British Journal of Applied Physics

You may also like

- Investigations of Ionic Transport
Transverse vibrations of power transmission Mechanism of Polyether-Based Polymer
Electrolytes for All-Solid-State Batteries
chains Yui Otake, Yuji Yokomaku, Yuna Okabe et
al.

- The theory of the chain fountain revisited

To cite this article: S Mahalingam 1957 Br. J. Appl. Phys. 8 145 Dragos-Victor Anghel

- Research on Sustainability of Supply

Chain Financial Model in Fujian Free
Trade Zone
Liu Wei and Song qianqian
View the article online for updates and enhancements.

This content was downloaded from IP address 196.203.130.7 on 09/02/2023 at 13:14

 Transverse vibrations of power transmission chains
By S. MAHALINGAM,
BSc., Ph.D., Department of Applied Mechanics, University of Sheffield*

[Papev received 10 September, 19561

In this paper the natural frequencies of a travelling chain are determined by treating the chain
as a uniform string. For a constant chain tension the natural frequencies decrease progressively
to zero as the chain speed approaches the velocity of transverse wave propagation, In practice
such a condition is not obtained since the chain tension increases with speed owing to centrifugal
effects. Formulae are obtained for the amplitude of forced vibration of the chain when excited
by a transverse displacement at one end. The effect of damping is considered and exact and
approximate methods of determining resonant amplitudes are given.

LIST OF SYMBOLS harmonics is very small provided the number of links in the
chain is large. The natural frequencies of a stationary uniform
m = mass of chain per unit length string are well known, but the effect of longitudinal velocity
7 = chain tension
has only recently been considered by Sack.(2)
T, = initial (static) tension of chain
p = pitch of chain
c = damping coefficient
V = speed of chain
L = length of chain
w, = circular frequency of free vibration -v
w = circular frequency of exciting force
Fig. 1. Displacements of travelling chain
The term “chain,” in general, would refer to the free
unsupported strand between sprockets.
In Fig. 1, let the deflexion at a point P be y. After an
interval at, the point will have moved to the position P’
IXTRODUCTION where the deflexion will be:
Transverse vibration of the strand of a chain drive may be
excited both by transverse and longitudinal displacements of
the ends of the chain. One source of excitation is polygonal
action,(’) i.e. the periodic fluctuation of the velocities of the Tinen the transverse velocity of the point is
ends of a chain caused by the fact that a chain lying on a
sprocket forms a polygon rather than a ckcle. When
sprockets with small numbers of teeth are used, these vibra-
(v-3xa f -3t)aY
tions are large and may reach resonant conditions. Vibra-
tions may also be excited by external forces, e.g. the torsional and the transverse acceleration =
vibration of a sprocket, and out-of-balance in the drive, etc.
In the case of the torsional vibration of a sprocket, the end
of the chain is given a periodical displacement which is
largely longitudinal and, to a small extent, transverse. In
practice, the most important condition is that of resonance,
in which the excitation frequency coincides with one of the The equation of motion is
natural frequencies of the chain. It is therefore necessary to
determine the natural frequencies, talcing account of the
longitudinal velocity of the chain, and the variation in tension
due to centrifugal effects. By assuming a damping force
Proportional to the transverse velocity, the resonant amplitude which may be written
of the chain may also be determined.

F R E E V I B R A T I O h-S
where V, =2/ (Tjm)is the velocity of propagation of trans-
A chain may be idealized as a light string carrying a number verse waves.
of equal equidistant particles. Although this idealization The general solution of this equation for vibrations of
Preserves the finite and discrete nature of each link, it is not circular frequency w is given by:
y y suitable for the study of a travelling chain. In the
l@llowinganalysis the chain will be idealized as a uniform y = ’4’ cos [ u t 4- uxl( v, - V ) - 541 T
h e a y string. Comparison of the critical frequencies of a
stationary chain as given by the two idealizations shows that
f A , cos [ut - WXII(V0 -I-
V) + 421 (4)
!he differencein the fundamental frequency and the first few Substituting the boundary conditions y = 0 at x = 0 and
x = L in equation (4), we get the natural frequencies
* Now at Department of Mechanical Engineering, University of
Ceylon. U, = (nrV,/L)[1 - ( V i V,)2] (5)
VOL. 8, APRIL 1957 145 BRITISII JOURNAL OF APPLIED PHYSICS
 S. Mahalingam
and the mode of vibration is given by and x=L, y = o
(lob)
x = 0, y = a1 cos w t
nrx V
L Substituting equation (loa) into equation (4), we have
where n = 1, 2, 3 , and .. is an arbitrary phase angle.
According to equation (5) the natural frequencies decrease y1 =
to zero as V approaches Vo.
(11)
and equation (lob) into equation (4), we have
EFFECT OF CENTRIFUGAL TENSION

If the tensions of the tight and slack strands of a stationary

chain drive are assumed to be T, and T,' respectively then,
when the chain is travelling at a speed V, the tensions become
+
(T, mV2) and (Ti mV2) respectively where mV' is
J- Amplitude of vibration at any point = y1 -y ~ .
the centrifugal tension. This has been considered by
Stametd3) and others. The velocity of wave propagation
Vo in the tight strand of the travelling chain is given by
FORCED V I B R A T I C N D U E TO L O S G I T U D I N A L
V,2 = T", = (T, - mV2)/m = (V: - V 2 ) EXCITATION
where V,2 = TJm. From equation (5) the natural frequencies When one of the sprockets is subject to torsional vibration,
are given by one end of the chain is given a transverse excitation
nrr
=-
V,Z a cos 2wt and a longitndinal excitation a' cos w? where w is
I? L (V,' V2)+ + the circular frequency of torsional vibration of the sprocket.
The response to the transverse excitation is given by equation
Equation (7) shows that owing to the presence of centrifugal (Il), while the response to the longitudinal component may
tension w, + 0 only as V + CO. be analysed as follows.
Forced vibrations due to polygonal action reach resonance Owing to the longitudinal excitation, the tension of the
when the frequency of tooth engagement is equal to the chain at any instant is (T 7 ATCOSut). The equation of
critical frequency of the chain. For this condition, motion (3) then becomes:
2T-
v =-
nr V,'
p L (V,' T V2)+

where p is the pitch of the chain. Putting y and w t =

= e'@."yo(t) 22 we have
Solving for V,
j;, t 2 i v p j 0 - [p*(V', - V* + h v $ c o s 2z)]yo = 0
Substituting yo = e-'V.uZ u(z) we get
U + (z f 29 cos 2z)u = 0 (141

FORCED VIBRATION D U E TO TRANSVERSE where a = p2Vt

EXCITATION 2q = p2AV;
When a chain drive runs at constant speed, polygonal Equation (14) is the well-known Mathieu's equation, a
action gives both ends of the chain periodic transverse dis- brief discussion of which is given by Den Hartog.c4) The
placements. Considering the fundamental components of response of the chain, in this case, has half the frequency of
these excitations, the boundary conditions for equation (4) the excitation.
may be written
x = 0, y = a, cos w t
x =L, y = a2 cos (ut - 0) , (9)
EFFECT OF SPEED ON THE AMPLITUDES AT
RESONASCE
where w = circular frequency of tooth engagement
= 2rV/p In order to determine the resonant amplitudes of the chain
0 = "phase difference" between the two ends of the it is necessary to take damping into account. If it is assumed
chain corresponding to the fractionai pitch in that there is a damping force proportional to the transverse
the tangent length of the chain. velocity, then the equation of motion may be written:
The response at any point on the chain will be the sum of
the responses due to the excitations given by:
x=o,
x = L,
y = o
y = a2 cos (ut f 0)
BRITISH JOURNAL OF APPLIED PHYSICS
I
, (loa)
where b = cjnz and c is the damping force coefficient. Let

146
one end of the chain be given a transverse displacement
VOL. 8, APRIL 1957
 Transverse vibrations of power transmission chains
a e'@'. Then, the real part of the response to this excitation difficult to make an accurate assessment of these factors but
!vi11 be that corresponding to a harmonic excitation a cos w t . experimental evidence, which indicates a decrease of resonant
Putting y = yo(x)e'@',equation (15) becomes amplitudes with increase of speed, suggests that a damping
force proportional to gives a fairly good
approximation.

24 a
I
I

Now, let yo(x) = eil*x. We then have

pLz(Vj- V 2 ) - p ( 2 V w - i b V ) - ( w 2 - i b w ) = O (16)
Let the solution of equation (16) be

and
Then the general solution of equation (15) is

For the boundary condition x = 0, y =0 for all t, we get

Chain rpeed Cft/rnin)

Fig. 2. Effect of speed on the critical fundamental

Substituting A i = k eim where k and 4 are real constants,
frequency and the resonant amplitude
we get
Resonant amplitudes: curve I, damping force cc
(exact); curve IA, damping force ct
curve 11, damping force cc (hy/ar) (exact).

An approximate but much simpler expression for resonant

amplitudes can be obtained by using the principle of work
done. Consider a displacement y = a cos w,t applied at the
end x = L. The major component of the response of the
chain may be written

where A is the amplitude at mid-span. This represents a

vibration of frequency w, in the fundamental natural mode,
at a phase angle 90" with the applied displacement. In
addition, a small component in phase with the excitation will
It will be seen from equation (16) that if b is small and be present. If the damping is small, this component may be
U = w, then a' and az will be such that (al - a2)L N 2.;;.
neglected.
Substituting in equation (17) From equation (19) we have
1
resonant amplitude at mid-span = ae-E1(L,2) - (18) [b y/3x],,, = @TIL) sin w,t

The transverse component of the chain tension at x = L is

A numerical example considering a 2 in. pitch simple ro!ler
chain therefore
(TA.ir/L)sin w,t (20)
= 0.066 lb per in.1386 L = 36 in. T, = 5'0 Ib from which the work done per cycle
b =2.5 sec-' units
is given in Fig. 2. Curve I, which represents resonant
amplitudes for the fundamental frequency, shows a steady
decrease with speed-a trend which has been verified experi-
" d y by the author. The resonant amplitudes for a
JO
TAT .
L =I d
sin w,t - (a cos w,t)dt
~

dt

damping force proportional to 3y/Jt is given by curve II. _-TaA7i2

-
This, however, shows an increase of resonant amplitudes with L
Wed. Now consider the energy dissipated in damping.
The damping force in a vibrating chain is provided by the
frictional resistance between the pins and bushes during the
afilculationof the links, the resistance of the air, etc. It is
VOL. 8, APRIL 1957 147
Damping force = bm V- - - y
;x
BRITISH JOURNAL OF APPLIED
( -3:) PHYSICS
 S. Mahalingam

Total energy dissipated per cycle the resonant amplitudes given by equation (24) are repre-
sented by curve IA in Fig. 2. Owing to the assumptions
made, the curve is necessarily approximate and shows
considerable discrepancy at high chain speeds.
J J
0 0
Substituting for y from equation (19), we have
ACKSOWLEDGEMENTS
energy lost Vr 2 r L V27; 2 iiL
=bi?iA2[(+) ,T(W+W,) G] (22) The above paper was part of a research project carried out
at this University, and the author wishes to express his
From equation ( 5 ) indebtedness to Prof. W. A. Tuplin and Dr. R. P. N. Jones
w, 4-(V2i;/V0L)= iivo/r for their valuable criticism and advice.
Substituting in equation (21) we get
energy lost = ~ ~ w I ~ ( ~ ~ /V,Z+
~ c oV')
,L)( (23) REFERENCES
Equating equations (21) and (23) (1) MORRISON, R. A. Machine Design, 24 (9), p. 155 (1952).
A = -Ta- - 2w, 1
(2) SACK:R. A. Brit. J. Appl. Phys., 5, p. 224 (1954).
nlb 7~ ( V , - V 2 )
(3) STAMETS, W. K. Trans Amer. Soc. Mech. Engrs, 13,
- 2v; (V,Z - V')+ p. 655 (1951).
-a.- (24)
bL (V,' - 2V') (4) DEN HARTOG.J. P. Mechanical Vibvations, 3rd Ed..
This shows that A decreases with increasing V , and A + 0 p. 415 (New York: McGraw-Hill Book Co. Inc.,
as V --f W. For the numerical example considered earlier, 1947).

The effect of sulphur and oxygen on the electrical properties

of oxide-coated cathodes
By G. S. HIGGISSON,B.A., Physics Department, University College of rjorth Staffordshire
[Paper received 16 November, 19561
Poisoning and recovery experiments using both oxygen and sulphur substantiate the Loosjes
and Vink theory of conduction. Sulphur is found to have a greater poisoning effect than oxygen.
A possible reason for this is suggested.

The deleterious effect of oxygen on thermionic emission from the vacuum tube in order that comparative measurements
oxide-coated cathodes has been subjected to extensive could be made on the same cathode.
investigation by Metson,(') Wagener") and others. Com- Oxygen and sulphur poisoning and recovery experiments
paratively little is known about the effect of sulphur, although were performed at a number of temperatures between 500
Stah1c3)has reported that cathodes subjected to atmospheric and 1CCO'K. Fig. 1 shows the behaviour of the emission
pollution by sulphur before assembly in a vacuum lilbe give and conductivity after poisoning in a typical experiment at
reduced emission on subseqnently undergoing the carbonate- high temperatures (in this case 900" K). Recovery from both
to-oxide conversion. sulphur and oxygen poisoning at this temperature was
Measurements have been made of the effect of oxygen and found to occur but recovery from sulphur poisoning was a
sulphur on barium-strontium oxide cathodes over a wide somewhat slower process. The emission and conductivity in
temperature range. Probe-diodes, of a type described by each case followed similar recovery patterns. This kind of
Shepherd!" were employed in order that the behaviour of behaviour was always observed at temperatures above 800$K
both the electrical conductivity of the oxide and the therm- and is consistent with the pore-conduction theory of Loosjes
ionic emission could be studied. The experimental procedure and Vink,(j) according to which conduction in this range of
described by Shepherd was used: briefly, the temperature was temperature is in the main a thermionic emission process in
set, the cathode was poisoned by the controlled admission of the interstices of the porous oxide coating.
sulphur vapour or oxygen to the space round it until the The behavionr of a cathode after poisoning at a tempera-
thermionic emission was reduced to about 10% of its original ture of 550" K is shown in Fig. 2. No recovery of emission
value, and then both the emission and electrical conductivity was observed at this temperature and the conductivity after
of the cathode were measured during the following ten an initially rapid decrease during poisoning, continued to
minutes or so. A tungsten filament coated with barium diminish during the subsequent ten-minute period. The
peroxide was used as a source of oxygen and a similar filament reduction in conductivity due to poisoning in the temperature
coated with molybdenum sulphide was found to be a suitable range 500-700" K was found to depend upon the temperature:
source of sulphur. A filament of each type was enclosed in the greater reduction occurring at the higher temperatures.
BRITISH JOURNAL OF APPLIED PHYSICS 148 VOL. 8, APRIL1957