
7.93 +.364
3.97 +.333

.36
3.95
3.73

3.34%
3.34%
3.04%

3.34%
3.33%

3333 +3.1%
5303 + 4.9%
7333 + 6.7%

13.55 +.123
12.61 +.201
14.03 +.279

3.53

3.34
3.34%

3.22

0.04%

3.33%
3.33%
3.03%

2.1 1%

0.31

13333 + 9.4%
15303 +13.9%
23303 +1P.5%

15.53 +.396
17.22 +.593
IS. 45 +.793

3.16

3.34%
3.34%
3.34%

3.33%
3.03%

6.5?%
9.63%
12.92%

333 3'! +23.0%

50333 +48.9%
70333 +74.3%

23.19 +1 .205
22.3S +2.1 13
23.34 +3.212

0.1

3.33
3.35
3.33
D.32

44

7.

0.34%
3.04%
3.34%

33 %

33

3.33%
3.33%
3.33%

,3S%

4.61%

19.66%
34.33%
54.41%

64 %

.4 6%
.

9|

3%

2.31%
4.17%
5.54%
3.29%
13.91%
19.73%

Table A-XVIII
TH =
TC:
DY

F(D3)

TE(K)

+
+

30
3

Y(DB)

0.15 +.032
0.22 +.033
0.29 +.033

5.37

33 + 8.5%
53 + 5.7%

0.43 +.035
0.69 +.037
0.94 +.039

4.77
4.3
3.92

.29 +.041

200 + 2.3%

+.046
2.28 +.050

323 + 2.6%
5 00 + 2.6%
700 + 2.8%

3.08 +.058
4.35 +.073
5.33 +.037

1.93
1 .40

1303 + 3.2%
1530 + 3.97.
2000 + 4.6%

6.43 +.107
7.90 +.141
3.97 +.174

0.31

0.57
0.44

3000 + 6.17.
5 300 + 9.0%
7033 +12.1%

10.55 +.243
12.61 +.371
14.00 +.503

0.33
0.13
0.13

10000 +16.6%
15030 +24.57.
20000 +32.6%

15.53 +.703
17.22 +1 .042
13.45 +1.396

0.09
0.06
0.35

30033 +50.37.
50030 +93.0%
73030 +197.17.

23.19 +2.165
22.39 +4.232
23.34 +3.526

133 + 3.77.
153 + 3.1%

1
1

.81

5.21

5.05

K
K

(0.33%)
(1.25 7.)

DG=.10%

DB

.01

10 +22.4%
15 +15.4%
2 5 +1 1 .9%

70 + 4.67.

1.03
1.33

ERROR CONTRIBUTIONS TO TE
ETH
ETC
EY
EG

4.39%
2.33%
2.27%

14.09%
9.55%
7.27%

2.92%
2.09%
1.67%

1.277.
3.917.

0.55
0.41%
0.36%

1.67%

5.03%

1.27%

1.187.

3.137.

0.97%

2.42%

0.957.
3.837.

3.47

0.82%

1.82%

2.91

0.737.
3 . 64 %

1.367.

1.14%

0.53%
0.53%
0.51%

0.91%
0.73%
0.65%

0.497.
0.437.

0.59%
0.55
0.52%

2.52

.03

0.03
0.02
0.01

45

3.4 77.

0.47%
3.4S7.

0.46%
46%
0.4 57.

J.

7.

7.

73

7.

3.33%
0.977.

1.17%
1.47%
1.99%
2.537.

7.

0.33%
0.31%
0.32%

0.35%
3.42%
0.51%
0.64%
0.36%
1.09%

0.50%
0.43%
0.47%

3.557.

5.65%
7.77%

2.457.

0.47%

10.997.

4.73%

0.4 6%

0.46%

0.467.

0.4 6%

0.4 6%

0.45%
0.46%

0.75
0.72%

0.73%

16.52%
22.34%

.54%

3.35%
7.0 2%

9.34%

35.35% 14.07%
0.4S% 72.957. 24.157.
0.46% 160.617. 35.6

Table A-XIX
TH=
TC =

303 +
S3 +

3.13
0.20

(3.03%)
(0.25%)

DG=.10%

DY=.31 DB

ERROR CONTRIBUTIONS TO TE
T(K)

10 +
15 +

F(DB)

7.4%
5.2%

20 + 4.17.

33 + 3.0%
50 + 2.1%
70 + 1.8%

0.15 +.011
3.22 +.311
0.29 +.01
0.4 3 +.012

3.69 +.014
2.94 +.215

Y(DB)

5.37

ETH

ETC

EY

EG

2.92%
2.39%
.67%

1.27%
3.91%
3.73%

5.21

0.41%
0.29%

5.35

23 %

1.91%
1.45%

4.77
4.33
3.92

0.17%
0.12%
0.10%

1.03%
0.64%
0.43%

1.27%
3.95%
0.33%

0.55%
0.41%
3.36%

0.36%
0.27%
0.23%

3.75%
0.72%
0.75%

3.33%
3.31%
0.32%
0.35%
3.42%
0.51%

Cm

<* /

133 + 1.5%
150 + 1.4%
2 33 + 1.3%

.29 +.017
1.81 +.020

3.47

2.28 +.024

2.52

0.08%
0.37%
0.06%

300 +
.4%
5 30 + 1.6%
700 +
.9%

3.38 +.033
4.35 +.044
5.33 +.357

.93
.40
.08

0.36%
0.35%
0.35%

3.13%
0.15%
3.13%

3.30%
3.97%
1.17%

1000 + 2.3%
1500 + 3.0%
2000 + 3.7%

6.48 +.077
7.90 +.139
3.97 +.142

3.81

3.35%
0.05%
0.05%

0.12%
3.11%
3.13%

1.4 7%

0.57
0.44

1.99%
2.53%

0.36%
1.09%

3000 + 5.2%
5300 + 3.27.
7003 +11 .3%

10.55 +.207
12.61 +.333
14.30 +.4 70

0.30
0.18
0.13

3.05%
0.35%
0.35%

3.10%
0.13%
0.09%

3.55%
5.65%
7.77%

1.54%
2.45%

10300 +15.9%
15030 +23.7%
20 030 +31 .3%

15.50 +.673
17.22 +1 .009
18.45 +1.362

3.39
3.06
3.05

0.35%
0.05%
0.05%

0.39%
10.99%
0.09%
16.52%
0.09% 22.34%

30330 +49.6%
53030 +97.2%
73030 +196.4%

20.19 +2.132
22.39 +4. 199
23.34 +3.492

2.91

3.33
0.02
0.31

46

0.05%
0.05%
0.05%

64 %

3.3 6%

4.73%
7.32%
9.34%

0.09% 35.35% 14.37%

3.09% 72.95% 24.15%
0.09% 163.61% 35.62

Table A-XX
TH=
TC=

390 +
4 +

DY=.01

F(DB)

TE(K)

DB

.33 K

0.50

DG=.13%

Y(DB)

ERROR CONTRIBUTIONS TO TE
ETH
ETC
EG
EY

0.43%
0.41%

15 +

4.4%
23 + 3.5%

0.15 +.339
3.22 +.309
0.29 +.313

13.45
12.23
I 1.25

33 + 2.7%
50 + 2.3%
70 + 1.7%

3.43 +.311
3.69 +.313
3.94 +.314

9.37
3.12
6.99

0.33%

130 + 1.5%
153 + 1.4%

1.29 +.317

10 + 6.2%

(3.33%)
(12.50%)

5.24%
3.55%
2.73%

0.34%
3.31%
0.30%

0.15%
0.13%
0.13%

0.36%

1.86%
1.13%
0.39%

3.29%
0.29%
3.33%

3.13%
3.13%
3.13%

5.35
4.66
3.39

0.35%
0.35%
3.34%

3.53%
3.51%
3.4 2%

3.3 2%
3.4 3%

3.14%
0.16%
3.17%

4.35 +.342
5.33 +.352

2.95
2.31
1.52

3.34%
0.34%
3.34%

0.34%
3.27%
0.24%

3.47%
3.63%
3.73%

3.21%
3.27%
3.34%

1033 + 2.0%
1533 + 2.5%
2033 + 3.1%

6.43 +.363
7.93 +.393
3.97 +.1 13

1.12
3.73
0.63

3.34%
0.34%
0.34%

3.22%
3.23%
3.19%

.02%
.43%
.79%

3.44%
0.61%
3.73%

3333 + 4.2%
5 333 + 6.4%
7 303 + 3.7%

13.55 +.167
12.61 +.264
14.33 +.352

3.4

3.25
3.13

3.34%
3.34%
3.34%

3.19%
0.13%
3.13%

2.57%
4.13%
5.73%

.12%
1.79%
2.4 7%

10333 +12.1%
15333 +17.3%
23033 +23.6%

15.53 + .5 ^9
17.22 +.75 7
13.45 +1.313

3.13
3.33
3.36

3.34%
3.34%
3.34%

3.17%
8.3 7%
0.17%
12.33%
0.17%
16.23%

33333 +35.8%
50030 +64.2%

23.19 +1 .543
22.39 +2.772
23.34 +4.435

200 + 1.3%

333 + 1.4%
533 + 1.5%
700 + 1.7%

7 3

333 +1^3.7%

+.323
2.28 +.024
1

.31

3. 33

+.333

0.34
3.33
3.32

47

3.4 7%

0.3 6%

3.34%
3.34%
0.34%

3.36%

1
1
1

3.49%
5.13%
5.39%

0.17% 24.9 6%
13.34%
3.17% 45.21%
17.49%
3.17% 73.32% 25.16%

Table A-XXI
TK=
TCr

333 +
4 +

DY=.01

F(DB)

TECK)

13 +
15 +

1.2%

23 +

.6%
.07.

33 + 3.3%
50 + 0.7%
73 + 0.7%

133 + 0.6%
153 + 0.7%

0.13
3.13

3.15 +.332
0.22 +.033
0.29 +.333

13.45
12.20

(0.33%)
(2.53%)

DG=.10%

DB

YOB)

ERROR CONTRIBUTIONS TO TE
ETH
ETC
EY
EG

.25

3.05%
3.34%
0.34%

1.05%
3.71%
3.54%

3.34%
0.31%
0.33%

3.15%
3.13%
0.13%

3.43 +.033
3.69 +.334
0.94 +.305

9.37
8.12
6.99

0.34%
0.04%
0.34%

0.37%
3.?4%
3.13%

3.29%
3.29%
3.30%

3.13%
0.13%
3.13%

1.29 +.007

5.35
4.66
3. 39

0.04%
0.03%
0.03%

0.14%
0.10%
3.33%

3.32%
3.36%
3.43%

0.14%
0.16%
3.17%

11

233 + 0.7%

+.013
2.23 +.012

300 + 0.3%
530 +
.0%
730 + 1 .2%

3.38 +.017
4.35 +.027
5.33 +.33 7

2.95
2.01
1.52

0.33%
0.33%
0.03%

3.37%
0.35%
0.35%

3.47%
3.78%

3.21%
0.27%
0.34%

1303 + 1 .5%
1533 + 2.1%
2:333 + 2.6%

6.48 +.052
7.90 +.076
3.97 +.100

1.12
3.78
0.60

0.03%
0.03%
3.03%

0.34%
3.34%
3.34%

1.32%
1.43%
1.79%

0.44%
3.61%
3.78%

3333 + 3.3%
5300 + 6.3%
7333 + 8.2%

13.55 +. 149
12.61 +.246
14.03 +.344

3.41
3.25
0.18

0.03%
33 %
3 . 33 %

3.34%
0.34%
0.04%

2.57%
4.13%
5.73%

1.12%
.79%
2.47%

+11.6%

15030 +17.3%
23333 +23.2%

15.53 +.493
17.22 +.733
13.45 +.991

3.13
3.23
0.06

0.33%
0.03%
3.33%

3.03%
3.33%
3.33%

3.07%
12.33%
16.23%

3.49%
5.18%
6.39%

30303 +35.4%
53330 +63.3%
70330 +103.2%

23.19 +1 .521
22.39 +2.753
23.34 +4.466

133 33

.81

3.04
3.33
3.02

48

0.33%
3.33%
0.33%

53 %

3.03% 24.96%
10.34%
3.03% 4 6.21% 17.49%
3.33% 73.32% 25.16%

'

Appendix
In this appendix,

'

Mismatch

B.

different mismatch errors are expressed

in a set of figures and tables.

The derivations of the equa-

tions used in the computer program are presented last.

The

computer program actually used is contained in Appendix D.

Mismatch Uncertainty

B.l

The following figures, referred to as the "Mismatch

Uncertainty" figures, are based on Eq.

(4)

in

the text.

This

set can be used to estimate the mismatch ambiguity discussed

in Section

or they can be used

to

estimate the error caused

by using noise standards that have a different reflection

The maximum uncertainty in

coefficient than the "antenna".

*

b,
'

versus
and

|r
'

|r
1

r
'

std

.,
std

ant

ant
r

ant

is

amp

is
'

graphed
&
r

for various values of

(3,

For estimating& mismatch ambiguity

I.
'

the magnitude of the uncertainty of

to

ant

An upper bound for the mismatch ambiguity is obtained using

the maximum value

for

|r
1

tainty1 of

ant

amp

consistent with uncer-

ant

The values of

and

for the various mismatch uncertainty

figures is listed in Table BI

49

Key to mismatch uncertainty figures

Table B-I.

F igure

Bl
B2
B3
B4
B5
B6
B7

.1
.1

.2
.3
.3

1
.

As an example of how to use these figures,

amplifier where

and

0.2

For this amplifier if

that figure B4 is appropriate.

<
Ir^J
std
I

0.01,

,.

ant

|r

'

amp

'

amp

< 0.05, and

I
1

|r

'

< 0.15 and

I
1

I
'

T^
std
,

From Table B-I we see

0.2.

consider an

'

ant

< 0.1, then

'

J <

ant

Thus using6

0.11.

figure B4 the maximum mismatch error caused by using

dard with
is

about

reflection coefficient different from the "antenna"

For this same example,

7%.

stan-

match ambiguity (because

estimate the mis-

to

can have any amplitude a phase

*

restricted onlv by
and

|r

^.j
'std
6%.

ant

'

'

ant

< 0.1.

B.2.

Mismatch error
Fjti

is

'

< 0.1) we use

r
'

ant

u.mp

'

< 0.15

The mismatch ambiguity

is about
'
&

Mismatch Error

the error in measuring& either T

or

caused when the reflection coefficients of the "antenna",

the hot standard, and the cold standard differ from each

other.

Part of the mismatch error, namely the error caused

because the "antenna" reflection coefficient differs from

50

the reflection coefficient of the two standards,

is

already

given in the mismatch uncertainty figures, and this contribution is NOT included in the mismatch error tables.

In

other

words, the tables only include the error because the impedance
of the hot and cold standards differ from each other.

of the error depends

magnitude
on the magnitudes
of
f
&
b
T

.,,
cold

3i

and bT /T
a

T,

hot'

A keyJ to the 19 mismatch error tables

provided in Table B-II.

is

The

are decibel errors to F, R

The errors listed in the tables

The meaning of the symbols used in

the mismatch error tables compared with the symbols used in

Eqs.

(1),

(3),

and

are as

(4)

follows:

Meaning

Symbo Is
T(HOT)

T(COLD)

BETA

BTA/TE

bT /T

hot
1 A
cold

ERR

r
1

ANT

1
1

F(DB)

a
r

hot

r'

ant

cold'

1
'
'

*
r

ant

amp

dB

As an example of how to use these tables, consider an am-

3=0.2

plifier where
10,000 K and

and b

0.2, and where a hot standard of

cold standard of 300 K will be used for the meas-

urement of noise figure.

Then from Table B-II (page 62) we note

that Table B-VI is appropriate for use with these given conditions
If the amplifier has a noise figure near

51

dB,

and

I\

-.,

and

0.05 and

<

amp

< 0.15

ant

so that

"ERR**

<

I\

>

hot

-r

,1

cold

0.1

"ANT"

r
1

ant
<_

0.2, the maximum mismatch error is

amp

0.281 dB.

Derivations

The output power, P

standard with temperature

T
r
is

attached to it

is

out o a linear amplifier when a

and reflection coefficient

std

is

M std

kGBM ,(T ^,+T J)

e
std
std

out,

^
std

where

the appropriate bandwidth

is

and M

'*

rB.ll
lu,iJ

the appropriate gain,

=

-,

std

[3]

where k is Boltzmann's constant,

G

the mismatch factor

is

defined in eq.
(5) except
n*.;
r

^j
std

with the subscript "ant" replaced by "std").

Using eq.

(4)

except with the subscript "ant" replaced by "std":

T
T

(std)
'
^

(1+bl
y

'

r
1

^
std
,

)
'

'

(B.2)
2

std

then
P

out

/(GBk)
c

T ,

std

-*

where Re

bT [|3|
a

a
2

[T

^J
std
1

'

-2Re(3*r^

(T
v

,-bT J)
std
a

(B.3)

td )]

implies the real part of the vector in parenthesis,

and the asterisk indicates the complex conjugate.

std

then using eq.

(B.3)
P

/CGBk)
v

out

std

(B.4)

(T ,,+T -AT _)
std
e
std^

(B.5)

ant

and eq.
=

If we let

(B.4)

ant

where M.
an

'

are the mismatch factor and effective

and T (ant)
e

"

input noise temperature when the "antenna" is connected to the

ampl if ier

^AT

ant

std

,(l-bT /T ,) [|e\ J
std^
a
std^
std
L

'

If G and B do not

2bT

[Re(,3V

change
as
6

hot+

td )].

(B .6)

changes,
then
6

ant

T,

v
Y

V*

+2Re(r'
v
ant std^)]

'

AT,

AT

hot

CB.7)

+ T

cold

cold

where the subscript "std" is changed to "hot" or "cold" as

Solving for T

appropriate.

(T,

hot,

-AT,

Y(T
y

aotJJ

,-AT

cold

(B.8)

For the computer calculations, eq.

so

,)
cold^

(B.6)

is

modified

that
ReCl

and
Re

std) "

&e'

std )

For mismatch error we assume

(B.8)

std

|e,

ro i ^

anc

'*"
*

~"~

computed for the eight sign combinations of the para-

is

meters

ant std'

ant

'

c,

hot^

cold
-,

and 6,

from the value of T

with

difference
and the greatest
"

e,

not

,
cold
,

used as the mis-

is

match error.
For mismatch uncertainty

eqs.

(B.2)

and

e,

hot

(B.4)

5 3

cold

so

that using&

ee

,,>
(std)

[1+blr
L

'

.-3|

ant
t

+b|

'

'

' J
ant.-S))l
J

2Re(r_c')
ant

ant

+2bRe(e'*(r
v
^
H5

'

CB.9)

Using eq.

(B.2)

and eq.

(B.9),
i

L+Lblr

(ant)-T
(std)J
^

'

ant

-3

+b

'

2
'

+2bRe

[e
L

(r
*-

'

'

d-|r ant

ant

-3)

)(i-L)

(B.10)

where
!

2ReCr;; e')
t

Equation (B.10)

is

ant

is

'

the mismatch uncertainty and unlike the

(B.8), the mismatch uncer

mismatch error calculated from eq.

tainty1

independent
of the values for
v

the computer calculation, eq.

(B.10)

Re(T a n .e") +

ant"
t"

is

TV

hot

and T

-,

cold

For

modified so that

_e
an t

and

Re[e'*(r'
v

ant

Then eq.

(B.10)

is

-6)]
J }

e (T
v

ant

-3)J

computed for the four different sign combina

tions of the parameters

r

ant^

c',

and

value is used as the mismatch uncertainty.

54

3',

and the greatest

|r

Fig.

Bl

ant

amp

Mismatch Uncertainty
Explanation on p. 49
55

i
UJ

O
o
I
<:
00

x
<

amp

ant

Fig.

B2.

Mismatch Uncertainty.
Explanation on p 49
.

56

0.1

.2

0.1

10

/////// /

g -

/ /

8 ~

<=>'

1 1

1 Y

1 1 1
1 1 1

F>

1 Y

cc
UJ

/
/

//

JSP

/
/
/
/
/

//

//

>

r-S

//

//

//

//

//

/ //
/
///////
//////

/ / /
****'//

?c71

/////// & // //
//////
11/ / / / /
&
/
/// / / / ^ /
// / / / /
y&
/
&
/
/
/ /
/
// / / X
S
&'

5
=>
jz

'

y/^^^^ ^^^^

/
1

r
l

0.2

!'

0.3

r
di!l

B3

Fig.

'

std

Mismatch Uncertainty.
13

xp Ian at ion
57

on

p.

49.

*^

0.2

0.1
r

ant
;
l

ig.

B4

amp

an at on
i

on

0.2

0.2

0.3

Mismatch Uncertainty
lixpl

49

i
q;
LlJ

<_>

|r

Fig.

B5

ant

amp

Mismatch Uncertainty
Explanation on p. 49
59

0.2

0.3
0.2

0.2

0.1
r

Fig.

B6.

ant

0.3

amp

Mismatch Uncertainty
Explanation on p 49
.

60

-III

10u

~r

'

/ /

/ /

/ / / &
/ /

0.3

&
i?

//

//

'

//

//

//

//

**

"7

/1

//

/1 /?
/ /

/
/

/
/ / / </ /
I / / f
/ /
/ / // /'& //
/ /
/
/
/
/ jP'&
/
/ / /
/

^^^

^^o

-^

0.2
ant
;
I

ij4.

B7.

0.3

amp

Mismatch Uncertainty,
lixplanat on on
i

61

49

Key to mismatch error tables

Table B-II.

Table

hot

cold

bT /T

III

10,000

300

IV

10,000

300

.1

10,000

300

.1

VI

10,000

300

VII

10,000

300

.2

VIII

10,000

300

IX

10,000

300

.3

373

80

XI

373

80

.1

XII

373

80

.1

XIII

373

80

.2

XIV

373

80

.2

XV

373

80

.3

XVI

373

80

.3

XVII

1,270

300

XVIII

1,270

300

.2

XIX

1,270

80

XX

1,270

80

.2

XXI

300

XXII

300

.2

62

.2
1

.2

3
1

.2
1

.2
1

.2

Table B-III
MAXIMUM MISMATCH

EP.ROF.

CDB)

TCHOT)= 100312 K
T(CCLD)= 333 X
3ETA= 3
STA/TE=

--GAMMA --

-FC

ERP.

AMT

,335

.32

.05

.31

,3

,332
,3 34

34

.338

.3 17

,338
,316

.35

.033

,33

.333

,331

,301

333

,333

33
.333
1

3 37

.3 36

.212
.323

.33 5
.311
.321

,333
.333

3 33

J03
536

.333
336

.302
.335

315
32
36

,313
327
,052

.312
.324

.311
,323
.34 3

.311

,331

.031

.331

307
,016

,336
,314

.336
.313

,031
.336

,312

.332
.306
.312

.2

.35

327
,355
134

.025

.023
.346
.387

.322
.044
.333

.339
.02
,336

,3

33

.3 17

,335
,066

.32
.05

.007
.017

.2

.35

.02
.35
.

.2

.35

.04 6

.334

,331

,364
121

,0'3 2

.034
,319
,342

.337

,333

,3 19

,02
.337

336
175
334

.33
.161

,372

,336

,265

3'39

.316
.346
.392

,3 19

.044

.32
.35

.37
.
33

.35

.335
.31

,32

.331

.333

.314
,326

.32

.32

332

34 2
,392
,

,177
.355
,673

.339

141

169
.33
.62
.

63

326
352

.091
,

157

297
545

.35

,094

,367
29

.322
.341

,32

,335

363

,361

,121

,24

,224

,116
.214

.333
,056
,391

.3

37
.353

.339
.059

,39

.09

15

.275
.498

145
.261

142
.252

.463

.4 5

Table B-IV
MAXIMUM MISMATCH ERROR CDB)
TCHOT)= 13302 K
T(COLD)= 333 K
BETA=
BTA/TE= .2
.

--GAMMA' -ERR
AMT

.335

.001

.33

.001

.301

.332
.334

.332
.333

.002
.003

.332
.333

.338
.315
.029

307
.313
.024

.306

.036

.306

.3 11

.31

.31

.35

.003
.017
.333

.321

.019

.02

.331
.304
.033

.032

.332
.004
.007

.302
304
.307
.

.333
.035
.037

.315
.331
.053

.313
.026
.348

.312
.0 23

.311
.321

.312
.022

.042

.037

.34

.302
.003
.017

.304
.303
.015

035
309

.01

.3 15

.036
.009
.315

.33

.027
.353
.398

.025
.346
.385

.023

.303
.323
.344

.0 14

.326
.343

.38
. 159

.373

.32
.35

034
.009
.

.2

.317
.034

.35

.065

.031

.32
.05

.338
.017

.2

.334
.369

.35

.131

.32
.35

.321

.335

.05

13

.332
.034

.2

.332
.034

.02

.32
.05

.31

F(DB)1

.345
.087

.062
.117

174
3 29

.296

.05

.315
.047
.096

.326
.355
.093

179

.172
.327
.634

.2

.35

.02

.2

.35

.353
.664

64

.3 34

.307

137
.25
.

.343
.367
.

132

162

.239
.517

.037
.015

.34 2

.324
.344

.376

.331

JI8
.023
.343

.021
.33
.043

.324
.332
.043

.063

.365
.112

.367
.118
.239

22

.219

.355
.075
35

156

.265
.46

197

.364
.381
.138
. 152
.243
.419

.072
.337
.1

.158
.261

.445

Table B-V
MAXIMUM MISMATCH EHROn (DQ)
TC!!OT)=

13300

T(C0LD)= 330 K
3ETA=
3TA/TE=
.

FCD5)-

GAMMA'
ERR
AMT

.005

.333
.304
.005

.334
.005
.037

.335
.335
.303

.036
.337
.308

.333
.014
.324

.31

.311
.319
.034

.012

.317

.33

.333
.314
.025

.33 2

334
.336

.307
.333
.311

.31
.3 14

.311
.311

.015

.313
.314
.316

.316
.323
.348

.32
.33 5
.36

.323
.039
.367

.342
.072

.319
.321

.323
.323
.332

.323
.329
.334

.332
.005

.01

.338
.316

.32
.05

.335
.039

.3 17

.2
.35

.333
.361

.32
.32
.05
.

.2
.35

.32
.35
.

.2

.35

.32
.35
.

.2
.35

.31

.316
.323
.351

.03

.039

.321
.336

It

/,

.005

.0 33

.3

.313
.02

,314
.317
.323

.332
.357
. 102

.334
.357
.398

.042
.37
. 122

.347
.079
. 137

.351
.335
. 146

.326
.037
.355

.343
.35
.366

.357
.362
.032

.069
.369
.092

.082
.384
.093

.038
163
.339

.334

.392
.253

.115
.136
.315

.209
.354

.224
.379

.337
.367

.367
.39
. 125

.113
126

174
75
.22

.203
.233
.235

.295
.457

.316
.469
.302

.319

.334
.366
.

.3

.05

13

.032
.303
.305

.001

.32
.35
.

.2
.35

22

.329
.

05

.111

147
.26
.

15

.
.

157

.345

.3 11

.211
.327

.6 23

533

.536

136

185

65

.329

.146
.

156
196

.263
.437
.667

129

.
.

.75

133

Table B-VI

MAXIMUM MISMATCH

EP.F.OP:

(D3)

T(H0T)= 13000 K
T(C0LD)= 333 K
3ETA= .2
3TA/TE= .2

--GAMMA* -ANT

FCD3)-

ERR

.005

13

.031

.331

.002
.004

.002
.304

.332
.033
.004

.039
.017
.033

.003
.016

.007
.012
.322

.007

.03

.007
.014
.025

.30
.304

.332
.334
.039

.3 33

04

.335
.039

.005
.339

.304
.005
.009

.005
.337
.339

.316
.032

.315
.327

.3 24

.0 59

.35

.344

.313
.023
.34

-.023

.35

.317
.035
.066

.032
.333
.018

.034
.309
.318

.306
.311
.018

.3 08

.31

.32
.05

.312
.318

.013
.319

.035

.03

.328

.055

.35

.027
.346

32

.033
.364
.119

.038

.33

.007
.023

.312
.326

.02
.332

.026
.336

.331
.34

.0/17

.34 8

24 9

.351

.053

.037
.344
.256

.335

.379

.376

143

.131

.3

.256

.223

.375
123
.209

.376
. 126
.213

.033
.362

37

.334

.39'

.112

32
123

.2

.35

.31

.02
.35
.

.02

.009

.2

.35

.02
.05
.

37
1

.339

.35

176
.331

.0 18

.32
.05

.05

.2

032
.334

.02
.35

.35

.35

133
357
.663

162

136

.055
. 379
.114

179

.334
.612

66

174

.014

.39
.121

.172

.332

.23

.531

.478

.03 2

.303
.304

.011
.32

73

.269
.442

.003
.003
.335

.307
.312
.321

.013

.342

.312
.315
.02

.327
.34 3
.384

136

175

.277
.463

Table B-VII
TAX I MUM

MISMATCH

EP.P.OP.

CD3)

T(H0T)= 13300 K
T(C0LD)= 330 K
3ETA= .2

3TA/TE=

GAMMA'

FCD3J4

AT IT

.005

332

.034

.303
.036

,33 5

336
007

033
.309

.31
.31

.313
.313

,337

3 33

.31

.311

.0 13

.32
.35

13

339

311

,016
,027

316
027

.313
.32
.334

.315
.023
.033

.016
.325

.004
.037
,311

.337

.317
.017
.32

.321
,021

.313

313
314
316

.323

.326
.326
.327

,019
.335
,063

,319
.332
.355

322
333
354

,03

.032

.367

,046
,076

.05
.031

.015

.35

,308
,014
.023

326
029
334

.335
.036
.042

.043
.344
.047

.353
.354
.356

344
363

.355
.334
136

.062
.394
153

.366
.101
-164

.097

.119

12

.145
147

123

.151

143

.35

.039
,017
.032

.32
.35

. 1

.2

.31

.2

.35

.02

.32

.2
.35

.05

.35

.02
.35
.

.31

.02

.027

333

.339

,07
27

.3 64

,024

.343
.355
,372

338
36

.32
.05

,098
,173
,32

32
166
1

,23

073
38

092
1

19

327

.34

14

233

.22
.352

177

,356
,036

102
126

173
137

.226
.233

13

161

21

.261

264
331

.323
.474

596

.74

235

.2

,365

.35

,65

,348
,573

67

166
,247
,396
,

.34

173
.265
.423
.

.275
.276
.293

.328
.333

.369
.533
.833

.395

,341

,57

391

Table B-VIII

MAXIMu.i MISMATCH ERROR CDB)

TCHOT)= 13330 K
T(COLD)= 333 X
BETA= .
BTA/TE= .2

--GAMMA' -ERR
ANT

.035
.32
.05

.002
.303
.005

.033
.004
.005

.303
.334
.305

.024
.035
.036

.038
.312
.321

.038
.012
.322

.338
.316

.038

.03

.026

.033
.013
.323

.02
.35

.031
.305
.339

.332
.335

.004
.006

.335
.337

.336
.333

.01

.31

.01

.01

.311

.012

.017
.032

.35

.013
.035
.066

-016
.029
.051

.315
.026
.046

.315
.025
.042

.316
.325
.043

.033
.009
.319

.005
.019

.003
.013
.02

.011
.315
.021

.3 14

.02
.05

.317
.323

.017
.02
.325

.036

.334
.065

.032
.358

.331
.353

.032
.351

.051

103

.392

.385

.338

.03
.351

.026
.033
.055

.034
.044
.059

.041
.05

.063

.349
.357
.069

.033

.035

.035

.035
133
.22

.088
.134
.227

134
122
143

.122
.137

.2

.02

.2

.35

.33

.033
.005

10

.039
.013
.033

.01

.031
.332
.035

.2
.35

.05

FCD3>4

.37
. 133

.02
.05

.003
.025
.049

'3

39
173

.36

.011

12

.015

166

.2

.35

.333

.304

.32
.35

.022
.355
. 104

.34
.069
.113

186
.361
.67 3

.342
.62

.2
.35

136

68

.3 14

149

.263

3 6L

139

.237

.037

.39
. 126

137
137

187

189

.315
.544

.293
.496

.
.
.

.033

.033

161

193
.29

.2

.465

.481

.293

Table B-IX
MAXIMUM MISMATCH ERROR (DB)
TCKOT)= 13033 X
T(COLD)= 333 K
3ETA= .3
BTA/TE=

FCDS)-

--GAM' 1A'--

ERR

ANT

.335

.02
.35
.

.01

.32
.329
.345

.325
.026
.027

.331
.331
.331

.333
.339
.34

.334
.348
.075

.333
.354
.384

.353
.39

.363
.364
364

.378
.379
.331

.377

.313,

.013
.319
.33

.017

.019
.02

.01

.311
.018

.31 9

.32

.036
.339
.313

.311

.313
.317

.022

.021
.037
.065

.323
.335
.059

.039
.36

.312
.013
.327

.322
.027
.334

.338
.34 1
.045

.051

.042
.074
.131

.046
.07 2

.355
.078

.113

122

.368
.098
.151

.334
.117
.
82

.033
.343

.361

133

137

17

.207

37 2

141

17

.37

.39

.32
.35
.1

.2
.35

.02
.05

.171

.21
.2 14

.23 3
.236
.437

.335
.468

.221

135
149

. 133
.161
. 196

.233
.247
.263

.3 14

.376
.377

.4 53

.326

.33

.466

.225
.33 5
.672

.256
.334
.619

.317
.436
.655

.394
.542

.44 3
.63 9
.916

.43 2
.651
.93

.-^5

.35

.147
.131

.376

.121

.254
.339

.02

.052
.355

.341

145

.2

.3 27

.234
.313

.35

.329

.3 24

34

133
188
.331

.2

.037

.319
.327
.34 2

.311

.039

.2
.35

.019

.312
.313

.33 5
.337
.003

.018
.033

.35

.315
.315
.215

.31

.033
.034
.037

13

.2
.35

.32
.35

.32

12

69

.307

.8 15

.453

Table B-X
MAXIMUM MISMATCH ERROR CDE)
T(HOT)= 373 K
T(COLD)= 30 K
BETA= .2
BTA/TE= .2

GAMMA' -ERR
ANT

.005
.02
.05
.

.2

.35

.01

.02
.05
.

.2
.35

.02

.02
.05
.

.2

.35

.05

.2

.35

.001
.001

.002
.002

.002

.0 04

.004
.008
.315

.035
.009
.316

.00 2

.004
.335
.037

.003
.005

.32
.35
.

.005
.035
.036

10

.01
.01
.01

.013
.018
.019

.03

.3 08

.311

.021

.011

.312
.017

.024
.031

.04
.05
.37

.02
.02
.021

.036
.337
.038

.065

.022
.325

.34

.379

.017

.01

.311

.313

.032
.035

.361

.07

.009
.016
.029

.311

.019
.032

.316
.022
.334

.333
.006

.008
.011
.015

.32
.023
.026

.342

.022
337
.065

.332
.345
.369

.345
.05
.07

.384
.098
125

.231
.28

.022
.029

.055

.107

.317
.3 27

.04

.361
.37

.112

187
193
.23

.314
.332
.359

.045

.353
.097

.216

.436

32

.251
.32

.511
.739

.233
237
.243

.396
.437
.424

.652

.254
232
.335

.454
.524
.665

.31

.017
.032
.059

.01

.32
.35

FCDE)4

.03 2
.143

.325
.339
.06

.09"

.2

.35

.303

17

166

.354
.363
.039
126
.233
.343
.

70

.03 5
.113
. 178

126
133
56

136

.
.

.252
.374

3b

.34
.04

139

.117
.131
.

.043
.362

.072
.374
.073

.
.

14

123

13

.141
.

16

.63 3

.742

36

.346
.444

Table B-XI
MAXIMUM .MISMATCH

EP.P.OP.

CD3)

T(HOT)= 373 X
T(C0LD)= 83 X
BETA=
3TA/TE= .2
.

FCDB)-

--GAr- !MA'-ER?.
AMT

.035

.02

.35

.032

10

.332
.003

.302
.333
.004

.005
.035
.305

.039
.039
.01

.017
.32

.304
.338
.315

.33 5
.339
.314

.006
.339
.316

.312
.015
.32

.324
.334
.352

.313
.319

.331

.321

.34

."3

15

.2

.334
.008

.35

.301

.32
.05

.332
.334

.032
.333
.00 5

.005
.036
.333

.31
.31
.311

.039
.317
.33

.011
.317
.329

.012

.024

.3 18

.33

.35

.003
.015
.328

.333

.342

.049
.369
. 105

.032
.335
.309

.034
.007

.011
.313
.317

021

.32
.35

.337
.339
.042

.062
.069
.38

.016

.023
.035
.353

.025
.037
.067

.04 3
361

.35

.319
.034
.061

.399

.331
.357

.02
.05

.336
.313
.324

.013
.321
.331

.332
.337
.046

.058
.06

.063

.113

. 163
.181
.238

.349
.038

.35

.341
.37 3
. 144

.061
.093
.151

127
159
.22

.253
.353
.536

.313
.332
.053

.037
.351
.372

.35

.02
.35
.

.338

.2

.32

.2

.35

12

031

.301

.01

.2

.35

31 A

162
.295

.01

156

.379
.39
. 108

133
185
.324

133

.20 2
.32

71

.322
.023

.363
.093
.

174

.136

.385

.39 9
134

.
.

.334

138
.211
.

14

.22
.231

146

.243

156

.276

.217
.371

.341

.531
.73

.465

.385
.439

.399

Table B-XII
MAXIMUM MISMATCH ERROR (DB)
TCHOT)= 37 3 K
TCCOLD)= 80 K
BETA= .
BTA/TE=

FCDB)-

GAMMA'
ERR
ANT

10

.034
.034
.034

.012
.313
.315

.3 24

.344

.376

.323
.034

.35 2

.09
.111

.034
.336

.319
.026

.35

.01

.305
.338
.015

.043

.2

.04

.131

.334
.035
.306

.0 39

.02

.024
.027

.049
.35 6

.031

.363

333

.337

.02
.05
.

.01

.35
.

.338
.312
.02

.2

.35

.003

.02
.02
.05
.

.01

.312
.016
.325

.2
.35

.041

.05
.02
.35
.

.2

.35

.032
.002
.003

.035

.02
.05

.31

.317

.33

.353
.33

.319
.319
.319

.35
.055
.363

.021

.377

.035
.06

.J 51

23

.233

.099
.114

.355

.119
.119

.549
.623
.735
.927
.34

13

.426

.095

139

.219

577

.2

.329

.857

72

179

.306

.21

.361

.776
.915
. 126

.414

.294
.447
.726

.459
.535

.447
.653
.329

12

166
.251
.434

.589
.396
.455

.258
.295
.351

18

.222

.334
.503
.839

132
145
165

.374

152

175

.257
.437

.29
.316
.357

134
127

107
162

.2
.27 5

.3 63

.446

.357
.392
. 155

.36

147

.223

.256

.043
.365
.134

.389

137

.352
.352

.033
125
.202

.023
.327
.033

.35

.039
.039

.0 64

.363

2.396

.65

.847
.268

.434

2,3^6

2.253
3.65

.954

136

.592
.37

.337

2.291

.73

3.308
4.547
7.345

2.574
4.113

Table B-XIII

MAXIMUM MISMATCH ERROR <DE>

TCHOT)= 373 X
TCCOLD)= 30 X
3ETA= .2
BTA/TE= .2

FCDE)-

--GA-: 'MA'--

ERR

AMT

.005

.32
.05
.

.2
.35

.31

.32
.35

.31
.01
.31

.018
.018
.019

.03

.304
.338
.015

.035
.309
.316

.333
.011
.017

.311

.312
.317

.021
.024
.031

.34
.35
.37

.33 2
.033
.305

.3 34

.3

.335
.337

.3 13

.32
.02
.021

.336
.337
.333

.065
.37

.316
.022
.334

.022
.325
.335

.34

.34
.041

.342

.372
.374
.078

.045
.05

.384
.098

.37

.311

.31

.015

.32
.023
.026

.2
.35

.017
.032
.059

.3 22
.337

.03 2
.045

.3 65

"",9

.31

.32
.35

.317
.327

.022
.329

;2
.35

.335
.335
.036

.308

.32
.05

.302
.302
.334

.333
.006

.32
.35
.

10

.301
.031
.032

.311
.319

.2
.35

.05

.039
.316
.029

.02

.045
.082
.

IAS

.325
.339
.36

.395

.2

.35

.333

17

.332

.011

.355

137
139

.361

.379

.043
.362

.332
.035

125

137
193

14

.123
.

13

.141
.

16

.231
.23

.314
.332
.359

.04

.361
.37

.112

.23

.353
.397

.035
.113

.117
.131

.216

.436
.511

166

354
.363
.339
.

126

.233
.343

73

173

132

.251
32

126
33
156

.233
237
243

.396
.437
.424

.652
.633
.742

136

.254
.23 2
.335

.454
.524
.665

.336
.346
.444

.
.

.252
.374

.739

Table B-XIV

MAXIMUM MISMATCH ERROR CDB)

T(HOT)= 373 K
T(COLD>= 33 K
BETA= .2
BTA/TE=

--GAMMA" -ERR
AMT

.035

.068
.039

.317

.05 2

.388
.396

.053

108

128

171

129

.215

.374

136
173

.255
.342

.447

.258

.5 34

.899

. 196
.211

.355
.336
.432

.639
.664
.749

.273
.358
.518

.512
.686

.395
1.212

1.3

.50

.9

.3 98

.31

.355

.135
.116

.012
.015
.024

.013
.32
.333

.362
.373

.315
.317
.319

.336
.336
.036

.097

.324
.032
.349

.337
.341
.067

.34

.32
.05

.345

.295
.395
.395

.25
.264
.234

.35
1

.35

.02
.35
.

.2

.35

.351

.2

.362
.334

.35

.02
.05

125

39

.399
.111

132

177

.35

.261

.396
.
.

109
174

.205
.23 6
.237

.239
.234
.365

74

137

102
.111
.

125
156

.216

.32
.398
.549

527
.554
594
.

.666
.824
1

.223
.332
.449

.303

.348
.351

.003
.333

.252

152
.165
.187

'.

177
192

01 8

.313
.313

.02

.2

.039
.054

336
.038
.012

.35

.331

.31

.35

.32

.30 9

13

.249

.35

.2

.324
.025
.327

.009
.039
.339

.32

.01

F<DB>2

.334
.334
.035

.234

.533
.594

.692
.935
.335
1

.234

103
.221

1.416
.345
2.64 6
1

.
.

.976
.092
1

.331

1
1

63 j

.3

.533
.672
.384

.292
.727

2.24 9
3.341

2.535

4.512

.836
.933

3.135
3.334
3.333

1
1

2.22
2.62
3.492
5.117

4.537
6.

23

9.369

>

Table B-XV
JAX

MISMATCH

<C3)

i?.?.OS

tchot>= 373
tccold)= 33

:<
:-:

3ETA= .3
3TA/TE= .2

3A" MA' --

FCDE)-

ANT

.335

.331

.02
.35

.R32
333

.2

.35

.31

.32
.35
.

.02

.315
.315
.315

.323

.316
.317

.33
333
.

.33"'

.335
.33 3
.315

.336

.31

.31
.3 17

.314
.32

.332
.334
.336

.3 35

.3 15

.3.37

.316

.339

.3 17

.3-39

.313
.32
334

.32
. 327
. 34

.311

.33

.335
.333
.312

.3 14

.33
.33 2

.313

.336

.326

.2
.35

.319
.33 A
.36

.341

.342
.355

.359

.3 14

.331

.32
.35

.321
.331

.338

.2

.34 9
.33 6

.35

153

.34

.367
.13 6
176

75

74

13d

34

.191

.113

.139

66
.131
.215

.34 3
.421

14 2

.235

174

35

.423

.33

.335

.573
534

34

.352
.33 3
.451

134
165

.334
.

.239

16

32

.275
.23

155

.157

.39
395

.379
.335
.394

.2 34
.331

.35

.14 3

.222
.363

.35 9

.333

133
177
.311
.

36
.36
.362

.355
J66

.357
.352

36

35

.119

.2

34

.346
.347

.353
354
.356
.

.332
.335
.

.2 33

329

34

37

37

.33
.33

27

364

.37

32

.333
.347
.367

.035

.327

.^2
.05

13

.337
.333
.339

.35

.333
.333
.334

.3 17

.2

.32
.35

.35

531

.2 32

.221

264
.349
.

.466
.4 34
.5 11

.559
.663
.33

.95 5
.991
1

.346

.532
.73 3

1.142
1.361

.866

1..733

Table B-XVI
IAXIMUM MISMATCH

(33)

E?.r.O?-

TCHOT)= 373
T(COLD)= 93 K

'<

BETA= .3
3 T A/TE=

FCDB)-

--GAM' 1A'-A JT
?

.3-35

.32
.35

.313

.35

.914

.311

.32

.312
.113

.37 3

.377
.332

132

.227

14

152

24 1
.2 63

172
.217
.332

.331

.367

.39 2
.114
. 156

.326
.326
.326

.37 2
.374

379

146
153
165

.265
.23
.334

.455
.433
.525

.3 27

.336
32

135

.229
.313

.344
.434
.635

.6

.35

.315
.32
.323

.323

.32
.35

.3 24

.353
.353

.327

.0 54

153

.293
.333
.331

.532
.562
.639

.911
.967
.353

.331
.34

.354
.356

173

.37

.2

.35

.3 57

.37

.69
.37
1.211

.36
.364
.37

.32
.35

.35

.33
. 133
. 146

.32

.2

.33 6
337
.33 9

.3 14
.3 13

.213
.313
.313

.333

.2

.35

.336
.336
.337

.31

.35

.02

.2

.31

23
136

35

.143

.2

.35

169

.215
.333

.023
.337

343
.351

134

144

15

.236
.27

.459
.623
.743
.73
.337

.34
.417
1.534

.3

.535

.763
I

.37

.23

.527
2. 145
1

2.29
2.43

.369
.332

133
133
139

14

145

.439
.522
.634

.936

.192

153
1.531

3.34

.292
.292
.293

.764
.79

1.513
.592

2.7 17
2.37

4.613

.332

34

5.325

.
.

.296
.337
.432

76

.4 33

.9 35
1

.371

1.393

.737

.936
2.349
3.2
1

.737

2. 136

3.51
4.4
1

i.

122

2. 643

3.313
3.33
5.375

4. 393

6.356
7.7
10 .794

Table B-XVII

"lAXrtUM MISMATCH

CDB)

EI'.riOH

TCHOT)= 1273 K
TCCOLD)= 333 K
3TA= 3
3TA/TE=

FCD3)-

-GAM'-TA'-;RR

AMI

005

.32
.35

.332
.335

31

.332
.335

.332
.334

.332

.31

334

.311
.323

.321
.34

.039
.319
.336

.333

.343

.333

.308
.316
.331

.32
.05

.334

.334

.31

.31

.334
.339

.334
.33 9

.3 17

.332
.334
.338
.316
.33

333
.333

.0 33

.316
.332
.36

.332
.338
.317

.338

.322

32

.3 13

.2

.34

.342

.337

.317
.33 5

.35

.336

.2?

.37

.366

.316
.333
.362

.33

.331

.20

.32
.35

.339
.322

.339

.333
.313

.333
.317

.335

.033
.366

.032

126

.121

.303
325

.339
325

.325

.34

34 3

.34 7

.33
.038
.319
1

.321

.344
.39

.384

.337
.375

1^3

.161

.32
.35

.332
324
.356

.304
.324
.354

.337
.324
.351

.111

.13 5
.213
.4 34

.397
.193
.363

.062
.116

.2
.35

34 f

.35

.227
.433

. ?

.32
.35

.352
.117

.316
.356
.116

.2

.332
.335

35

13

.35

32

39

.227

.219

.2

.4 59

.35

.87

.43-1
.3 17

143

.3 25

.236
.399
.741

11

.37
. 133

39

13

.337

.33 3
. 172
.321

.3 64

.31

336

166

.31

365
.115

.037
.063
.115

.23

.199
.376
.693

.194
.362
.662

.191

.3 33

.269
.114

.353
.643

Table B-XVIII

MAXIM'iv MISMATCH ERF.O

CDB)

T(HOT)= 127 3 :<

T<COLD)= 333 K
-3ETA=

.2

'

3TA/TE=

iA'TMAV* -ETIR
A'.'T

.335

FCD3)-

32

.35

.322

.333
.335
.333

.336
.333

.313
.314

.31

.3 15

.3 23
.3 24

.355
.359
.365

.345
.357
.379

.399
142

.313

.34

.335

.316
.319
.327

.329
.037

.313

.34 5

.37

.046
.343

.375
.081

337

.32

.2125

.31

.3 16

.35

.316

.32

327
.023
.329

.2

325
.345

.35

.331

.327
342
.069

.332
.333
354
.

.313

.3 27

.32
.35

.321
.33 2

.332
.34

.355
.357
.36

.355
.335

.365
.37 6

.2

.35

.32
.35
.

.2

.35

n
.

.336
.35 5

.332
.

35

14

.37

.33
.118
174
.

.263
.4 63

.827

145

.385

15

36

17

.14

13

.229
433

.32
.35
.

.351
.391
. 162

.217
.35

153
86
.227
.
.

.29 3
.45
.728
1

78

57

.19 9
.23 3

.316
325

.34
.3 66

.426
596
.

.351
. 359

.335
.337
.34

.313
.323

.35

.2

.'3

.92

.35

.31

r>

.37

391

.118
.

13

.114

.151
198

15S

.234

.091
.393
.397

. 143
.151
. 163

.222
.233
.262

133
.119

. 134
.23
.313

.334
.397

149

.237
.243
.252

.369
.333
.413

.567
637
663

.269
.337
384

.471

.537

.773
.3 34
.433

37 4

.53 7

.519
.537
.57
647
.

.8 55

.3 39

.774
.813
.873

.57

1.173
.253
.375

1
1

97 3

1.586

1.212
.659

2. 349
2. 92

Table B-XIX

[AMIMUM MISMATCH EP.30R CD3)

TCH0T)= 1273 X
TCCOLD)= 33 X
3ETA= 3
BTA/TE=

--GAMMA' -AMT

F(DE)

TP.?-

.335

.32

.03
.332

.332

.333
.336
.312

.334
.337
.314

35

.2

.35

.33

.032

.305

.03
.33 2

332

.335

34

.334

.335

3C 3

.3 39

.316

.317

.313

.319

31

.3

.335
.305

.332
.3 35

.332
.035

.3 37

.3 33

.31

.01

.314
.323

.317
.032

.3 13

.319
.337

.333

.33
.333
.033

.331

.33
.33 5

.3 15

.317
.334
.365

337

371

.037
.315

.317

.317

.334
.311
.021

.3 27

.2

.333
.366

.339
.276

.35

.347
.39
.
66

.352
.399
.135

.3 39

.316

.32
. 35

.351
.33 9

.33
.35

.326
.042
.366

.333
.35
.37 6

.33 7
.35 5

12

.129
.223
.435

.35

.3'3

.033
.33 6
.313
3 24

.32
.35

.(533

.336

.2

.313
.326

.35

34 9

.332

.32
.35

.031

.03

.332
.334

.2

.332
.334

.35

33

.32
.35

.32

13

3
1

3
.

.31

.35

33 3

124

.37

136
.253

.329
.35

.14

.736
.

.292

79

.334
.339

.13 7
. 193
.343

2H

.335

.31
.3 19

.3 33

33

.214
.333

.331
.33 5
.311

.32
.339
.374

.32

.3 32

.335
.311
.021
.34

.077

039
.313
.332

.31

.355

.357

.
.

134
192

. .3

32

.319
.333

.
.

136
193

.339
.058
.

33 6

134

.236
.419

Table B-XX

MAXIMUM MISMATCH

TCK0T)= 1273
TCCCLD)= 33 K
3ETA= .2
3TA/TE =

FCD3)-

AM!

.361
.33

.336

.337
.347
.067

.334
.336
.333

.359
.364

343
.353
.372

.37 4
.395
. 135

.37

.112
.119

.334
.306

.337

.311

.31

.3 15

.321
.3 26

.35

.31

.317

.327

.035
.33 5
.337

.039
.311

.32
.32
.221

.2

.339
.313

.35

.021

.313
.32
.333

.322
.333
.354

.311
.314

.319
.32

.34

.32
.35
.

.3 13

.32

.31

.35

32
.35
.

.34 3
.347
.352

.0 35

.327
.329
.332

.31

.327
.341

.345
.363

337

.3 67

137
145

.327
.33
.337

.3 54

.111

135

.356
.359

.113
.116

.347
.363

.373
. 139

.121
. 177

174

.325
.342

.363

.32
.35

.37

.35

.322

137

123
128
135

.332

132
145
.225

162
.234
.365

80

.117

.367
.394

.35 5

.373
.373

.313
.319

.342
.343

.35

.3 17

13

.01
.3
.31

.32
.35

.35

.035
.335

.32
.35

'

.332
.333
.333

')

.31

--GAM MA'-

.305

CD3)

EP.SOP.

34

122
.161
.

,234

176
189

.21

13

149
192
.27 2

.246
.324
.47

93

.29
.3 29

.235

.336

.451
.485
.537

.233

.223
.277
.373

.385
.491
.692

.626
.322
1.133

243
.252
.253

.432
.413
.444

.617
.654

.269
.331
.596

.433
.533
.733

.94.

.71

1.339
1.113

.636
.32
.426

.292
.634
2.42

1
1

Table B-XXI

MAXIMUM MISMATCH

E?.?.0?.

CDE)

TCKOT)= 300 K
T(COLD)= 4 K
BETA=
3TA/TE=

--GAMMA' -ANT

FCDB)-

EP.R

.005

.02
.05
.

.2

.35

.01

.331
.33 2
.304

.2

.35

.02

.031
.331

.301

.032

.301
.23 2

.332

.303
.037
.014

.234
.033
.315

.334
.303

.33 2

.031

.002

.3 11

.001

.032

.001
.303

.002
.3 34

.33 2
.3 24

.33 2
.304
.333

.304
.007
.314

.336
.312
.022

.037
.314
.327

.338
.316

.339

.33

.333

.30 1
.002
.004

.331

.33

.033
.037

.3 34

.0 33
.0 15

.012

.3 15

.0 24

.323

.045

.029
.355

.032
.034
.337

.004
.007
.013

.007
.012
.32

.003
.015
.025

.309

.01

.3 17

.313

.323

.33

.312
.323
.043

.321
.34

.3 34

.342

.364
.116

.07 9
. 144

.347
.333
.161

.05
.3 94
.
72

.009
.313

.3 16
.0 23

.326

.3 19

.333

.354

.03 3
.04 6
.366

.037
.352
. 374

.339
355
.03

.113
.195
.343

.121

..33

.0 04

.02
.05
.

.2

.35

.02
.05

.039
.017

373

.3 37

.003

.03

.051

.3 32

.2

.052

.0 33

.141

.101
174

.35

.091

56

.24?

.306

.3 16

.332

.05

1'

.007

.05

.35

.03

303
.336

.031

10

.0 04

.32

.2

.32
.05

.30

81

.331

.334
.339
.317
.033
.362

.334

.31 7

.302
.005
.31

.313
.335

.366

.203
.366

Table B-XXII
MAXIMUM MISMATCH ERROR <CB)

TCHOT)= 333 K
TCCOLD)= 4 K
3ETA= .2
3TA/TE=

FCD3)-

--GAM" !A'--

ERR

AMT

.335

.32
.05
.

.2

.35

.31

.32
.05
.

.2
.35

.32
.35
.-

.2

.35

.321

.334
.336
.339

.339

.329
.033
355

336
.336

.3 14
.3 15

.0 11
.0 17

.334

.391
. 103

.364
.356

.
.

.293
.32
.362

.246
.333
.494

.435
592
.383

.337
.363
.414

.537
.642
.726

.494
66 3
.991

.871

.473
1.617
323

.313
.322
.033

.353
.376

.312
.312
.012

.329

.336
.392

.131

.033

.131

.22

.315
.323
.037

.337
.345

.117

.3 67

.22

.347
.539

.332
.333
.334

.07 3
. 33

.222
.237
.261

.463
.531
.559

.354

.341
.36
. 397

.396
.118

.3 32

.6 59

174

.391
.56

.877
234

.246
.631
2.4 39

.171

.469

.953

.173

.5
54 3

32

^5

.373
. 374
.376
.39

35

.33

. 1

.037

. 1

39

153

.239
.252

.63

13

234

.3 65

1.149

82

109
129
173

.254

197

.26

.217
.296

163
133
.236

09

.093

16

.131

.342
.345
.35

146

.44

.247

127

123
166

.316

345
.349
.354

.323
.025

.33 6

.35

.05

.037
.337
.333

13

.337
.311
.319

.2

.32

.333
.333
.333

.32
.35

.93
1 .346

1.185
.768
1

139

2.

2. 576

.744

4.434

2.996
3.273
3.694

335
1.151

.3 97
2. 129

1.351
1.739

2.529
3.4

4 . 4 S
5. 992

2.635

5.319

3.912

Appendix

C.

Effect of Front End Loss on Noise Figure and

Effective Input Noise Temperature

If the reference plane used to define the input of the

amplifier

is

changed, then the noise figure, F, R

and the

effective input noise temperature, T

changes if there is any

loss between these reference planes.

These changes in ampli-

fier noise are calculated in this appendix.

One application

of the results of this appendix is to estimate the uncer-

tainty in amplifier noise because of the uncertainties in

the connector loss.

Using figure C-l, the effective input noise temperature

at

reference plane
r

is T

The two-port between

and

and at reference plane

is

has the physical temperature T

and an absorption loss of A so that the thermal radiation from

the two-port expressed relative to port

is A T

effective input noise temperature at reference plane

is

The

[8].
2,

the amplifier noise and the two-port noise expressed rela-

tive to port 2, namely

T*

(A
T
v

conn

+T )/(l-A)
J
K
J
e

(C.l)
y
J

or
T*

=
e

-
i

(T

conn

+T ).
e

(C.2)

Expressing the absorption loss in decibels,

A

dB

-10 log

(l-A)
1()

83

(C3)

o
o

84

then

A
T

For small losses

=
e

(i.e.,
T

A-,

-1) fT
*

<<

To express the difference

2

from plane 1, recall that

(T

dFi R

/T

(10 log,

and where T

(C

+T

F-,

A, D

)
J

dB

CC.5)

figure at reference
10 log,. F, where

Because

290 K.

=
Q

"

ah

dB

4.343

fj

dB

(T
*

-T J)/(T
+T
v
o

conn

(C.6)

e J)

e
v

dB

T
*

(C7)

dB'

If the tivo-port's physical

temperature,
r J
v
'

(C4)

eJ

5)

equal
T
n
o

+T

e)dF/F, then for small losses

?ah
dB
or using eq

conn

in noise

plane
F

conn

10)

0.2303 (T
*

/10

(10

conn

is

nearlyJ

290 K, then
F

dB

dB

Thus an amplifier which includes

absorption loss at 290

two-port with 0.1 dB

(16.8 C or 62.3 F)

0.1 dB greater than the same

(C.8)

dB'

has

noise figure

amplitude excluding this two-port

85

Appendix

D.

Computer Programs

This appendix contains the computer programs used for the

figures and tables in this paper.

language

86

The programs are in BASIC

Figures

D.l.

For figures 1-4,

Through

the program DW11

in table D-I was used.

A typical output of this program is shown in table D-II.

The basic relationships used for this table are as

using eq.

of reference

(1)

follows

and changing the notation to

correspond with the remark statements in Table D-I, and adding

the gain function G

[cf step 4400

Y

or rearranging

in Table D-I]

G*(TH+TE)/(TC+TE)

[cf steps 4300,

(D.l)

4440, 4450, 4480, 4490, 4520,

4530, 4560, 4570]

(G*TH-Y*TC)/(Y-G)

TE =

This eq.

(D.2)

from TH

TC

is

used to calculate the error contributions

Equation

Y and G.

4340 and 4330)

is

and the derivative of eq.

to errors

(2)

of reference

(cf steps

used to convert from F(DB) to TE

F(DB)

F(DB)

(D.2)

in TE

10*CLG(l+TE/290)
(D.3)

(D.3)

used to convert errors in

is

(cf step 4427 which is also divided by

100 to convert to percent)

DF(DB)

10[(F-l)/(F*Log(10))] DTE/TE,

where the symbol D denotes the derivative, and

(D.4)
is

defined via

(cf step 4425)

F(DB)

The error caused by loss

lowing relationships.
a

10 CLG(F)

in the connector uses the

(D.5)

fol-

First converting loss in decibels to

ratio (step 4435) uses

CONNECTOR LOSS (DB)

-10 CLG(A).

(D.6)

The changed noise temperature TH

by a loss at 300 K is given by

TH

DTH

A*TH

For mismatch error, eq.

(or similar

[8]
+

(A. 6)

for TC)

caused

(cf step 4640)

(1-A)*300.
is

(D.7)

used as explained in

appendix A (of steps 1000 through 1150)

Table D-I.

11:03

DW11
5

10

Computer program

CSS

for figures 1-4.

FRI. 11/24/72

REM MODIFIED STEPS 74,95, 96, 43 15, 4597-4633, 1033+, FROM TNES 5391
REM
USES NBS 5S53 FOR MISMATCH

20 N=3
30 M = 3
43
42
45
53
63
65
73

C=62
V=0
REM 1=1

FOR TE,I=2 FOR F(DB),I=3 FOR TE RECYCLE, 1=4 FOR FCDB) RECYCLE

PRINT
PRINT

REM
B2=BETA
PRINT
71 REM
T = TE
TlrTH
72 REM
73 REM
T2=TC
74 REM
T3=BTA
75 REM
T4=TA
76 REM
77 REM
D1=DTH
D2 = DTC
73 REM
D3 = DY(D3)
79 REM
D4=DG(Z)
B0 REM
D5=CONNECTOR LOSS (DB)
81 REM
D6=CLIPPING (7.)
82 REM
83 REM
EUETH
84 REM
85 REM
E2 = ETC
E3 = EY
86 REM
87 REM
E4=E3
E5=E(MISMATCH)
88 REM
E6=E(C0NMECTOR>
89 REM
E7=E(CLIPPING)
90 REM
91 REM
F=F
92 REM
F1 = F(DB)
93 REM
94 REM
G1=GAMMA'(ANT)
95 REM
G2=GAMMA'(H0T)-GAMMA'CANT)
96 REM
G5=GAMMA'(COLD)-GAMMA'(ANT)
97 REM
93 REM
103 REM
Y=Y
102 REM
Y2=DY
104 REM
105 REM
T5=PERTURBED TH
106 REM
T6=PERTURBED TC
108 REM
150 GOSUB 4333
193 READ G1,G2,T3,D3,D4,D5,D6
195 B2=3

200
205
245
247
250
263

READ Tl ,D1,T2,D2
READ I,K1,H2,H3
GOSUB 5033
ON I GOTO 253,810,253,813
PRINT USING 5550
PRINT
89

Table D-I
300
333
343
353

(Continued)

FOR T=H1 TO H2 STEP H3

G0SU3 4343
PRINT USING 5553, T, El , E1+E2, E1 + E2+E3 , E3, E3+E6, ES+E6+E7, Eg, P
GOSUB 3000
NEXT T
IF 1:3 GOTO 205
GOSUB 3330
GOTO 190
PRINT USING 5570
PRINT
M=M+2
FOR F1=H1 TO H2 STEP H3
GOSUB 4330
PRINT USING 5530, Fl ,R ,R 1+R2,R 1+R2+R3,R8, RS+R6, RS+R6+R7,R9, P*B
GOSUB 3000
NEXT Fl
IF 1=4 GOTO 205
GOSUB 3033
GOTO 193

770
775
800
802
310
815
820
830
840
890
895
900
931
905
910
1000
1005
1010
1320
1033
1043
1345
1050
1060
1070
1083
1090
1100
1110
1120
1130
1143
1153
1160

3003
3310
3020
3030
3040
3350
3060
3070
3330
3093
3100
3110
3120
4300
4310
4315
4320

K=0
G5=G2
FOR J:l TO 8
B2=-B2
X=K+l
IF K=2 GOTO 1060
K=l

G5:-G5
IF J<>5 GOTO 1083
G2=-G2
T5=T1+FNM(T1,G2,B2)
T6=T2+FNM(T2,G5,B2)
W=ABS(FNT(T5,T6,Y,1)-T)
IF J<>1 GOTO 1130
X=W
IF X>W GOTO 1150
X=W
NEXT J
RETURN
M = fH-l

V=Vfl
IF V<>3 THEN 3370
V=3
PRINT
PRINT
M=fH-2

IF M<C-3 THEN 3123

FOR Q=M TO C+S
PRINT
NEXT Q
M=3

RETURN
DEF FNT(H,C,Y,G)=(G*H-Y*C)/(Y-G>
DEF FNE(W,X,Z)=50*ABS((W-X>/Z)
DEF FNM(X,Y,Z)=UX-T*T3)*(Y*Y+2*G1*Y)+2*T*T3*Z*Y>/C1-G1*G1)
RETURN

90

Table D-I

(Continued)

4333 T=290*(10T(F1/10)-1)
4343 Fl=10*CLG(l+T/290)
4430 Y=CT1+T)/CT2+T)
4410 Yl =13*CLG(Y)
4415 Y2 : Y*L0G(13)* D3/10
4425 F=10T(F1/10)
4427 B=3.I*CF-1)/(F* LOGC10))
4430 T4=T-G1t2*CT+T3)
4435 A:13t(-D5/13)
4440
W=FNTCT1+D1,T2,Y, i:
4453
X=FNTCT1-D1,T2,Y,
4460
Z=T
4473 E1=FNECW,X,Z)
4475 R1=E1*B
4483
W=FNTCTl,T2+D2,Y,i:
4493
X=FNT(T1,T2-D2,Y, l:
4510 E2=FNECW,X,Z)
4515 R2=E2*B
4520
W=FNT(T1,T2,Y+Y2, I!
4533
X=FNTCT1,T2,Y-Y2,i:
4550 E3 = FNE(W,X,Z)
4555 R3=E3*B
4563
W=FNT(T1,T2,Y, 1+D4/130)
4573
X=FNT(T1,T2,Y,1-D4/103)
4590 E4=FNE(W,X,Z>
4595 R4=E4*B
4633 GOSUB 1303
4613 E5=103*X/T
4635 R5=E5*B
WrFNT(A*Tl+(l-A)*330,A*T2+Cl-A)*333,Y, 1)
4640
4650
X=T
4660
Z=T
4673 ES=2*FNE(W,X,Z)
4675 R6=E6*3
4630
W=((1-D6/100)*T1-Y*T2)/(Y-1)
4713 E7:2*Ft)E(W,X,Z)
4715 R7=E7*B
4717 E9=E1 + E2+E3+E4+E5+E6+E7
4713 R9=E9*3
4720 E8:E1 + E2+E3 + E4
4725 R8=B*S3
4727 PrSQRCEl T2+E2 T2+E3 T2+E4T2+E5t2+E6r2+E7t2 )
4733 RETURN
5303 PRINT USING 5533, Tl , Dl , 03*D1 /Tl
5035 PRINT USING 55 13, T2, D2, 00*D2/T2
5010 PRINT
5023 PRINT USING 5523, T3 , G2, Gl ,B2
5053 PRINT USING 5533, D4, D3 , D6, D5
5060 PRINT
5065 MrM+8
5370 RETURN
5333 PRINT USING 5550
5093 PRINT
5100 PRINT
5113 RETURN
5533:
TH=##### + fit.* K (#.##%)
1Cz44444 + 4.4 K (4.447.)
5513:
BTA/TEr##.# G( STD- ANT) =#.###
GCANT):#.###
3ETA:.##
5523:
D\z. 44 DB
T>Gz44.47.
LOSZz. 444 DB
5533:
CLIPPING:*** .47.
EG
LOSS
MATCH QUAD-E
5553: TEC K)
ETC
EY
CLIP
ETH
5563:#/ #i? 444.44 444.44 444.44 444.44 444.44 444.44 444.44 444.44
ETC
EG
LOSS
CLIP
MATCH QUAD-E
5573: FCDB)
EY
ETH
,44ft
.4444
.4444
.4444
.4444
.4444
.4444
.4444
5580:##.#
1

91

Table D-I (Continued)

8000
8313
8015
8020
8033
8040
8053
8063
8070
8072
8074
8075

DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
END

.05, .01,3, .01, .1, .01,0

1 0030, 1 50,303, .5
2,2, 16.5, .5
.05, .01, .3,. 01,. 1, .01,.
373, .5,80,1

3,100,330,20
1,330,803,50
.05, .31 ,0, .31

.31 ,0

1270,3,330, .2
4, .1,1,.
2,1,14, .5

READY

92

Table D-II.

Print out of D-I

program

READY
RUN
DW11

13:44

FRI. 11/24/72

CSS

TH:10030 + 150.0 K (1 .537.)

TC = 300 + 0.5 K (0.1 7Z)
BTA/TE: 0.0

G(STD- ANT) =0.0

DG= 0.17.

DY

.01

G(ANT)=0.050
CLIPPING: 0.07.

DB

CLIP

BETA:. 33
LOSS:. 010 DB

F(DB)

ETH

2.0
2.5
3.0

.0686
.0685
.0683

.0736
.0729
.0723

.0843
.0836
.0831

.0389
.0883
.0878

.0991
.0985
.09 79

.3991
.0985
.0979

.1359
.1359
.1051

.0709
.0707
.0735

3.5
4.0
4.5

.0682
.0681
.0680

.0713
.0713
.0709

.0826
.0822
.0818

.03 73

.0975
.3971

.0975
.0971
.0967

.1044
. 1333
.1033

.3704
.0702
.3701

5.0
5.5
6.0

.0679
.3678
.0677

.0705
.0701
.0698

.0816
.0813

.0864
.0362
.0361

.0965
.0963
.0961

.0965

.1029
.1025
.1022

.0700
.3700
.0699

6.5
7.0
7.5

.0677
.0676
.0676

.0696
.0693

.0810
.0839

.0360
.0360

.0963
.0963
.3960

.1323
.1319
.1019

.0698
36S3
.0698

8.0
8.5
9.0

.0675
.0675
.0675

.0689
.0688
.0686

.0809
.0813
.0311

.0361
.08 62

.1019
.1019
.1321

.3698
.0698
.0698

9.5

.0674
.0674
.0674

.0685
.0634
.0633

.0312
.0814
.0817

.0367

.0693
.3698
.9699

.0673
.0673
.0673

.0682

11.5
12.0

.0680

.0820
.0823
.0828

.0880
.0335
.0892

12.5
13.0
13.5

.0673
.0673
.0673

.0679
.0679
.0678

.0833
.0839
.0846

.0900
.0908
.0918

14.0
14.5
15.0

.0673
.0672
.0672

.0678
.0677
.0677

.3353
.0362
.0872

15.5

.0672

.0677

.3883

10.0
10.5

11 .0

ETC

EG

EY

.081

.0691

.0681

.0869
.3866

.03 60

LOSS

.0967

.09 63
.3961

.39 63

.3963
.39 60

MATCH

QUAD-E

.0961
.0963
.0965

.0961
.09 63

.0968
.3971

.0968
.3971

.0975

.09 75

.1023
.1326
.1329

.0983
.0986
.0992

.0933
.0986
.0992

.1334
.1039
.1345

.0599
.3700
.0701

!000
1008
.1013

.1003
.1003
.1313

.1053
.1361
.1371

.0703
.0734
.0706

.0929
.0942
.0956

.1329
.1042
.1056

.1329
.1042
.1056

1082
.1394
.1133

.0708

.39 73

.1073

.1373

.1124

.0718

.0865

.0871
.0375

93

.0965

.071

.0714

Table D-II

(Continued]

0.5 K (0.13%)
TH= 373 +
TC=
80 + 1.0 K (1 .257.)

G(STD-ANT): 0.010

BTA/TE: 0. ,3
DG : 0.1%

DY :.0!

BETA:. 00
LOSS:. 010 DB
MATCH QUAD-E

3.80

3.80

4.17

2.05

3.42
3.15
2.96

3.75
3.45
3.25

1.82
1.65

2.29

3.42
3.15
2.96

.92
1.85
1.79

2.20
2.12
2.06

2.81
2.70
2.61

2.81

2.61

3.09
2.96
2.87

2.02

1.74
.70
1.68

2.54
2.48
2.44

2.54
2.48
2.44

2.79
2.73
2.68

1.28
1.24
.22

0.98

!.65

1.94
2.40
2.40
2.64
TH:
373 +
0.5 K (0.137.)
TC:
80 + 1.0 K (1 .25%)

1.20

EG

1.92

2.59

2.88

.69

2.33
2.15
2.02

2.61
2.43

ETC

100

0.31

120
140
160

0.28
0.27
0.26

180
200
220

0.25
0.24
0.23

240
260
280

0.23
0.22
0.22

300

0.22

BTA/TE: 0, ,0
DG : 0.1%

G(AHT) =0.050
CLIPPING: 0. 0%

CLIP

EY

ETH

TECK)

DB

1.52
1,39

1.30
1.22
1.15

.10
1.05
.02

1.99
.96

G(STD-ANT): -.010
DY :.01

DB

!(K)

ETH

ETC

300
350

0.22
0.21

0.98
0.91

400
450
500

0.20
0.20
0.20

0.86
0.83
0.79

1.59
1.59
1.59

550
600
650

0.20
0.19
0.19

0.77
0.75
0.73

1.60

1.96

.61

1.99

1.63

700
750
800

0.19
0.19
0.19

0.71

.65

0.70
0.69

1.68
.70

EG

EY

.65
1.61
1

.94

1.92

LOSS

2.70

2.40
2.34

2.40
2.34

2.64
2.57

1.20
1.16

1.15
1.14
1.15

'

2.53

2.30
2.30

2.51
2.51

2.02

2.32
2.33
2.36

2.32
2.33
2.36

2.52
2.53
2.56

2.06
2.10
2.14

2.39
2.42
2.46

2.39
2.42
2.46

2.58
2.62
2.65

94

1.32

MATCH QUAD-E

'

2.31

.94

.44
.37

CLIP

LOSS

2.31

.53

BETA:. 00
LOSS: .010 DB

G(ANT) :0.050
CLIPPING: 0. 0%

2.30
2.30

1.91
.92

.16
1.18
1.20

1.23
1.25
1.28

D.2.

Figures Bl through B7

For figures Bl through B7

table D-III was used.

front panel for

the program DW21.3 listed in

A minicomputer was used for this calcula-

in step 1052

tion and call 11

reads

diagnostic routine.

this program is listed in table D-IV.

eq.

(B.9)

is

switch position on the

A typical

For the calculation,

used as explained in appendix

95

output from

B.

Table D-III.

p^TMT

^w

Computer program

for figures

(23 NOV 72)"

MI "WATCH UNCERTAINTY; SWITCH
'FROM NBS PAGE 5927, PROGRAM

B1-B7.

*"Drf21.3

UP FOR DIAGNOSTICS
MODIFICATION FO 533S
REM
B1=3ETA
REM
GUGAMMA' (ANT)-GAMMA' (STD)
REM
14
G2 = GAMMA* (AWT)
16
REM
1^
REM
3=REVERSE RADIATION PARAMETER
<J3
DEF FNA(C) = INT ( 1 2l(233*C+.5>/l 33
133
DIM D[25]
11'?
FOR 1=1 TO 23
120
READ D[ I]
133
NEXT I

13
12

143
20:?

213
223
233
?43
253
253
273
233
23 3
3 -3 3

313
323
325
343
35*
3S3
373
33 3
375
39 3

433
435
410
423
433
44 3

453
453
47 ^
4' 7

HI
S0
52?

.3

READ 31
PRINT
PRINT
PRINT
PRINT TAB (13);
PRINT "MAXIMUM MISMATCH UNCERTAINTY a OF TF) M
PRINT
PRINT
PRINT ,,"FOR BETA=";B1
PRINT
PRINT
PRINT
PRINT
PRINT " --GAMMA'--"'
-j
PRINT "
VER
RADIATI0N PARAMETERS
PRINT ""fr""""t :

FOR K=l TO S
?3INT TA3
NEXT K

D[

7+K ]+

;D[

1+X]

PRINT
PRINT
PRINT
1=1 TO 5
FOR J=l TO 6

F03.

LET G = D I
LET G2 = D[ 5+J
IF J<>1 GOTO 45^
PRINT Glj
prist TAB (7),G2;
FOR K=l TO 6
LET B=D[I1+K]
GOSUS
1303
LET X= FNA(E) + .0001
IF X> 3 GOTO
513
1

^JZf'-P?

(Dti7+ "- LOG (X)/ LOG (13)),

96

FNA(E);

Table D-III
53'?

54 3

553
5S3
57"'

539
5?3
S03
si g

520
533
64 3
65 "1

NEXT K
PRINT
LET V=VH
IF V=3 GOTO
IF "v/=6 GOTO
GOTO
623
LET V:
PRINT
PRINT
NEXT J
NEXT I
FOR M = l TO 17
PRINT

(Continued)

61:3

553

653
N EX T [1
673
GOTO
143
1330
LET E= 3
1335
FOR Q =
TO A
1010
LET GU-G1
134 3
1323
IF Q<>3 GOTO
133 3
LET BU-31
1343
L ET L = ( G 1 * G 1+2* G2 * G 1 ) / ( 1 -32* G2 )
l

LET E2=L*B*(G2-B1)*(G2-B1)
LET E3 = B*G1*G1
LET E4=2*3*G1*(G2-B1)
LET E5=(1-G2*G2)*(1-L)
LET E1=(L+E2+E3+E4)/E5
CALL 11, 3,P
IF P= 3 GOTO
1363

1341

1^42
1343
1344
1353
1352
1353
13 54

PRIi'JT

1355
1356
1357
13 63
373

PRINT "E=" Ej " ETa;"="El "G1 = "G1

PRINT
PRINT " E2 = " E2;"E3 = "E3 " E4 = "E4
IF
AB5 (E1)<E 30TO
1333
LET E= ABS (El)
;

imn

N Ev T

139 3
53 33

RETURN

5310
5323
5333
5343
5353

DATA

53 73

5375
53 3 3

53P3

EM

DATA

"G?

'3- "E5

G1=GAMMA' (AMT)^GAMMA' (STD) FOLLO

.335,
32

.01,

= G AM M A "

.32,

3,

.32,

.05,

AN T )

FOLLO WS

.35,

.1,

.1

.35

.2,

REVERSE RADIATION PARAMETE

REi*

DATA
REM
DATA
REM
DATA

"

"G2

RET"
R

3,

.2,

.5,

1,

2,

TAB POSITIONS FOLLOW

17,

27,

36,

45,

64

55,

BURET A FOLLOWS
3,

.1,

.2,

.3,

.4,

.5

97

FOL

"31="31;"3:"B:"L="L:

Table D-IV.

Print out of D-III program.

RUN
D\'2

MBS

(30 MOV 72)

PAGE 5940, RECHECK 5928

.3

VJ

MAXIMUM MISMATCH UNCERTAINTY

(%

wU i

OF TE)

FOR BETA=

--GAMMA'-ERR
ANT

P.EVE^.SE

.09

,33

.21

,05

.39

.71

.2
.35

1.42
2.92 %

16

.32

7=

.57

.02
35
.

.2

.99 %
.95
7

.35

3.93

.46
.96
2.73

.32
.35

4.33

.14

13 5.32

18

C7

/o

%
%

.42
.79

.42

2. 13

4.33

%
%

2.34
5.83

%
Z

4.26
3.75

.32 ?
.64
1.13 %

19

%
%

.24
.43
.85

1.19

2.33
4.77

%
7

%
%
%

%
%

%
%

4.89
3.69

%
g"

14.39 %

17. 27

2.3 6
2.7

3.72

5.49
9.67
19.17

4. 53

3.35
15.93

%
%

z
'

j>

2.4

3.

4.19

.32
.35

2.59
3.49

5.1

6.12

=7

3.9

13. 69

.35

17. 64

?..

21.16

98

6.

f.

%
07

'10.36 %
21 . 59 %

2. 53

3.39
4.65
6.36
12.03
23.97

3
%

.43

"

2.92
5.97

2. 19

1.3 6

.39
.63

.27
.63
1.13

2. 13

2.94
4.39

3.

.35

14

.32
.59

3.5

.05
1

.7

%
%

.85
1

1.72
2.26

.2

.75

.02

2.35
3.27

7. 24

.35

.5

.25
.47

,04

12

.32

RADIATION PARAMETERS

.2

.65

13.36
26.4 5

2.92
3.92
5.45

%
%

3. 15

8.51
17.5

2.97
5.84

z
1

%
z

.93 %

.96 7
.93 %
3.4 %
1

1.7%

=7

4.25

or

.54 %
.26 ?

2.36

.48 Z
.96 z

1.98
3.89
7.96

<*t

5.93

.67

23.87

3.77

4.33
5.83

%
7

.76

8. IS

16.35

12.23
21.72

24.46
43.45
36.36

14.43 %
23.79 %

4 3. 18

<v

%
"

%
Z

"

5.16

4. 52 %

6. 79

6.2

9.29

%
%

27.4 5

43 .33

j7

5.37

3.44
3

10.31 7
13.57 %
18.59 %

9. 15

16.11

"

31.96

13.73
24.16
47.93

%
5. 17

6 Z
7. 76

"

15.52

6.

99

fo

10.43

15.3

33.61
53.43

3.3 3
5.24

7,

10.2
17.31

35.27

12 %

26.71

52. 31

fo

97

%
%

D.3.

Table

For table 3, the program DW13 listed in table D-V was

used.

Essentially eq.

(D.4)

is used

99

in step 320.

Table D-V.

Computer program

for table 3.

LIST
CSS

15:02

DW13

THU. 05/04/72

53 N=l
S3 V=3
103 PRINT
113 PRIST
123 PRINT
133 PRINT USING 143
140:
TECK)
=
F(DB)
153 PRINT
150 FOR J=l TO 4
173 FOR K=l TO S
133 ON K GOTO 233,213,220,23*3,243,250

T:13TJ
GOTO 333
T=1.5*13TJ
GOTO 333
T=2*13tJ
GOTO 333
T=3*13tJ
GOTO 333
T=5*l*Tj
GOTO 333
T=7*13tJ
GOTO 333
Fl=13*CLG(T/293+l)
F=l+T/290
B=.1*(F-1)/(F* LOGU3))
T5=290*(13f< . l*r. )-l)
34 3 F5=l+T5/29 3
353 TS: r 5*LOG(13)/CF5-l)
3S3 PRINT USING 3 73, T, Fl ,B, N, T5, T5
37i3:##### + 1%
##.## + .#### DB

203
235
213
215
220
225
233
235
240
245
253
255
303
313
323
333

33

V- "HI

43-*

RINT
NEXT K

44?

>

5'>3

EXT J
FOR 1:1

51J

PKINT

52

999

\!EXT

TO 13

END

R E A DY
5.i:

T(K)

393 VzJ+l
433 IF V<>3 THEf! 433
41
V:^
423 PRINT
42 c

FCD8)

FOR 1=1

TO

2'

100

*# +

.1

DB

##### + ##.##%

D.4.

Tables AI--AXXI

For table A-I through A-XXI the program DW10 listed in

table D-VI was used.

This program is very similar to parts of

DW11 listed in table D-I.

101

Table D-VI.

Computei- program for tables A-I through A-XXI,

LIST

08:50

DW10

CSS

THU. 04/20/72

20 N = 3
30 M=3
40 C=62
42 V:0
45 1=1
50 PRINT
60 PRINT
70 PRINT
71

72
73
74
75

76
77
78
79

S3
81

82
83
84
85
86
87
88
89
90
91

92
93
94
95

96
97
98

REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM
REM

T = TE

T1=TH
T2 = TC
T3 = BTA
T4 = TA

DUDTH
D2 = DTC
D3 =

DY(DB)

D4 = DG(7.)

D5=C0NNECT0R LOSS (DB)

D6=CLIPPING (Z)

EUETH
E2=ETC
E3 = EY

E4=EG
E5=E(MISMATCH)
E6= EC CONK ECTOR)
E7=E(CLIPPING)

F=F
F1=F(DB>

GUGAMMA

(AMP)

G2=GAMMA (STDS)
Y=Y
Y1=Y(DB)
Y2 = DY

99
100 DEF FNTCH, C, Y, G) = (G*H-Y*C) /CY-G)
110 DEF FNE(W,X,Z)=53*ABSUW-X)/Z)
120 DEF FNF(T)=10*CLG(l+T/290)
190 READ G1,G2,T3,D3,D4,D5,D6

200
245
300
310
320
330
343
350
355
360
370
380
390

READ T1,D1,T2,D2
G0SU5 5030
FOR J=l TO 4
FOR K:l TO 6
ON K GOTO 1000,1023,1043,1060,1383,1103

G0SU3 4433
PRINT USING 55 63, T, ES,

Fl ,R8, Yl

El

M=M+1
V:V+1
IF V<>5 THEN 760
V=0

PRINT
PRINT

102

E2, E3

E4

Table D-VI

(Continued)

400 MrPH-2
760 IF M<= C-3 THEN 77<3
763 IF M> C-3 THEN 765
765 GOSUB 3300
770 NEXT K
7S0 NEXT J
800 GOSUB 3000
802 GOTO 200
803 GOTO 200
804 GOSUB 5000
1000 T=10tJ
1010 GOTO 330
1020 T=1.5*10TJ
1030 GO TO 330
1043 T=2*10tJ
1050 GOTO 330
1060 T=3*10TJ
1070 GOTO 330
1080 T=5*10fJ
1090 GOTO 330
1100 T=7*10tJ
1110 GOTO 55<3
3000 FOR Q=M TO C+6

3010
3020
3025
3040
4430
4410
4415
4420
4425
4427
4430
4435
4440
4450
4460
4470
4480
4490
4510
4520
4530
4550
4560
4570
4590
4600
4610
4630
4640
4650
4663
4673
4683
4710
4723
4725
4733

PRINT
NEXT Q
M=3
RETURN
Y=(T1+T)/(T2+T>
Yl =13*CLGCY)
Y2 = Y*LOG(13)*D3/13
F1 = FNF(T)
F=10t(Fl/10)

B=0.1*(F-1)/(F*LOG(10))
T4:(1-G1 T2)*T-T3*G1
A=l0t(-D5/10)
W=FNT(T1+D1,T2,Y,1)
X = FNT(T1-D1,T2,Y, 1)
Z=T
E1=FNE(W,X,Z)
W=FNT(T1,T2+D2,Y,1)
X = FNT(T1,T2-D'2,Y, 1)

E2=FNE(W,X,Z)
W=FNT(T1,T2,Y+Y2, 1)
X=FNTT1,T2,Y-Y2,1)
E3 =FNE(W,X,Z)
W=FNT(T1,T2,Y, 1+D4/103)
X=FNT(T1,T2,Y, 1-D4/130)
E4=FNE(W,X,Z)
W=(T4+T3*((Gl+G2)/Cl-Gl*G2))>/<l-<<Gl+G2)/(l-Gl*G2))t2)
X=(T4+T3*ASS(G1-G2))/(1-(A3S(G1-G2))T2)
E5=FNE(W,X,Z)
V=FNT(A*T1+<1-A>*300,A*T2+(1-A)*300,Y,1>
X=T
Z-T
ES=2*FNE(W,X,Z>
W=((1-D6/133)*T1-Y*T2)/(Y-1>
E7r2*FNE(W,X,Z)
E8=E1 + E2+E3 + E4
R3:B*E3
RETURN

103

Table D-vi
5030 PRINT USING
5005 PRINT USING
5010 PRINT
5020 PRINT USING
5050 PRINT
5060 PRINT
5065 M=Mf8
5070 PRINT USING
5080 PRINT USING
5090 PRINT
5100 PRINT
5110 RETURN
5500:
5510:
5520:
5543;
5550:
TE(X)

5500, Tl , Dl , 00*D1 /Tl

5510, T2, D2, 100* D2/T2
1

5520,D3,D4

5540
5550

TH=##### + ###
TC:##### + #.#
DY=.## D3
FCDB)

5560:##### +##.#%
##.# + .##
8010 DATA .1 ,.1,300, .01, .1,. 02, .1
8020 DATA 18000,270,300,1
8030 DATA 18030,270,80, .5
8040 DATA 18000,270,4, .5
8050 DATA 10000,150,300,1
8060 DATA 10030, 153,80, .5
8070 DATA 10300, 153,4, .5
8080 DATA 1270,3,333,1
8090 DATA 1270,3,83, .5
8100 DATA 1270,3,4, .5
81 10

8120
8130
8143
8150
9999

DATA
DATA
DATA
DATA
DATA
END

373,
373,
373,
300,
300,

(Continued)

.1 ,290, .1
.1,80, .5
.1,4, .5
1,80, .5
1,4,.

READY
RUN

104

Y(DB)
##.#

(#.##7.)

(#./#Z)

DG=.##%
ERROR CONTRIBUTIONS TO TE
ETH
ETC
EY
B
##.##% ##.##% ##.##% /|

D.5.

For tables Bill

table D-VII was used.

--

Tables BIII--BXXII
BXXI

the program DW22 listed in

This uses the same mismatch error calcu-

lation from appendix B that was used in DW11 listed in table

D-I.

105

Computer program

Table D-VII.

for tables B-III

LIST
1

2
5

10
12
14
16
17
18

23
21

22
24
26
23
33
33

43

"D' 22(3 DEC 723"

PRIMT "-PAGE 5856, CK 5391"
:MISMATCH
ERROR..! SMITCi
5
u? fo; DIAGMOSTICS
REM
P.EM
(PA !E 5935), DV 2 1.3 (PAGE 5 97J
MOD D'.'l
REM
B1=BETA
REM
Gl =GAMX,V (HOT) - GA :riA" (and
REM
G2=GAMMA' (AMT
P.EM
G3=GAMMA* (COLD) - GAMMA' (AMT)
PEM
B= REVERSE RADIATION PARAMETER
REM
T = T3
REM
T1=T(H0T)
REM
T2=T(C9LD3
REM
T5=PERT'.;P.3ED T(HOT)
REM
T6=?ERTUR3ED T(COLD)
P.EM
T7=?ERTUREED TE

PP.IMT

P.EM
pr-,

PI

=F(D33

REM
53
IMT C1323*> +.53/1303
DEE FMA(X3
93
103
DIM DC25]
13
FOR 1=1 TO 23
123
Ri AD D C I
133
Mr XT I
,T2,31 ,3
READ
143
233 PR IMT
1

213
223
233
243
253
263
273
272
274
276
273
233

293
333
313
3 23

325
333
343
353
363
373
333
335
393

PP.IMT
PP.IMT

TAB (133

PRIMT

PP.IMT " MAXIMUM

1ISMATCH ERROR CDS)'

PRIMT
PP.IMT

PRIMT
PRIMT
PRIMT
PRIMT
P.P. I

TCiIOT) = "TU

t(cold3=-t;

JIT

0ETA="31
3TA/TE="!

PRIMT
PRIMT
PRIMT
PRIMT
PRIMT " --GAMMA' --'
PRIMT "
PRIMT "
PRIMT "
AMT";
FOR K=l ro 6
PRIMT TAB (DC i7+:<] + 3;dc
i

ME XT

'+:<]:

!<

PRIMT
PRIMT
PRIMT

106

through B-XXII

>

Table D-VII (Continued)

430
435

FOR 1=1 TO 5
fop J=l TO

423
433
443
453
463
473
472
473
474
476
483
435

PP.IMT Gl ;
PP.IHT
TAP C7>,G2J
FOP X=l TO 6

LET
LET
LET
LET
LET

TJEXT
r;EXT i

64 3

FOP.

1332
1305
1306
1307
1013
1023
1033
1043
104 5

1350
1353
1354
1056
10 63

1073
1 133
1
1
I

1
1
1

123
125
130
135
150
160

1173
1232
123?
1213
1215
1223
1225
1233
124 3
1250

F=13' CF1/1 ~>

Y=CT1+T)/<T2+T)
E4=l 3*(F-1 )/(F* LOG
1300

13

LET X= FMACE)+.3C0I
510
IF X> 3 GOTO

530
513
533
540
550
560
573
580
593
633
613
623
633

15103

F1=DC t+KJ
T=293*C 3' CFI/13}-

GOSUO

4 93

653
663
673

LET Gl=Dd]
LET G2 = DC5+.J1
453
IF J"l GOTO

4 13

LET

:<=

33

TA3 CDC17+K3), FNACE);

PP.IMT
NEXT X

PRINT
LET V=V+1
IF V=3 GOTO
IF V=6 GOTO
GOTO
623
LET V= 3
PP. I MT

613
593

PP.IMT
J

TO

'1=1

17

pr.niT
r.'EXT

r;

143
GOTO
LET 3= 3
LET E3= 3
LET G3=G1
DEE FME<X)=CX*X+2*32*X)
LET E1=1-G2*G2
FOR P.= TO 3
=-31
LET
LET S = S +
363
IF S=2 GOTO
LET S=l
LET G3=-G3 .
1053
IF ?= 3 GOTO
PP. I NT
l

PP.IMT

IF*P.<>5 GOTO
G = -G1

LET
DEF
LET
LET
LET
LET
LET
LET

133

FXFCX)=2*3*T*X*31
T5 = C<T1 -3*T>* E"ECG1>+ FNFCG1))/E1

T5=T1-T5
T6=C<72-3*7)* FNE<G3)+ F?JFCG3))/E1
T6=T2-T6
T7=CT5-Y*T6)/(Y-1
3 2 = A3S C(T7-T)/T)

CALL
IF P=

11,

3,

GC x O

1233

PPIT
pp.imt "f= f; fi ="fi ;"T="T;"Ti="Ti ;"T2="T2
PP.IMT 'T5="T5;"T6 = "T6;"T7 = "T7;-'Y="Y
PP.INT "Gl = "G1 ;"G3 = "G3;"G2 = "G2; ,- D1="31 ;"3="3
pp.imt "e<"~.;")="e;"ei ="Ei ; ,, e2 = "E2; ,, e2 = "E2;"E4 = E4
1253
IF E2<E3 GOTO
,-

,,

'

LET E3=E2
LET E=E3*E4

107

Table D-VII (Continued)

1261
1271

NEMT

5035!
5015!

"EM
DATA
pem
DATA
"EM
DATA

5020
5030
5351

?.etu:

37

t.em

5075
508 3

DATA
HEM
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA

5-39 3

5103
5113
5123
5133
5143
5150
5160
5173
5

130

5193
5230
5210
5220
5233
5240
5333
5310
5433
54 10

GI=GAMMA* f AMD-GAMMA* C5TD3 FOLLOWS

.305, .31, .02, .05, .1

g2=ga;:ma**cahtj follows
0, .32, .35, .1, .2, .35
l=F<DE) FOLLOWS
1

2,

4,

6,

3,

13

TAB POSITIONS FOLLOW

17,
:

27,

36, 45, 55,

64

SETS OF Tl ,T2,0ETA,BTA/TE FOLLOW

13093, 300, 0,
10303, 300,
2
,
10000, 3 30,
,
10003, 330, .2,
13333, 300, .2,
130 30,

33-3,

10003, 330,
373, 83, 0,
373, S3,
i,
373, 80,
,
373, 33, .2,
373, 30, 2,
373, 30, .3,
373, 33, .3,

.3,
.3,
;

.
.

3 33, 3, 3
1273, 300, .2,
1273, S3, 3,
1270, 33, .2,
303, 4, 3, 3

127-3,

303, 4, .2,

10

Appendix
References

[1,

E.
3,

Reprints of Useful Papers

and 8]

are reprinted here.

109

Reference

Measurement
mismatch

Reprinted with Permission of the Publisher

Technical Feature

[1]

of Amplifier Noise
measurements are

errors and amplifier noise

interrelated

--

other errors are exposed

D. F.

WAIT

National Bureau of Standards. Boulder,

CO

Measurement of amplifier noise, in particular a state-of-the-art

measurement using hot-cold noise standards of 10,000 K 300 K
or 373 K 80 K, is discussed. An error that is especially difficult
to estimate is the mismatch error. Measurements are suggested to

Reference Standards

help with this problem. Also, the traditional

services are briefly sketched.

NBS

thermal noise

NBS

All

noise standards are resistive elements main-

known temperature T 3 The

power P a available from these blackbody radiators is P = kTB where k is Boltzmann's constant, and 8
a
is the bandwidth being viewed. The cold standards maintained at a uniform and

'

theoretical

tained by NBS are liquid-nitrogen cooled and have noise

temperature near 80 K. The temperatures of the hot

standards vary: the

RF and microwave

In its

thermal-noise services, the

National Bureau of Standards has

in

the past concentrated

on developing means for transferring the fundamental

standards of noise power to the scientific and engineering
community of the Nation. With substantial capability for
doing this

now on

hand,

NBS

has turned

its

operates

at

NBS 14mm coaxial hot standard

WR284 standard (2.60 3.95 GHzi

373 K; the

18.0 GHz), and

WR90 (8.2 12.5 GHz), WR62 (11.9WR15 (50 75 GHz) standards near

1250

at*9lK;and

the

K. Figure

NBS

indicates rreauencies at which

maintains the hot and cold primary standards.

attention to

the application of these standards, and of related techniques, to

measurements of more

One

"ultimate user."
recent efforts
his

is

Calibration Services

direct concern to the

of the key objectives of these more

to transfer to the user the ability to make

own measurements

and/or to verify that these are

The comparison radiometer, used to

known

deed valid. In special instances NBS may simply make

measurements for the user; in others it may repeat the
customer's measurement, if necessary at the customer's
own laboratory, to validate his technique and results.

10 T

*&*
'

-GH 2
15

WR

new

directions

in

NBS

tz

|
I

18

GHz

8.2

GHz

62
\VR 90

figure that has been used by NBS in the validation process

and that, most important, can be used by others. It is typi-

of the

is

ROUTINE SERVICE

This paper outlines and discusses the amplifier noise

cal

calibrate an un-

noise source relative to the primary standard,

in-

|WR 28 4|

noise services.

3.95
2.60

GHz
GHz \*fi

The emphasis in this paper is on how to make amplifier

measurements and on sources of measurement error

noise

and uncertainty. But because these and so many other

similar measurements critically depend on the availability
of noise standards, the standards that can be made available through the

more

traditional

NBS

calibration services

are first briefly sketched.

TRADITIONAt CAPABItlTIES

IN

NOISE

60

MHz

30

MHz

'

60

MHz

30

MHz

The National Bureau of Standards maintains reference

noise standards and calibration radiometers, and has been
active in developing the theory to support these areas.

The most

utilized noise service

bration, although

is

that of noise-source cali-

more and more frequently

NBS

is

asked

1973

MHz

is

a contribution of the National

COLD STANDARDS

Bureau of Standards

Fig.

110

NBS

CALIBRATION
SERVICES

calibration services

"

HOT

CALIBRATION
SERVICES

not subject to copyright.

January

COI.D

source.
This paper

MHz

to analyze a radiometer rather than to calibrate a noise

STANDARDS

and standards.

most elaborate

the

part of

any calibration

service,

and the

need to understand its operation spawns most of the NBS

experimental and theoretical noise studies''"' 6 The fre.

which both reference standards and compari-

Te

<K)

son radiometers exist are the frequencies of the available

10

shown

1G
20

1%
1%
1%

quencies

at

noise calibration services

noise temperatures between

brated to within

2%

is

Figure

1.

Typically
are cali-

total uncertainty (excess noise ratios

between 4 and 30 dB
uncertainty

in

1000 to 300,000 K

to within 0.1 dB)

degraded somewhat

if

.1 dB
2 .1 dB
3 .1 dB

30 1%
50 1%'
70 1 % =

0.43 .0041 dB
0.69 .0064 dB
0.94 .0084 dB

4 .1 dB
5 .1 dB
6 .1 dB

7
'

Te
0.15 .0014 dB
0.22 .0021 dB
0.29 .0028 dB

The

calibration

the connector of the

IK)

75 11.20%
169 6.24%
283 4.62%

438 3.83%
627 3.37%
864 3.07%

"
=

source to be calibrated differs from that of the correspond-

NBS

ing

standard. For frequencies at which cold-source

calibrations are indicated in Figure 1, noise sources are

typically calibrated to within

75

1%

at

MEASUREMENT OF AMPLIFIER NOISE

300 1%
500 1 %
700 1 %

dB

7 .1 dB

.0148 dB

8.1dB

dB

9 .1 dB

3.03 .0221 dB
4.35 .0275dB
5.33 .0307 dB

10 .1 dB
11 .1 dB
12 .1 dB

dB
dB
dB

13 .1 dB

14.1dB
15.1dB

1.81

2.28 .0177

greater than the coverage of the noise-source

calibration services, which are limited

by the frequency

An Overview
The most

crisply defined

measure oi amplifier noise

the effective input noise temperature.*

If

two

T hot and
are sequentially connected to the input of an ampli-

then the change in output power of the amplifier,

measured as a ratio Y, is

3000
5000
7000

1%=
1%1%=

10.55

.0393dB
dB
14.00 .0417 dB

16 .1 dB = 1 1255 2.36%
17 .1 dB = 14244 2.35%
18 .1 dB = 18007 2.34%

10000
15000
20000

1%=
1% =
31%=

15.50 .0422 dB
17.22 .0426 dB
18.45 .0423 dB

21 .1

30000
SO0OO
70000

1%

6.48 .0337
7.90 .0364
8.97 .0379

=
=

12.61 .0410

We

<T hot

*Te )/(T cold+ T e

Te

see that

is

(1)

19 .1 dB = 22745
20 .1 d8 = 23709
dB = 36218

= 20.19 .0430 dB
22.39 .0432 dB
23.84 .0433 dB

2.33%
2.33%
2.32%

22 .1 dB = 45671 2.32%
23 -A dB = 57572 2.31%
24 r.1 d8 = 72554 2.31%

1%=
1%=

Conversion between effective input noise temperature, T (K),

and noise figure, F
dB The asymmetry in the error statements resulting from the logarithmic non-linearity of F^,- is avoided by
using the slope of F ~ at the corresponding T
d
.

the signal channel, or through both channels. Nevertheless,

noise figure

measure of the amplifier's

terms of the available** power

essentially a

internal noise expressed in

549G 2.42%
6394 z 2.40%
8880 2.33%

fier,

Y=

is

noise

standards with corresponding noise temperatures

T co

2609 2.56%
3360 2.50%
4306 2.46%

1%
1%
1%

ranges of the comparison radiometers.

Te

1163 2.88%
1539 2.74%
2013 2.63%

1000
1500
2000

At present, the National Bureau of Standards can

measure effective input noise temperature only in WR15
waveguide but, in principle, the noise figure service can be
readily implemented at any frequency where an NBS primary noise source exists (see Figure 1 ). Thus, the potential frequency coverage of NBS noise-figure-measurement
is

1.29 +.0111

temperature above

K.

capability

=
100 1% =
150 ri% =
200 1 % =

is

density of a noise generator at the input o.f the amplifier.

Another popular measure of amplifier noise is noise
19 26
8
with
figure'
but the ambiguities and difficulties

used as

is still

To bypass the

measure of amplifier noise.

past difficulties an

unorthodox definition

used here, namely

F dB

10

+Te /290].

log [1

(2)

"

the

IEEE

definition of noise figure limits the usefulness

of noise figure. The most serious ambiguity

is

related to

This definition corresponds approximately to double-

channel noise figure

24

and

superheterodyne amplifiers for which two different input

channels contribute to the same output channel. The
noise figure for the same amplifier differs depending on

Sheibe

whether the

conversions

user expects useful information through only

"Effective Input Noise Temperature, Average T (of a multiport transducer with one port designated as the output port).
The noise temperature m degrees Kelvin which, assigned simultaneously to the specified impedance terminations at all
frequencies at all accessible ports except ttie designated output

as

Mumford and
IEEE definition

discussed by

-equivalent to the

way, F dB can be

as a

Te and

cnange
F

in scale

dS are given

* The noise temperature of a noise source is defined to be the

physical temperature of a blackbodv radiator which has the
lame available power density as the noise source (at the frequestion).

111

of

T e The
.

Table

is

ment of Te

1.

The
a

of

are

Y because
(3)

in a

measure-

in the value of T
hot uncertaint
uncertainty in the measurement of

uncertainty

T^^,

Te

effective input noise tempera-

measurement

-YT cold )/(Y-1)

the value of

1.

G-.al

known rrom

(T hot

in

of

in

most accurate and easiest measurements of

based on Equation
ture T.

ewed

bet'.-.esn

The Accuracy
The

:
.

There are seven important sources of error

except the output port."

in

is

noise figure for a ingle-response amplifier. Defined this

port of a noisefree equivalent of the transducer, would yield

the same total noise power in a specified output bond delivered
to the output termination as that of the actual transducer connected to noisefree equivalents of the terminations at all ports

quency

Y; the amplifier-cjin instability during the time required

to meas".-? tH ratio Y; the inequality of the three reflec-

the

microwave

jo urn.

tion coefficients, vis., the hot and cold noise standards,

and the reflection coefficient "seen" by the amplifier input port when it is in actual use (mismatch error); difference in losses associated with the connector of the hot
and cold standards and the amplifier under operating con-

ditions are either not reproducible or are

unknown;

un-

certainty in the noise generated in the measuring system

known

not accurately

is

300

200

500 700 Ik

T HriT

fllSMATCH

2k

30

5k

3k

7k

10k

0S K
150 K

= 10.000

rANT -rsTD <0.01;

ir AM2 .-r aMP '

do.o;

< 0.05)

-(o.Ol drj)

"antenna"

gain -cp.i^;
Y-KACTOR

COLD<017

The State of the Art In establishing an accuracy goal,

one of the best guides is an awareness of how closely one
wishes to approach the state-of-the-art.
Figures 2 and 3 show the accumulated errors for stateof-the-art measurements of F
dB and T e using two frequently employed hot-cold standard combinations; namely, 10,000 K and 300 K, or 373 K and 80 K. With one of
these combinations, measurement uncertainties of about
0.1 dB in noise figure or effective temperature uncertainties of about 2.5% are state of the art.
Mismatch Uncertainty The effective noise temperature
T e is a function of the reflection coefficient (i.e., impedance) of the "antenna" (components attached to the input
of the amplifier).

CONNECTOR LOSS

be expected, and the quality of available connectors.

(cascade error).

1.5 k

T COLD

the art, the magnitude of mismatch uncertainty that can

known exactly, then the T e that the

have when in use cannot be known exactly.

amplifier will
-(p.01

the reflection coefficient of this

If

not

is

dB)

TABLE

(PRELIMINARY! CONNECTOR ABSORPTION LOSS

!>.)

ABSORPTION COEFFICIENT A

T HOT

dB

IN

AND

s
s
There

<1.57)

is

very

IN

PERCENT

evaluated connector loss literature. The

little

values in this table should be treated as preliminary data suggesting

the general magnitude of losses to be expected.

FREQ.

TYPE

Fig.

1.

COMMERCIAL TYPE N

2.

PRECISION TYPE N

24

2 Physical parameter contributions to noise-figure measure-

ment

limit of error.
3.

COMMERCIAL WR90'
COMMERCIAL WR9u d
NBSTYPE WR90*

373 0.5 K
80 1

LOSS

(GHz)

ib

0.01.

Al%]

(dB]

<0.1

0.23,

0.1

0.01

0.02

0.23
0.46

0.4

0.04

0.9

1.0

0.04

20

008

1.8

4.0

0.11

2.6

8.0

0.20

46

0.9

<0 01

8-12

< 2.3

<0.23

9 4

0.003

0.07

(dirtyl

0.003

0.07

(lappedl

0.001

0.023

0.0004
0.0002

CMR WR90*
(lapped or not!

SPECIAL CPR WR112'

SPECIAL WR430'
1
PRECISION COAX

4.
5.
6.

14
7

mm
mm

0.01-8.5

001

COMMERCIAL WR15'

7.

in

0.009

00046

^OOOOOl

23

17

GO

0.00023

0.003 v-f
007^/f
0.007 to 0.045

0.07^1
O.IBx/f
0.16 to 1.04

CHi

IJPU pn.ateci >mmonrcaoo

Robert T Ada.r (NBSI pn.are com monicanon
R W B-atty. Proc IEEE Vol SI PP 642 642 1965
Amon, IRE Tranr. on Iminffrr, eon.
R W Beany. G F Enqen. and W

f.

IEEE Standard

g.

B Yetei 1NBS).

a.

Q
e.

Charlei T. Stllcrftid

219 226, 1960.

Utile IIVBS) private common t^tion.

Vol. 1-9. pp
E.

Vol 1M-

7.

Predion Coa-ia- -onnecion IEEE

204 222. 968

lor

pp.

Irani,

on Imtr. and Meat..

Pr.,a:e

common, ca r

:r-.

This uncertainty in effec;:ve input noise temperature

referred to here as the ..ismatch uncertainty.

be hesitant to

set

an accuracy goal better than

is

One should
this mis-

match uncertainty.

T COLD

T OT

^
200

100
Fig.

C1 - 25

"")

Figure 4, the

-100

500
T e (K)

600

700

800

3 Physical parameter contributions to effective input noisetemperature measurement limit of error>

nector loss

when
In setting

an accuracy goal for the measurement of

amplifier noise,

it

is

useful to

keep

in

am"

:fier as

'

-(0.13^J
^

300

maxmum

mismatch uncertainty is
described by Engen
It
is clear from Figure 4 t"Jt the mismatch uncertainty bethe magnitude of the difference
comes very important
between the "antenna's'' and standard's reflection coeffiJ
cient is expected to di' :r by more than 0.01
Connector Quality Table 2 is a list of connector loss
measurement results. Tbs change in F^g due to the conIn

plotted for an ideal

mind the

it

is

guide has

is

equal to the decibel loss of the connector

at the temper;;'

shown

that

duplicate these values.

state of

112

',,

a g'-.'J!

290

K. Experience in wave-

deal of care

A carelessly

is

required to

marred flange, or

happen as these standards are exchanged for each other:

in gain, in gain stability, in bandwidth, and in
linearity. In Figure 2 and 3 these ugly possibilities were

poorly mating pair can easily be worse by a factor of ten.

good practice to remeasure the noise figure after a

break and a remake of some of the connectors. Bad connectors can sometimes be spotted in this way.

changes

It is

ignored, and only the changes for the best possible ideal

by Engen"' were considered. It is

mismatch error from the measurements

amplifier described

Measurement Technique

safer to estimate

suggested below.
In principle the

standard

measurement of Te

is

simple.

connected, then the cold standard

is

is

The hot
connected

and Y, the ratio of output powers, is measured. The details

of changing from hot to cold (e.g., by changing current
through a solid-state noise source), of measuring Y, and
displaying the results depend on the equipment used. The

The change of T e with

measured by

cient can be

iniiimjiiimiiiim

a slide-

3^^i

very helpful but often the information needed to establish

is

and

screw tuner (or the appropriate equivalent) between the

"
manufacturer's literature and the prior literature 23 33 are

error limits

a change of reflection coeffi-

inserting an isolator

difficult to obtain, especially for the auto-

matic noise-figure meters. Often the straightforward (connect a hot standard and measure the output power, then

and measure the output power)

it has been
found 34 at NBS that, for 1-dB gain compression in the
power measuring system, the power level needs to be
connect

a cold standard

technique

about

is

best.

a factor of

Fig.

as related to reflection coefficient.

10 greater than the

mean

noise

power

noise standards and the amplifier as

From
if

systems. Systems operating with too

35 37

much

measurement hazard is "mismatch"

The problem of using standards having a
different reflection coefficient from the "antenna" has
already been discussed as mismatch uncertainty. The additional problem of the hot-standards and the cold standards
having different reflection coefficients from each other is
the mismatch error displayed in Figure 2 and 3. Here the
insidious

"

limits indicated.

many

in

Figure 5.

is

not exact because the

the tuner-isolator combination as the

re-

290 K) increases the noise figure by 0.015 dB.

One estimate of mismatch error is obtained by measuring T with and without the isolator-tuner shown in
e
(at

Figure 5.

reflection coefficients have arbitrary ampli-

flection coefficients are different

in

shown

the mismatch uncertainty can

change is usually
small. For example, for commercial X-band slide-screw
tuners, measured changes in absorption loss are less than
0.015 dB as reflection coefficient is changed from zero to
0.5 (any phase). Because of the definition of the noise
figure (Equation 2), a 0.015-dB increase in absorption loss

23,38 40

tude and phase within the

Te

flection coefficient changes, but this loss

gain can pro-

"

two standards

change

losses

duce serious errors

The most

this variation of

be estimated. This measurement

the error from clipping the noise is to be less than 0.1%.

This is also true for the commercial noise-figure measuring

error,

5 Block diagram for measurement of mismatch uncertainty

One precaution however:

When

the

If

the isolator-tuner

impedance, then the change

re-

in

is

adjusted to the "antenna"

noise figure should be

equal to the absorption loss or the isolator-tuner.

ugly things can

If

this

absorption loss can be estimated (insertion loss for low-

lOr

reflection elements

is

nearly equal to the absorption loss),

then any serious mismatch errors

method,

if it

will

can be implemented,

is

be spotted.

better

to attach a slide-screw

tuner to the hot standard only, and to observe the change

in

T e with the change of

cedure

is

reflection coefficient. This pro-

then repeated with the slide-screw tuner

at-

tached to the cold standard only.

Estimating Measurement Error
Figures 2 and 3 are useful for estimating

measurement

dB or T e are
direct proportion to the magnitudes of the

errors because the error contributions to F

essentially in

respective error source parameters. Therefore,

if

is

meas-

ured to within 0.02 dB. the Y-factor contribution error

bands

in

Figures 2 and 3 double. However, this propor-

good approximation of the mismatchmeasurement of T e versus rstd

should be used'for estimates of the mismatch errors. Thus,
Figures 2 and 3 can be used to translate from uncertainties
in the various measurement parameters into the error induced in Te
The cascade error is not shown in Figures 2 or 3. The
25 41
desired T e can be deduced
from the measured T |otal
tionality

is

not

error estimate, but direct

OO 5

0.2

rANT VAMP '

1

Fig.

4 Moxiit
coeffic

'

when

'

natch uncertainty

"antenna"

amplifier input reflection

reflection

the effective input noise temperature of the meas-

uring system

complex conjugate.

113

is

T e2 from

T e2 /G,

=T.

T.

where G,

is

Bureau of Standards,"
18

the available power gain of the amplifier being

measured. The cascade error depends on

how

19.

CONCLUSIONS
An

accurate measurement of amplifier noise requires

that particular attention be paid to the

mismatch

error.

"On the Definition of Noise Performance,"

Proc IRE. Vol 49 No 1 p 376, Jan 1 961
Engen, Glenn F., "A New Method of Characterizing Amplifier Noise Performance," IEEE Trans Instr and Meas. Vol
IM-19 No 4, pp 344-349, Nov 1970.
Mumford. W. W. and E. H. Scheibe, Noise Performance Factors in Communication Systems, Horizon House-Microwave,
Inc., Dedham, Mass.. 1968.
Sleven, Robert L "A Guide to Accurate Noise Measurement,"
Microwaves, pp 14-21, July 1964.
Letzter, S. and N. Webster, "Noise in Amplifiers," IEEE Spectrum, Tech Notes T-243, pp 67-75, Aug 1970.
27. Topics in Noise Figure Measurements, Airborne Instruments
Lab.. 1963.
Noise Figure Primer," Application Note 57, June 22, 1962,
28.
Hewlett-Packard.
22. Rhodes, D. R..

This error depends on several amplifier characteristics and

changes with the different terminations on the ampli-

their

fier. It is

23.

therefore important to make measurements of the

the change in amplifier noise with a change in terminating

reflection coefficient, in order to estimate this

mismatch

error.

From
80 K r

is

it

measurements
300 K or 373 K

a discussion of the state-of-the-art

for hot-cold standards of 10,000

seen that measurements of noise figure to with-

about 0.1 dB, or effective input noise temperature to

within about 2.5%, is the best that can be done. Such
in

small uncertainties are possible only at those frequencies

at

which

NBS

Spec Publ 250.

1963.
Cohn, Seymour B., "The Noise Figure Muddle," microwave
journal. Vol 2. pp 7-1 1 March 1959.
20. Greene, J. C, "Noisemanship - The Art of Measuring Noise
Figures Nc3rly Independent of Device Performance," Proc
IRE. Vol 49 No 7, pp 1223-1224.
Hendler, A. J., "Noise-Figure Measurements Definitions,
21
Techniques, and Pitfalls," Topics in Noise. Airborne Inst Lab,
Deer Park, NY. pp 1-4, Aug 1960.

accurately

G, are known.

Te2 and

NBS

"IRE Standards on Electron Tubes Definitions of Terms,

1962," (62 IRE 7.S2). Proc IEEE, Vol 51. pp 436-442, March

calibration of noise sources exist.

J. J. and G. A. Deschamps, "A Wave Amplitude

Approach to Noise Analysis with Application to a New
Method of Noise Measurement," Tech Report AF AL-TR-70-

29. Sweeney,

REFERENCES

20, Air Force Avionics Lab. AFSC, Wright-Patterson

Daywitt. W. C. W. J. Foote, and E. Campbell, "WR15
Thermal Noise Standard," NBS Tech Note 615. March 1972.
Trembath,
C. L., D. F. Wait, G. F. Engen, and W. J. Foote.
2.
"A Low-Temperature Microwave Noise Standard," IEEE
Trans MTT, Vol 16 No 9, pp 704-714. Sept 1963.
3. Trembath, C. L.. W. J. Foote. and D. F. Wait, "A Liquid
Nitrogen Cooled Microwave Noise Standard," Rev Sci fnstr.
Vol 42 No8, PP 1261-1262, Aug 1971.
4. Wells, J. S.. W. C. DaYwitt, and C. K. S. Miller, "Measurement of Effective Temperature of Microwave Noise Sources,"
IEEE Trans Inst Meas, IM-13, pp 17-28, March 1964.
5. Wait, D. F.', Seminar Coordinator, Notes for High- Frequency
and Microwave Noise Seminar, NBS, Boulder, Colo., 21 Apr 1 May, 1970. (Unpublished).
6. Miller, C. K. S., W. C. Davwitt.and M. G. Arthur, "Noise
Standards, Measurements, and Receiver Noise Definitions,"
Proc IEEE, Vol 55 No 6, pp 865-877, June 1967.
7. Miller, C. K. S.. W. C. Daywitt, and E. Campbell, "A Waveguide Noise-tube Mount for Use as an interlaboratory Noise
Standard," Acta Imeko III, pp 371-381, 1964. .
8. Wait, D. F., "Sensitivity of the Dicke Radiometer," J Res
NBS, 71C (Engand Instrl, No.2, pp 127-152, April-June

30.

11.

33.

Trans

36. Cohn, C.-E., "Errors in Noise Measurements due to the Finite

Amplitude Range of the Measuring Instrument," Rev Sci

Vol 35, pp 701-703, June 1964.
and H. V. Hance, "A Note on Receivers for Use
in Studies of Signal Statistics." IRE Convention Record.
Part I. Information Theory, Vol 1 po 7-13, 1953.
38. Imboldi. E.. "Transmission-Line Mismatches and System
Noise Figure," Electrical Communication, Vol 40 No 2. 1965.
39. Brady, M. M., "The Influence of Mismatch Error in Noise
Performance Measurements," Proc IEEE. Vol 52 No 9, pp
1075-1076. Sept 1967.
40 Pastori. W. E., "New Equation Analyzes Mismatch Errors in
Noise Measurements," Microwaves. Vol 7 No 4, pp 53-53,
Instr.

37. Deutsch, R.

April 1968.

41

pp 636-640. Sept 1969.

17. Mason, H.. Ed., "Calibration and Test Services

January

9,

1973

Temoerature Contriin a Low-Noise Receiving Svstem," Proc IEEE Vol

76-77.
Jan
1964.
52. pp
David F. Wait received the
Ph.D. degree in Physics from
the University of Michigan in
1962. In 1963 he was a senior
scientist with Lear Siegler, Inc.,
Ann Arbor, Michigan. From
1964 to the present he has
'been with the National Bureau
of Standards. Boulder, Colorado,
and was the coordinator of the
Stelzried, C. T., "Post-Amplifier Noise

bution

1970 "High Frequency and

Microwave Noise Seminar."
IEEE. MTT Group (Chairman 1970.
Vice-chairman 1969 of Cwnvcr-Boulder Chapter).

in

Dr. Wait

of the National

r~\

13.

No

Reeve, G. R.. (Nat Bur Stand. Boulder). Private

communication.

MTT,

Vol 16 No 10. pp 866-873, Oct 1963.

AMred, C. M., "A Precision N"cise Spectral Density Comparator." J Res NBS. S6C I Engand Instrl, No 4, pp 323-330.
Oct - Dec 1962.
14. Arthur, M. G.. C. M. Allred, and M. K. Cannon, "A Precision
Noise-Power Comparator," IEEE Trans Instr and Meas, IM13. No 4. pp 301-305. Dec 1964.
15. Arthur, M. G., "A Precision HF-Noise Power Measurement
System." ISA Trans. Vol 10 No 3. pp 264-268, 1971.
16. Engen, G. F., "A Method of Calibrating Coaxial Noise Sources
in Terms of a Waveguide Standard." IEEE Trans MTT, Vol 16

Lebenbaum, M., "Measurement of Noise Figure," Handbook

of Electronic Measurements, Vol 2 Chap 1 pp 1-51, Poly-

35. Bell, R. L., "Linearity Range of Noise-Measuring Amplifiers,"

Wireless Engineer, pp 1 19-122. Apr 1947.

and T. Nemoto. "Measurement of the Noise

Temperature of a Mismatched Noise Source," IEEE Trans
Vol 16 No 9. pp 670-675, Sept. 196S.
Wait, D. F., "The Precision Measurement of Noise Temperature of Mismatched Noise Generators," IEEE Trans MTT,
Vol 18, pp 715-724, Oct 1970.
Wait, D. F., "Thermal Noise from a Passive Linear Multiport."
IEEE Trans MTT. Vol 18 No 9, pp 637-691 Sept 1968.
Nemoto T. and D. F. W ait, "Microwave C.rcuit Analysis

IEEE

9.

Wait, D. F.

using the Equivalent Generator Concept,"

ESSA Tech

technic Inst of Brooklyn, 1956.

34.

12.

Factor Measurement,"

Haus, H. A., et al, "Standards on Methods of Measuring

Noise in Linear Twoports," Proc IRE, Vol 48, pp 60-68, Jan
I960.
32. Haus, i1. A., et al, "Representation of Noise in Linear Twoports," Proc IRE, Vol 43, op 69-74, Jan 1960.

MTT,
10.

"A Note on Noise

Report ERL133-WPL

Hildner, E. G.,

31.

1967.
9.

AFB,

Ohir-.

1.

114

is

member

of

tJ'e

29

Reference

Tr TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT,

A New

Method

VOL. IM-19, NO. 4,

NOVEMBER l970

of Characterizing Amplifier

Noise Performance
GLENN

Although the use

of noise figure or noise temperature

performance is a well-established practice,
this
parameter provides only a partial
also recognized that

Abstract

this

to characterize amplifier
is

It

ENGEN

F.

that

paper to put this description in an alternative form

provides additional insights for certain problem

areas.

description of the amplifier noise properties. In general, the noise

depends upon the generator impedance and

Is thus a function of the signal-source and amplifier combination.
Typically, the noise figure is measured by the Y-factor method
using hot and cold noise sources that are nominally matched (reflectionless). The result of this measurement is of value as a figure of
merit; however, if optimum performance is to be realized, the applications engineer must know whether to adjust the signal source impedance for maximum power transfer, minimum noise figure, or
some other criterion, and he must know the deterioration in perforfigure (or temperature)

mance

that results

if

Undoubtedly

method

of characterizing amplifier

number

in

terms

prove more instructive,

from an elementary
this provides the additional benefit that the
of earlier results. It will

however,
model;

give

to

presentation

noise per-

is

I.

"self-contained."

that

linear

is

and includes noise

it is

of the amplifier noise properties.

More

where
1 and

recently, the use

has been proposed, but again, this

[1]

provides only a partial description. As usually defined

made

of a given amplifier,

ports

<Sl2<*3

&1.

(1)

62

S2iai

S72 <h

&2

(2)

respectively,

power

is

for two amplifiers to hi.ve the same noise

and yet have markedly different sensitivities to

bi n

i 3

au~the

a&i.

b3

one of putting this description in the most convenient

S =

|Si 2

certainly not to suggest that this point of view

is

negligibly small.

it will

In addition,

be
it

for the present discussion to

assume

condition imposed on b^ and a

by the

convenient
0.

The boundary

form.

of

(3)

In keeping with the usual design objectives,

assumed that

sum

completely correlated
uncorrected. Thus,

first aft,, is

with & ln the second b i0

generator

is new, and indeed a number of useful descriptions have

been given (for example [3], [4]). It is the purpose of

Manuscript received June 4, 1970.

The author is with the National Bureau of Standards, Boulder,

6 2 , such that the

two components: the

operation in arbitrary systems, and

is also

Colo.

prove convenient to express

available in attempting to optimize the

system performance. It is quite possible, for

is

a? are the corresponding

are the scattering coefficients

and matched (reflcctionless) terminations on

2
and 2 is given by |&i| and |&j| r respectively.

changes in source impedance. In order to account for

this and other phenomena, it is necessary to begin with
a "complete" description. The remaining problem is

This

and a u

The S

example,
figures

emergent wave amplitudes at ports

6 3 are the

2,

It will

the options
overall

SllOl

represented by the terms 6 ln

performance characteristics must be known in sufficient

detail, to predict its

6i,

passive
to be

in Fig. 1,

spectral density (per cycle of bandwidth) delivered to

combination.
is

shown

of the two-port. Finally, the internal noise sources are

[2],

a function of the signal-source and amplifier

optimum use

sources). This device,

61

incident waves.

the noise figure depends upon the generator impedance

its

a two-port

provides gain

also recognized

that this parameter provides only a partial description

If

is

it

of noise figure or noise

a well-established practice,

of noise measure

in that

conveniently described by the equations

temperature to describe amplifier performance is

the amplifier

for

and active (both

Introduction

ALTHOUGH the use

is

derivation

of

is

and thus

its

The model chosen

parameters that provide ready answers to these

questions. In addition, the measurement of these parameters via a
simple extension of the Y-factor method will be described.

formance

would be possible to obtain the material

it

that follows as an extension or reinterpretation of a

this is not done. It is the purpose of this paper to

present an alternative

General Theory

II.

is

conveniently expressed in the form:

a,

+ 6,r.

(4)

1
If Sn = 0, the term Sn is only of interest in attempting to
optimize the coupling between the amplifier and the system that

follows

115

it.

ENOEN: CHARACTERIZING AMPLIFIER NOISE PERFORMANCE

r

sn

5,

345

r
1I

i>..r

T.

"r<

minimum

amplifier noise temperature (as

a function of source impedance) referred

to amplifier input,

1_ /

b,

the source term and r,

is

is

the generator re-

In order to examine the dependence

of the amplifier noise upon the source impedance, it is
convenient initially to assume that b e = 0.
These equations and assumed conditions may now

flection coefficient.

be combined to yield the following expression for the

noise power delivered to a (matched) load at the amplifier output terminals:

16,1'

This

result

<*

- sa f,\
displays

explicitly

source impedance V, and this

I6..I'

may

physical significance of

be interpreted as

what extent the conditions for maximum

power transfer and minimum amplifier noise differ.
expressing to

In (7), the parameters T G, b, and |/3| are dependent

on the amplifier properties only, and moreover their
values are "terminal invariant" (i.e., their values do
not change if a "lossless" tuner is added to the amplifier
input). The entire dependence upon the generator impedance is found in the term TJ.

Evaluation of Parameters

(5)

|6,.r*.

The terms Ta
dependence upon
dependence is a rather
the

may

b,

and

|/S|

may

be evaluated by an

extension of the F-factor method.

Let a temperature T, be assigned to the generator of

Fig.

substantial simplification in this expression

- |M)

gfl

-s?i

III.

complicated one.

to T.,

/Sji

The
Here, 6,

of the noise temperature of

iput port*
theinpu

|S|V

IB

o,

Basic circuit for discussion of amplifier noise.

Fig. 1.

_ ratio

|6,.|*

T. VI

be

1.

t.

Then

(7)

T,G(i

by introducing "terminal invariant" parameters.

(Tor an earlier discussion see [5].) The desired trans-

As

formation begins with the substitution^

calls for the

becomes

|r;|

TjG{\

b |r;

f).

(8)

effected

method, the measurement procedure

in the ^-factor

use of "hot" and "cold" loads (noise sources).

will be designated T, h
T, respectively, and
T. c the corresponding outputs when the noise sources
are connected to the amplifier input. As a further condition it is assumed that the impedances of these loads
(see the Appendix). After the obis such that T'a =

These

rj

(r.

so that (5) can be written

T.

|&|'

S*)/(i

- Su r.)

(6)

277(1

+b

|rj

(7).

0fj

?",

servation of T, K and

Tac

a sliding short

where T

is

the noise power output (to the matched load),

is

connected to

maximum and minimum

the amplifier input and the

values of T observed in response to motion of the short.

These will de designated T, u and T. m respectively.

These four observations provide a set of simultaneous
equations, which can be solved to yield
,

G =

amplifier

power gain,

i..r

v.* -

101

2|(r

This transformation has the following properties: 1) T Q

when the conditions

for

maximum power

r.J(r.
2T. t )(T, k -

+ T '

t)
2T(T.>

'

(T.
4

T.b

transfer are satisfied

3) the value of !l7! is

(r = Sii*), 2) jr/l = 1 when |r| = 1,

"invariant" to the choice of terminal surface between generator and
amplifier, provided th:it this choice is limited to a lossless region.
For a more complete discussion, see |G|.
1
Although the functional form of this equation is equivalent to
that obtained by a number of earlier authors, it is important to
tecognize the following distinctions. I) The emphasis is on noise
temperatures instead of noi.se figures. 2) In addition Ut "characas r o
terizing the rapidity with which T a increases ab^ve Ta
departs from 0," the term 6 is also the ratio between the noise
temperature of the input port and T '.i) The complete dependence
upon source impedance is contained in the single term IV- A) The
behavior with either power matching or noise matching is imtnedu t'v evident from inspection.
fl

TJ +

|/3

0)

T. c )\

- t.<)
- TJ

r.)(r
1

(r.

- {TJT)T _
{TJT) - 1

T.,

T,

(10)

|P|

T.

4
This is the available noise power (per cycle of bandwidth)
from the input terminals expressed as a temperature.
6

In obtaining this

result, it is

helpful to note that

116

(ID

(i

(i

ir.i'Xi

|r,'|)
|i

isp

-sr,|=

ON INSTEUMENTATION AND MEASUREMENT, NOVEMBER 1970

IEEE TRANSACTIONS

The

last

two equations nan

By

also be solved explicitly

and 6 :n terras of the observed quantities; however,

the suggested computation follows the indicated stepj.
[Note that (10) requires the value of \ff\, and (11) the
value of both |/3| and 2'6.] Finally, 6 is determined by
taking the ratio of (10) and (11).

use of the calculus, this

maximum

occurs for

for T

D =

term on the right

in (11) is just the usual one for determining amplifier
noise temperature via the F-factor method. In addition
it may be recognized that these equations call (implicitly)
only for ratios between the different output temperatures
(e.g., T ok /T. c etc.). If the absolute value of these terms
is measured, it is also possible to obtain the amplifier
It

is

|r;|

- Vi -

(\/D)[i

d'}

(is)

W)}.

(16)

where

of interest to note that the first

(26 |0|)/[1

shown that

It can be

D <

+
1

6(1

and

can be expanded

(15)

to yield

|r:i

AM.
6(1

|/

to complete the descrip-

is the argument of 0. This may be obtained by

computing the argument of T't that provides r. This
argument may, in turn, be obtained from (6) provided
that S n and the corresponding argument of r, are known.

tion

These

signal-to-noise ratio at the amplifier output)

to

know whether

for

maximum power

or according to
that results

if

transfer,

this

minimum
criterion,
8

not done.

is

and the penalty

will

now

be assumed that the signal

impedance and that its available
power includes a signal component S and a noise component kT,B. The signal power S at the amplifier output

source

is

\fi\>)]>

+ I

(17)

complexity, this result will not be

its

small, however, the following approximation

S.

kB[T,

r.(l

6 |017(1

It is of interest to evaluate the potential

over matching oa a
noise basis.

amplifier noise,

These topics

is

N.

The

from

in turn,

some other

be investigated.
Returning to Fig.

necessary

it is

adjust the generator impedance

to

|/3|

S.

maximum

the

useful:

Assuming that the applications engineer wishes to

(i.e.,

can now be substituted back into (14) to

Because of

ratio.

given. If

Power Matching Versus Noise Matching

obtain the best overall performance,

results

6(1

obtain an expression for the optimum signal-to-noise

is

IV.

is

[1

The only remaining quantity

M)

(6

gain.

maximum power
by

letting rj

improvement

transfer or

may

required expressions

(11)

(18)
6))]

minimum

be obtained,

and r;

0. If

the

subscripts p (for power) and n (for noise) are used to

identify these two cases, a little manipulation leads to

the following approximate expressions:

1, it will

of variable

(SJN ,)\
(SJN.)\ m

and

given by

S.

S,G(1

(S./N.)\.

(12)

\T'f)

(S/iV)L
while the noise power output 7V

is

given

by
It is

N. = kT,BG(l

The

|rj|

s.(i

N
By

kT,B(l

|r;i

inspection, for

the

at

lr;i

|r;! )/[i

).

(13)

output

b |r;

(14)

2
/3|

small

worth noting that the potential improvement is

|/3| is small, but increases rapidly as |0| increases.

if

V. .Experimental Results

is

fcr.S(i

In an effort to explore the method further, the measurement procedures described in Section III were applied

number of different amplifiers.

The first item to be evaluated was an X-band

to a
signal-to-noise ratio, the

argument of T'a should equal that

problem is that of maximizing
(i

amplifier

/S|

maximum

kT.BG{\ +b\T',-

ratio

signal-to-noise

&

6(|r;|

of

jS.

The remaining

tunneldiode amplifier. In this case the measurement was straight-

forward and yielded the results: \&\ = 0.03, 6 = 0.35,

!T = 825 K. It should be noted that if the circulator,

|/3|)

which

].

8
The point of view adopted here is that there are many situations
which thfi applications engineer might hi willing to adjust the
impoditr.ee inarching bar. unwilling to tackle the more ambitious
problem of also adding inverse feedback. For a discussion of this
more general problrm, see (H, see also, J. S. Engberc, "Simultaneous
input power match and noise optimization using feedback," M.S.
thesis, Syracuse Universltv, Syracuse, N. Y., July 1960.

above

in

The
|/5|

is "ideal," this would

and bT = ambient temperature. The

results

agree with the theoretical expectations.

is

lead to

required in such devices,

|/3|

evaluation
0.13, 6

of

an A'-band

0.65, T

crystal

mixer gave

= 496 K. Again the

it may be noted that

results

67",
appear plausible enough;
only slightly above ambient while the value for T,

117

is
is

ngen:

characterizing amplifier noise performance

half the usually

method

utilized

347

quoted value since the measurement

both the signal and image channels.

with "available" noise power.

used, no change in (7)

is

With regard

In spite of this plausibility, however, these results are

some question in that a

bias was observed in response

subject to

to the short motion.

TJ
is,

The source

of this

shift in bias resulted. This, in turn, gives rise to

An

application of the
|/3|

method

0.22, b

method

some

0.59, T

were believed to be in error in that

an input-port temperature of 94 K. Some
further effort was expended in attempting to refine the
impedance matching, but this proved to be a difficult
broad-band problem in that the bandwidth of the deInitially these results

\t:\

tection

r.

system was a significant fraction

An

frequency.
finally

2((r.

2r)(r t

(t.

by an evaluation

the

of

b,

a more general

prescribe

to

and

|/3|

than that outlined

it is convenient to

response can be substituted for the sliding-short operation

The appropriate equations

described in Section III.

r)

r.)(r. t

2[t

- t)
- t..(i -

\v'.\

(T.

T.b

was

T. h

T.

)](T^

r;

input-port

temperature using a narrow-bandwidth radiometer. This

gave the value 120 K, which may be regarded as a "confirmation" even if the agreement is not very good. It is
unfortunate that it has not, as yet, been possible to
repeat the entire evaluation with this narrow-band

By

are

(22)

t.,)\

- T)
- T)

T,)(T, h
|/3

(T. h

a proper choice of \T'S

it

(23)

(T.JT.JT,.

(T.JT)

T.b.

(24)

should be possible to avoid

noted earlier. On the

other hand, the realization of the required noise source,
the

nonlinear operating

region

which the argument of TJ is adjustable without

changing its magnitude, is somewhat more of a problem.
One possible solution is to add a tuning transformer
to the amplifier input terminals and adjust for an impedance match. To the extent that it may be considered

radiometer.

for

Although these results represent only a cursory evaluation of the method, they do call attention to certain
practical problems that must be considered in its application.

VI. Extension to

lossless,

More General

Sn =

In the preceding analysis, a number of simplifying

were introduced. These will now be

is

b,

r.

now

and

An

|/3|.

For the tuned

amplifier,

adjustable generator (at ambient

conveniently provided by a matched

attenuator followed by a sliding short.

At

in greater detail.
easily generalized to

and V, =

temperature)

approximations

The theory can be

the presence of this transformer will not affect

the values of

Operating Conditions

examined

possible

is

for evaluating

of the operating

indirect confirmation of this result

obtained

it

Returning to Fig. 1,
assume for the generator a noise source of temperature
T whose impedance can be adjusted in such a way
as to vary the argument of T't while keeping its magnitude
constant. Inspection of (S) shows that the first term
on the right will remain constant while the second will
go through maxima and minima. An observation of this

this implies

can be shown

upon IV

T',)

in Section III.

30-MHz vacuum= 161 K.

to a

it

in the form of (7), even when S 13 9^ 0. There

however, a dependence of the different parameters

Finally,

a linear manner.

tube amplifier gave

postulate,

is still

(including

question as to whether or not the device was behaving

in

<S, 2

the second interpretation

that the functional dependence of output power upon

substantial shift in

problem was due to a substantial

coupling of local-oscillator power into the input arm,
and when this was reflected back into the system, a

crystal

to the

If

required.

is

account for

this point

it is

possible to consider the tuner either

as part of the amplifier or part of the adjustable noise

arbitrary impedance conditions ( 22 r 0, etc.) at the

output port. In order to retain the terminal-invariant

generator.

character of G,

impedance requirement on the "hot" and "cold" loads

(required to complete the evaluation of T b, |j}|) is
now that these are also matched. In the projected evalua-

it

redefined as the ratio of the net

is

power delivered (to the amplifier input) to the available

power at the output port. (That is, the definition that
follows

(7)

should contain the factor

The right-hand side of

2
by (1 |r;| ) where

(12)

and

(1

- l^iV)-

(13) should be multiplied

when

The

former

be negligible and further study

VII.

and
load.

T,

is

|i

noise

the reflection coefficient of the terminating

The same

is

true of

(7),

provided that T

An

(21)

- S n r,\
is

as-

sociated with the "delivered" noise power as contrasted

is

convenient

tion of cryogenic systems, however, these losses

st
in

interpretation

the tuner losses are negligible since the resulting

alternative

is

may

not

indicated.

Summary

method

has been presented,

of

characterizing

which

amplifier

(assuming linearity)

completely accounts for variations in source impedance

and in which all the parameters have a simple physical

meaning. With the exception of the argument of

118

J,

these

IEEE TRANSACTIONS

ON INSTRUMENTATION AND MEASUREMENT, NOVEMBER 1970

parameters are "terminal invariant." In addition, methods

for measuring these terms have been described and an
analysis, leading to optimum source impedance, presented.

The

experimental

preliminary

evaluation

suggests

that the methods should prove of value in system optimization, although confirmation of this awaits further

work.

In particular

it

appears that the method

real value in the evaluation of parametric-

and

ticipate larger values of b

On

is

|/3|

amplifier at least, this expectation

is

by recognition that these devices typically exhibit a

of linear operation

Measurement

Fig. 2.

system, for recognizing conjugate

impedance

match.

seriously qualified

substantial sensitivity to changes in source impedance,

and the region

H^EH

of

some reason to anthan those reported

the other hand, in the case of the parametric

type amplifiers, since here there

above.

may be

and maser-

may

be difficult to

r d or T g thus there
assuming T d = V = 0.

either

Combining these

no

is

loss in generality

results with (25)

and

(26)

in

leads to

establish.

The

moving short alone

possible use of a

uation of maser noise has

From

appears that

it

(7)

it

determine an upper limit to

At the

provided that

is

known.

appears

be

to

the

in

assumed amplifier

changes in source impedance,

not output versus input). However, if this method puts
this aspect in better focus, even this may prove to be a
linearity

(in

respect

= K

&i

(iS 12 02

13022)!^
1

~f~

Q 13
(30)

- s 22 r.

indeed possible thus to

is

the most serious limitation to this

present,

description

for the eval-

been previously suggested.

to

Next

it is

assumed that

Tx

has been adjusted such that

when port 2 is terminated by the load of

interest (r, = r), and T, has been adjusted such that
|6,| is constant in response to a moving short at port 2.
The conditions imposed on the scattering parameters
by these adjustments are, respectively,
vanishes

6,

contribution.

S +

(S, 2 S :3

S 13 S 12 )V. =

(3D

(32)

Appendix

The measurement procedure

outlined in Section III

adjustment of the hot and cold load impedances to equal the complex conjugate of the amplifier
input reflection coefficient r. In principle, this can be
effected by measuring this impedance (by use of a slotted
line, for example) and then using the impedance meter
to recognize when the hot and cold load impedances
are the conjugate of this value. An alternative procedure

and*

calls for the

"13

"T"

(S l2 S 23

By

inspection,

s =
Equation

(26)

may now
bt

it is

<S l lS 22 )lS*2

follows.

Referring to Fig. 2

r*.

(33)

be written

= S 23 6

T*a,

(34)

convenient to assume a system

comprised of a directional coupler, tuning transformers

(T T y ), load isolators, generator, and detector as shown.
Scattering equations can be written for this system as

or, stated in

words, the equivalent generator that obtains

at port 2 has the source reflection coefficient

obtain a termination with this reflection,

r*.

To

only neces-

it is

arm 3 by a passive load

same impedance. The purpose of the isolator

sary to replace the generator on

follows:

of the
6,

= Sa,

Sa,

S 13 a3

62

= S 21 a,

S 22 a 2

S 23 a 3

The boundary

conditions

imposed

by

(25)
(26)

the

different

terminations are

on arm 3 is to minimize the requirement that this impedance be the same as for the generator and to avoid
possible pulling effects with the short motion.

on arm 1 avoids the requirement

impedance with changes in power
Finally,

a,

fc,r d

(27)

by anj

Oi

& 2 Tj

a3

b,

(28)

b 3 T,.
is

level.

(29)

no occasion to change

119

may

(e.g.,

be effected
of the procedures for recognizing the equality

of two impedances.

In the operation to follow, there

isolator

the adjustment, of other impedances

the hot and cold loads) to equal this value

r

The

of a constant detector

For a discussion of this point, see

See, for example, [6].

[7].

IEEE TRANSACTIONS

ON INSTRUMENTATION AND MEASUREMENT,

VOL. IM-19, NO. 4,

Acknowledgment
The author expresses his appreciation to O. L. Patty
and M. G. Arthur for providing the experimental results

[4j

NOVEMBER 1970

349

H. A. Hans, W. R. Atkinson. G. M. Branch, W. B. Davenport,

Jr., W. H. Fonger, W. A. Harris, 8. W. Harrison, W. W. McLeod.
E. K. Stodola, and T. E. Talpey, "Representation of noise in
linear twoports," Proc. IRE, vol. 43, pp. 69-74, January I960.
Lange, "Noise characterization of linear twoDorts in terms of
invariant parameters," IEEE J. Solid-State Circuits, vol. SC-2,
pp. 37^40, June 1967.

[5] J.

described.

References
[1]

[2]

H. A. Haus and R. B. Adler, Circuit Theory of Linear Noisy

Networks. New York: Wiley, 1959.
"IRE Standards on electron tubes: Definitions of terms, 1957"
(57 IRE 7.S2), Proc. IRE, vol. 43, pp. 933-1010, Julv 1957.
Proc.
H. Rothe and W. Dahlke, "Theorv of noisy fourpoles,
IRE, vol. 44, pp. 811-818, June 1956.

[6]

[7]

[3]

"An introduction to the description and evaluation

microwave systems using terminal invariant parameters,"

G. F. Engen,
of

NB3 Mono.

112.

G. F. Engen,

"A

variable impedance power meter

TRANSACTIONS ON
INSTRUMENTATION AND MEASUREMENT

Reprinted bv permission from IEEE

Vol. IM-19, No.

4,

November 1970

Copyright < 1970, by the Institute of Electrical an.] Electronics Engineers

PRINTED IN THE
NRC 11S48

120

and adjustable

reflection coefficient standard," J. Res. Nat. Bur. Stand., Sect. C,

vol. 68, pp. 7-24, January-March 1964.

'

U.S.A.

Inc

Reference

[8]

Thermal Noise from a Passive Linear Multiport

DAVID
Abstract

The

thermal noise from passive multiports

from fundamentals so that

it

discussed

is

In Section Injunctions with uniform physical temperature

can be understood, measured, or calculated

by a microwave engineer. The multiports are assumed

to

are considered.

have a uniform

The
= AT where
T is its physical temperature and A is its absorption coefficient. An approximate method of measuring A, and a method of measuring an A as small
is

For them, the

noise contribution

is

directly

proportional to their physical temperature. The constant of

temperature, but with no restrictions on reciprocity or mismatch.

noise temperature, TV, contributed by such a multiport

WAIT

F.

proportionality

7\

is

referred to as the absorption coefficient

reflection coefficients are derived. Finally, the crosscorrelation of the noise

which may either be measured (Section III) or calculated

(Section IV). The calculated form is stated in terms of scattering matrix elements and reflection coefficients of the port
terminations. In Section V, the correlation of noise from
different ports is discussed. Multiports with almost reflec-

from

tionless terminations are treated in Section VI. In Section

as 0.008 within 5 percent, are pointed out. Also, exact and approximate
expressions for

terms of scattering matrix elements and termination

in

different ports

is

briefly considered.

VII, the absorption coefficient

ORE THAN

known

Introduction

I.

40 percent of the noise

M:

in low-noise

port switchable circulator,' or a four-port hybrid

Neglecting
perature

is

sometimes inconvenient to

language, but

may

more

it

quired to reexpress the results

is

similar

in

more

useful terms.

is re-

What

is

if ail

is

at a different temperature, then the

known

expressed as a temperature

is
/.

coefficients of the other port terminations,

is

available output of port

j,

196S; revised

May

where

elFect of the

T,

is

generator at

the available
its

own

aij Tj

power

in

(i)

spectral density of the

output port and TV

proportionality factor a u

is
i

the-

multiport

the ratio of the

to the

y'th

the remaining

is

contribution due to the thermal sources within

The

total

r.v.,

density available from the generator on port_/

n. Ilffi.

power

power

spectral

when only

the

This re-

time dependence of the noise on system sensitivity

The

available noise power spectral density is detined as the maxinoise power spectral density that could be delivered to a toad.
maximum power when the load has the
proper rejection coelticient. No manipulation of the rellection eoelli2

mum

The

Colo. S0302

The

shown

so

spectral density available out of port

is

Because the sources are independent (thus uncorre-

I.

not belabored.

search was supported by the Advanced Research Projecls Agency of

the Department of Helcnsc. under I'rojccl 1)1 T.NDI.R.
The author is with the National Uureau of Standards. Boulder,

is

the reflection

multiport with independent port generators

Fig.

itself.

not known, this

2,

i.e.,

coefficient "'looking*' out of the multiport.

cted) they each make an additive contribution to the

tee.

Since a tractable extension to mis-

Manuscript received Ft.bru.uy

.Thus defined, the value of the noise temperature depends on

which port is being considered and also upon the reflection

might be useful to a
microwave engineer. The results are derived from fundamentals to make them more understandable. An exception
is the discussion of the correlation of noise from dill'erent
ports where the results can be obtained from the literature

is

due only to thermal

is

as the noise temperature of the multiport at port

spectral density- at, say. port

linear multiports in various terms that

difficulty.

is

sources within the multiport. The acaitable noise power

This paper will discuss the thermal noise from passive

little

a uniform tem-

ports of a multiport are termi-

radiation emerging from any port

a circulator, but relatively useless in estimating the noise

matched multiports

at

Boltzman's constant and

nated with loads at the absolute zero of temperature, while

depends on the details of the system involved. Thus, a

laboratory measurement of the scattering matrix parameters
might prove helpful in estimating the noise contributed from

with

is

available in each

the available noise from a "passive"' multiport

the multiport itself

useful

contribution from a hybrid

way

described. Specifically,

use, especially if the

because a conversion

thermal source

TB. where k

Multiport

Passive

power P

a temperature, and conversely, a temperature can be used to

be due to unfamiliar mathematical

likely,

the

express an available power spectral density. In a somewhat

elements used to terminate the multiport are mismatched.

This inconvenience

T is P =k

effects,

B is the bandwidth. Thus, the available power per unit bandwidth (available power spectral density) may be expressed as

or

tee,

quantum

waveguide mode from

other multiport element contributes to the system noise, the

literature

The Noise Temperature of a

II.

communication systems and 95 percent of the

noise in low-noise radio astronomy receivers is due
to "black body" radiation from passive linear elements.
Likewise, this type of noise accounts for most of the noise
in a Blum radiometer [I J, [2] used at the National Bureau of
Standards to compare cryogenic noise sources [3].
The literature on the subject of noise is very extensive.
Some of it applies to the problem of describing the noise
from multiports like those used in modern low-noise communications and measuring systems [4]-[7]. But if one
wants to determine how much of the noise power a three-

related to loss concepts

is

for two-ports. Conclusions are given in Section VIII.

circuit will deliver the

cient of

treated by [8 J.

121

any of the port generators

is

implied.

ON MICROWAVE THEORY AND TECHNIQUES, SEPTEMBER

IEEE TRANSACTIONS

where Tis the physical temperature of the junction. Thus to

measure A one needs to adjust a set of noise generators 4 to
have the same known noise temperature T, attach them to

T,

r.

the ports of the junction to be measured, and then measure

the available power spectral density [11],

it,)
j

Tj

r,

1968

TM *\,i

from the

output port. Because the absorption coefficient depends upon

T(

the reflection coefficients of the port terminations, the noise

generators used

in

the measurements must have the appropri-

ate reflection coefficients.

Fig.

1.

multiporl with independent thermal noise generators.

generator^' has nonzero amplitude.

tors

ay depend on

terminations.

The proportionality

fac-

the reflection coefficients of the other port

should be intuitively obvious that a

It

also

is

2K when T

power from the generator on porty when a

used on porty. 3 Thus a, ;

is

power

ratio.

The

will

CW source

be referred to as the available

noise temperature of the multiport

TXni

is

a characteristic of the multiport and the reflection coefficients

of

terminations but independent of the amplitudes of the

its

port generators.

To

determine

7",v.

consider the multiport

Then from

as well as

7"totai..

all

the

when

is

it

in

port generators at tempera-

all its

second law of thermodynamics,

the T's

must equal

T,v.,

T.

cient

be measured within 5 percent.

is

needed, a simpler measurement will

mate thermal contribution of

suffice.

For

a two-port placed between

antenna and a low-noise receiver as shown

ex-

in Fig. 2.

an

The

two-port might be a directional coupler terminated with a

when

fired,

adds a known amount of noise

When

to the system for calibration purposes.

the terminated directional coupler

may

the tube

is off,

be treated as a two-

port at uniform temperature.

For small Ai and

7"

r,

(4)

becomes

A.~0.230AP dB (l

(2)

where iPdB

where

and (Ta -T) can be known to

as small as 0.008 (a two-port loss

ample, consider the problem of determining the approxi-

Thus

= A,T

[8]

can be measured within

In case only an approximate measure of absorption coeffi-

gas tube which,

thermal equilibrium with

ture T.

may

of 0.03 dB)

800K
Thus A,

rtota i,)

10

is

2.5 percent [12].

equal to the ratio of available power out of port /to the available

than one tenth, then

If the reflection coefficients are less

the small difference of (Ta

is

T/T)

(5)

power ex-

the relative change in available

pressed in decibels

when

the two-port

is

inserted into the

system.

-{'-4-

A,=

Ignoring the receiver noise,

(3)

measuring the

relative

change

APd B
in

can be determined by

power when

a noise source

feeds the receiver directly (from reference plane

The constant A,
port

is

here called the absorption coefficient for

of the multiport. For a two-port. (1) reduces to the

form given by Daywitt

is

the ratio of available

power

at the input,

HI.

The

[10]:

7".

power

and Ts

is

ot3l

= a7s+(l a)T,

where a

at the output to the available

the temperature of the source.

Measurement of the Absorption Coefficient

losses within a junction

can change the power avail-

able at the output port, and the

amount of the change can be

loss. For example, if the

and each port of the junction is termi-

used to determine the absorption

junction

is

lossless

nated by a noise generator with the noise temperature T.

then the output from the junction
if

7",

ora

i.,

would be T

But

the junction has a nonzero absorption coefficient Ai, then

from Section

II

the difference between T

and

7"

compared with when a source with

same

A.(T,

2).

The

noise sources used must have the

tion coefficients, amplitude,

i

typically less than 0.15 percent at A"-band for reflection

is

neglecting the receiver noise

T)

in Fig. 2)

same reflecand phase as the circuit they

replace 5 (f and I\, respectively) so that the same fraction
of the available power is delivered to the receiver for both
parts of the measurement.
The precision of measuring 1P.:b depends on the precision
of the readout of the receiver, on the stability and resolution
of the receiver, on the noise temperature of the receiver, and
on the relative accuracy of the noise sources. If a single noise
source with adjustable reflection coefficient, like that shown
in Fig. 3, is used for both parts of the measurement, the
change in noise temperature due to the change in tuner losses
plane

coefficients in the range of zero to 0.3 [14].

noise tempera-

ture feeds the receiver through the two-port (from reference

is

tma i.,

the

(4)

where

rrec

is

is

The

on the order of

error due to
(.4,

T /T),
r

the receiver noise temperature.

that the value a,-> is the same for

beyond the scope of this paper. The foundations for such a proof are in Middleton [')], however, the route to the
proof may be more obvious from the appendices of the paper by the

1
An "internal" attenuator is the usual way to modify the noise
temperature of a noise source.
1
A simple procedure for adiusting the reflection coefficient of one
circuit to be equal in amplitude and phase to another is described by

author

Engen

The rigorous mathematical proof

noise as the

CW

signals

is

[8J.

122

[13].

THERMAL NOISE FROM A PASSIVE LINEAR MULTIPORT

WAIT:

689

deleted.

Thus

for is*/

(1"

(1-

receiving system with a two-port between

the antenna and the low-noise receiver.

(10)

\D,

so that

At-l-Z

^=^

For example,

CIS OISCWKCE
nose sounce

An adjustable reflection coefficient noise source. If necessary,

the noise temperature of the source is reduced by the level set attenuator so that the source does not saturate the receiving system.
(1

Calculation of the Absorption Coefficient

know

useful to

it is

the relationship between the

(l- |r*h|z> s
(i- |r.h|D,.o|'
.-

(11)

for the three-port, 7

Fig. 3.

At times

|Z>r.-,

if.

Fig. 2.

IV.

rj|)

\Sij

SxkTk

\sts

(i-sr*)l

- SuTi)
-skj v,-

(12)

-SitTt

(13)

(i-Swrol

and

absorption coefficient and the parameters that describe the

multiport and

power

To do

terminations.

its

f,

[/)]-' |5

-S*Tt

-5r,

\Sa

the available

this,

terms of the scattering

ratios a,, are first expressed in

(l-5r

(H)

-SjtTt

;)

matrix elements of the multiport and the reflection coeffi-

and then the absorption co-

cients of the port terminations,

efficient is

The
the

obtained using

available

power

V. Crosscorrelations for

(3).

ratio will be calculated as the ratio of

CW power available at port

i,

>

b,

ivai

from the generator on port/,

available

to the

',

>

/ a /"" 1

'

CW

power

available from port

the noise emitted at another port.

lation

is

the Matched Multiport

interesting to

junction noise emitted from one port

(6)

equivalent "generator" wave 6 amplitude 6,

it is

is

where f

is

in

fcVU

power

rh

the frequency domain.

where I\

is

<3,

and

the internal

measure of crosscorrethis

where

).,

paper

In the frequency

*{(n,V) av

(K,,

is

is

entirely

domain, the
appropriate,

[17].

(n*rij)*

(15)

denotes the ensemble average and the asterisk

denotes the complex conjugate.

If//,

and

//>

are the voltages

waves
due to the noise sources within a matched multiport, then

|r,h

(8)

:'

and port/,

respectively, of the generator

[5]

The

<n,V>. v

NUT

(16)

can be established by matrix

a,

techniques (see the Appendix).

expression of this relationship

->

front port

the reflection coefficient of the generator.

relationship between

;'.

is

Nva-

Wa=

(7)

available from the generator in tern"

"generator" wave amplitude a,

Fourier transform of the crosscorrelation

[16]

the reflection coefficient "looking" into port

Similarly, the
its

if

crosscorrelated to

the time average of the product of the time depen-

namely, the cross-spectral density

iV

know

is

dent voltages. However, the analysis of

expressed in terms of the

some applications

P*r"'/Pa.

The power

In

where T is the physical temperature of the multiport and

is an element of

compact and convenient

is

S,= \D liSi) /D

N=

- SSj

(17)

(9)

where the dagger indicates the Hermitian conjugate (transwhere Dusfi

its ith

is

column

matrix S,

is

the determinant of the matrix (1

is

replaced by the /th

Sr)

column of

when

scattering

the diagonal matrix with elements (T),, equal

to the reflection coefficient of the load

the determinant of (1

Sr) when

/. and Duo is
row and column are

on port

its ith

complex conjugate). Bosma [5] refers to TV as the noise

For the matched multiport, a "generator" wave is the same as the emergent wave.
The special case of /=/ may also be obtained from (2) and
(11) by setting all T's equal to zero, that is
pose,

distribution matrix.

The "generator" wave amplitude is denned as the amplitude the

generator would deliver to a reflectionless load [15], [16]. In this paper,

a basis has been chosen so that the characteristic impedance Z. associated with a port / will not be expressed explicitly. For those who wish
an 'explicit statement, each voltage such as
should be replaced by
a /^/2 ot and each scattering matrix element S,, should be replaced by

(n,.*>.v

= (l-|f,|=)r,v.,=

{l

-Zw.j

T.

(18)

}i

' When actually writing out D

u s/). one should take advantage ofthe
properly that Dasfl is unchanged when the y'th row and column of the
determinant are deleted. This property is expressed by D,,s,i = DdSiiu/).

123

IEEE TRANSACTIONS

690

An

interesting special case

= -

(nxn,*),T

Thus,

the two-port

if

Sn and S22

that

the two-port

is

(SiiS 21 *

+S

matched to

is

ON MICROWAVE THEORY AND TECHNIQUES, SEPTEMBER

are zero, then there

I2

loss

S 2 o*)2'.

(19)

rellectionless loads so

no correlation 5

is

Ld= 10

logic n (in decibels),

The absorption

in the

In

situations the port terminations have small reflec-

many

For

tion coefficients.

this situation, the results

of Section

111

simplify. In particular, neglecting the squares of the reflec-

D(iSj)/D(.ii)

= (Mj/Mdn

aii

where
ports,

M, and M,

(28)

are the mismatch factors of the /th andy'th

respectively.*

matched

way.

Thus,

if

[16], dissipative loss

both ports are

and absorptive

Sij +

2-i

S m TS

(20)

is a loss that depends on the output load impedance, but independent of the
input generator impedance, while absorptive loss is a loss

is

negligible, then the absorption coeffi-

depends on the input generator impedance, but

pendent of the output load impedance.

cient simplifies to

T's

all

A?

is

and

the absorption coefficient of the junction

22
where

Nml

(S im r m N,

it is

when

s, m *r a *N mi *)

are defined in (17). If the elements

known, then

(22)

Nmi

are not

of interest to determine an upper limit for

From

the magnitude of the N,fs.

it is

(21)

the definition of

t,

in (17)

/V-

The noise from the thermal sources within a multiport element is equal to the absorption coefficient of the element,
times the physical temperature of the element. The absorption coefficient may be measured or calculated. The calculated absorption coefficient is simply the sum of the available power ratios from all ports to the output port. In turn,
the available power ratios are expressible in terms of scattering

matrix elements and

was assumed; no

linearity

The

;V--

coefficients.

reflection

restrictions

on

reciprocity,

Only
sym-

correlation of the junction noise emitted from differ-

(23)

ent ports

N=(l SSt)

passive multiport

is

a positive semidefinite matrix for a

[5], [16],

then [18]

\Nv \*^N
This

inde-

metry, loss, or condition of match are needed.

clear that

Because

is

Conclusions

VIII.

Ai^Af+ti
where

conjugate

loss are indis-

that
|f,-|

ab-

tinguishable. Otherwise, dissipative loss

tion coefficients [15],

If in addition

i.e.,

would be (/4,-,)dB= 10 logn a,-,-. Efficiency is a

ratio of delivered powers in the same way that a,, is the ratio
of available powers. The efficiency and the available power
sorptive loss,

Multiports with Low-Reflection Port Terminations

VI.

v is the efficiency.

coefficient expressed in decibels,

ratio are related in the following

thermal noise emitted out the two ports.

where

1968

jr i ,.

(24)

when

the junction has refleetionless terminations

depends on the appropriate matrix elements of N=l SSt

and the junction temperature. When all ports of the junction have nearly refleetionless terminations, the expression
for the absorption coefficient greatly simplifies, and a limit
of error can be established

in turn implies

Na + Nn < Nu + Njs

N are

The diagonal elements of

Nu =

the port reflections are merely

Appendix
directly related to the ab-

sorption coefficients of the multiport,

if

ignored.
(25)

Equivalent Wave Generators

viz.,

The mathematics and

the notation in this appendix have

(20)

been influenced by N'emoto. Further details and interpreta-

where A," is defined in (21). Thus, because 2Z m |S im 2 <l

and all |r| 2 and A," are less than some maximums |rmax

tion of the concepts introduced here are discussed else-

(1

IS,,!'-).!,

where

[15].

A multiport

and A max ", then

is

described by a scattering matrix that relates

the emersinu waves b, to the incident waves

i|

where n

is

the

<

2(n

number of

- l)U m |

rm |

a,-.

(27)

b =Sa.

(29)

ports.

The boundary conditions imposed by terminating each port

VII.

Many

The Aksorption Coefficient

types

literature.

of losses have

Among

those

absorption coefficient

is

been distinguished

described
related

by

with a load or a generator

as a Loss

Bculty

most nearly to

in

the

[19],

the

dissipative

3
One should remember that zero correlation does not imply that
the signals arc unrelated. For example, lor the related signals iii = v+r

v are uncorrelated, C/ 'i"-.v=-''.v.v* av

n~ = x
v, where x and
yy*'** s zero when the power spectral density in v equals v. II' the
spectral densities of x and v are equal at all frequencies, then n and /rj
are uncorrelated. This principle is used in the IJIum radiometer [1 ], [2].

and
~~

'

124

is

described bv

+ rb

(30)

where d, is a generator wave, that is, a wave a generator

would deliver to a refleetionless load, and r is a diagonal
matrix whose diagonal elements are the reflection coefficients
9
Mismatch factor is used here in accordance with Miller el nt. [10],
-(I r,l-)(l-| P.;-)/ l-r,f',|-, where 1\ is the reflection coellicient "looking" out of the multiport and 1\ is the rellection coellicient
"looking" into the multiport at reference plane

A/,

/'.

WAIT:

THERMAL NOISE FROM A PASSIVE LLNEAR MULTIPORT

looking out of the multiport. The generator wave a, characterizes the

generator on port

sense that

in the

its

Using (3S) we obtain

amplitude

depends only on the output of the generator and that

it

Gn =
To

nating a between (29) and (30),

when only they'th

wave dj is nonzero.

fa

Sa

obtain

(40)

obtained from (40) and (35) for the case

generator, and thus only they'th generator

i, is

(9),

(31)

Acknowledgment

where

Se(J-

J. J. Fari?, W. Daywitt, G. F.
Engen, and T. Nemoto for helpful discussions and to M. B.
Hall for his support and encouragement.

The author

Sr)-'S.

(32)

C = A~'B, where
B are square and have the same dimension, is
is the determinant of
C,, = det A (xB /det A, where det A uh
A after the ith column of A has been replaced by they'th colBecause a matrix element of a product

D (lSi) /D

is

independent of the load that terminates the generator. Elimi-

is

indebted to

and

umn

References

,)

j,

[1] E. J.

where

= D iis/D

(33)

[3]

stands for the det (1 Sr).

same way that the inputs a, were each divided into

a generator wave plus a reflected wave, the outputs 6, can
be divided into an equivalent generator wave plus a reflected
In the

and recepteurs a cor-

NBS.

D. F. Wait and C. L. Trembath. "A comparison between a

switched and a correlation radiometer for cryogenic noise source
measurements," 1965 Proc. 20th Ann. ISA Con/., pt. 1, vol. 20, no.
14.4-3-65.

[4]

[5]

wave.

A. E. Siegman. "Thermal noise in microwave systems Part I:

Fundamentals," Microwave J., vol. 4, pp. 81-90, March 1961.
H. Bosma, "On the theory of linear noisy systems," Philips Res.
Repis. Suppi, no. 10, 1967.

fa

where

"Sensibilite des radioteUscopes

relation," Ann. Asiwphys., vol. 22. pp. 140-163. 1959.

[2] J. J. Faris, "Sensivitiy of a correlation radiometer." J. Res.
vol. 71C, pp. 153-170. April-June 1967.

of B;

Sn

Blum.

is

fa

a diagonal matrix

[6]

ta

(34)

whose elements

are the reflec-

[7]

W. R.Bennett, Electrical Noise. New York: McGraw-Hill, 1960.

H. A. Has and R. B. Adler, Circuit Theory of Linear Noisy Networks.
New York: Wiley, 1959.

D. F. Wait, "Sensitivitv of the Dicke radiometer," J. Res. NBS,

71C, pp. 127-152,' April-June 1967.
D. Middleton. An Introduction to Statistical Communication
Theory.
New York: McGraw-Hill, 1960.
[10] C. K. S. Miller, W. C. Daywitt. and M. G. Arthur. "Noise standards, measurements, and receiver noise definitions," Proc. IEEE
vol. 55, pp. 865-877, June 1967.
[11] D. F. Wait and T. Nemoto, "Measurement of the noise temper;
ture of a mismatched noise source," this issue, page 670.
[12] J. S. Wells, W. C. Daywitt, and C. K. S. Miller, "Measurement
effective temperatures of microwave noise sources," IEEE Tra
Instrumentation and Measurement, vol. IM-13, pp. 17-28, Mai
[8]

vol.

tion coefficients looking into the multiport.

[9]

Eliminating

fa

from (31) using (30) and

fa

G
G =

where one form for

From

(oo)

is

|(1

- rf)s -

the nature of the definition of

elements of

(36)

and

a, the diagonal

the determinant of (/

D=
ofT

G.

F.

Engen,

"A

variable

impedance power meter, and

(i

(37)

ST)

r.-roD, w

diagonal element of (36)

G =

(1

reflection coefficient standard," J. Res.

after the

j'th

row

(38)

is

- TitdDusn/D.

[14]

G.

F. Engen, private

discussed on p. 50.)
G. M. Korn and T.
lists and Engineers.

[18]

Ibid., sec.

[19] R.

communication.

W.

M. Korn, Mathematical Handbook fa

New York: McGraw-Hill, 1961, sec.

13.5-6.

Beattv, "Insertion loss concepts," Proc.

663-671, June 1964.

IEEE TRANSACTIONS

Volume MTT-16, Number

u,

September, 1P6S

pp. 6S7-691

106S

adjust;

C, pp. 7

Nemoto and

[17]

(39)
Reprinted from

vol. 68

D. F. Wait, "Microwave circuit analysis

the equivalent generator concept," lEEETrans. Microwave T
and Techniques {to be published).
[16] D. M. Kerns and R. W. Bcatty, Basic Theory of Waveguide
lions and Introductory Microwave Network Analysis.
New
Pergamon, 1967. ("'Generator" waves are denoted as bo f
[15] T.

ON MICROWAVE THEORY AM) TECHMQCES

Copyright

NBS,

1964.

= Dusa/Dwt

and /th column have been deleted.

Again looking at a diagonal element of (36) and solving
for (1 r,f ,), replacing the remaining l\ by (37), we obtain

An

1964.
[13]

Pi
is

rj.

fa

must be zero. Setting the diagonal elements of

and solving for f we obtain

(36) equal to zero

where D<)

(34),

= Ga

Tin: Institute or Klhtuic.m. and Electronics Engineers, Inc.

PRINTED IN

125

THE

U.S.A.

IEEE, vc

FORM NBS-114A

(1-71)

DEPT. OF COMM.

U.S.

1.

PUBLICATION OR REPORT NO.

BIBLIOGRAPHIC DATA

SHEET
4.

2.

Gov't Accession

TITLE AND SUBTITLE

9.

8.

Wait
PERFORMING ORGANIZATION NAME AND ADDRESS

Performing Organization Code

Performing Organization

national bureau OF standards,

DEPARTMENT OF COMMERCE

Boulder,
12.

August 1973
6.

AUTHOR(S)

David F

Publication Date

5.

Considerations for the Precise Measurement of

Amplifier Noise
7.

Recipient's Accession No.

3.

No.

NBS TN-640

CO

Project/Task/Work Unit No.

11.

Contract/Grant No.

13.

Type of Report & Period

Covered

2726203

Boulder Labs,

80302

Sponsoring Organization Name and Address

Same as

10.

#9.
1

15.

SUPPLEMENTARY NOTES

16.

ABSTRACT

(A 200-word or less factual summary of most significant information.

bibliography or literature survey, mention it here.)

If

4.

Sponsoring Agency Code

document includes a significant

For the best accuracy in measuring noise figure, attention needs to be

given to the choice of the hot and the cold noise standards and to mismatch
problems. Tables and graphs are presented to aid in choosing the proper
measurement conditions, and an example is given to demonstrate their use.
This paper essentially supplements a previous paper (included in an appendix)
treating in more detail topics that become important when state-of-the-art
measurements are required.

17.

18.

KEY WORDS

(Alphabetical order, separated by semicolons)

Amplifier noise; effective input noise temperature; mismatch error; mismatch

uncertainty: noise figure
AVAILABILITY STATEMENT
19. SECURITY CLASS
21. NO. OF PAGES
(THIS REPORT)

129

fxl UNLIMITED.

UNCLASSIFIED
I

FOR OFFICIAL DISTRIBUTION. DO NOT RELEASE

TO

NTIS.

20.

SECURITY CLASS

22. Price

(THIS PAGE)

11.25
UNCLASSIFIED
'

U. S.

GOVERNMENT PRINTING OFFICE

'572 784

568/

12

12REGION NO

USCOMM-DC

66244-P71

Iki

NBS TECHNICAL PUBLICATIONS

NONPERIODICALS

PERIODICALS

JOURNAL OF RESEARCH

reports National

Bureau of Standards research and development in

and chemistry. Comprehensive
scientific papers give complete details of the work,
physics, mathematics,

including laboratory data, experimental procedures,

and

theoretical

and mathematical

analyses. Illustrated

photographs, drawings, and charts.

listings of other NBS papers as issued.
with

Published

in

two

Includes

Applied Mathematics Series. Mathematical

manuals, and studies.

Building Science Series. Research results, test

methods, and performance criteria of building materials, components, systems, and structures.

Handbooks.

Recommended

codes of engineering
(including safety codes') developed in cooperation with interested industries,
professional organizations, and regulatory bodies.

and

industrial practice

sections, available separately:

Special Publications.

ences,

Physics and Chemistry (Section A)

and properties of matter. Issued six times

Annual subscription: Domestic, $17.00: For-

constants,

eign, $21.25.

bibliographies,

pamphlets,

Papers of interest primarily to scientists working in

these fields. This section covers a broad range of
physical and chemical research, with major emphasis
on standards of physical measurement, fundamental
a year.

tables,

etc.

Monographs. Major
literature
scientific

Proceedings of NBS conferannual reports, wall charts,

contributions to the technical

to the Bureau's

on various subjects related

and technical activities.

National

Data

Reference

Standard

Series.

NSRDS

provides quantitative data on the physical

and chemical properties of materials, compiled from
the world's literature and critically evaluated.

Product Standards. Provide requirements

Mathematical Sciences (Section B)

for sizes,

and methods for testing various industrial products. These standards are developed cooperatively with interested Government and industry
groups and provide the basis for common understanding of product characteristics for both buyers and
types, quality,

Studies and compilations designed mainlv for the

mathematician and theoretical physicist. Topics in
mathematical statistics, theory of experiment design,
numerical analysis, theoretical physics and chemistry.
logical design and programming of computers and
computer systems. Short numerical tables. Issued quarterly.

Annual subscription: Domestic, $9.00; Foreign,

$11.25.

Their use

voluntary.

is

Federal

TECHNICAL NEWS BULLETIN

The

sellers.

Technical Notes. This series consists of communications and reports

covering both other-agency and
NBS-sponsored work") of limited or transitory interest.

Information

Processing

Standards

Publications. This

best single source of information concerning the

Bureau's measurement, research, developmental, coand publication activities, this monthly

publication is designed for the industry-oriented
individual whose daily work involves intimate contact
with science and technology for engineers, chemists,
physicists, research managers, product-development
managers, and company executives. Includes listing of
all NBS papers as issued. Annual subscription
Domestic, $6.50; Foreign, $8.25.
operative,

series is the official publication

within the Federal Government for information on
standards adopted and promulgated under the Public
Law 89306. and Bureau of the Budget Circular A 86
entitled. Standardization of Data Elements and Codes
in Data Systems.

Consumer Information Series.

based on NBS research and

Practical informa-

experience, covering areas of interest to the consumer. Easily understandable language and illustrations provide useful
tion,

background knowledge for shopping

in today's tech-

nological marketplace.

BIBLIOGRAPHIC SUBSCRIPTION SERVICES

The following current-awareness and literature-survey
Bureau

bibliographies are issued periodically by the

Cryogenic Data Center Current Awareness Service 'Publications and Reports

A literature survey issued weekly. Annual subscription: Domestic, $20.00; foreign,
Liquefied Natural Gas.

of Interest in Cryogenics).

$25.00.

literature survey issued quarterly. Annual subscription: $20.00.

Superconducting Devices and Materials. A literature survey issued quarterly. Annual subscription $20 00.
Send subscription orders and remittances for the preceding bibliographic services to the U.S. Department
of Commerce, National Technical Information Service, Springfield, Va. 22151.
:

Electromagnetic Metrology Current Awareness Service (Abstracts of Selected Articles on Measurement

Techniques and Standards of Electromagnetic Quantities from D-C to Millimeter- Wave Frequencies). Issued
Send subscription order and
monthly. Annual subscription: $100.00 (Special rates for multi-subscriptions
1

remittance to the Electromagnetic Metrology Information Center. Electromagnetics Division. National Bureau
of Standards, Boulder, Colo. 80302.

NBS publications (except Bibliographic Subscription Services')

from: Superintendent of Documents, Government Printing Office. Washington, D.C. 20402.

Order

U.S.

DEPARTMENT OF COMMERCE

National Bureau of Standards

Washington, D.C.

20234
U-S.

OFFICIAL BUSINESS
Penalty for Private Use.

POSTAGE AND FEES PAID .

DEPARTMENT OF COMME'RCE*

COM-2I5
$300

1298