Paper No.

67-WAIDE-8

P. C. RENZO
Director of Engineering.

S. KAUFMAN
Chief Engineer,
Coupling Division.
Gear Couplings
A discussion and mathematical analysis of the operation of gear couplings at angular
D. E. DE ROCKER misalignment; transmission of uniform motion, tooth separation, both load distribu-
Chief Engineer, tion, coupling load capacity, tooth bearing, and special tooth forms.
Product Developmen!.

Sier-Bath Gear Co., Inc.

North Bergen, N. J.

THERE are dilferences of opinion as t o how flexible should also have a constant distance from each other in the direc-
gear couplings act, whether they transmit true uniform motion tion of the surface normals to match the sleeve teeth. This re-
a t substantial angles, how many teeth are in contact, and what quirement is no different on cor~plingsthan it is on gears. Teeth
kind of crown should be applied. with unequal normal distance p could not be brought to match
We shall now try to establish the facts pertinent to this and take over load smoothly from one another.
subject. As a result of this requirement, the tooth profiles of the hub, in
Gear couplings are usually arranged in pairs, each individual planes g perpendicular to the hub axis, should change increas-
coupling comprising a sleeve with straight teeth on its inside, and
ingly with increasing distance from the hub axis, a t least when the
a hub with teeth crowned to cooperate with the sleeve a t the coupling is designed for a substantial running angle. This will be
range of angles specified. further described.
A sleeve, S,and hub, H, are shown in aligned position in a frag- Fig. 2 shows a conventional uniformly crowned hub. The hub
mentary cross section in Fig. 1 and in an axial section in Fig. 2.looks like an excessively crowned gear. As on a gear, its shape is
Generally, the sleeve teeth have involute profiles, am, rising best defined by the shape of the rack teeth with which it can
from a base circle, b, as on conventional gears. Adjacent in- mesh and run so that its entire tooth sides get into contact.
volute tooth surfaces have a constant distance p from each other This rack can be considered a n extremely large, infinitely large,
anywhere, taken in the direction of the surface normal q. I n gear. A hob produces the straight profile of the involute rack in
consequence, the adjacent crowned tooth surfaces of the hub the midplane G. If now the hob is fed about an axis C as if
- turned about this axis, it will produce the same straight profile in
Contributed by the Design Engineering Division for presentation all planes containing axis C and envelop the rack tooth shape.
at the Winter Annual Meeting, Pittsburgh, Pa., November 12-17,
1967, of THE AMERICAN SOCIETYOF MECHANICAL ENQINEERB.Each tooth surface of the rack contains straight profiles that
Manuscript received at ASME Headquarters, July 26, 1967. Paper intersect axis C a t the same point and that have a constant in-
No. 67-WA/DE8. clination to axis C. I n other words, the surface that would be
produced on the rack is a conical surface with axis C.

Fig. 1 Fig. 2
 Fig. 3

I n its feed motion about axis C, the rotating hob, so to say,

represents a moving rack t h a t is the counterpart of the rack or
infinitely large gear produced by the hob. Fig. 3 shows part of
the represented rack with axis C in perspect,ive. The straight
rack profiles show up in all nxial sect,ions, as in midplane G.
They come to a point or apex on axis C. The cone apex of side
Fig. 4
SIis a t CI, the'one of side SZa t CZ. Adjacent sides S,, S2 are iden-
tical conical surfaces merely displaced along the cone axis C.
They have a constant distance from each other everywhere in
the direction of the surface normals.
The tooth sides of the hub are produced in a generating motion To describe the coupling action, we shall first look a t the tooth
whereby the rack represented by the hob rolls on the hub. This contact without appreciable load and without ally elastic deflec-
rolling motion is as if a cylindrical pitch surface k of the hub tion of the contacting tooth surfaces and teeth. The effect of
would roll on the pitch plane of the rack in contact therewith. elastic yielding will be introduced later. We shall also show t h a t
The pitch surfaces intersect the tooth sides in curves called pitch involute gear couplings can transmit true uniform motion, even
lines. As the Pitch surfaces roll on each other without sliding, a t very large coupling angles..
the pitch lines of the hub are as if printed on its cylindrical pitch Fig. 4 is an axial section showing a sleeve S and a hub H a t an
surface by the pitch lines of the rack. I n development of the angle i. Here the tooth contact has shifted toward the tooth
cylindrical pitch surface to a plane, the pitch lines of the hub are ends. Fig. 5 is an axial end view of the sleeve that has straight
identical with the pitch lines of the rack. involute teeth with base circle b.
At any intersection point Q of a rack pitch line with the contact The uniform motion of the sleeve and hub is like the motion of a
line P Q of the pitch surfaces, the contacting tooth surfaces have pair of bevel gears whose axes coincide with the sleeve and hub
the same direction. They have a common surface normal QC axes a., a, that intersect a t 0. The sleeve and hub move as if two
(Fig. 2), whose inclination t o axis C is constant and equal to the imagined conical pitch surfaces of the gear members roll on each
pressure angle in the midplane G. other without sliding. These rolling surfaces contact along the
Normal QC remains normal t o the conical i.ack tooth sides in its instant axis I t h a t bisects the angle between the coupling axes.
fixed position even as the rack moves along axis C. I n each rack The instant axis I is very useful to show up the relat,ive moliol~
positioii, it intersects t,he rack-tooth side a t a poirit of col~tactwit,h of oiie member with respect t o the ot.her. At any one instant, tllc
the hub meshing therewith. I t is a path of contact. As this is hub moves briefly as if turned about the instant axis relatively 1.0
true for all points Q and their surface riormnls, these suilace the sleeve. This defines the direction of relative motion of each
riormals determine and define the surface of progressive contact. point of the hub. The velocity of this motioli depends on the
As in gearing, this surface can be used to establish the required turning velocity w i about the instant axis. I t is known to be ob-
tooth shape of the hub. tainable from the turning velocity w of the sleeve and hub by
As the conical rack-tooth sides, such as St, St, have a constant geometric addition, as expressed in formula (2):
normal distance from each other, the so-generated hub-tooth
sides also have a constant normal distance from each other, as re-
quired for smooth power transmission.
At any point Q, the profile inclination or pressure angle is con-
ui = % sin (s)
stant in plane QC (Fig. 2 ) t h a t contains the cone axis C. Hence it 'The relative velocity of any point is the product of its distance
is bound lo be different in plarie g t h a t is parallel to midplane G from the instant axis and of wi. At a poirit of tooth contact, it
and perpendicular to the hub axis. I t can be shown t h a t the pro- defines the sliding velocity. I t increases with increasing angle i.
file inclinat.ion 4' or pressure angle in plane g is related to the As the instant,aneous relative motion is a turning motion about
pressure angle 4 in midplane G and to angle QCP = e as follows: the instant axis, the surface normal at,any point of tooth contact
is bound to intersect the instant axis I. With involute sleeve
tan 4' = tan 4 cos e (1) teeth, a surface normal fixed ill space stays a surface normal a t all
The t,erm $' is smaller than $. It drops off first very slowly turning angles of the sleeve. It becomes a path of contact.
mid then more rapidly with increasing angle e. Uniform motion is transmitted. The contact point moving along

Transactions of the ASME

 Fig. 6

Fig. 5 Fig. 7

its straight path describes a line on the tooth sides as the cor~pling Tooth Bearing
turns. This line coincides with an involute profile on the sleeve
Fig. 6 is a fragmentary axial section of a sleeve, wherein i t s
teeth.
involute tooth profiles appear as straight lines. In the aligned
A pat,h of contact n lies in a plane perpendicrllar to the sleeve
position of the sleeve and hub, the tooth contact is along profile yo
axis a, and when extended is tangent to base circle b (Fig. 5 ) .
a t zero load, all the teeth contacting simultaneously. On both
I t also intersects the instant axis I, a t N,.
sides of line yo, the contacting tooth surfaces gradually separate
The crown of the hub teeth can be determined so as tJo place
from each other a t a rate depending on the amount of crowning.
the path of contact a t a desired axial distance z from the inter-
They separate a t first very slowly. The cross-hatched area
section 0 of the axes, a t the design misalignment angle i. The
around point P has a separation within a fixed, very small amount
foregoing requirement defines the location of the contact point PC
z', such as 0.001 in. Such a separation might be overcome by
a t the pitch circle k. PCgenerally does not lie directly on the
elastic deflection under heavy load. The area then becomes a
vertical V t,hrough 0, but close enough to it that its distance from
tooth bearing area.
the instant axis I does not differ much from pitch radius R. For
At a coupling angle i, the contact hasshifted away from central
this reason, the sliding velocity v, a t mean contact point P, can
position to two mesh zones. A tooth contactsonly a t onepoint a t
be put down as
a time, a t zero load, a t a point such as Q in one turning position;
and after half a turn a t point Q'. The cross-hatched elliptical or
i
o, = - o (approx) (3) oval area has a separation within a given small amount z'. I t is
60 smaller than the cross-hatched area around point P. As the
coupling turns, the contact point moves along profile y. The
where o is the peripheral velocity R . w , and io is the coupling rectangular area around Q or Q' is within a separation z' of getting
angle i measured in degrees. into tooth contact a t zero load. Under load, it may become the
Each of the two sides of the teeth has two diametrically oppo- area swept by tooth coritact.
site paths of contact. One is along normal n that intersects the With the conventional uniform crowning the width of these
instant axis I a t N,. The other is along normal n' that inter- areas aror~ndpointas P a ~ Q, ~ Q'
d is approximately equal. When
sects instant axis I a t Nif on the opposite side from 0. The the cor~plingruns a t an angle, however, there are fewer teeth in
cont,act normals n, n' intersect the cylindrical inside surface of the contact, o111ytwo at times a t the maximum design angle, and
sleeve teeth and a spherical outer s~trfaceof t,he h11b teeth. The these fewer t,eeth have less intimate contact. Moreover, sliding
path of contact. is between the two intersection points. I t s length increases with increasing angle i. 111 consequence, the sustained
determines the duration of contact. If it were exactly equal to load capacity a t the design angle is only a small fraction of the
the normal distance p (Fig. 1 ) of adjacent tooth sides, then each capacity of the coripling in alignment or near-alignment.
tooth starts contact when the preceding tooth leaves off. As on Fig. 7 shows the kind of tooth bearing obtained a t a substantial
gears, profile overlap is desired, a length of 1.2p to 1.6p or more. angle i when the profile inclination of the hub teeth is constant in
This length and the duration of contact depend on the tooth depth planes g, Fig. 2, perpendicrllar to the hub axis rather than being
and on the pressure angle or profile i~lclination. constant in planes QC containing axis C . The profile inclination
After passing through the contact, a tooth separates from its is then too large in planes g, so that the tooth bearing is displaced
mate, to contact it again a t a different spot after about half a toward the top of the sleeve teeth when the couplhg runs a t the
turn. The maximum separation attained depends on the coupling design angle. This affects the smoothness of the transmitted
angle i. motion and causes early wear.

Journal of Engineering for Industry 3

 ih4
Fig. 8 Fig. 9

Contact Cycles and Backlash R ctn 9

tance 7 from P.
Let us look a t the relative motion of the hub with respect to a cos z
sleeve maintained stationary. Instead of turning both the sleeve The distance of curvature center from projected circle k' is
and hub on their axes, a n opposite turning motion about the found to amount to R ctn 9 ,cos j and the distance z of circle k'
sleeve axis is added to the system comprising sleeve and hub, so cos i
that the sleeve turning motion is cancelled out and the sleeve from curve p . is
stands still. The hub axis then describes a conical surface about
t,he sleeve axis. I t s apex is a t the intersection point 0 of the z = R ctn Q (*%'- 1 )
cos i
axes.
We shall first consider the case where the crowning axis C in- In the more general cases where the distance R, of mean point,
tersects the hub,&is, a t - 0 , and look a t a spherical surface S P ~ from the crowning axis C-C (Fig. 2) from R, a sphere
centered at O and thrdugh mean pointlP.of the with radius Re is considered. The foregoing formula for z applies
Fig. 8 is a radial view taken in direction PO. Fig. 9 is a side view also when R, is substituted for R.
taken in the direction of the sleeve axis a,. The spherical 'surface At small angles i, as in common use, the formrlla can be t,rans-
S p h intersects the hub-tooth surface passing through P sub- formed into
stantially in a circle ( k ' ) centered a t 0 . In projection, Fig. 8, it
z = l/,R, ctn Q tan2 i sin2 8
aDpears as a straight line t h a t coincides with the hub axis ah.
The same sphere S p h intersects the contacting tooth surface of The maximum separation 20 is attained when 8 is 90 deg where
the sleeve in a curve p, whose mean curvature radius in projection, sin 8 = 1. Hence
Fig. 8, can be shown to amount to R ctn #I on curves having only a
small distance zu from 0 . 20 = I/1Re ctn 9 tan2 i (4)
The circle and curve p, contact or nearly contact a t point P'.
In the relative motion, the hub axis describes a conical surface
about sleeve axis a#,whereby a point H o of the hub axis describes The coupling runs a t minimum backlash a t the maximum angle
a circle k'. At a turning angle 0, point H o reaches a position H. i. The backlash is increased by Ab when the coupling is set in
And a t turning angle 8 = 90 deg and 0 = 180 deg, i t reaches alignment, whereby the separation zo is added on each side:
positions H' and H', respectively. I n the view in Fig. 8, the
projected hub axis OH appears inclined a t an angle j to sleeve axis
a,. Tan j can be readily computed a s The foregoing figures apply to uniformly crowned teeth.
t a n i= tan i cos 0
f
At a turning angle of 90 deg, when Ho is a t H', the circle k' o
the hub-tooth side again appears projected a s a straight line in
Fig. 8, a line coinciding with the projected hub axis OH'. And
a t a turning angle of 0 = 180 deg, it appears projected as a
straight line OH'. I t appears in Fig. 8 as if swinging about radial
line P O between end positions O H o and OH'.
We shall now compute the distances z between circle k' and
curve p, as if circle k' would swing about radial line P O whereby
its plane always contains the hub axis. Although this assumption
is not exactly fulfilled, it provides a close enough result a t the
~ n o d e r ~angles
le i now considered.
The curv-ature center M of projected curve p, lies in the plane
passing through P a t right angles t o the sleeve axis a,, at a dis-

Transactions of the ASME

 The variation of separation z with the turning angle 8 is directly formula (4) for no appreciable load and uniformly crowned teeth.
shown in diagram, Fig. 10. I t shows a circle k, with center Oo The separation z a t any turning angle 8 from contact position is
and diameter zo = 0'0'. The term z shows up as the distance defined in formula (5).
of any point Q' of the circle from the straight-line element X-X To estimate the number of teeth in contact, we consider average
tangent to the circle a t 0'. Point Q' corresponds to a turning conditions, without such irregularities in contact pattern as may
angle 8 = QfO'O'. I t can also be obtained by plotting an angle occur when a new tooth gets into contact or a tooth gets out of
28'from center Oo. Separation z is seen to vary harmonically contact. In the case considered here, moreover, the individual
with the double turning angle 28. tooth load is in direct proportion to the elastic tooth deflection,
Example: With 4 = 20 deg, i = 1 deg and R, = 21/2 in.; zo = the added deflection of the sleeve tooth and hub tooth, both sur-
0.0010 in. from (4). face deflection and bending. This proportionality is a t least
This is a quantity small enough that it compares with the approximately fulfilled.
elastic deflection of the teeth under load. Under load, then, more Those teeth are in contact whose separation z is smaller than
teeth get into simultaneous contact, especially a t small angles i. the deflection z, of the teeth that carry the largest individual load
We shall now determine the number of teeth in contact under Pi, where
load. P; = cz,
The proportionality factor C depends on the material used and
on the tooth design. I t s computation is involved and omitted
Load Distribution here. I t can also be determined reliably by test. In thetest, all
The number of teeth that carry the load depends on the angle i but two diametrically opposite teeth of the test hub are removed.
between the coupling axes, on the tooth design, and also on the Contact with the sleeve teeth is established a t zero coupling
load. angle. Then torque is applied which results in a slight relative
._- In operation a t an angle at very small loads, each tooth gets into turning displacement. The displacement is measured close to the
contact, separates, and contacts again after half a turn of the contacting teeth. C is the proportion of the tooth load applied
coupling. We have given the maximum tooth separation zo in to the displacement s t the pitch point.

Journal of Engineering for Industry 5

 The elastic deflection z, of loaded teeth allows adjacent mating mined from the maximum safe load Pi per tooth, based on surface
teeth to move toward each other. The teeth whose,separation z stress as well as bending stress. If factor C is already known,
was equal to the elastic deflect,ion z, of the most loaded teeth t.hen 2. = Pi/C. Let it be assumed that, z, was determined as
bheri move into cont,act wit,h each ot,her b l ~ carry
t no load. The
t,eeth whose separat,ion z was less than z,areloaded i r ~proport.ion z, = 0.00026. Then = 0.250 a t i = 1 deg.
Zo
lo t.he difference (z, - z). Their load is Read from the q-graph, Fig. 12, q = 0.218.
If the maximum safe tooth load at a given tooth pitch and
crowning is, for instance, 1OOO lb, and the hub has 50 teeth while
the accuracy factor K is 0.80, then the coupling could cariy a total
The term z is plotted in Fig. 11 in ternis of the a~igle0, the load of 40,000 Ib when in alignment, and q(40,OOO lb) = 8720 lb
angular distance from the mean contacting tooth where z = 0. when running a t an angle i = 1 deg.
It is the vertical distance from X-X of any point (Q) whose Example 2: How much load call the Yame coupling carry a t
horizontal distance from 0 is proportional to angle 0. The half the angle i of Example 1, a t i = I/, deg?
curve s so obtained is a sine-curve. I t repeats with every half Pi is the same as before, and 2, remains 0.00025. 20,however, is
turn. smaller, as tan i in formula (4) is approximately one half of the
After determining z, for a load Pi that, can be carried by a single former amount, and tan2 i is one quarter thereof. Hence 20 =
tooth with a margin of safety, z, is plotted in Fig. 11 from X-X
0.00025 in close approximation, and Z" = 1.
up, and a line R-B is drawn parallel to X-X a t a distance z, Zo
therefrom. z, should be plotted a t the scale used for 20, the maxi- Read from the q-graph, q = 0.500. This results in a total
mum separat.ion a t the now considered coupling angle i. toot,h load of q(40,000 lb) = 20,000 lb that the coupling can
Line B-B cuts o f the bottom of curve s, between end points safely carry a t a coupling angle i = deg.
Q. 9,'. All the teeth within the spread QrQel carry some load, Example 3: How m~lchload can the same conpliiig carry a t
that fades out and beromes zero a t these end points. The load an angle i double that of the first example, at i = 2 deg?
a t any point is proportional to the vertical distance within the
Here zo is approximately = 0.004 and = 0.0625. The graph
cross-hatched area a t that point. The total load carried in one Zo
engagement zone is proportional to the cross-hatched area,f below shows q = 0.107, so that the total toothload figuresq(40,000lb) =
the spread QrQ,'. If all the teeth within 90 deg to both 4280 1b.
sides of 0 would carry the maxim\\m load P,, then the total load These figures are based on contact a t a mean tooth depth.
within that range would be proportional to the area of the At the larger augles i, profile action should be considered as well.
rectangle 1-2-3-4. The said total load amounts to Elastic tooth deflection then decreases in importance. The
couplings act more and more like gears, with increasing angles i.
We are aware of but have not directly introduced tooth sliding
or double that amount all around the periphery, N = tooth num- in the foregoing computations, which increases with increasing
ber. The safe load that call be actually carried on perfectly angles i; nor intimacy of tooth contact, which decreases with in-
accurate couplings is a fraction q of the load NP,, where q is the creasing angles i and thereby increases the surface stresses.
rat,io of the cross-hatched area f to the area of rectangle 1-2-3-4. Thus we have treated a somewhat simplified concept based on
These areas can be readily computed. Using radian or arc elastic tooth deflection.
measure for the angles, area f can be shown to amount t,o There will hardly be any disagreement with ishe general oon-
clusion that the usual couplings can carry much more load at,
f = ' / ~ z ocos 20. (tan 20, - arc 20,) small running angles than a t large ones.
= 1/2z0 cos 20, inv 20, (7)
while the area of redangle 1-2-3-4 is in arc measltre
The Vari-Crown1
To make the drop in load capacity less drastic, Sier-Bath has
nz, = nzo sinz 0, = 1/2nzo(l- cos 20,) developed the Vari-Crown. Fig. 13 is an axial section taken
through a hub. I n conventional crowning, as produced by
Hence proportion q amounts t o
hobbing, the feed path of the hob and the tooth bottom B pro-
f cos 20, (tan 20, - arc 20,) duced thereby is ordinarily a circular arc. I t may be centered a t
q = = for z , S z o (8)
Tz, n(1 -cos 20.) 0 on the hub axis, or a t C or C', depending on how much of a
crown is required. Other methods also aim to produce the same
The load capacity of an actual coupling comes t.he closer to the kind of crowning.
figure qNPi at the co~~pling angle i; the more accurate it is, the With the Vari-Crown (Trade Mark), the feed path of the hob
closer its tolerances. We may introduce a coefficient K to express and the tooth bottom B' are more curved in the midplane G and
the tolerances; K = 1 for absolutely accurate couplings. I t is less curved further toward the tooth ends. An axial section of the
somewhat smaller than 1 on commercially accurate couplings, the tooth surface through pitch point P shows a curve m almost
difference from 1 increasing with increasing tolerances. Thus identical with the feed path there. The varying curvature of the
the load capacity P, a t the coupling angle i is t,ooth surfaces widens the tooth contacts with increasing distance
from m i d p l a ~ ~G,e while a t t,he midplane it has the smallest width.
I t tends t,o more equalize t.he load capacity a t different angulari-
ze ties, taking away from the excessive load capacity at zero allgu-
Proportion q is'plotted in Fig. 1 2 in terms of the proportion - . larity and distributing this excess to larger angularities. Thus
Zo
the capacity a t the largest angularity is more than twice as large
z. as with the uniform crown.
Line B-B in Fig. 11 corresponds to = 0.250. If z, = 20, then
zo
q = 0.500. If z. > zo, then q = ['/ZZO + (z. - zo)] + z. = 1
End Round
I n some cases, a range of angularities is specified for action
.under load, and an additional range for action without ap-
A feikxamples will illustrate the use of the q-graph in Fig. 12.
preciable load. There are two common ways of meeting this
Example 1: In the example previously given for a coupling
angle i = 1 deg, 20 was computed a t 20 = 0.001 (in.). z, is deter- 1 Made under patent No. 2,922,294;other patents pending.

Transactions of the ASME

specification. One is designing the coupling for the whole range without binding. Neither of these ways is quite satisfactory.
of angularities. This means increased crowning and increased For such cases, Sier-Bath introduces the End Round. I t is
stress. The other is to design the coupling for the angularities illustrated in Fig. 14. Crowning is sharply increased a t the tooth
under load, and letting the tooth contact shift to the ends and ends, a t E.
edges of the teeth for the no-load angularities, allowing sufficient The main crowning is figured for the angularities under load.
backlash so that the coupling can take the largest angularity At the maximum load-angularity, the center of the tooth contact is
placed a t Q, where i t would be normally placed. The End
Round starts further out, a t a distance from Q about half the
width of the contact area. I t is sufficiently curved to prevent
contact to shift t o the end edges, keeping it on the tooth-side
surfaces. This is tangent contact capable of carrying moderate
load, without any tendency to chew up the sleeve tooth sides.

Extended Contact Design

With spindle-type couplings, sometimes large angularities are
specified, b u t each coupling of a group is run only within a re-
stricted angular range which is different for different couplings.
I n general, all these couplings are made alike and are designed to
run up t o the largest angularity. For simplicity of replacement,
this arrangement is ideal. B u t we can make much stronger
couplings and multiply the coupling life perhaps 10 times if the
couplings do not all have t o interchange with each other.
For instance, if the range of adjustment of a coupling with a
maximum angularity of 4 deg is cut down to 1 deg, so that
the coupling may be designed to run only between 3 and 4 deg
angularity, the load capacity based on surface stress can be in-
creased three to four times, and the coupling life a t a given load
increased many more times. The tooth contact can be much
widened. I s the extra simplicity of the present setup worth
wasting coupling life t h a t much? I t is not difficult to have two
or three different coupling designs t o take the place of a single
design for the whole range. They could be marked, for in-
stance, in different colors.

Wearing I n and Wearing Out

I t might be thought t h a t the couplings wear into the right shape
anyhow, and t h a t there is no need for refinements or even for
Fig. 13 accuracy. B u t this we consider no more valid for couplings than
i t is for gears. Improper shape causes excessive surface stresses
and tends to damage the tooth surfaces. I t is the first stage
Q\ of wearing out. We suggest t h a t the tooth shape should be as
nearly correct and adequate as possible.

History and Conclusion

Gear c o ~ ~ p l i n gstarted
s o u t with straight teeth on the hub.
And they worked a t the small angularit.ies where they were used.
As the need for large angularities arose and grew, i t was recog-
11ized that the tooth ends of the hub had to beeasedoff. Crown-
ing was i n v e ~ ~ t e d .I t was also recogt~ized that a t the larger
i~ngularitiesthe tooth contact is confilled to two diametrically
opposite zones and t h a t then centering was desirable or required.
A spherical outside surface was introduced on the hub, centered
on the hub axis, to let the hub teeth bear against the tooth
bottoms of the sleeve for centering. The contact between the
spherical outside surface and the cylindrical inside surface of the
tooth bottoms provides accurate centering a t all angularities.
This may have been suggestive of the thought t h a t the sides
of the hub teeth should also be crowned about center 0 (Fig. 2).
A while later came the realization t h a t there was no compulsion
or natural law for crowning about center 0. Other crowning
centers were used as well to achieve different amounts of crown-
ing. And now we know t h a t we do not even need a center for
crowning, t h a t crowning may be varied along the tooth.
The stepwise progress was accompanied by a n increased
~~nderstanding of how the teeth act. At large angularities, they
are like g e a n with internal mesh. Like gears, they transmit true
uniform motion. They have, however, the peculiarity that a
Fig. 14 tooth gets into contact in two zones of mesh. Between these

Journal of Engineering for Industry

zones, the teeth separate. Their separation is much smaller than therefor. We have confirmed mathematically that the number
on gears even with internal mesh. And a t small angularities, it of teeth in contact increases with decreasing angularity and have
is very small, so small t h a t the elastic deflection of the teeth described a simplified and approximate computation of the num-
plays a n important part. ber of teeth in contact a t different angularities.
I n this article, we have treated couplings like gears, starting out
by assuming rigid bodies, and formulated the separation of the Acknowledgment
teeth for rigid bodies. And then we have considered the elastic We express our thanks to Ernest Wildhaber, our consultant,
deflections of the contacting teeth, obtaining modified results for his valuable advice and help in preparing this paper.

Printed in U. 9. A.

0 Transactions of the ASME