Journa l of Research of the Nationa l Burea u of Standards
6, December 1956 Research Paper 2721
Scavenging Characteristics of a Two-Stroke-Cycle
Engine as Determined by Skip-Cycle Operation
P. M. Ku and T. F. Trimble
A m ethod for determin ing t he m ass fraction of fr esh charge in t he cylinder of t he two-
sLrok e-cycle engin e, from m eas uremen ts of engine p ower u nder norm al op eration and wi th
engine fired only once in several cycles of operation is described. T he r esults obtain ed from
a crankcase-scav en ged en gine are presented .
1. Introduction
The Lwo-stroke cy cle engine, in order to develop
one power cycle for every r evolution of th e crank-
sh aft , mu sLrely on th e acLion of an auxiliary device,
known as the scavenging pump , to expel th e combus-
t ion products from th e cylind er and to fill the cyl-
inder wi th fresh charge. Th e scavenging pump may
b e and frequenLly is ph ysically separate from the
engin e cylinder proper. In th e crankcase-scavenge d
engine, th e cylinder-crankcase-piston combination
serves as th e scavenging pump .
The ideal obj ective of the scavengin g proccss is, of
course, to r eplace compleLel.v th e combustion prod-
u cts in Lh e engine cylinder wiLh fresh cbarge deliver ed
throu gh Lh e inleLpor Ls, wiLhou t m ean t im e losing any
of th e fresh ch arge throu gh the exh aust ports. In
this manner , maximum power is derived from th e
cylinder wi th th e leasL exp enditure of scavenging-
pump power . I n practice, su ch an ob.i ective is never
attainable, h ence the cxcellence of the scave nging
process must b e judged by t he mass of fr esh ch arge
r etained in th e engin e cylinder p er uni t time, in
r ela tion to the mass of fresh charge supplied p er uni t
t ime by th e scavenging pump .
It follows from the above tha t two mass r ates per
uni t time, viz ., th at suppli ed by th e scavenging pump
and th at retained in the engine cylinder , are r equired
to defin e th e scavenging characteristics of an engine.
Of th ese two quantities, th e mass supplie d b~r the
scavenging pump p er unit time can be directly
m easured by the use of a flow me ter placed in th e
inlet system. The mass retained in the engine c~-l­
incler per un it time is not, h owever, subj ect to a
direct measuremen t. Fo r its determina tion, a num-
b er of indirect me th ods h ave bee n devi sed and
studied [1,2).1 In gen eral, such met.h ods ar e ei ther
ex tremely laboriolls, or li a bl e to co nsiderable errors.
This paper describ es a rath er simple method for the
determination of a n importa nt rela led qu antit.\" th e
mass fraction of fresh ch arge in the engin e cylin der
after port closure. On the basis of tes ts perfonnecl
on a crankcase-scavenged engine, this m eth od ap-
peared to be quit e practi cal and reliable.
1 Figures in brackets indicate t he literature refere nces at the end of th is paper.
4080G3- 56--2
Vol. 57, No.
2. Nomenclature
For analy t ical purposes, it is convenient to r edu ce
th e two mass rates m en tion ed above to nondimen-
sional quantities. In this connection, several cEf-
feren t systems of nondimensionalization [1 ,3] h ave
b een proposed. E ach proposed sys tem h as cer tain
a t tractive fea tures, t hu s the par ticular choice is, at
least to som e extent, a ma tter of personal prefer ence.
A s~Ts tem r a ther similar to tha t employed by
Taylor and Rogowski [1] is used . To nondimen-
sionalize th e m ass of fr esh charge supplied by th e
scavenging pump per unit time, th e term "scaveng-
ing ratio" , R s , as defin ed belOW, is introduced,
R_ ]o.£g •
(1)
s- PsVdN
T o nonclimens ionali ze the mass of fresh ch arge r e-
tain ed in th e engine cylinder p er uni t tim e, th e t erm
"scavenging e ffic i enc~-", es , is in trodu ced,
Af T
es = V 7o. T (2)
Ps ai '
where
Mg=mass of fresh ch arge supplier] b~' Lh e
scavengin g pump pel' uni t time,
Mr=mass of fr esh charge retain ed in th e engine
cylinder after por t closure, per uni t of
time,
ps= density of fresh charge compu te d on Lb e
basis of inlet tpm pc rature and exh aust
pressure,
Va= displacemenl volume,
N = engin e sp eed , rev olu tions p CI' uni t ti me.
I n th e above expressio ns, di splacement volume
Ildis used b asically as a m a LLer of con ve nie nce, for
this is th e volume used in definin g the mean crfective
pressm e of the engine. The introdu ction of exh au st
pressure into density term Ps is especiall conveni en t
for engines with la Le exh aust closing, in which t he
exhaust pressure rather than the inlet pressure
exer cises controlling infl uence on th e density of the
cylinder conte nts at the time of port closure.
325
 For the purpose of the present paper, it is helpful
to note that the quantity Mr is related to the cylinder
volume at the instant of port closure , as follows:
(3)
where
e= mass fraction of fresh charge in the engine
cylinder after port closure,
px= average density of the cylinder contents at
the instant of port closure,
Vx= cylinder volume at the instant of port
closure.
Combining eq (2) and (3), the scavenging efficiency
may then be written
(4)
This paper deals with a method for the determina-
tion of the quantity e, and eq (4) shows the relation
between this quantity and scavenging efficiency.
The ratio Vx/V d is obviously determined by port
timing. No attempt has been made to determine
the ratio Px/ Ps in the present investigation. How-
ever, some qualitative remarks may be in order here.
The ratio Px/ Ps is, in general, influenced by heat
transfer between the fresh charge and the engine
parts, flow resistance at the inlet and exhaust ports,
and the dynamics of flow in the scavenging system.
For engines with late exhaust closing, if the heat
transfer, flow resistance, and flow dynamics effects
were negligible, then the ratio Px/ Ps obviously would
become unity. Thus, Px/Ps provides a measure of
the combined effect of heat transfer, flow resistance
and flow dynamics in this instance.
For engines with late inlet closing, the relation
may best be examined by writing
where
Pi= density of fresh charge at inlet temperature
and inlet pressure,
pt= inlet pressure, absolute,
p .=exhaust pressure, absolute.
Applying this expression to eq (4),
Here the ratio p dp. accounts for the static effect of
inlet pressure, as against the exhaust pressure used
in the definition of scavenging efficiency. For
engines with late inlet closing, the ratio Px/ Pi provides
a measure of the combined effect of heat transfer,
flow resistance, and flow dynamics . In the ideal
case where these effects are negligible, the ratio
Px/Pi reduces to unity.
3. Principle of Skip-Cycle Method
In the normal operation of a two-stroke cycle
engine, when the engine fires every cycle, the
cylinder is full of combustion products at the begin-
326
ning of the scavenging process. At the end of the
scavenging process, only a fraction of the combustion
products remains in the cylinder. Now, consider
the case when the engine is allowed to fire for one
cycle and then skip, or run without firing, for several
cycles. Obviously, any combustion products that
remain in the cylinder must come from the firing
cycle. Accordingly, the progressive scavenging ac-
tion of the skipped cycles results in progressively less
combustion products, or progressively more fresh
charge, in the cylinder after each skipped cycle.
The above condition can be rather satisfactorily
represented by a simple mathematical relationship .
Let j denote the mass fraction of residual gas re-
tained in the cylinder during normal operation.
The mass fraction of fresh charge during normal
operation is then
e= l - .f. (6)
Let it be assumed that the conditions of heat trans-
fer, flow resistance and flow dynamics are not signifi-
cantly influenced by skip cycling- an assumption
which appears to be justified by the experimental
results reported herein. In that case, the total
mass of retained cylinder contents would remain
substantially constant, and likewise the mass
fraction of cylinder contents to be passed from one
cycle to the next. Consequently, if the engine is
allowed to fire only once in n cycles of operation,
then for each of the firing cycles, the mass fraction
of combustion products remaining in the cylinder
would be jn, and the mass fraction of fresh charge
in the cylinder after port closure would be
(7)
Figure 1 gives a plot of (l - jn) j (l - j) versus f,
for several values of n. The curve for n = ex:>
represents the limiting case where the cylinder con-
tains nothing but fresh charge.
0 .I
,.J~
I
"
"
1
I 1lL
60
I 1/1/
0
I
1/// .I
.0
/;V/v
Jt~ / v /
1/
0
~~
L----------I--- ' ~
t=:==
...--. ~ r:::::---
" I ,-
10
ID
FIGURE 1. Plot of (1 -fn) / (1 - f) versus f.
 The skip-cycle method for the determination of
E is based on the premise that at constant fuel-air
ratio and b es t-power ignition timing, the indicated
thermal effi ciency of a given internal combustion
engine is very n early constan t over a wide range of
operating conditions. The indicated mean effective
pressure of a two-stroke cycle engin e supplied with
a premi;{ed fuel-air charge, a nd operating normally,
can be exp rC'ssed, in terms of scave nging effic i e n c~T
defin ed pre v iou sl~-, as follows,
(8)
where
imep = illdicated m ean effective pressure (nor-
mal operation),
F = fuel-air ratio
Ec = heating valu ~ of the fuel ,
1) ;= indi cated thermal effi ciency based on
fu el quantity retained in t he cylind er,
J = m ec han ical equivalent of h eat.
Appl~'ing eq (4) to thi s express ion ,
Th e engine employs conventional loop scavengin g.
The total transfer port area at t he erank:ca e is
1.27 in .2, the total inlet port arca is ].41 in. 2, and
the total exhaust port area is 0.89 in .2 Th e rotary
valve has a maximum opening of 1.32 in. 2 The
port timing of the engin e is shown in figure 2.
The engine was connec ted to a small electric dy-
.n amom etcr. The enginc speed was measured by an
electric tachometer, accurate sp eed con trol bein g ac-
complished with the aid of a stroboscop e directed
upon a striped fly wheel. An MIT hydnwlic scale
[4] was u sed for dynamom eter torque measurem ent.
Figure 3 is a schematic diagram of the experimental
set-up . As indi cated , air was supplied by a centrif-
ugal blower, throu gh appropriate controll ing and
measuring devices, to the engine rotar~T valve inlet
(herein defmed as the engine inlet ). Fucl (a 20:1
gasolin e-oil mixture) was supplied under pressure,
through appropriate controlling and m easurin g de-
v ices, to the inlet surge tank, where it was mixed
T DC

(9)

Under skip-c.\Tcle operation, eq (9) can be modified

to read
(10)

wh ere ime pn is t he imep of tJIC firing cycle wiLh firing

occurring once every n cycles of operation. The
quantities Px, F, and 1) i a re constant, under the as-
sumptions made previously.
Combini ng eq (9) and (J 0), and introdu cin g t he
values for e and en as given b~- eq (6) a nd (7),
imepn e 1- f "
(11)
imep = en = 1- f'
It is seen that as n approach es infinity, en approaches
unity for any value of f less than unity, in which case
e= imep!imep m as a limi t. This r elationship,
though simpl e, is unfortunately difficult to apply in BDC

practical determination of E, because it is difficult to Fw l ' RE 2. Port timing oj the engine used 'i n the ex periment.
extrapolate with any reasonable accuracy for the
value of imep m from r eadings taken with the neces-
sarily limited values of n in any practical experiment. Flo _ meters

Experience indicates that it is far more sa tisfactory Throttle

to solve for t he value of f from the ratio of imepn! Air
imep for filliLe valu es of n, and then obtain e= I - f.
Figure 1 can convenien tly be used for this operation .

4. Experimental Equipment ASME

Ori f ices

A small single-cylinder, air-cooled, crall kcase- Cooli ng

Air
scavenged engin e was u sed for the exp erimental
investigation. Th e principal dimensions of t his
engine are as follows:
Bore __________________________________ 2. 500 in .
Stroke __ ____________________________ 2.000 in.
Dis place men t vo lum e __ _______________ 9.818 in .3 Elhausl
Blo we,
l
Cy linder compression ratio ______________ 7. 05
Crank case co mpression raLio _____________ l. 44 FI GU R E 3. Schematic diagram of the eng'ine setup.

327
with the inlet air. The air rate was measured by
calibrated sharp-edged orifices of the ASME flange
taps design [5] . Two orifices were used to cover the
range of the experiment. Calibrated float-type flow-
meters were used for fuel rate m easurement. Again,
two instruments were used to cover the entire ex-
perimental range.
The engine was enclosed in a duct, and was cooled
by air supplied b y a blower. T h e stations for the
measurement of pressures and t emperatures are indi-
cated in the figure. All temperatures were m easured
by m eans of thermocouples. The pressures wer e
measured by mercury and water manometers.
Figure 4 shows the ignition circuit used in th e ex-
periment. The circuit was designed b asically to
operate as a conventional automotive ignition sys-
tem . There were four sets of breakers: the main
breaker, a, oper ating at crankshaft speed , and the
auxiliary breakers, b , c, and el, operating at reduced
sp eeds. The reduced speeds were obtair ed by mount-
ing the cam shafts , b , c, and d , in a gearbox giving
speeds in th e progression of 1, K and X. The whole
assembly was then driven through chain and sproc-
kets b y th e crankshaft. The sprockets could b e
changed to furnish a speed ratio of either 2 or 3
between the crankshaft and shaft b . In this fashio n ,
the auxiliary breakers could operate at either ~, X,
a nd Ys of crankshaft speed , or ?~ , ?t, and 7\2 of crank-
shaft speed.
The main breaker was used to control the spark
advan ce at all times . The spark advance was r ead
by an automotive timin g light. To obtain normal
operation (n = l ), switch A should be closed, in which
case a spark would b e produced every time the main
breaker was opened. For skip-cycle operation, the
switch giving the desired value of n should b e closed
and all other switches opened. In tha t case, th e
prim ary circuit would be completed only for the
cycle for which ignition was d esired . For th at cycle,
a spark would be produced at the instant the main
breaker was opened.
The indicated m ean effective pressure, i mep, was
computed from dynamometer readings taken with
en gine firin g and motoring, b y the familiar relation
Breaker
Spark
"'""
Timing
Light
FIGURE 4. Circuit diagram of the i gnition system.
where
0 = a constant,
P b= dynamometer reading taken with engine
firing,
P r = dy n amometer reading taken with engine
motoring.
Under skip-cycle operation, the indicated m ean effec-
tive pressure of the firing cycle, imep n, was computed
from the dynamometer readings as follows
imep n= n C(P b+ Pf)'
where n, as defined earlier , is the numb er of cycles
of operation p er firin g cycle.
5 . Results and Discussion
Three series of runs wer e made, with the engin e
speed, inlet pressure and exhau st pressure varied in-
dependently. With the exception of the variable
whose effect was being m easured, all operating vari-
ables were maintained constant at values shown
below:
N, engine sp cecl ____ _____ ___ _rplll __ 1, 800
pi, inlet prcssure (i nches of mercury,
absolute) ______ _____ __ " H ga __ 29.92 ± 0. 0l
Pe, ex haust press ure __ __ ____ _" H ga_ _ 29. 92 ± o. 03
T i , inl et te mperature ___ __ ___ _ _°F __ 100 ± l
T a , coolin g air outlet t emperature
°F __ 130 ± 3
Ji', fu el-air ratio ______ ___________ __ 0.080 ± 0. 001
The high est engine sp eed used in this experiment
was 3,000 rpm. Limitation of the dynamometer, as
well as vibration s of the installation, did no t p ermit
operation at higher sp eeds.
Figure 5 shows the variation of imep, ihp, and air-
flow rate, M g , of the engine, under best-power spark-
advance conditions, over a fairly wide range of en -
gin e sp eed, inlet pressure, and exhaust pressure.
Figure 6 shows the variation of imep n versus spark:
advance under skip-cycle conditions, at 1,800 rpm .
Note the progressive clecrease in best-power spark
advance, and th e pro gressive increase in imep n, as
the value of n was increased. 2 An indication of the
degree of r eproducibili ty of th e experiment can be
had by noting th at results of three sets of runs, m ade
Reducl ior, Normol
on three different occasions, have been included in
Gear Ralio Position this figure.
,
I close
open
From data presented in the form of figure 6, the
2,
4,
open
open
values of b est-power spark advance and imep n under
b est-power spark-advance conditions, could b e de-
rived for each set of op erating conditions. In this
manner, figures 7 and 8 were constructed. Kote
hO"lv the imepn curves ·with varied exhaust pressure
cross each other, and cross the curves with varied
inlet pressure (fig. 8). This b ehavior clearly shows
that a v ariation of exhaust pressure affected not
only ~, but also Px, in eq (9 ).
The imep n/imep curves shown in figures 9 and 10
were derived from the imep n data presented in figures
2 'I'be variation of ihe best· power spark ad vance reflects ihe elYect of , on flame
speed. The variation of ;.mePn is expected on the basis of eq (11) .
328
 E
"-
.0
'"a.
~
c;, a.
='E .c '"E
60 6 80

70

50 5 60 . ..-x
Ime p '-.......... ............. x --'
50 ...----
4 0 4 40

30

30 3 20

10

20 2 0

10

0 0
o 500 1000 1500 2000 2500 3000 27 28 29 29.92 29.92 31 32 33
N, rpm Pj , " Hgo
FIGURE 5. P e1jonnance characteristics of the engine.
Operating conditions (unless otherwise specified): 1,800 rpJl1 ,·p .= p. = 2D.92")Tga,~T.= l OO° Y, fuel·air ratio = 0.080. best·power spark advance.

120
n' I 'pm

80 4 • 3000
100 ~+ -+---+ 0
2400
1800
60 ~ ~/o "_ "- 0 ~x

~:==:-.
1200

100 '"a.
c
80

60
i 60n

n. 0- /"

V'
80 <>
E
60 Symbol 'pm
40
.~
600
40 x 1200
f 100
n=4 20
0
+
180 0
2 400
4 3000
80

0
60

40

1_
::s::: lI'J 40
100 0:",
it'"
<f)ffi J rpm
0:'" 2;3~ J
'f~~O~~t~~§!~.:::~~~~
80 4
?t°. 20
0'"
60
Q.
,
U
Z • :~gg
'_ . _ x
f- "
<f»
. . -- 600
", 0

40
CD" 0
I ! ! ! !

0 10 20 4050 01020 40 50 3 5 6 8 9 10 II 12
SPARK ADVANCE, DEGREES n
FIG UR E 6. I mepn versus spark advance, for varying degrees of FIGURE 7. Best-power spark advance and imepn at best-power
skip cycling. spark advance as affected by skip cycling, at several engine
Operating conditions: 1,800 rpm, p.=p.=2D.92" llga, Ti= lOOOF, fuel·airratio = speeds.
0.080. 'fhe three different symbols denote resnlts of three runs under identical
conditions, made on three different occasions. Operating conditiom: ]Ji=p.=29.D2" lIga, T i= 100 0 F, fuel·air ratio = O.080.

329

L
 P j / Pe
2 .4

;:-:
100 ______0
0 . 92

2 .2 0 . 925
80
0-

2.0
'"a. 60
1
-o?: --
a.
c a. -0 0.96
OJ
E -l
E 40 Sym bo l

0 -
Pi, "Hgo

29 . 92
Pf!' "Hgo

29.92 "c:
1. 8
f 0.9 75
Cl.
0-
-0 ---
Vo r ie d '
29 .92
29.92
Var i ed
E1.6 0- _ _ 0 -- -0
0
1.00
20 /0
0

1.4
0
I Sy mbol Pi. w Hga Pe , "Hga
o
0 29.92 2 9.92
0- Varied 29.92
1. 2 -0 29.92 Var i ed
60

~
<>
1.0
L' ,

o
9"~..... _
1.'-..

-
~

8t:~;~><~~---
-g---=.-o- --

I
- 1·.,·,
/ 0.925
~~0.92
_i} - 0.975
1'::.0·96
1.00
\JJ
0.7

0.6
- 0 - - 0 - - 0 - -- 0 -

- - 0-
- - -0
0-
-0
_-0

g
0
1
Pi I P e

1. 00

0 .975
0.96
2 4 6
n
7 9 10 II 12
0 .5
- - 0-
0-
if 0 . 925
FIG U RE 8. B est-power spa/'k advance and i m ep n at best-power ---0 -0
-0 0 .9 2
spa/'k advance as affec ted by skip cycling, at several inlet and 0 .4
exhaust p ressures. 3 6 8 9 10 II 12
Operating cond itio ns: 1,800 rpm , '1';= 100° F , fuel-air ratio=0.080. n
FIG U R E 10. M"ass fraction of fresh charge in the cylinder, at
several values of p d p ,.
Operatin.g condit ions: see fig u re 8.

rpm

t
1.8
600

i 1200
0

a. 1. 6
OJ
E
r 1800
2400
3000
0.7

"a.c:
Q)
1.4
0.6

E rpm
600 0.5
1200
1. 2 o 1800
+ 2400 0 .4
3000
\JJ
1.0
0.3

Sy mbol rpm Pj."Hgo Pel "Hgo

0 .7 0.2 0 1800 29.92 29 .92
0- 1800 Vari ed 29 .92
-0 1800 29.92 Vari ed
0.1 X Vari ed 29.92 29.92
\U 0 . 6

a
0 .5 a 0. 1 0 .2 0 .3 0 .4 0 .5 0.6 0 .7 0.8
2 3 4 5 6 7 9 10 II 12 Rs
n
F IGUR E 11. Nlass frac tion of fresh charge in the cyli nder as
F I GU RE 9. lvlass fraction of fresh charge i n the cylindel', at a function of scavenging ratio, for range of engine speed,
several engin e speeds. inlet pressure and exhaust pressure investigated.
Operating conditions: ~E'e figure 7. Operating conditions: Ti=lOOo P , fuel-air raLio=0.08~ .

330

L
7 and 8. Note that because the variable Px was
eliminated in this method of presentation, the imepn/
imep curves in figure 10 appeared in a systematic
order with respect to the ratio p dp •.
FigUl'es 9 and 10 also present the values of f , as
derived from the im epn/imep data given in the same
figures, by the method outlined in section 3. Note
that the value of f for each case remained substan-
tially constant over the range of n investigated.
This constancy of the value of f is gratifying, in view
of the complexity of the scavenging process and the
simplifying assumptions made. By averaging the
results obtained at several values of n, a rather
reliable value for f may thus be derived .
The results presented in figures 9 and 10 are SUlll-
marized in figure 11. It is interesting to note that
the mass fraction of fresh charge in the cylinder was,
within experimental accuracy, a function of scaveng-
ing ratio only, regardless of the individual val ues of
engine sp eed, inlet pressure, or exhaust pressure,
within the range of the variables inves tigated .
331
The assistance provided by Edward Bryant in
setting up the experimental eq uipment and in oper-
ating the engine is gratefully acknowledged.
6 . References
[1] C. F. Tay lor and A. R . Rogowski , Scavcnging the two-
stroke en gine, SAE Trans. 62, 486 (1954).
[2] P. H . Schweitzer, Scavenging of t wo-stroke cycle diesel
engines, ch. 16 (The Macmillan Co., New York, 1 . Y .,
1949) .
[3] P . H. Schweitzer, Scavenging of two-strok e cycle diesel
engines, ch. 3 (The Macmillan Co., New York, N. Y. ,
1949).
[4] H ydraulic scale for measurement of e ngi ne d ynamometer
forces, Laboratory Equipment Bulletin No.6 , Diesel
Engine Manufacturers Association (January 3, 19'19).
[5] Fluid meters-their theory and application, Report of
ASME Special R esearch Committee on Fluid Meters,
pt I (1937) .
WASHINGTON, April 5, 1956 .