Simultaneous Inspection and Correction of Gears in Production
Simultaneous Inspection and Correction of Gears in Production
Simultaneous Inspection and Correction of Gears in Production
of Gears in Production
By CHARLES H. LOGUE! ,.\:--;D R. B. FEHR2
MEETI:>:G PAPEn
B
y means of the gear-correcting process described.
spur and helical gears are corrected to give a
high degree of uniformity in spacing and profile so
that the gears become practically interchangeable.
They acquire a "crown face" which enables them to
run with unusual quietness under practical conditions.
This is essentially an inspection-correction process. as
it automatically finds and eliminates the errors_
The lap is the important item in the proces!'l_ It is
of chilled cast-iron, gray cast-iron, or type metal.
and is made by casting in a mold around a !!teel chill
cut to approximate the gear to be corrected but has
a face-width several times that of the gear. The lap.
when completed. looks like a wide-faced internal gear.
The gear-correcting machine is of relatively simple
construction and performs three main functions:
(a) it provides reciprocating motion for the gear;
(b) it causes mutual pressure between the faces of
the gear teeth and thl' lap; (c) it advances the gear
G
RADUAL improvement an,j refinement of a basic-
ally sound idea in transportation has been re-
sponsible for the evolution of the automobile.
Quiet gears constitute one refinement that has been
given close attention lately. because of increasing de-
mand for noise elimination. The automobile user has
for this reason been educated to demand a car with a
quiet rear axle; but this demand was not made until it
had been demonstrated that rear axles could be made
quiet. He also expects a relatively quiet engine having
noiseless or nearly noiseless timing gears. and. because
of the difficulties of making timing gears quiet. most
automobile builders have adapted chain drive to the
timing shaft.
Quietness of the transmission is not expected. prob-
ably because everyone has been educated to believe that
the production of a relatively quiet transmission is
more or less an accidental occurrence. Almost the only
requirement insisted upon is that the gears shall not
rattle when the car is running in direct or third speed.
or when the car is standing at the curb and the engine
idling. In both the same four transmi'lsion-
gears are engaged: the driveshaft pinion with the con-
stant-mesh countershaft-gear. and the countershaft
reverse-gear with the reverse idler_ We have all been
led to expect more or less noise when the transmission
gears are transmitting power. All that is expected is
reasonable quietness, for the car user always has asso-
ciated noise with transmission gears. But when it is
discovered that relatively quiet transmissions can be
produced regularly and not accidentally. as insistent a
I M_S_A.E_-CoMultlng O!nginO!O!r_ Syracuse_ N_ Y.
DirO!ctor of thO! !abt)ratory_ Copland GO!ar Lapping SyndicatO!.
DO!tr'llt.
/llu8trated with PHOTOGRAPHS, DRAWINGS A:-<D CHAitT
one tooth with respect to the lap at the completion of
every cycle so as to assure uniformity of spacing and
contour.
Unlike the many instruments and methods that
have been devised for merely detecting errors in
gears, the authors state that the so-called inspector-
corrector device not only inspects the gears but also
corrects them after all other operations. including
heat-treating. Further. they assert that the device
will do this work without adding to the total cost of
gears, for the feeds of gear-cutting machines can be
increased to compensate for the time requinc! by the
gear-correcting operation_ The resultant gears will
be interchangeable. and will eliminate virtually all
matching-up. as well as rejections in the final tests_
In the discussion. a further description is given
of how the conjugate tooth-form of gear and lap is
obtained. and additional details of so-called "multi-
ple-series" lapping are presented.
demand will be made for quieter transmissions as there
is at present for quiet rE'ar-axles and engines.
Because of this growing demand for noiseless gears.
gear departments have had problems and troubles re-
lating to heat-treating. cutting tools, gear-cutting ma-
("hines. inspection. the matching-lip of sets for assembly
and dismantling them after unsatisfactory final tests.
and salvaging. all combined with difficulties due to the
personal human element. In addition. gear makers dis-
cover faults in transmission-gear and gearcase design
of such character that a satisfactory job cannot be pro-
cuced by the best t:urrent shop methods. The design
and the method of manufacture must both be right.
But this is not always the case. and the cost of timing-
gear and transmission-gear manufacture has expanded
out of all proportion to the results. Additional opera-
tions and all kinds of refinement have been adopted_
These improvements. accompanied by constant vigilance.
have produced better gears; but exceptionally good sets
of transmission and timing gears are not yet produced
with reasonable consistency.
PRIMARY FACTORS OF QUIET-GEAR DESIGN
Primary considerations for the production of quiet
are: design. accuracy of spacing. uniformity and
accuracy of profile. finish. concentricity. and operating
clearance or backlash. The self -alignment of gears is
also of fundamental importance.
Design enters into the problem to an extent not ordi-
narily realized. Proper design increases the allowable
errors. automatically reduces thE' cost of manufacture
and tends toward interchangeability. For example.
mating gears with insufficient overlap must be held
within extremely close limits to assure continuity of
478
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
VOL. XXIII
SIMULTANEOUS INSPECTION AND CORRECTION OF GEARS 479
action and consequently quiet operation; but, with a
sufficient degree of overlap, these limits can be ex
tended somewhat. Conventional design, which too often
means improper design, is the source of many of the
present difficulties, and is really suitable only for ordi
nary gear-applications as, for example, lathe or milling
machine change-gears.
No other mechanical feature is so important as uni
form tooth-spacing. Unequal spacing means a poor and
therefore a noisy gear, no matter how accurate the gear
may be in all other respects.
Uniform profiles, accurately spaced, result in the uni
formity of engagement necessary for quiet operation,
because the teeth of a gear are a series of cams for
producing constant angular-velocity. It should be re
membered that tooth deflection under load may equal
or exceed the error in tooth profile. Without this slight
flexibility of gear teeth, the production of quiet gears
would in some cases be an impossibility.
Accuracy of profile is of vital importance; but, with
correct design, accurate spacing, and uniformity of
profile, the working limits as to profile accuracy are
somewhat widened and this becomes, by comparison, of
secondary importance. Experience has demonstrated
the impossibility of securing quiet gear-operation by
close attention to accuracy of profile alone.
The finish of gear-tooth surfaces becomes increas
ingly important as the base circle is approached. The
smoothness of the profile should increase inversely as
the radii of tooth curvature, which become zero at the
base circle. For example, in the case of a pinion having
a small number of teeth, an almost infinitesimal amount
of misplaced material near the base circle will render
the gear exceptionally noisy. Such small inaccuracies
are difficult to detect and are impossible to remove by
current methods. They have an effect on gear opera
tion that is evidently out of all proportion to their
magnitude.
Eccentricity in the driving or the driven gear will
cause the velocity of the latter to vary according to a
sine curve and, as the total variation is distributed over
one-half the number of teeth, the tooth-to-tooth error
ordinarily is slight, and its effect is not to be compared
with errors in spacing and profile of the same magni
tude. Excessive eccentricity causes a sound which,
gradually rising and falling in intensity, may be ob
jectionable; but the principal objection to eccentricity
is the extra backlash that must be allowed to prevent
the mating teeth from running tight at one point.
Quiet operation requires, in many cases, minimum
operating clearance or backlash, which necessitates ac
curate spacing, minimum eccentricity, uniformity, and
reasonable accuracy of profile.
CROWNED FACES DESIRABLE
Suppose that all the conditions outlined as essential
to quiet operation have been fulfilled and that the
mating gears will revolve with a backlash of say 0.0005
in. Is there any assurance that they will operate
quietly? If such a pair of gears is correctly aligned,
and the housing or casing as well as the shafts are
sufficiently rigid to maintain this correct alignment, it
is reasonable to assume that the pair will operate satis
factorily. Actually, however, the alignment cannot be
perfect, nor can any alignment, perfect or imperfect,
be maintained under all conditions of load, because of
shaft deflections, variable conditions of sliding gears on
splined shafts, and the like. Either the shafts must
hold the gears in proper alignment or the gears will
attempt to align the shafts and be noisy.
It is doubtful if the spiral-bevel gears in the rear axle
could have attained their present quietness with a full
face contact such as has been attempted for transmis
sion and timing gears. In spiral-bevel gears the driving
and the driven faces have slightly different curvatures,
which result in a full-profile bearing only in the central
portion of the tooth face and taper off toward each end
of the face, the effect being known as a "crowned
face." In the spiral-bevel gear the position of the full
profile-bearing contact will vary with each change in
the alignment of the shafts. No cramping can occur
unless the change in alignment is greater than that pro
vided for by the difference in the face curvatures. It
is thus evident that the crowned face of spiral-bevel
gears has been an exceedingly important factor in their
quiet operation. Since perfection of gear alignment is
beyond the range of practical possibility, it seems obvi
ous that a slightly crowned face is essential to the quiet
operation of all gears and is, therefore, just as impor
tant in timing gears and transmission gears as in rear
axle gears.
DESCRIPTION OF THE PROCESS
The essentials for carrying out the Copland gear
correcting process consist of a lap, a gear-correcting
machine, and a continuous flow of abrasive through the
lap, which is of chilled cast-iron, gray cast-iron, or type
metal.
The lap, shown at the right in Fig. 1, is made by
casting in a mold around the steel chill shown in the
central view, which is cut approximately to the dimen
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
480 S. A. E. TRANSACTIONS
sions of the gear to be corrected except that it has a 6-in .. face. The
chill is removed from the casting by hydraulic or mechamcal
and it can be used repeatedly. No special effort is made t? obtaIn
great accuracy in the making of the lap, as it will be
trued-up in from 10 to 30 min. during the normal operatIon of the
machine, and will assume the average form of the gears that are put
through it.
The gear-correcting machine is relatively simple In
performs three main functions. First, it provides a recIprocatl.ng
motion for either spur or helical gears with respect to the .lap,
is suitably fixed on the machine; second, It prOVIdes
for mutual pressure between the faces of the gear
teeth and the lap by the application of torque as
soon as the gear has entered the lap. The torque
provides not only pressure for rapid .but
a positive means for correcting errors In
for as SOON as a new gear enters the lap, the hIgh-
est' three spots on the tooth faces will come into
contact with the teeth of
the lap, since the gear on
its arbor is free to find its
own center. Not only
will the errors be cor-
rected in a positive way,
but the wear of the lap
will be automatically com-
pensated.
Fig. 2 shows the gear
as it enters the lap. All
teeth of the gear clear all
teeth of the lap. In Fig.
3, .after the application of
torque, the highest three
spots on the gear are in
contact with the lap. In
Fig. 4 the correcting
operation is complete and all teeth of the gear
make full contact with all teeth of the lap.
Fig. '6 is a diagram of the cycle of operation.
The third function is that the machine ad-
vances the gear one tooth in the lap at the
completion of every cycle, that is, at every
top position. This feature automatically as-
sures uniformity of contour and spacing of
the teeth, which will be well within 0.0002 in.,
or less than can be measured accurately with
the instruments commonly employed for the
purpose.
The success of the process does not depend
upon a "charged" lap, but rather upon the
continuous application of a suitable abrasive-
mixture which is circulated through the lap
by a small vertical centrifugal pump located
in the base of the machine. Numerous abra-
sive-mixtures have been found effective, the
one most commonly used and which produces a
satisfactory finish being 10 cc. of FFF car-
borundum to 15 cc. of kerosene.
Fig. 5 shows the essential mechanisms of
the machine for spur and helical gears. The
crankshaft a and connecting-rod b give re-
ciprocation to the rack segment c meshing
with rack d, which carries the spindle e up
VOL. XXIII
and down. The helical guide f has the
same lead as the gear g to be corrected
and the lap h. The helical guide f moves
through split nut i, which is mounted on
worm-wheel hub j.
Bell-crank lever k is connected to
worm shaft I. Spring scale m pulls
downward on the arm of lever k.
Pressure-control cam n prevents the
arm of lever k from falling when
gear g is leaving the lap to be ad-
vanced or "indexed" one tooth and re-
turned to the lap h. But as soon as gear
g has safely entered lap h, pressure-con-
trol cam n moves past the roller at the
end of lever k, which is then free to
move under tension of the spring scale
m This moves in an axial direction
0 which, acting as a rack, gives
a rotation to worm wheel j and
therefore to gear g, until the teeth of
gear g press against the teeth of lap h.
The gear g goes to the bottom of lap h
and returns to the top, when pressure-
. control cam n lifts lever k and, moves
the teeth of gear g away from the top
teeth of lap h.
When gear g is clear of lap h, index
cam p gives partial rotation to pin-wheel
q, which gives rotation to index gear-
train r and indexes gear g one tooth.
After indexing, gear g moves toward
and enters lap h, and the cycle is re-
peated until an automatic mechanism
stops the machine with the gear at the
top of the stroke ready for the operator
to take out the ar1:)or holding gear g and
put in another arbor. The only work
the operator does is that of loading,
changing, and unloading arbors.
The accuracy of the results, as re-
gards uniformity of spacing and contour
of the gear teeth, does not depend upon
the accuracy of the so-called "indexing"
mechanism, which serves merely to ad-
vance the gear one tooth at a time with
respect to the lap. This mechanism
New Text
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
VOL. XXIII SIMULTANEOUS INSPECTION AND CORRECTION OF GEAR 481
needs only to be accurate enough to cause the gear to
enter the lap, there being an initial clearance of several
thousandths of an inch between the gear and the lap.
The accuracy of the process
depends upon the "averag
ing" effect produced by each
new application of torque,
which occurs as soon as the
gear teeth register with a
new set of lap teeth. The
more times this cycle is re
peated, the more accurate
will be the results up to the
point when all the errors
are eliminated; that is, when
all the teeth have been
"cleaned up" as shown by
visual inspection.
Fig. 7 shows a battery of
four gear-correcting ma
chines that can be run by
one or two operators, de
pending upon the footage re
quired to correct the gears.
Each machine is driven by a
11/2-hp. motor mounted on
the rear of the frame, the
maximum load being in gen
eral less than 1 hp. Each
machine, or one machine in a
battery of machines, is pro
vided with a locating fixture
such as is shown in Fig. 8.
This enables the operator
to place each gear on the
arbor in the same angular
relation with the drive pin of the arbor, so that there
will be no difficulty in having the gear enter the lap
when starting the machine. The lap is mounted in its
holder on the machine in such way that it can be ad
justed or removed easily.
A gear is set on its arbor by means of the locating fix
ture shown in Fig. 8, the arbor being then placed and
held in position on the vertical spindle of the machine
by the chuck, which is aiso shown. The starting lever
is then lifted and a friction clutch puts the spindle into
operation at 120 double strokes per min., each giving a
total of 1 ft. of contact between the gear and the lap.
HOW "CROWNED FACE" IS OBTAINED
Fig. 9 shows a helical gear about to enter the lap.
As the gear enters, the lap teeth and the gear teeth
clear each other as indicated in Fig. 2, but torque is
immediately applied by a suitable spring and lever
mechanism controlled by a cam. A spring scale enables
a predetermined pressure to be exerted between the
gear teeth and lap. This pressure usually is kept on
until the gear completes a down and an up stroke and
returns to the top of the lap, as shown in Fig. 6, when
the pressure is removed as the gear emerges from the
lap to be advanced one tooth. The arrows in Fig. 3 indi
cate how the highest three spots on the gear make initial
contact with the lap. These high spots are found auto
matically and are eliminated simultaneously by the re
ciprocation of the gear through the lap, which action, in
conjunction with, the advancing of the gear one tooth at
each top position, reduces these inaccuracies until all
faces of all the gear teeth are in contact with the con
jugate faces of the teeth in the lap. This two-stroke
cycle is repeated until the required footage of lapping
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
482 S. A. E. TRANSACTI ONS VOL. XXI I I
has been r un, t he machi ne bei ng t hen st opped aut omat i c-
al l y wi t h t he gear i n t he t op posi t i on f or r emoval of t he
ar bor f r om t he chuck.
I t i s evi dent t hat a " cr owned f ace" i s obt ai ned by
keepi ng t he f ul l pr essur e on whi l e t he gear i s par t l y out
of t he l ap at t he t op and t he bot t om, t hus gi vi ng a hi gher
uni t - pr essur e at each end of t he t eet h. The amount of
" cr owni ng" obt ai ned af t er 200 f t. of gear cor r ect i ng,
whi ch i s t he maxi mum amount usual l y r equi r ed per
si de, i s 0. 0010 t o 0. 0015 i n. , as det er mi ned by hal f t he
di f f er ence i n t hi ckness of t he t oot h bet ween t he mi ddl e
and edges.
PRI NCI PAL FEATURES SUMMARI ZED
Fr omt he f oregoi ng descri pti on i t i s evi dent that the
pr ocess i s essent i al l y an i nspect i on- cor r ect i on oper at i on,
f or i t aut omat i cal l y l ocat es er r or s and i rregul ari t i es
and el i mi nat es t hem. Er r or s i n spaci ng can be r educed
wi t hi n t he er r or of avai l abl e measur i ng i nst r ument s,
or t o pr act i cal per f ect i on, whi ch i s an i mpor t ant f eat ur e.
Uni f or mi t y of prof i l e f r om t oot h t o t oot h and f r om
gear t o gear i s assur ed, and t he act ual prof i l e obt ai ned
wi l l be t he aver age of t he prof i l es whi ch t he gear s had
bef or e bei ng put t hr ough t he pr ocess. I f t hi s aver age
prof i l e shoul d not be wi t hi n t he r equi r ed l i mi t s, i t i s
possi bl e t o al ter i t by a rel at i vel y si mpl e means, but
t hi s pr ocedur e has been f ound necessar y i n onl y a f ew
i nst ances.
A hi gh degr ee of concent r i ci t y can be obt ai ned by
cor r ect i ng t he gear s before f i ni sh- gr i ndi ng t he hol e. A
speci al chuck l ocat i ng f r om t he accur at el y spaced and
uni f or m f aces of t he t eet h wi l l t hen enabl e t he hol e t o
be gr ound wi t h mi ni mum eccent ri ci t y. I n gener al , how-
ever , when t he desi r ed degr ee of concent r i ci t y can be
obt ai ned by t he usual met hods, t he gear - cor r ect i ng
oper at i on i s per f or med after t he f i ni sh- gr i ndi ng of t he
hol e. A uni f or ml y smoot h sur f ace i s secur ed on al l
t eet h, i ni ti al i rregul ari t i es t oo sl i ght t o be measur ed
ar e aut omat i cal l y el i mi nat ed, and t he pr ocess i s as easi l y
appl i ed t o har dened gear s as t o sof t- steel or cast - i ron
gear s of ei t her spur or hel i cal t ype.
Not onl y ar e uni f or mi t y and i nt er changeabi l i t y as-
sur ed, wi t h consequent el i mi nat i on of t he sel ect i ve as-
sembl y of gear set s, but t he i mpor t ant " cr owned f ace"
t hat wi l l enabl e t he gear s t o be sel f - al i gni ng i s obt ai ned
r eadi l y, sel f - al i gnment bei ng a pr oper t y t hat gi ves good
gear s a chance t o r un qui et l y. These resul t s ar e ob-
t ai ned wi t h smal l ef f ort and expense. For exampl e, t he
oper at or ' s t i me on each gear i s onl y 1 t o 2 mi n. , and
t hi s t i me can be compensat ed by i ncr easi ng t he f eed
of t he gear - cut t i ng machi nes. I t i s expect ed t hat t he
i mmedi at e r esul t s of empl oyi ng t hi s pr ocess i n a gear -
maki ng depar t ment wi l l be t o pr oduce uni f or ml y bet t er
and qui et er gear s, t o r educe t he total of al l cost s pr op-
erl y char ged t o gear s, and t o el i mi nat e uncer t ai nt y i n
t he maki ng of good gear s.
PROCESS AS APPLI ED TO QUANTI TY PRODUCTI ON
Regar di ng t he pract i cabi l i t y of t he pr ocess i n quan-
ti ty pr oduct i on, one of t he quest i ons r ai sed by t hose
who l ear n of i t f or t he fi rst t i me r ef er s t o t he l ap, or
" mast er t ool " as i t has been cal l ed. I t i s not al ways
under st ood at f i r st t hat t hi s mast er tool i s a r esul t of
t he i nher ent accur acy of t he pr ocess and does not i n-
vol ve any gr eat di f f i cul ty i n i ts maki ng. When t he l ap
i s f i r st pl aced i n posi t i on on t he gear - cor r ect i ng ma-
chi ne, i t i s a rel at i vel y cr ude tool . Gear s f r om t he
r egul ar r un, t o t he number of 20 or 30, ar e t hen put
t hr ough t he pr ocess, but wi t h onl y about 25 f t. of l ap-
pi ng f or each gear i nst ead of t he usual 100 t o 200 f t.
per si de. The r esul t of t hi s oper at i on, whi ch r equi r es
f r om 10 t o 30 mi n. , i s t hat t he l ap gr adual l y assumes
t he aver age cont our of t he gear s and each i ndi vi dual
gear i s but sl i ght l y af f ect ed. As soon as t he hi gh spot s
of t he l ap have been r emoved and t he gear t eet h show
evi dences of bei ng t ouched al l over , t he l ap can be con-
si der ed r eady f or r egul ar pr oduct i on. The s ame gear s
t hat wer e used f or put t i ng t he l ap i nt o condi t i on can
now be put t hr ough i t agai n, gi vi ng t o each t he f oot age
of l appi ng r equi r ed t o cl ean up al l t he t eet h.
The i mpor t ant el ement , t he l ap, i s i mpr oved wi t h use;
t he mor e gear s t hat ar e put t hr ough i t, t he mor e accu-
r at e t he l ap becomes, unt i l i t r eaches t he pract i cal l i mi t
of accur acy, whi ch i s t hen mai nt ai ned t hr oughout i ts
l i f e. The i nher ent nat ur e of t he pr ocess makes i t un-
necessar y t o shar pen or dr ess t he l ap or ot her wi se t o
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
VOL. XXI I I SI MULTANEOUS I NSPECTI ON AND CORRECTI ON OF GEARS 483
keep i t i n a t r ued- up condi t i on. Al l t hat i s r equi r ed
i s t hat t he gear s be cut wi t hi n t he nor mal l i mi t s as t o
si ze and accur acy of prof i l e. Uni f or mi t y of spaci ng
and prof i l e ar e t hen car ed f or aut omat i cal l y.
MODI F I CATI ON OF T OOT H F OR M
Our exper i ence i n t he cor r ect i ng of t housands of
hel i cal t i mi ng- gear s and t r ansmi ssi on gear s i s t hat t he
aver age prof i l e i s gener al l y wi t hi n t he l i mi t s t hat wi l l
gi ve rel at i vel y qui et gear s. However , i nst ances exi st i n
whi ch i t may be desi r ed to modi f y t he r esul t ant t oot h-
f or m. Suppose a gr oup of gear s t o have a posi t i ve
t oot h- f or msuch t hat t he prof i l e near t he t i p of t he t oot h
i s one or mor e t housandt hs of an i nch t hi cker t han t he
i nvol ut e and t hat i t i s desi r ed t o br i ng t hi s prof i l e down
to a f or m mor e near l y appr oachi ng t he i nvol ut e. The
gr oup of gear s wi l l be put t hr ough what we t er mt he
"mul t i pl e- ser i es" pr ocess. Al l t he gear s wi l l be put
t hr ough, say f or 50 f t. of l appi ng on each si de. They
wi l l t hen be put t hr ough agai n f or 50 f t. of l appi ng on
each si de, and yet agai n, i f mor e f oot age i s r equi r ed t o
cl ean up al l t he t eet h.
We wi l l t hen have a gr oup of gear s i n whi ch t he
prof i l es ar e so near l y al i ke t hat t hei r di f f erences cannot
be measur ed accur at el y by any means of whi ch we
know, but t hese prof i l es ar e al l i n er r or to t he same
degr ee, say 0. 002 i n. They ar e t hen put t hr ough a
modi f i ed l ap such as i s shown i n Fi g. 10, i n whi ch t he
ti ps of t he t eet h wi l l under go mor e abr asi on t han t he
root s of t he t eet h, and t he t oot h prof i l es wi l l be br ought
down wi t hi n any desi r ed l i mi t s. I f t he aver age prof i l e
of t he r awgear s i s det er mi ned by car ef ul measur e-
ment s, i t i s possi bl e to cor r ect i t by put t i ng t he gear s
t hr ough t he modi f i ed l ap i n t he f i r st pl ace and t hus t o
mi ni mi ze t he amount of st ock t hat i s r emoved. Thi s
resort t o t he modi f i ed l ap i s not pr act i sed except i n
speci al cases, and i t has been our exper i ence t hat t he
aver age prof i l e of l ar ge l ots of gear s can, by pr oper
at t ent i on, be mai nt ai ned wi t hi n r equi r ed l i mi t s on t he
gear - cut t i ng machi nes.
L I F E OF L AP S
Based on act ual dat a obt ai ned, t he l i fe of l aps f or
gear s of 7 t o 10 di amet r al pi t ch can be expr essed i n
t er ms of t he total f oot age. For har d and semi - har d
cast - i ron l aps t he l i f e i s 300, 000 t o 600, 000 f t . ; f or
sof t cast - i ron l aps, 150, 000 t o 200, 000 f t . ; and f or t ype-
met al l aps, 25, 000 t o 35, 000 f t.
. Some gear s r equi r e 200 f t. of l appi ng on each si de of
t he t eet h, or a t ot al of 400 f t. Ot her gear s, such as
cr ankshaf t t i mi ng- gear s, may r equi r e 100 f t. or l ess on
one si de onl y. I t i s t hus seen t hat t he l i fe of a l ap
depends on t he f oot age r equi r ed to f i ni sh a gear . I n
t he case of t he cr ankshaf t gear s r ef er r ed t o, a l ap may
do 250 t o 6000 gear s accor di ng t o t he mat er i al . I n
gener al , chi l l ed cast - i ron l aps ar e pr ef er r ed, as t hey
have a much l onger l i f e; but , when onl y a f ewgear s ar e
to be cor r ect ed, or when an except i onal l y f i ne f i ni sh i s
r equi r ed, t he t ype- met al l ap i s t he best choi ce.
DECREASE I N TOOTH THI CKNESS
Fi g. 11 shows a charact eri st i c cur ve gi vi ng t he rel a-
ti on bet ween t he st ock r emoved f r omt he gear and t he
amount i n f eet of gear cor r ect i ng per f or med by t he
machi ne. I t wi l l be not ed t hat t he gr eat est appar ent
r emoval of st ock t akes pl ace wi t hi n t he f i r st 40 f t.
of gear cor r ect i ng, t he decr ease i n t oot h t hi ckness be-
i ng about 67 per cent of t hat at 100 f t . ; and t hat , be-
yond 100 f t. , t he cur ve becomes pract i cal l y a st r ai ght
l i ne wi t h onl y a gr adual sl ope. The r eason f or t hi s
i s t hat onl y t he i rregul ari t i es and hi gh spot s of t he
gear ar e t ouched at f i r st , and t her ef or e t he uni t pr es-
sur e i s much hi gher t han t hat l ater on. Thi s measur e
of t he decr ease i n t oot h t hi ckness was det er mi ned by
pl aci ng t wo gear s i n met al - t o- met al cont act on a gear -
t est i ng st and and not i ng t he r eadi ngs of t he di al i ndi -
cat or . I n gener al , i t can be st at ed t hat t he decr ease i n
t oot h t hi ckness wi l l be 0. 001 t o 0. 003 i n. , dependi ng
upon t he amount of cor r ect i on necessar y t o br i ng t he
gear wi t hi n t he r equi r ed l i mi t s of si ze and oper at i ng
charact eri st i cs. However , f or a gi ven set t i ng of t he
aut omat i c st op on t he gear - cor r ect i ng machi ne, say at
100 f t. , t he decr ease i n t oot h t hi ckness, and t her ef or e
t he amount of st ock r emoved f r omal l gear s put t hr ough
t he pr ocess, i s so near l y uni f or m t hat i t i s di ffi cul t to
measur e di f f erences f r omgear t o gear . One advant age
of t he pr ocess i s t hat onl y a ver y smal l amount of st ock
i s r emoved, t he resul t bei ng t hat ver y l i ttl e addi t i onal
backl ash i s i nt r oduced.
EFFECT OF PRESSURE
Suf f i ci ent dat a have not yet been accumul at ed t o de-
t er mi ne preci sel y t he ef f ect of t he amount of pr essur e
on t he r at e of r emoval of st ock f r omgear s. Wi t hi n t he
l i mi t s of uni t pr essur e t hat have been empl oyed10 t o
25 l b. per sq. i n. of t oot h f ace as cal cul at ed f r omt he
scal e pr essur e, l ever syst em, and t oot h di mensi onsi t
seems t hat t he r at e of abr asi on i s by no means di rectl y
pr opor t i onal to t he cal cul at ed uni t - pr essur e but var i es,
per haps, as t he squar e r oot of t he uni t pr essur e. Many
f act or s ar e i nvol ved, such as t he speed of t he machi ne,
t he i nert i a of r ot at i ng par t s, t he ki nd and amount of
abr asi ve, and t he f ri ct i on of t he pr essur e mechani sm.
A det ai l ed st udy of t hi s subj ect can be made onl y by
pr ol onged i nvest i gat i on. Pract i cal l y, i t has been f ound
t hat t he pr essur e shoul d be i ncr eased to t he poi nt above
whi ch t her e may be a t endency t o scr at ch or scor e t he
t oot h sur f aces, and shoul d t hen be mai nt ai ned sl i ghtl y
bel owt hi s poi nt . For t unat el y, wi t hi n r at her wi de l i mi t s
of uni t pr essur e, t her e i s no gr eat di f f erence i n t he re-
sul ts obt ai ned as t o qual i t y of pr oduct and r at e of abr a-
si on, whi ch seem t o depend mor e upon t he f oot age
t han upon any ot her si ngl e f act or .
The pr ocess shoul d be r egar ded as a "pr eci si oni ng"
or f i nal f i ni shi ng- pr ocess i n whi ch mi nut e i rregul ari t i es
and i naccur aci es, whi ch ar e i nevi t abl e i n t he most car e-
f ul l y cut gear s, can be r emoved r api dl y and aut omat i c-
al l y. The gear - cut t i ng machi nes ar e expect ed t o do t he
real cut t i ng, and i t i s bel i eved t hat t hei r f eeds can, i n
gener al , be i ncr eased subst ant i al l y wi t hout i ncr easi ng
er r or s i n t he gear s t oo muc h f or t hem t o be r emoved
economi cal l y by t he i nspect or - cor r ect or machi ne de-
scr i bed.
APPL I CATI ON T OHEL I CAL GEARS
Our exper i ence i n t he f i ni shi ng of t housands of hel i cal
gear s i ndi cat es what can be expect ed i n l arge- scal e
pr oduct i on. These gear s wer e cor r ect ed i n l ots r angi ng
f r om50 t o 400 gear s of each ki nd. The camshaf t gear
was r un f or 100 f t. of l appi ng on each si de of t he t eet h,
si nce i t was bot h a dr i vi ng and a dr i ven gear . Wi t h t he
gear - cor r ect i ng machi ne oper at i ng at t he r at e of 120 f t.
per mi n. , t he oper at or ' s t i me was about 2 mi n. i n t he
case of t he camshaf t gear , and about 1 mi n. i n t he case
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
484
S. A. E. TRANSACTIONS
VOL. XXIII
of the crankshaft and the accessory gears, which were
run for 100 ft. of lapping on one side only. In large
scale production this time could be cut in half, as the
same operator easily could run two machines and there
by double his output.
Fig. 12 shows the fixture that has been used for test
ing assembled timing-gear trains. It was originally
called an inspection fixture, but it was found that almost
all these gears were interchangeable and could be as
sembled without prior inspection, except as to appear
ance. Therefore the fixture is more properly called
an assembly test-fixture. It is belt-driven from a 1-hp.
motor and has a small hand-brake by which the operator
can apply a load to check whether, at certain speeds,
the gears are within the established limits of quietness.
The use of a battery of gear-correcting machines in
conjunction with a simple assembly test-fixture un
questionably will result in a remarkable uniformity of
product, the virtual elimination of engine tear-downs,
and a general economy in the production of quiet gears.
From the data obtained, the indications are that spur
gears offer no problems that have not been met with
helical gears. Several thousand gears for transmissions
have been corrected so satisfactorily that they have
passed final assembly-tests under conditions which can
be regarded as strongly indicative of the possibilities
of the process. Fig. 13 shows a gear-testing stand
adapted for the comparative testing of spur gears. A
small motor is used to spin the gears, which are placed
on studs at the correct center-distance. This simple
device has been very effective and is of real value in
enabling us to compare gears before and after correct
ing.
MACHINE ARRANGEMENT FOR LARGE-SCALE PRODUCTION
If only 100 ft. of lapping is required per gear, and on
one side only, the following scheme will be found to
produce the desired quality of gears economically. Two
machines will be operated by one man, four arbors being
in use when two different gears are being finished, or
three arbors when one kind of gear is being finished in
both machines. In the case of two different gears and
four arbors, the time schedule is as follows, A, with its
subscript number, representing the particular arbor
and M, with its subscript, the machine:
It will be noted from the foregoing schedule, which
is based on time studies of the actual operations, that
120 sec. is required to complete the cycle and that four
gears are finished, or one every 30 sec. With due al
lowance for contingencies, it is reasonable to expect
that one operator with two machines could correct 80
gears per hr., if 100 ft. of lapping is required per gear.
Once the process has been properly started under com
petent supervision, the direct-labor cost will be ex
tremely low.
If the other extreme of 200 ft. of correcting on each
side is considered, or a total of 400 ft. per gear, it will
be found advantageous in some cases to use one oper
ator with four machines and six arbors. Two of the
machines will operate on one side of the gear teeth and
the other two will operate on the other side, 100 ft. per
machine. Three of the arbors are arranged to hold the
gear with the hub up; the remaining three hold the gear
with the hub down. Each lap will thus assume the aver
age profile of both sides of the teeth, and any lack of
symmetry of the two profiles will thus be corrected.
The time schedule is as follows:
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
VOL. XXI I I SI MULTANEOUS I NSPECTI ON AND CORRECTI ON OP GEARS 485
I t wi l l be not ed t hat 64 sec. i s r equi r ed to compl et e
t he cycl e and t hat one gear i s f i ni shed dur i ng thi s t i me.
Under t hese condi t i ons, i t i s f ai r t o expect t hat , i n ac-
tual practi ce, one oper at or wi t h f our machi nes coul d
aver age 30 gear s per hr . Ot her ar r angement s and
schedul es wi l l suggest t hemsel ves when a st udy has been
made of t he speci al condi t i ons of any part i cul ar j ob.
The i mpor t ant t hi ngs to r emember ar e t hat uni f or mi t y
of spaci ng and prof i l e wi t h a cr owned f ace wi l l resul t
aut omat i cal l y, but t hat t he accur acy of t he prof i l e, wi l l
depend upon t he average of t he prof i l es of t he gear s as
t hey come f r omt he gear - gener at i ng machi nes and heat -
t r eat ment . I t i s t her ef or e i mpor t ant t hat t he gear s be
cut so as t o have t he gener al prof i l e wi t hi n r easonabl e
l i mi ts of accur acy.
The pr ocess al so can be used t o sal vage gear s. Near -
l y al l t he gear s t hus f ar correct ed have been rej ect ed
gear s. The f act s t hat such gear s have been sal vaged
and put t hr ough fi nal assembl y and test wi t h a negl i gi bl e
number of rej ecti ons, and t hat t hey have i n numer ous
THE DI SCUSSI ON
F. G. WHI T T I NGT ON
3
: Of what met al i s t he l ap con-
st ruct ed ?
R. B. F E HR: Wh e n correct i ng a l arge number of
gear s, we make t he l ap of chi l l ed cast- i ron by not pr e-
heat i ng t he chi l l bef or e pour i ng. When i t i s an ex-
per i ment al j ob i n whi ch we may have to modi f y t he
cont our , t he l ap i s made of sof t cast- i ron so t hat we
can easi l y use a cut t i ng tool , wi t hout resort i ng to a
gr i ndi ng oper at i on, f or modi f yi ng t he l ap. When onl y
a f ewgear s ar e to be correct ed or a parti cul arl y hi gh
pol i sh i s desi red, we use t ype- met al , whi ch has t he
added advant age t hat t he met al can be pour ed easi l y i n
t he l aborat ory.
CHAI RMAN D. D. ORMSBY
1
: Ho w i s t he conj ugat e
t oot h- f or mof gear and l ap obt ai ned?
MR. F EHR: T he wi de gear or "chi l l " shown at t he
cent er i n Fi g. 1 i s j ust l i ke t he gear t hat i s to be cor-
rected. I t i s cut by t he or di nar y met hods wi t h t he
teeth a tri fl e t hi cker t han t hose on t he gear , so t hat ,
when t he gear ent ers t he l ap, t here wi l l be a sl i ght
cl earance of say 0. 005 t o 0. 006 i n. As to howwe mai n-
tai n a pr oper gear - cont our wi t h a condi t i on of t hi s
sort, t he l ap ref erred t o i s t he count er par t of t he ex-
ternal gear because i t i s cast ar ound t he chi l l , whi ch i s
practi cal l y t he same as t he ext ernal gear t hat i s t o be
correct ed. Ther ef or e, when t he gear s to be correct ed
ent er t he l ap and emer ge, thi s act i on at fi rst t akes off
mer el y t he hi gh spot s of bot h gear and l ap, and t hus
t he l ap, whi ch l ooks l i ke an i nternal gear , i s gr adual l y
br ought to a f or m i denti cal wi t h t he aver age cont our
of t he gear s goi ng t hr ough. The pr essur e t hat I re-
f erred t o previ ousl y comes on at ever y st roke and, as
t he l ap wear s down, thi s appl i cati on of pr essur e com-
pensat es f or t he wear .
MR. WHI T T I NGT ON: Wha t i s t he pr opor t i on of t he
l engt h of l ap t o t he f ace of t he gear ? I n ot her wor ds,
f or a gear of a cert ai n wi dt h, what woul d be t he l engt h
of t he l ap?
* M. S. A. E. Chi ef engi neer , St ewar t - War ner Speedomet er Cor p. ,
Chi cago.
* M. S. A. E. Consul t i ng sal es engi neer, Br own- Li pe- Chapi n Co. ,
Syr acuse, N. Y.
i nst ances been used to repl ace gear s t hat had been re-
j ect ed i n f i nal road- t est s, ar e st rongl y i ndi cati ve of t he
ef f i ci ency of t he pr ocess i n correct i ng gear s posi ti vel y.
But t he real f i el d f or t he pr ocess i s i n l arge- scal e pr o-
duct i on.
The pr ocess does not depend f or i ts savi ngs i n pr o-
duct i on costs upon t he subst i t ut i on of a l ess expensi ve
mat er i al f or gear s. For exampl e, whi l e i t i s possi bl e to
pr oduce by t he pr ocess a qui et t rai n of rel ati vel y i nex-
pensi ve al l - metal t i mi ng- gear s, i t i s al so possi bl e to re-
duce t he cost of a t rai n i n whi ch t her e i s one composi -
t i on gear by appl yi ng t he pr ocess to t he met al gear s,
t hus r educi ng t he total cost of pr oduct i on and at t he
same t i me el i mi nat i ng t he er r or s usual l y f ound i n such
gear s. Wi t h gear s havi ng uni f or mi t y of spaci ng and
prof i l e i t i s r easonabl e to expect gr eat er durabi l i ty and,
t heref ore, a l onger l i fe wi t h qui et operat i on. The use
of t he pr ocess assur es t he mi ni mum r emoval of st ock,
and consequent l y t he mi ni mum amount of backl ash be-
t ween mat i ng gear s.
MR. F EHR: Ther e must be a cl earance of 11/ 2 i n. at
t he t op of t he 7%- i n. st roke f or al l t he l aps t hat we
make, so t hat , wi t h t he except i on of shor t er l aps f or
ver y smal l gear s, t he l engt h of l ap i s usual l y 6 i n. f or
t i mi ng and t r ansmi ssi on gear s.
ACCURACY OF T OOT H SPACI NG
A ME MB E R : I shal l not pr esume t o say t hat absol ut e
spaci ng actual l y exi sts. Someone has sai d t hat al l di -
mensi ons ar e myt hs, and t hat t he onl y quest i on i s how
near l y we can appr oxi mat e t hem. Hownear , i n act ual
measur ement s, does t he operat i on pr oduce "absol ut e
spaci ng"?
I n deal i ng wi t h case- har dened gear s, I have f ound
t hat t here i s i n some cases a necessi t y f or an al l owance
of somet hi ng l i ke 0. 010 to 0. 012 i n. to al l owf or har den-
i ng di storti on so t hat t he t eet h wi l l cl ean up and gi ve
a correct cont our . Howdoes Mr . Fehr deal wi t h such
condi t i ons ?
MR. F EHR: Our devi ce f or measur i ng spaci ng i s
cl ai med by t he maker s to be accur at e to 0. 0001 i n, We
t hi nk t hat under usual condi t i ons i t can be rel i ed upon
to wi t hi n 0. 0002 i n. The gear s t hat have been meas-
ur ed on t hi s i nst r ument certai nl y come wel l wi t hi n thi s
l i mi t.
The pr ocess i s not a gear- cut t i ng but r at her a gear -
correct i ng process. The gear s ar e expect ed to come
wi t hi n t he l i mi ts of si ze t hat ar e cust omar y, and we do
not make any cl ai mf or r emovi ng economi cal l y a l arge
amount of st ock; but st ock can be r emoved i f t i me
enough i s t aken t o do i t. The means t hat we have f or
keepi ng uni f or mcont our s when r emovi ng l ar ge amount s
of st ock ar e rel ati vel y si mpl e. We do not f i ni sh t he
gear s al l at one t i me, thi s bei ng a ver y i mpor t ant f ea-
t ur e i n such cases. Suppose we have 100 gear s. We
put each gear t hr ough t he pr ocess f or 100 f t. Then
we put t hemal l t hr ough agai n and agai n i n seri es unti l
we have r emoved t he amount of st ock requi red. Thus,
by doi ng what we cal l "mul t i pl e- seri es" l appi ng, we ar e
abl e t o mai nt ai n al l t he t oot h cont our s of al l t he gear s
wi t hi n ver y cl ose l i mi ts of t he aver age cont our of t he
gear s we put i n. I f we kept one gear goi ng t hr ough
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014
486 S. A. E. TRANSACTI ONS . VOL. XXI I I
unti l compl et ed, t her e mi ght be a sl i ght t endency f or
t he cont our s t o become ei t her mor e posi t i ve or mor e
negat i ve, especi al l y i f t he l ap or gear happened t o be
of somewhat var i abl e har dness f r om ti p t o r oot of
t oot h. But , so l ong as we keep put t i ng t he gear s
t hr ough i n mul t i pl e- seri es, never doi ng mor e t han 100
or 200 f t. on one gear at a t i me, we f i nd t hat uni f or m
cont our s can be mai nt ai ned. The f act i s, however , t hat
we har dl y ever r equi r e mor e t han a total of 200 f t. on
one si de of an aver age gear to cl ean i t up; t hat i s, t o
cor r ect spaci ng and to pr oduce uni f or mcont our s.
C. L. KNOP F
5
: I f t he l ap i s i n cont i nued use, does i t
not devel op shoul der s and i s i t not necessar y t o shor t en
t he st r oke t o avoi d excessi ve cr owni ng of t he f ace of
t he gear t eet h?
MR. FEHR: Ref er r i ng t o Fi g. 6, t he gear i s br ought
out hal f way f r omt he l ap at t he bot t om; at t he t op i t
comes out wi t h t he pr essur e sti l l on; so t he t endency
5
M. S. A. E. Lubr i cat i on engi neer , Si ncl ai r Ref i ni ng Co. , New
Yor k Ci t y.
i s t o bel l - mout h t he l ap r at her t han t o cr eat e any shoul -
der . That i s howt he "cr owned- f ace" ef f ect i s obt ai ned.
CHAI RMAN ORMS BY : Wh a t i s t he l engt h of l i f e of
such a l ap?
MR. F E HR: T h e chi l l ed cast - i ron l ap has a l i f e of
f r om 1000 t o 3000 gear s, i f bot h si des of t he gear ar e
put t hr ough t he pr ocess.
A ME MBE R: I s t hi s pr ocess consi der ed pract i cabl e
f or case- har dened gear s of f r om 85 t o 90 Shor e har d-
ness?
MR. FEHR: Cas e- har dened gear s of as hi gh as 85
Shor e har dness have been cor r ect ed. We have not yet
f ound any gear of any degr ee of har dness t hat woul d
not yi el d readi l y t o t he cor r ect i ng ef f ect of t he pr ocess.
MR. WHI TTI NGTON: Conc er ni ng t he pr essur e on t he
gear when goi ng t hr ough t he l ap, i s t he l ap hel d ri gi dl y
agai nst t he gear ?
MR. F E HR : The l ap i s hel d i n a f i xed posi t i on on t he
machi ne, but t he gear on t he f r ee end of i ts ar bor i s
al l owed t o f l oat t o i ts pr oper posi t i on i n t he l ap when
t he t or que i s appl i ed by spr i ng t ensi on.
Downloaded from SAE International by University of Michigan, Wednesday, October 29, 2014