Sie sind auf Seite 1von 8

Downloaded from SAE International by BP America Inc.

, Thursday, July 30, 2015

247

#150013
UNIVERSAL JOINT EFFICIENCY
By C. W. Spicer
(Member of the Society)
A very large majority of the motor cars now being manufactured
for either pleasure or business include universal joints in their con
struction. The efficiency of these parts is therefore of interest to
every one connected with the motor car industry as either producer
or user. Until recently, so far as I am aware, there has been little or
no information available on this subject. It has been known, how
ever, by those who have given the subject careful thought that the
efficiency of a properly designed and applied universal joint is only
a little below 100 per cent. This knowledge, coupled with the numer
ous difficulties incident to an accurate determination of the compara
tively small losses involved, and the consequent refinements required
for the determination, probably accounts for the meagerness of the
available information.
About three years ago Professor P. F. Walker, Dean of the Me
chanical Engineering Department of the University of Kansas at
Lawrence, undertook experiments which were carefully planned and
worked out. The experiments, which were conducted on sets of uni
versal joints furnished by the Spicer Manufacturing Company and
by Blood Brothers, have been in progress much of the time since,
having been completed only recently. The method of conducting the
experiments, the data obtained and some of the results derived, are
well described in a paper* presented before the American Society of
Mechanical Engineers.
I have three purposes in mind in drawing this subject to your
attention::
First, to emphasize some of the points brought out in the paper
referred to;
Second, to develop some of the data given into some interesting
curves not given in the other paper;; and,
Third, to place the valuable information given in the records of
the Society of Automobile Engineers.
Pursuant to the first purpose, it may be mentioned that·
(a)When the joints are properly designed and applied the power
losses under usual working conditions are negligible-far less than
is commonly supposed, judging by the number of applications to the
U. S. Patent Office on complicated and freak designs with both plain
and so-called ball or roller bearings.
(b)To secure the highest efficiency and absence of vibration, care
should be observed in two-joint construction to have the yokes on
*By Prof. P. F. Walker and W. J. Malcolmson, Dec. 2, 1912.
Downloaded from SAE International by BP America Inc., Thursday, July 30, 2015

248 THE SOCIETY OF AUTOMOBILE ENGINEERS

the two ends of the intermediate shaft lie in the same plane. The
effect on the efficiency is plainly shown by the curves in Fig. 1. The
design should preferably be such that both joints will operate at ap
proximately equal angles at all times. This can be accomplished by
pivoting the torque arm midway between the joints and otherwise
arranging the parts as outlined in Fig. 2. The pivoting point may
be anywhere on line AA.
(c) To secure the minimum of vibration, two joints disposed as
Downloaded from SAE International by BP America Inc., Thursday, July 30, 2015

UNIVERSAL JOINT EFFICIENCY 249

indicated in paragraph (b) are better than a single joint.


(d)The efficiency is practically constant above about 10 per cent,
of full load. The curves, Fig. 3 (A.S. M. E. Fig. 8), are shown ex
tended to only approximately one-half the continuous full load
capacity of the joints tested, when using very ordinary commercial
grades of lubricant, the full load capacity being determined largely
by the character of lubricant used.
(e)Between 10 and 100 per cent. of full load, the efficiency is
Downloaded from SAE International by BP America Inc., Thursday, July 30, 2015

250 THE SOCIETY OF AUTOMOBILE ENGINEERS

above 99 per cent. for all angles less than 7 degrees. This is shown
in Fig. 1 the curves being plotted from the data of Table 1. The
column "Total Losses" in Table 1 refers to the total electrical and
mechanical losses of the entire testing outfit. It is correctly assumed
that there is no measurable loss at O° deflection and that therefore

the watts lost when running at an angle, less the watts lost when
running straight, is the loss chargeable to the joints.
Permission was very kindly given by Mr. Lester G. French, Editor
of the A. S. M. E. Journal, to use the data on which this paper is
based.

DISCUSSION

R. H. Rosenberg:-I would like to know whether the joints used


in the test were prepared specially to insure perfect bearing surface
and the proper admission of lubricant, or stock tightly fitted joints
as delivered to the consumer. The majority of the joints are fitted
so tightly that they cannot be manipulated by hand. In view of the
bearing surfaces and their area, the subsequent friction set up by
the addition of the torque load, the high rate of oscillatory speed the
joint parts have to travel, the starts and stops at the completion of
each revolution, the power required to break the grip of each of the
eight bushings on the eight pins twice in each revolution, and the
friction necessary on the dust cover, it is hard to understand the high
efficiency claimed in the paper.
It is obvious that the normal speeds of the motor and the com
ponent driving parts are a great deal higher than those employed in
Downloaded from SAE International by BP America Inc., Thursday, July 30, 2015

DISCUSSION OF UNIVERSAL JOINTS 251

the test. The results at 1000 r.p.m. and over are not to be compared
with the results at speeds below 1000 r.p.m. The critical speed of
drive shafts is between 1200 and 2000 r.p.m. The company with
which I am associated passes no drive unless it has run at least 2200
r.p.m. without showing appreciable vibration.
Mr. Spicer's suggestion of means of eliminating vibration dis
closes certain weaknesses involved in the two-clevis or cross-toggle
type of joint. After an engineer has decided to use the two-joint
drive, if he must have an angle, he will put it at the forward joint,
for he reasons that he will lose a certain amount of efficiency in but
one joint instead of in both, the rear joint taking care of the frame
weave, spring deflection and the deflection caused by compressing the
torque-arm shock-deflecting springs. If the rear joint be in line with
the driveshaft and take only the deflection mentioned, it will outwear
the forward joint or may be smaller.
With the Hotchkiss type of drive, the engineer would certainly
not mount the rear joint at an angle, as this would put undue strain
on the whole axle structure and transmit vibration to the spring clips,
causing them to become loose prematurely. The proposed lay-out and
torque arm suspension would add expense, as it requires an additional
frame cross member. If two joints placed at approximately the same
angle are required for a vibrationless drive, how shall we maintain
this uniformity of angularity with the constantly varying deflection
of the springs? I venture to say that Mr. Spicer does not approve of
a single-joint drive, as it would have no counter vibration producer
as in a two-joint drive, and would therefore soon wear itself out or
cause undue strain of surrounding parts such as the bearing at the
end of the torque tube or rear end of the transmission. This drive
is, however, used extensively and must, therefore, be successful.
Owing to the ever-increasing speed of motors, all the parts that
rotate at a high rate of speed must be absolutely concentric and well
balanced; they must either be machined all over or have parts of
uniform section. This is not permissible with the clevis type of
joint, owing to the fact that the parts are drop forged, and the de
sign such that they cannot be machined over all economically.
Considering the action of the slip sleeve of spline or square hole
construction, we find faster wear, with the Hotchkiss drive, if the
joint is too rigid; a certain amount of looseness is required to facili
tate the slip on the downward Or upward motion of the car under full
torque load ; the drive-shaft will bind on not only the spline or square,
but on the extreme end of the sliding head, so much as to soon wear
out the slip sleeve or the bearings on the joint, or both. This can
not, of course, occur with a ball and socket type of joint, as no end
thrust can reach the driving mechanism. The ball floats freely in
its housing and adapts itself to the motion of the slip member.
After looking over the various types of joint you will find a lot
of clever new work. The joints are all good, some better than others,
Downloaded from SAE International by BP America Inc., Thursday, July 30, 2015

252 THE SOCIETY OF AUTOMOBILE ENGINEERS

but the most efficient is among those so unfortunate as to be classed


in Mr. Spicer's paper as complicated and freak designs, with both
ball and roller bearings.
C. W. Spicer: I am sure it was Professor Walker's intention to
investigate the most commonly used type, not make, of universal joint,
as a purely engineering proposition and in enlarging on and em
phasizing some of his conclusions I had no thought of taking any other
position. I believe Professor Walker selected the two kinds of joint
he did because they were representative of the most common American
practice of the day, the general type being what is known as the Hooke
joint by English-speaking people, and as the Cardan joint, which is
the same thing, among French-speaking people. The two kinds of
joint, i. e. the Blood joint and the Spicer joint, were selected as being
representative of the type; and only discussion of that type was in
tended by Professor Walker and surely by me. There was no intention
of going into the relative merits of different kinds of mechanical con
struction.
First, as to the joints being specially prepared for the test, I
wish to answer emphatically that the Spicer joints tested were not
specially prepared, but were standard stock joints; and I have no
reason to believe that the Blood joints had special preparation.
Second, as to the value or reliability of the curves and data given.
The joints not having been specially prepared and the test having
covered the usual working range of the joints, no good reason seems
apparent why the results are materially different than would be se
cured under usual working conditions; in other words, why the in
formation secured is not just as valuable and reliable, in its way, as
the information regarding a motor which is obtained from a carefully
conducted block test.
Third. Tests up to 6,000 r.p.m., aside from those made by Profes
sor Walker, seem to show that the efficiency of the type under dis
cussion is practically constant at all speeds up to the point where
lubrication becomes difficult or impossible due to centrifugal force
tending to throw the lubricant off the bearings and at high speeds to
split up the lubricant into its component parts. This obviously will
depend largely on the design of the joints. As to the tests having
covered the uptodate working range, it need only be said further
that 1,000 r.p.m. of the propeller-shaft correspond to a speed of a
little more than 25 m.p.h. of a car having 34-inch tires and 4:1 gear
ratio in the axle. This speed is the limit of the law in many States,
is above the average for town driving and certainly a high average
for ordinary touring.
Fourth. The representative of the Universal Machine Company
appears to confuse in his discussion of paragraph (6) of my paper,
vibration in the car due to variation in angular velocity in the pro
peller-shaft, with which I assume every engineer is more or less
familiar, and vibration caused by the "whipping" of propeller-shafts
Downloaded from SAE International by BP America Inc., Thursday, July 30, 2015

DISCUSSION OF UNIVERSAL JOINTS 253

at high speeds, which is quite another subject and not discussed in


the paper. Variation in angular velocity between the driving and
driven members is characteristic of all the usual types of universal
joints but may be easily neutralized by properly combining two single
joints into a set with connecting shaft or into a double universal joint,
both of which are common practice, although in the case of the set the
assembling workman is sometimes careless or not informed as to the
proper relation that should exist between the two yokes or parts con
nected by the set.
The vibration mentioned in paragraph (b) refers only to the vibra
tion which sometimes exists due to unneutralized variation in angular
velocity. This vibration is not noticeable in most cases, except in those
highly refined, smooth-running cars which have a relatively short pro
peller-shaft and consequently large angle in one joint, especially when
the springs are in their extreme upper position. If, in this case, the
joints are assembled so as to add the variation of one joint to that of
the other, instead of being assembled so that the variations will neu
tralize each other, the vibration referred to becomes quite pronounced.
The vibration due to this cause is of major importance only in con
nection with extreme refinements. This is proved by the thousands of
cars which are giving satisfactory operation with only one joint.
There is no perceptible vibration under any usual conditions due to the
inertia of reciprocating parts.
As to vibration or loss of efficiency caused by friction in the joint
due to either bad workmanship or poor design, I submit that if the
friction is sufficient to produce perceptible vibration in the car, the
joint will be so injured by running the first mile at moderate speed
as not to be worth thirty cents when the mile post is reached.
Fifth. Mr. Rosenberg is quite correct in his statement that a num
ber of patents have been issued within the past few years on meri
torious improvements in universal joints. In the reference in my
paper to applications for patents on freak construction, I had no
intention of belittling any one's endeavors in that line, and should
perhaps have avoided any such reference. But the fact remains that
for every meritorious device upon which a patent is applied for, there
are perhaps ten applications for patents on devices that most experi
enced mechanical engineers would regard as impractical from a
commercial standpoint and therefore remarkable only for the ingen
uity exhibited. There are probably an equal number of applications
of which the public never hears. This seems to prove the statement
in my paper that much thought is being given the subject and for
which there must be some reason. As stated, I believe one of the
important reasons is the erroneous notion in the minds of the mechan
ical public that considerable power is lost in the universal joint, since
very many of the designs are described by applicants as "friction
less," "ball-bearing," "roller-bearing," etc. Also, it should be noted
that I did not say that all or any particular ones of the numerous new
designs were complicated or freaks.
Downloaded from SAE International by BP America Inc., Thursday, July 30, 2015

254 THE SOCIETY OF AUTOMOBILE ENGINEERS

Chairman Wall: I- i have noticed that on several cars, especially


foreign racing cars, ball bearings are used on the universal joints.
Do you think there is any material gain in doing that?
C. W. Spicer: I do not. I think that the results outlined in this
paper show that the loss is very, very slight in the universal joints
when properly applied. I might add that the popular impression that
there is so much power lost in the universal joints comes from ex
perience with designs and applications that were not suited to the
work required of them. The bearings must be designed for the work
which they have to do. They must be properly spaced, and ample
bearing area must be provided. This is true of any universal joint,
regardless of the name it carries.

Das könnte Ihnen auch gefallen