You are on page 1of 32







F. M. D aw son


A. A. K a l in s k e


Issued semi-monthly throughout the year. E ntered at the post office at Iowa City,
Iowa, as second class m atter under the Act of October 3, 1917.
Offers complete undergraduate courses in Civil, Electrical, Mechan
ical, Chemical, Commercial, and Hydraulic Engineering, also graduate
courses in these fields leading to advanced degrees.
For detailed information application may be made to
H . C . D orcas , R egistrar,
Iowa City, Iowa


The Iowa Institute of Hydraulic Research has been organized to
afford an agency for the co-ordination of the talent, facilities, and the
resources that may be made available at the University of Iowa for
undertaking projects of unusual magnitude, scope, or complexity in
the field of hydrology and hydraulic engineering.
The Institute affords a connection, through which technical societies,
governmental departments, industrial corporations, and other interest
ed parties may effectively co-operate with the University in the field of
hydraulic research. Correspondence regarding the work of the Insti
tute should be addressed to
E. W. L a n e
Associate Director in
Charge of Laboratory.

B U L L E T IN 10




F . M . D aw so n

Dean, College of Engineering

D irector, Io w a In stitu te o f H ydraulic Research


A. A. K a l in s k e

I n s t r u c t o r ,H ydraulic and S anitary Engineering

Research Assistant Engineer
I o w a In stitu te o f H ydraulic Research


I. I n t r o d u c t io n 5
1. Synopsis S
2. Acknowledgments 6
3. Previous Work 6

II. F l o w o f W ater i n P artly F u l l V ertical P ip e s 6

4. Statement of Problem 6
5. Apparatus 7
6. Procedure 8
7. Analysis of Data 9
a. Rational Analysis 9
b. Analysis of Test Data 10
8. Summary and Applications 13

III. P n e u m a t i c E f f e c t s of W ater F l o w in g D o w n P artly F u l l

V ertical P i p e s 14
9. Statement of Problem 14
10. Apparatus IS
11. Procedure IS
12. Analysis of Data 16
13. Summary and Applications 27


Fig. 1. Experimental Vertical Pipe Installation 7

Fig. 2. Velocity Determination Apparatus 8
Fig. 3. Velocity of Water in 3-in., 4-in., and 6-in. Stacks with Different
Rates of Flow 11
Fig. 4. Pressure Variation in Open Unvented 3-in. and 6-in. Stacks 16
Fig. S. Water Flow, Air Flow, and Air Pressure Data for 4-in. Stack 17


Table 1. Maximum Velocity and Friction Constants in 3-in., 4-in., and

6-in. Stacks 12
Table 2. Air Flow Data for 3-in. Stack 18
Table 3. Air Flow Data for 4-in. Stack 20
Table 4. Air Flow Data for 6-in. Stack 22

1. Synopsis . T he general problem of non-uniform flow in conduits

of various types is a fascinating and fertile field for hydraulic research.
Experimenters have so far, except in a few instances, overlooked the
hydraulics involved in the flow of liquids in various parts of the plumb
ing drainage system. Accelerated and retarded flow in the vertical
pipes and drains with steep and flat slopes, as found in every plumbing
drainage system, present complicated hydraulic problems which merit
further study by hydraulicians.

Besides the hydraulic problems, the plumbing drainage system pre

sents interesting pneumatic problems. Since the pipes flow only p a rt
ly full the flowing water causes pressures to be developed within the
closed plumbing drainage system which m ay be detrim ental to the
w ater seals of plumbing fixture traps. Therefore air m ust be supplied
a t certain points and removed at others by a system of vents.

After considerable time spent in reviewing existing information, pre

liminary test work, and development of test techniques, it became
quite clear th at the hydraulic and pneumatic problems involved in the
flow of water down vertical plumbing pipes and the prevention of re
sultant excessive atmospheric disturbances in the drainage system, were
far from simple. As the work progressed the complexity of the many
hydraulic problems encountered became apparent, and it was evident
th at the work could not be completed in a short time.

The purpose of these studies is to establish definite minimum drain,

stack, and vent sizes for various plumbing installations, and to ana
lyze the various parts of the plumbing drainage system so th at they
can be designed on a sound rational basis. Although tables of re
quired stack, drain, and vent sizes could be made up from the infor
mation contained in the report, it was not thought advisable to present
them here. However, the information given can be of considerable
value to designers of plumbing drainage systems of large buildings.

The report is divided into two parts. The first deals with the hy
draulics of steady flow in partly full vertical pipes of plumbing sys
tems; besides considerable experimental data, a rational analysis of
the flow is presented. T he second deals with the pneumatic disturb
ances caused by the flow of water in vertical pipes, and with the factors
effecting the am ount of air necessary to prevent excessive negative
pressures in adjoining drains.

2. Acknowledgements . The authors wish to acknowledge the many

helpful suggestions given in the preparation of this report by F. T.
M a v is , Professor of Hydraulic Engineering, the State University of
Iowa, and Consulting Engineer, Iowa Institute of Hydraulic Research.
The experimental work on which this report is based was per
formed a t the Hydraulic and Sanitary Engineering Laboratory of the
University of Wisconsin.

3. Previous Work . H unter and Snyder of the United States Bur

eau of Standards1 mention certain characteristics they observed in flow
of water down partly full pipes and discuss the factors affecting the
velocity of the water. In the report2 they present a sound basis for
design of vents, bu t do not give very extensive data on air-flow.
B abbitt3 has presented some good information on tests of different
parts of the plumbing drainage system but presents little experimental
information on the hydraulics of vertical pipes.
From the standpoint of pneumatics the vertical pipes (or stacks)
in a plumbing drainage system are the most im portant parts of the
system, because flow of water in them produces by far the worst pneu
matic effects. A considerable p a rt of this report is devoted to the
hydraulics and pneumatics of stacks.


4. Statement of Problem . The purpose of this investigation was to

determine how water flows down partly full vertical pipes, what veloci
ty it attains in different heights of fall, and how th at velocity varies
with rate of discharge and size of pipe.
The tests were made with water entering vertical stacks from hori
zontal drains similar to plumbing system installations. No tests were
made with water entering stacks vertically, such as discharges down
l R ecom m ended M inim um R e q u ire m en ts for P lu m b in g , 1929, pp. 104-108. R eport
of Sub-C om m ittee on P lu m b in g , B u ild in g Code C om m ittee, U. S. D e p a rtm e n t of Com
2Loc. cit., pp. 110-120.
3H. E. B a b b itt, H y d ra u lic s a n d P n e u m a tic s of H ouse P lu m b in g , Bulletins 143
and 178, U n iv e rs ity of Illin o is E n g in e e rin g E x p e rim e n t S ta tio n , 1924 a n d 1928, respec
tiv e ly .
pipes from tanks, vats, and roof rain collectors. Tests on such flow
conditions are to be undertaken later. The theory presented may be
applied to any type of entrance condition, as long as the vertical pipe
flows only partly full.
S. Apparatus. Fig. 1 shows a diagrammatic sketch of vertical 3-in.,
4-in., and 6-in. pipe set-ups in which the velocity of water was de-

Fig. 1. E x p e r im e n t a l V ertical P ip e I nst a l l a t io n
termined for various constant rates of discharge. The water entered
the vertical pipes from a large tank through a valve the opening of
which determined the rate of flow. Between the entrance point and
bottom of the vertical pipe there were various openings through which
a pitot tube could be inserted for determining the velocity head. The
water was discharged a t the bottom through a horizontal drain. In
each of the three installations a vertical length of approximately 30
ft. was available in which the velocity could be observed.

6. Procedure. The velocity was measured with a pitot tube, the

velocity head being determined by a mercury U-tube gage as shown
in Fig. 2.

F ig . 2. V elocity D e t e r m in a t io n A ppa r a t u s

A velocity reading was taken by recording the average mercury

column difference reading of the pitot tube gage during the period of
constant discharge. At any one level, readings were taken a t various
points inside of the pipe in order th at an average velocity value could
be computed. T he velocity reading as finally computed represents
about ten different observations for any single rate of flow. I t was
necessary to read the pressure head in the vertical pipe in order
to obtain the correct velocity head from the observed pitot tube gage
pressure difference. In virtually all instances this pressure head was
negative (below atm ospheric). A correction was also made for height
of water column above the pitot tube opening in the connection lead
ing to the mercury gage. Referring to Fig. 2 the velocity was deter
mined in the following manner:
hi velocity head as measured by pitot tube in in. of mercury,
fe = static pressure in stack, usually a vacuum, ft. of water.
hi = height of water above pitot tube opening, ft.
13 6
H = ( ' hi) -f- hi -(- hs = velocity head in ft. of water.
V = \/2 g H = velocity in ft. per sec.
As shown diagrammatically in Fig. 2, the water tends to flow along
the walls of the pipe a few feet below the entrance. The direction of
the average velocity is parallel to the pipe except a t very low rates
of flow. At the very low rates of flow the water spirals around the
walls of the stack. No velocity determinations were made under such

7. Analysis of Data .

(a ) R ational A nalysis

A rational analysis of the flow down a p artly full vertical pipe in

volves the establishment of a relationship between the velocity a t any
point below entrance and the initial velocity and height of fall.

Let A I be an accelerating elemental disc of water a distance L be

low the point of entry into the vertical pipe. The disc is accelerating,
and the force acting on it is equal to the difference between gravita
tional force and frictional force.
nia = mg k' x D (A L) wV2 (1)
in which k' the friction resistance expressed in lbs. per sq. ft. per
unit density of liquid flowing per unit velocity.
D = diameter of stack, feet.
w = unit weight of water, lb s./ft.:i
V average velocity of water in ft./sec. (Frictional force
varies approximately as square of velocity.)
g = gravitational constant = 32.2 ft./sec.2
a = acceleration of water a t any point.
m mass of water = (A L ) A
Since the stack is not flowing full, A can best be expressed as ~
where Q is the rate of discharge in cubic feet per sec. Also, in order
that the friction constant used be dimensionless k' is replaced by
k X D V3
Therefore: a = g ---------^ ---- - (2)
T , 7 k- D A V dV
Let z = , and as a =
V dL
then, dL = (3)

Integrating and letting V = V0 when L = O:

1 r, / \. I. 3VmV
3KiK \ _ / ,. 3 F WF \-i
6ZF [ gfi( + ( 7 * F)V ~ gc( + (FW 7 0)2)]
------ = --------( arctan----------= ------ a rc ta n -------------= ] (4)
V3 ZVm\ Vm \/l Vm\Jl J
Since the velocity V is a maximum when the acceleration a is zero,
it is evident from Equation (2) th a t maximum or terminal velocity
of water in the stack is

According to Equation (4 ), the maximum velocity will be attained

when L is infinite, b u t practically this is not the case. Changes in
frictional resistance on the walls of the pipe cause the water to attain
a maximum velocity in a definite length and Equation (4) can readily
be used for computing L. Assume D = 3 in., Q 90 gpm, V0
3 ft. per sec., and find the distance in which maximum velocity is a t
tained. From experimental data shown in Table 1, k = .00225,
therefore from Equation (5 ), Vm = 15.4 ft. per sec. Substituting in
Equation (4) and letting ( Vm V) = .01 ft. per sec., L = 16.7
ft.; when (Vm V) = .1 , L 11.3 ft. For all practical purposes
it can be said th at the maximum velocity is attained in approximately
12 to IS ft. for conditions stated.

(b) Analysis of Test Data

Fig. 3 shows the variation of velocity with height of fall in the 3-in.,
4-in., and 6-in. vertical pipes, as obtained experimentally. Velocities
are plotted as ordinates and heights of fall as abscissas. Note th at
the velocity a t any point increases as the rate of discharge increases,
and also th at the velocity becomes constant a relatively short distance
below the inlet. T he maximum velocity for a given rate of discharge
in any size pipe is independent of the initial velocity and is definitely
determined for any given discharge. In the 3-in. stack installation it
was possible to maintain the initial velocity a t about 3 ft. per sec. for
H e ig h t of Fall in S tack in Feet

F ig . 3. V e l o c it y of W ater i n 3 - i n ., 4 - i n ., a n d 6 - i n ., S ta c k s w i t h D if f e r e n t
R ates of F lo w

all discharges. On the 4-in. and 6-in. installations the initial velocity
was not constant for all discharges; varying from about 6 to 14 ft. per
sec. for the 4-in. stack, and from 2 to 10 ft. per sec. for the 6-in. stack.
The friction coefficient k for each discharge and size of pipe was
computed from quation (5) and the values are shown in Table 1.
In these tests k was usually higher for the low rates of flow and de
creased as the rate of flow increased becoming fairly constant for the
higher rates of flow, when water filled the stack to at least one-fourth
the cross-sectional area.
The maximum velocity for any discharge in a given size stack can
be computed by M annings formula, in which the hydraulic radius is
and the slope is 1.0.
Experimental values of n are given in Table 1. N ote that the rela
tion between k and n for uniform velocity is:

/1 .4 8 6 \ % / Q y /s
\ n ) n)
The experimental values of n generally fall in the range .0055 to
.0065. The values for the 4-in. stack are below this range. In the
tests on the 4-in. stack the water entered the vertical pipe through a
3-in. opening, and for the lower rates of flow the stream of water
probably did not touch the walls of the 4-in. pipe. In the 3-in. stack
tests the water entered through a 4-in. opening, and the values of n
for the various rates of discharge differed less than 3 per cent from
the average .0064. In the 6-in. stack tests the water entered the ver
tical pipe through a 6-in. opening. D irect visual observation through
openings in the 6-in. stack indicated th at a t the higher rates of flow
the stream of water did not touch the walls a t all points, thus account
ing for the smaller values of n at the high rates of discharge. For

T able 1
M a x im u m V el o cit y and F r ic t io n C o n st a n t s
iisr 3 - i n ., 4 - i n ., and 6 -in . S tac ks

Diam. of Water Maximum

Stack Discharge Velocity k n
in. gpm ft./sec.
3 45 11.5 .00268 .0063
90 15.4 .00225 .0062
135 17.6 .00226 .0064
180 19.5 .00220 .0066

4 90 16.8 .00129 .0044

135 19.0 .00134 .0047
190 20.4 .00152 .0053
240 21.2 .00170 .0057
300 22.0 .00190 .0062

6 115 12.3 .00280 .0067

165 14.4 .00252 .0066
220 17.4 .00190 .0059
345 21.7 .00155 .0055
4 50 25.0 .00132 .0052
560 26.5 .00139 .0054
675 27.5 .00148 .0057
rates of flow between 200 and 700 gpm in the 6-in. stack the average
value of n was .0055 and the maximum deviation was 6.5 per cent.

In the foregoing calculations it was assumed th at the wetted peri

meter was equal to x D. Obviously if actual conditions do not approxi
m ate this, the roughness index n, as computed from formulas in which
it D is used as the wetted perimeter, will vary with size of pipe and
discharge. Relative size of entrance opening and vertical alignment
of the stack are im portant factors in determining whether the stream
will flow along the entire inside perimeter of the stack. The only test
in which visual observation definitely indicated th at the stream of water
touched the entire perimeter a t all times was th at for the 3-in. stack.

Since k varies with velocity for any given stack and discharge, it
is necessary to compute L by Equation (4) for increments of velocity
for which the average value k can be assigned. If the velocity is
changing the value of k can be computed from the expression:


Using experimental values of initial velocity and friction constants

n for the 3-in. stack, curves were drawn b y computing values of L
from Equation (4 ). These curves had the same general shape as those
in Fig. 3 for the experimental data.

8. Summary and Applications . From the experimental data and

the theoretical relations presented, the maximum velocity in a pipe
flowing partly full and the height needed to produce that velocity can
be determined for any of the common sizes of vertical pipes used in
plumbing and industrial installations.

The maximum velocity attained in vertical pipes flowing partly full,

with the water entering the pipe through an opening equal to or great
er in size than the pipe itself, can be computed within 10 per cent by
M annings formula in which n .0060. This value of n, determined
by experiments on 3-in., 4-in., and 6-in. cast iron vertical pipe, is sug
gested also for 2-in., 5-in., and 8-in. vertical pipes.

The length in which the maximum velocity is attained varies with

the initial velocity, however a general idea of the magnitude can be ob
tained from the experimental data presented. The method for com
puting this length by use of the theoretical relation between length
and velocity is fairly simple if the initial and maximum velocities are

The theoretical relation between height of fall L and velocity V can

be used in computing the initial vertical velocity Vo by experimentally
measuring the velocity a t a known distance below entrance. Although
it is difficult to determine the initial vertical velocity by direct ob
servation, it is an im portant quantity when the maximum possible
discharge a t any one level into a vertical pipe is considered.

The maximum velocity obtained is of importance in studying effects

a t bottom of vertical pipes where the flow turns 90 and flows in a
horizontal drain. I t is also to be used in the pneumatic studies in
order to determine how much air can be carried by the rapidly flowing


9. Statement of Problem . W ater flowing down vertical pipes such

as plumbing stacks, roof leaders, and tank drains, creates pneumatic
disturbances of various degree depending on the installation and rate
of flow. These pneumatic disturbances are im portant in plumbing
systems since the pressure differentials may not be greater than about
1 in. of water in the drains to which plumbing fixtures are connected.
Although the water seal is about 2 in. in most plumbing fixture traps,
some are as small as 1 in.

As water enters a vertical pipe from a horizontal drain and flows

down, partial vacuums are formed under the point of entrance, and
positive pressures near the bottom of the down-flow pipe. The falling
water entrains air and compresses it near the bottom of the stack to
a degree depending on the ease with which the air can escape a t the
bottom. A partial vacuum or positive pressure is maintained a t various
points in the stack for the duration of the water flow. The magnitude
of the pressures or vacuums depends on the rate of flow and installa
tion conditions.

To prevent the formation of excessive pressure differentials in the

drains to which plumbing fixtures are connected, air must be supplied
or removed at definite rates. The problem is to determine w hat these
rates of air flow are for various rates of water flow in different size
vertical pipes a t various distances below water entrance. T he deter
mination of these rates of air flow will provide a basis for proper vent

10. Apparatus . The 3-in., 4-in., and 6-in. vertical pipe apparatus
(Fig. 1) used in the velocity studies was also used in these tests. Air
pressure measurements were made with ordinary U-tube water or mer
cury gages. A tee was installed in the vertical pipe approximately
one floor level below the water entrance and most of the measurements
of air flow were made a t this level.

The air flow was measured with a 3-in. anemometer which was cali
brated by means of standard sharp-edge orifices. The pressure differ
ence across the orifices used for calibration did not exceed 6 in. of
water, thus permitting the use of ordinary hydraulic formulas in cal
culating the air flow. The instrum ent was calibrated for different sizes
of pipes.

An anemometer was used for convenience, and also because the pres
sure just outside of the vertical pipe in the vent, could be maintained
constant much more readily than if orifices or other types of meters
were used directly. The pressure ju st outside the vertical pipe in the
vent was maintained a t -1 in. of water corresponding to th at which
can exist in a plumbing installation. A head of 1 in. of water would
be the maximum head for producing air flow. A negative pressure
greater than 1 in. would produce too much siphonage of plumbing
fixture traps.

Intervals of time were measured by an electric clock reading to

1/100 second.

11. Procedure. Two separate studies were made on each of the

vertical pipe installations. First, for different rates of water flow down
the pipes, pressures were read along the entire length of the pipe with
no air entering except at the top of the vertical pipe. This test indicated
the variation of the pressure inside the pipe flowing partly full of
water, from the point of water entrance to the bottom.

The second study was concerned with air-flow measurement for vari
ous water discharges. T hree quantities were observed for each air
flow measurement: (1) the rate of air flow, (2) the pressure inside the
F ig . 4 . P r e s s u r e V a r i a t i o n i n O p en U n v e n t e d 3 - in . a n d 6 - in . S t a c k s
3-in. S ta c k C onnected to 70 ft. of 4-in. Sewer.
6-in. S tack C onnected to 25 ft. of 6-in. Sewer.

unvented stack without air flow, and (3) the pressure inside the vented
stack with air flow.

Some tests were also made to determine the am ount of air entering
through the top of the vertical pipes for different rates of w ater flow,
both with and w ithout venting below the water entrance.

12. Analysis of Data . The data for the first study are shown on
Fig. 4 in which the distances below water entrance are plotted as or
dinates and the pressure differentials for the 3-in. and 6-in. pipe in
stallations as abscissas. The pressure shown is th a t produced inside
a stack open a t the top but having no vents below w ater entrance
This pressure reaches a maximum negative value a short distance be
low water entrance and maximum positive value a t the bottom, the
magnitude for any installation depending upon the rate of discharge.
N ote that the point where pressures change from negative to positive
is the same for all rates of flow. The location of this point of zero
i 500

(a ) 4 V?nt - StacV n (b) 3"Yerit - S t a c k Oper +

1 400 c
L> f 10
-f- __ n * is
|- 6 'N y 300
-4 200
prej: o
P re& o
-2 100 5
-10 500 c
/ V
-8 / 400|
/ V. ( d h >" Veri t - :itacV Ope n Li-
300 J.
-4 jy 200 <+.
t o
-2 r rVSSW'L-'O' a>
100 +-
CO0 3 ' \ /ent- stack Clost s
0 100 200 300 100 200 300
Rate o f W a te r F lo w in G a l l o n s pe r M in u te
F ig . S. W ater F low , A ir F l o w , and A ir P r e ssu r e D ata e o r 4 -i n . Sta c k
For A, 33, a n d C a ir flow w as m e a su re d 12.5 ft. below w a ter e n tra n c e. For
D a ir flow w as m easured 24.5 ft. below w a te r en tra n c e.

gage pressure, the magnitude of the positive and negative pressures,

and the location of the point of maximum negative pressure depend
on conditions a t the bottom of the vertical pipe, type of base fitting,
and length of drain at the bottom.
Representative data for the second study are shown in Fig. 5. Rates
of air flow and the corresponding pressures inside the 4-in. stack are
plotted as ordinates and rates of water flow are plotted as abscissas.
Fig. 5(a) shows these data for a 4-in. sanitary tee vent connection
and top of stack entirely open. In Fig. 5(b) the conditions are sim
ilar except th a t the vent opening is 3 in. The data in Fig. 5(c) are
for a 3-in. vent opening but with top of stack entirely closed. In
Fig. 5(d ) the data were .obtained two floor levels below water entrance
instead of one floor level. T h e data for the 4-in. stack indicate in
general all the various items th at affected the variation of air flow
and pressure inside of stack with water discharge. The test results on
the 3-in. and 6-in. stacks do not indicate anything different from w hat
O 00 O O'
'O 't
O 't


1P0 O O O

H N 'lO O to to o
N. <Oo
qi!- *0 t-H
ui uinnDUA

SupuaA N o o p o o
ijlTM 3pB}S g O' to N O ' V)
UI Uinn3By\ *

T abl e

MOIJ j;v S CN rO M tN rsi ro c<j cn
aiqissoa ? O to to to
(N N N M O to to to
M (N M M
um uiprej\[ Cd

dox 0}i O'

o o o o
MOIJ jjv

qSnojqj O
M o u jry

0O ' ^^ 0M to
ri- O' ^

fi H O M g H O
C a
H V in H
can be concluded from the 4-in. stack data. Complete test data on
the 3-in., 4-in., and 6-in. stacks are shown in Tables 2, 3, and 4. The
recorded values are the average of some five separate observations.

A study of the experimental data brought out the following facts:

(1) For a given stack the maximum rate of air flow occurred a t the
same rate of water discharge for all conditions except when the top
of the stack was closed. For open stacks the maximum rate of air
inflow was caused by water rates of 130 gpm, 190 gpm, and 550 gpm
in 3-in., 4-in., and 6-in. stacks respectively. W ith the top of the
stack closed, a slightly greater maximum rate of air inflow was caused
by water rates of 40 gpm, 90 gpm, and 350 gpm in 3-in., 4-in., and
6-in. stacks respectively. (2) The negative pressure inside the stack
with air flowing in increased almost uniformly with increase in rate
of water discharge. A definite pressure difference existed between the
vent, just outside of the stack, and inside of the stack itself. (3)
When the top of the vertical pipe was closed, high partial vacuums
were developed, reaching about 13 ft. of water for the 6-in. stack,
6 ft. for the 4-in. stack, and 2 ft. for the 3-in. stack. However, the
rate of air inflow was not greatly increased above th a t existing when
the top was open, though the partial vacuum inside of the pipe re
mained much higher as shown in Fig. 5 (c). (4) T he type and size
of vent opening through which the air flows into the vertical pipe af
fects the air inflow. (5) The maximum air inflow for each stack
occurred on the floor immediately below the water entrance. Toward
the bottom of the stack the air inflow decreased and became air out
flow in the regions of positive pressure. For instance, on the 4-in.
stack which is similar to a 3-story building installation, when water
entered on the third floor, the air inflow a t the first floor level was
about one-half th at a t the second floor level. No air entered the stack
at the second floor when the air flow measurement was made a t the
first floor level. If the stack received air at the second floor no great
am ount of air was required a t the first floor level.

In Fig. 5, note the curves indicating the pressure inside the stack
a t the point of air entrance. In all the tests the negative pressure in
the vent just outside the stack was m aintained a t 1 in. of water. Ex
cept for the smallest rates of water discharge, the negative pressure
inside of the stack exceeded 1 in. of water. In other words a pres
sure differential existed between a point in the vent just outside the
stack and a point inside the stack two points but a few inches apart.

H v> CO H V ) fO
"LO t o 't I O 'O

;U3A O
co O
o ^O o
CO p
co PC cm CO
ui xunnD-B^

SupuaA o_^ 5 o o o o
tfliM 3]OB5g
ui uinno-BA.

to 'O O VO

o * IJ J!V O ^ o O O io O io O O io
M aiqissoj g in o*
I vp
o io
ON u)
in \oI \o
o in
O' V
uinraixBj\| 0>

doX oiu] to O O O O
Mou j;v CM CM CM CO co

*?U3A 5 o O mO oTl- io Oo
qSnojqx O 'O <N
CO rt
CO Cv| O
C0 O CO 1-H
CO co co co
mou iiy

o ^ o o o

H oa &
m g Ho
9 > ri a
V V co H
vq 10 P p lO CO O

O O' vj lo


p p (N N N O

IT , O'

p o o o o o o

00 vO
00 *"


O O *0 o *0 o O VO

VO 'O to lO VO VO *0 lO

in vo O *o o

iI vO f" vO '3-

O o to O



Tt O
<N o cm
3" Vent to

Top Closed

San. Tee.

5 H


s O ' 00 o in t>- to to ^
N fO N CO CO rj- IT) vo vd

}U d j \
q ^ iM 3pK } (
P p p O'
u t u in n 3 B y \ ts > CO to .

Supua^ ON o p q o o O o o o-
I fllM jp B J S s
"3- ri o co
rs) co d o ^
ui umnDuy\ h
6 -In .

io O O IO O
CO Cvl co 00 'O
D ata

m o ij J iv O O O O o o O O O O
1-H CO O CS -cj- rH
aiqissoj 1I H CO O' CS)
F low

CO * 0 t'~. I'* r-l CO *0

uirraiixBj^ O
A ir

doX lnI <- to to to O

"O co O'
rf r j- \n m

*?U3A 5 o o o O o o O O
IO to
qSnoaqx 0> t '-
O O '
oo O e) co
CO 00 O
O' O

j ;v

O ^ to O *o O O 'o to o i n O O to
A lO [ j N VO N Tf U) VOo vO ro tJ- to vO i
CO -t IO VC-
c# ____i cm CO rf to

, <v<
ai C
a a H
H o
q g- v c rPS'
u tj- co H
o n o ^ o (N p o o p o o
10 O ' N N r
o (N
Tj- vO oo C
(S H H
M O'

to o \to o vo) vo) C

h o o MI IO t
1 Oo O O q
oo ^ _
rH rH r-H CNJ CO ^ yI rI CO f ' (v* i '"

o in o o o o p p p p p p
Th VO -H

cm co CO CM t
co to O vC) 'O^
O ' IO ^

O to o o
vO O t (N O ' ^
CM co co i - ' 00

On C O
cO VO H CO VO i>>

o o o o o o

O O O to O O O Q
vO cm to CM to Q 'o
t J- 00 O CM O O' ON

Vo OM to I
*'o t o O to Oo VOO t
** V O CM Tft
h M co rf to vO


vo in H c/3 H
The resistance to the flow of air producing this pressure differential is
the blanket of flowing water covering the vent opening. The higher
the rate of discharge the thicker was this wall of water and the greater
the pressure differential. Hence, a partial vacuum whose intensity de
pends on the rate of water flow must exist inside the vertical pipe
even if there is a direct opening to the air a t the vent or drain con
nection tee.' If the tee through which the air flowed into the stack
had been open directly to the air, the negative pressures inside the
stack as shown for various discharges would have been simply 1 in. of
water less than the value indicated on the curves in Fig. 5.

The fact th at there is a pressure differential as mentioned above has a

practical significance in the venting of plumbing systems. If the pressure
differential is to be kept, say, below 1 in. of water, the permissible
rate of discharge is quite low. If the allowable differential is 1 in. of
water and if the negative pressure in the vent just outside the stack
is Yz in. of water, the pressure inside the stack will be -l.S in. The
Yi in. of water would be the pressure producing the necessary air flow
in the vent, from the outside air to the stack. Tables 2A, 3/1, and 4A
show the water flow, air inflow, and stack air pressure data for the
3-in., 4-in., and 6-in. stack installations respectively, when the stack
top was open and a sanitary tee vent connection was used equal in
size to the diameter of the stack. Interpolating between the experi
mental values given it is seen that a rate of water flow above 500 gpm
for the 6-in. stack, 170 gpm for the 4-in. stack, and 100 gpm for the
3-in. stack produced a pressure differential greater than 1 in. of water.
N ote th a t in these experiments the pressure in the vent ju st outside
of the stack was -1 in. of water. Therefore when there was a pres
sure differential of 1 in. of water between inside and outside of stack
the pressure inside the stack (hv ) was -2 in. If the vent opening is
smaller than the stack diameter the allowable rate of water discharge
will be less. Table 3B shows th at for a 1-in, pressure differential be
tween inside of stack and vent, on a 4-in. stack with a 3-in. vent open
ing, the discharge was ISO gpm. Table 4 B shows correspondingly on
a 6-in. stack with a 4-in. vent, the discharge was 350 gpm.

However, there is no apparent reason why the pressure inside of a

plumbing stack should be kept a t any particular limit, as long as
there is proper and sufficient venting to keep the pressure in the ad
joining drains, to which fixtures are connected, above -1 in. of water.
If the negative pressure inside of the stack exceeds 1 in. of water,
every drain or fixture connected into the stack through a multiple
opening fitting at th at particular level should have a separate vent.
Consider a specific case on the 4-in. stack and refer to the data shown
in Fig. 5 (a ). If the pressure in the vent is not to exceed -J4 in. of
water, and if the water rate is greater than. 110 gpm, the pressure in
the stack will be greater than -1 in. of water, and every drain or fix
ture connected to the stack a t that point must have a vent leading to
the outside air. The value of water discharge obtained from Fig.
5(a) is th at for which the pressure inside of the stack was -1.5 in. In
the tests the pressure in the vent was kept a t -1 in. of water and a
pressure of -1.5 in. inside the stack therefore, means th a t there was
a pressure differential of Yi in. between a point inside the stack and
one just outside the stack in the vent.

Tables 2, 3, and 4 show water flow, air flow, and air pressure for the
3-in., 4-in., and 6-in. stacks. In order to be able to predict roughly the
maximum air inflow for a particular size of stack, it is necessary to de
termine the factors upon which the actual maximum air flow depends.
Since the air is carried along in the stack by the rapidly flowing water,
the theoretical maximum air flow for any rate of discharge is equal
to the product of the area of the stack not occupied by the water and
maximum water velocity:

Qn:ax. = ( ^ - ^ ) Vm (9)
4 Vm
in which:

Qn ax. = theoretical maximum rate of air flow, cu. ft. per sec.
Qw = rate of water flow, cu. ft. per sec.
Vm = maximum velocity of water, ft. per sec.

The actual air inflow will undoubtedly be less than this theoretical
maximum rate. For one test on the 4-in. stack and one on the 6-in.
stack (see Table 3B and Table 4B) the air flow into the top of the
stack was determined for rates of flow a t which the air inflow in the
vent below the water entrance had also been measured. The sum of
the air flow into the top and th a t through the vent was found to agree
quite closely with the computed theoretical maximum rate of air flow
when the vacuum in the stack was small, indicating that the stack
was getting all the air it required.

Theoretical maximum rates of air flow ( Qmax.) computed from
the water discharge and velocity data (Column 5, Tables 2, 3, and 4)
reached a maximum and then gradually decreased. N ote th at the values
of actual air flow, shown in Column 3 of Tables 2, 3, and 4, decrease
relatively much more after reaching a maximum than do the computed
values of theoretical air flow. The reason for this is th a t the actual
air flow, besides being a function of the theoretical air flow, also de
pends on the conditions at the vent opening into the stack. At high
rates of water discharge the blanket of water across the vent opening
produces enough resistance to the inflow of air to decrease the air flow
appreciably. In other words for any stack installation there is a defi
nite maximum rate of air inflow coming within a range of water dis
charge that stacks are ordinarily designed to carry. If vents are pro
vided to supply this maximum rate of air flow the worst possible con
dition of venting will be provided for.

In an attem pt to find some criterion that would indicate the rate

of water flow producing maximum air flow for a given stack installa
tion, a study was made of the pressure data inside the stack, both
with and without air inflow, for any given discharge. The relative
amount of air inflow will depend on the amount the negative pressure
is reduced at the point on the stack where the air enters; in other
words on the difference between the pressure reading with no air and
with air, (knv hv). (See Columns 7, 8, and 9 of Tables 2, 3, and
4 ). Also, if the remaining negative pressure in the stack with air
flowing in is high, the air inflow will be relatively low. Therefore,
the actual air flow should be a function of the ratio, JL. This
quantity was computed for various rates of water discharge, and in
every case its maximum value coincided with the water discharge pro
ducing maximum air inflow. The rate of water discharge producing
the maximum rate of air inflow can, therefore, be determined for any
given installation by experimentally determining the maximum value
of the above indicated quantity.

The last column of Tables 2, 3, and 4 shows the ratio of actual to

theoretical air flow. The maximum value of this ratio for open stacks
with sanitary tee connections was .54 for 3-in. stack and 4-in. vent,
.61 for 4-in. stack and 3-in. vent, .58 for 4-in. stack and 4-in. vent,
.63 for 6-in. stack and 4-in. vent, and .79 for 6-in. stack and 6-in.
vent. These ratios correspond to the water flow producing maximum
air flow or to the design rate. Thus if the water discharge produc
ing the maximum rate of air inflow has been determined, the actual
maximum rate of air inflow can easily be calculated for common sizes
of stacks from the above data. For example, suppose for a given
4-in. stack installation it is found th at a water rate of 200 gpm will
produce the maximum rate of air flow. W hat will be the maximum
possible rate of air inflow through a 3-in. vent opening?* For 200
gpm the maximum water velocity is about 20.5 ft. per sec. (See
Column 3, Table 1). Therefore from Equation (9) the maximum
possible rate of air flow is,

0 . = . > ( i - i s o ^ a s ) - 20' 5
= .81 cu. ft. per sec., or 36S gpm.

Referring to Columns 1 and 2 of Tables 2, 3, and 4, note th at for

the test installations with sanitary tee vent connections and open stacks
the rates of water flow producing maximum air inflow are: 130 gpm
for the 3-in. stack, 190 gpm for the 4-in. stack, and 550 gpm for the
6-in. stack. For other installations these rates may be slightly differ
ent bu t there seems no reason to suppose the variation will be great.
Also, these rates should apply to 45 wye vent connections. However,
these rates will be considerably less for stacks with tops closed or
with top openings much smaller than the stack diameter.

13. Summary and Applications. This investigation may be sum

marized as follows:

(1) W ater flowing down a vertical pipe creates pressure differentials

from the point of water entrance to the bottom of the stack. These
pressures vary in magnitude and sign depending on the distance be
low the entrance and on installation details. The curves showing pres
sure variations along the stack are similar in shape for all rates of
water flow.

(2) The pressure differentials for the greater portion of the stack
are negative and relatively large, approximately 30 in. of water. In
any plumbing system these large pressure differentials will break
the water seals of fixtures and they must be relieved by supplying air.

*This does not im ply t h a t the v e n t pipe its e lf m u st be 3-in.
(3) Where fixtures are connected to a stack and where a negative
pressure exists, air must be supplied for the duration of the water flow
a t a definite rate depending on the water discharge and the ability of
the water to carry the air along.
(4) For any stack there is one rate of water discharge which pro
duces the maximum rate of air flow. For a given installation this
water discharge can be determined experimentally by finding the

maximum value of the ratio, ^ nv, ^V' > where hnv is the negative
pressure inside of the stack w ithout air inflow (vent closed), and hv
the negative pressure inside of the stack with air inflow (vent open).
(5) The rate of air inflow cannot exceed the product of the area
of the stack occupied by air and the velocity of the falling water.
The actual maximum rate of air inflow observed in these tests was
approximately 50 per cent of this hypothetical maximum for the 3-in.
stack, 60 per cent for the 4-in. stack, and 75 per cent for the 6-in.
stack. These values are for sanitary tee connections and vent open
ings of size equal to the stack diameter. Other types of openings may
give values slightly different depending on how the water flows by the
vent opening; however, there is no reason to believe th at these values
would be different for the common 45 wye connections.

Bulletin 1. The Flow of W ater Through Culverts, by D . L. Yarnell,

F. A. Nagler, and S. M . Woodward, 1926. 128 pages, 26 figures, 23
plates, price $1.00.

Bulletin 2. Laboratory Tests on H ydraulic Models of the Hastings

Dam, by M artin E. Nelson, 1932. 72 pages, 40 figures, price $0.75.

Bulletin 3. Tests of Anchorages for Reinforcing Bars, by Chesley

J. Posey, 1933. 32 pages, 18 figures, price $0.50.

Bulletin 4. T he Physical and Anti-Knock Properties of Gasoline

Alcohol Blends, by Theodore R. Thoren, 1934. 32 pages, 13 figures,
price $0.35.

Bulletin 5. The T ransportation of D etritus by Flowing W ater, by

F. T . Mavis, C hitty Ho, and Yun-Cheng T u, 1935. 56 pages, 15
figures, price $0.50.

Bulletin 6. An investigation of Some H and Motions Used in Fac

tory W ork, by Ralph M . Barnes, 1936. 60 pages, 22 figures, price

Bulletin 7. A Study of the Perm eability of Sand, by F. T . Mavis

and Edw ard F. Wilsey, 1936. 32 pages, 12 figures, price $0.35.

Bulletin 8. Radiation Intensities and H eat-Transfer in Boiler F u r

naces, by H uber O. Croft and C. F. Schmarje, 1936. 32 pages, 17
figures, price $0.35.

Bulletin 9. A Summary of Hydrologic D ata, Ralston Creek W ater

shed, 1924-35, by F. T . Mavis and Edward Soucek, 1936. 72 pages,
25 figures, price $0.50.

Bulletin 10. R eport on Hydraulics and Pneumatics of Plumbing

Drainage Systems I, by F . M . Dawson and A. A. Kalinske, 1937.
32 pages, 5 figures, price $0.35.