CNC Bevel Gear

Generators and Flared

Cup Gear Grinding
Theodolre Kr,enzer
The Gleason Works. [Link], INY

New freedom of motion available with CNC Starting with a brief background of the
generator make po sible improved tooth con- flared cup proces , the article proceeds to de-
tact on bevel and hypoid gears. Mechanical scribe the possible linear and angular motion
machines by their nature are inflexible and re- variations and their effect on the gear tooth
quire a special mechanism for every desired surface. The article concludes with the use of
motion. These mechanisms are [Link] exotic Tooth Contact AnaJysis (TCA) to evaluate the
andexpensive, As a result, it was not until the enhancement of this duplex process made pos-
introduction of CNC generators that engineers sible by applying these motions.
started exploring motion possibilities ami. their [Link] Cup Process
effect on tooth contact. When cutting face milled gears usingthis
This article covers the exploitation of new operation, the cutter is po itioned relative to the
motion freedoms to improve tooth contact pat- gear blank. 0 that the correct piral and pre sure
terns on gear et manufactured by a face- angle will be produced ..The gear blank is held
milling duplex process. a manufacturing stationary, and a 'tooth slot is form-cut by
method where both flanks are completed in a infeeding the cutter. The part is indexed one
single operation. pitch. and the process is repeated. See Fig. 1.
When the cutter is replaced bya grmding wheel,
eontactexists over the entire length and depth of
/ the tooth surface ..Heat buildup results, cau ing a
tendency for surface damage because ofburning.
work gear The flared cup grinding proce sis u ed to over-
come thi problem.
The flared cup process uses a wheel which is
/ machine root angle
tilted out of the work (30"s of till is commonily
tooth 101
used). The out ide wheelsurface ha a norma]
radius of curvature less than the conventional
tool, and the inside wheel surface has a normal
radius of curvature greater than the can venti anal
tool, Line contact exists between the work and
the wheel. The wheelis positioned relative tothe
gear blank so that atthe calculated mean position
on the gear tooth surface, the correct spiral aund
pressure angles are produced. The tilled axis of
Fig. ]- Conventlenal face milling duplex setup. the whee] is in a plane normal to the tooth
18 Il EAR TECHNOLOGY
surface. In achieving this setup the tilted axis is
offset from the conventional tool axis. The tooth
is swept out by rotating the flared wheel about /
the axis of the conventional tool axis. See Fig. 2.
As the wheel is dres ed its radii change.
which require compensating machine changes
to maintain the proper tooth geometry. Mak-
ing these setting changes manuall'y on me-
chanical. machines is a problem. Tooth geom-
/ conventional tool axjs,

etry often varies from part to part. Fun CNC flared tool axis A .',
machine. where wheel. radius can be accu-
rately determined, are programmed to auto- Fig..2 -[Link] cup setup.
matically compensate for wheel size changes
re ulting from dre sing ..
Wheel life i a function of the radius change
which OCCUJS as a re Lilt of dressing. Over the \-~.

u eful life of a wheel, the relative curvature

~
..
.,
\.1.

decreases between the convex tooth surface and

I .~ \.' ..
~:~'I
th.e inside wheel surface and increases between .
''{':.'I
')
.J

the concave 'tooth surface and the outside wheel ideal wheel new wheel spent wheel
surface. Although the final tooth surface is pro-
Fig. 3,- Contact area variation ..
duced by line contact, at any instant surface
contact exists between the wheel and work in
proportion to the depth of grind. The contact area
is dependent on the relative curvature between
the wheeland work and the variation in the
contact area between the two tooth ides is used
to determine wheel life. As arule ofthumb, good
re ults are obtained when the difference in con-
tact area between the two ide of the tool, does
aot exceed the ratio of 2: 1. Fig. 3 show ection
of a. tooth and grinding wheel at three stages of
I
wheel life - idea], new whee], and spent wheel. Aa
pivot point
New Freedoms ~ .-'

Three angular and three linear motions de-

fine the relativ - motions that can exist between
any two bodies. in this ca e between the flared \ n~
cup tool and the work gear. One of the angular
freedoms i u ed to weep out the tooth urface ;
therefore, effecti vely onJy two angular freedoms
are available for contact pattern control. Fig. 4 - Flared ,cuP'motions.
At any instant in sweeping out the tooth I. Rotational motion in the instantaneous
surfaces, the CNC generator has the capability radial plane:
to. change the relative orientation between the 2. Rotational motion about the instanta-
contact line and the gear tooth. Moti,ons to neous radial line;
achieve a change could be defined in any num- 3. Liaear motion along the coaventional
ber of reference systems. For thi case all of the tool axis;
motions are defined ba ed on the in tantaneous 4. Linear motien along the instantaneous
radial. plane; thai is. the plane containing the radial Ii11 e:
conventinnal tool axis and the radial. line to the 5. Linear motion perpendicular to. the in-
mid-height point on the contact tine. stantaneous radial line.
Tile freedoms are defined as follows: Fig. 4 .is a sketch of a flared cup setup
JU LV/AUGUST IU3 19
 showing these motions, The angular [Link]
Program 'Fest
821.001 pivot about a point at mid-tooth depth and
Rer (solid) Surface - GTWRtur801IPAR
Base mid-slot width .. A timed relationship exists
Sec (Dasbed) Surface - GTWRl<tnIO I fPARA between the rnotions and the angular po [Link]
DehaA2· ,02

Vertical Distance at
of the wheel as it is swept through the tooth
'FTip 0.00015" slot. Although a number of function could be
H Tip 0,0003,1'
T Root -.0000 I " Root
used to define the relationships, polynomial
H Root -.00004"
Toe
expressions were selected.
Scule Factors
X Mal!
V Mag
1,00
1.00
Radial Tilt. This angular motion i a tilting
Thel 60.00 of the tool in the instantaneous radial plane as
Dif 5000.00
the tooth is swept out. The effect 1S to change
Tip Convex Gear Sun ce DiffJArrow
Scale: Metal on .001" the pressure angle on both flank of the rooth
as the grind line moves from the tooth center
Program Te~t
821.001 section. The change increases the pres ure
Ref (solid) Surface -GTWRKfT801IPAR
Base angle on one flank, and decrease the pres ure
Sec (Dashed) Surface - GTWRKtr801fPARA angle 0111 the other flank as compared with the
Dealta A2 " .02

Root
conventional tooth. At any tool phase angle
Vertical Distance at
T Tip -.00023" position designed by Aa. the radial tilt. of the
H Tip -,000(5"
T Root 0.0000 I" tool is given by:
H Root 0,00004"

Scale Factor'S
lla = A I Lio:... A2 tlu? ... A) Aa3 + A4 Aa4
X Mag 1,00 where A I' A2• A3, A4are the coefficients that
V Mag 1.00
Thet, 60.00 contrnl the motion.
Dif 5000.00
Tip Concave Gear Surface DiffJAlTow
Scale: Metal on .001" Fig. 5 schematically illustrates the cbange
in surface topology on the convex and concave
Fig. 5 - T,opologygraph of [Link]!order Aa ,change. flanks of a gear tooth, The solid lines represent
the baseline surface, and the dashed line repre-
Program Test sent the surface resulting from a second order
821.001
Ref (solid) Surfuce - GTWRKtr80 1182100 I change in Lia. It can be seen that. metal is
Program Test
removed 011 each side of the center eerion at the
Sec (Dashed) urface - OTWRKlfl!OlfPARB 1
Delta BI •. 01 bottom of the 'convex flank and at the top ef'the
Vertical Distance at concave flank of the gear teeth. The oppo ite
T Tip -..00020'
IITip -.00042" enect occurs at the top of the convex side and at
T Root 0.00011"
H Root 0.00022" 'the bottom of the concave side.
Scale Factors The Al coefficient produces a velocity in
X Mag 1.00
V Mog 1.00 the normal direction atAO: = 0, the setup must be
Thel 60.00
Dif 50(KHJG altered to accom:modate the velocity when this
coefficient is used.
1'.angemiat Tilt. This angular motion is a
Program Test
821.001
tiUing of the tool around the in taotaneou
Rd (solid) Surface - GTWRKfI'30 1/821'00'1
Program Test
radial line as the tooth is swept out. Again the
Sec (Da hed) Surface- GTWRtur80IIPI\RJiII
effect is to change the pressure angle on both
Delli 81, • .oJ
flanks of the tooth as the grind line moves from
Vertical Dlstance ,al
T Tip -.00039"
the tooth center section. In this ca e the pres-
H Tip -.00025" ure angle is increased or decreased on both
T Root 0.00014"
Hi Root 0.00011."
flanks,a compared with the conventional tooth.
Scale Factors,
X Mag 1.00
At Blnytool phase angle position designated by
V MUI! 1.00 6., the tangential tilt of the tool. is given by:
The! 60.00
[Link] 5000.00
=
A~ B I 6.0: of- B2 6.0;2 ... B3 Aa3 ... B4 6.0.4
Tip encave Gear Surface DifffArrow where B], 82, B3• B4 are the coefficients thai
Scale: Metal 011 .001"
control the motion.
Fig. 6 schematically Illustrate the change
Fig ..(i - Topology graph of second order AP change. in the surface topology due te a second order
201 ileA'R rECIiNOLOGV
change in Ap. It can be een that metal is Program Te.~l
821.001
removed on each : ide of the center ection at Ref [ olid) Surface - GTWRKff801f82 1001
Program Test
the top of both tooth flank . The oppo ite effect
Sec (Dashed) Surface - CilTWRKff801IPARL
occurs at the bottom of both tooth flank . Della L2 -",02
.Axiat Motion. Thi Linear motion is one that Vertical Distance al

has been described previously in Refs, 1 and 2. J( T Tip 0.00023"

H Tip 0.00052" 'ROOI
T Root 0.00025"
is an advance or retreat of the work along the H Root 0.00043"
conventional tool axis. The effect i to mod!ifythe Seale Factors
X Mag 1.00
surface topology m the same manner on both V Mag 1.00
The! 60.00
tooth flanks ..More metal is either removed or left Dif 5000.00
on the tooth flank: in the lengthwise direction
than in conventionally formed teeth. At any too] Convex Gear Surface DiCr/Arrow
Scale: Mew on .001"
phase angle position de ..ignated by An. the change
in position of'lhe tool relative to the work in the PmgramT~"
821.001
Ref (solid) Su rface - GTWRKff801/82 I 00 I.
direction of the cutter axis is given by: Program Test
M = Ll Au + ~ Aa2 + L3 a3 + L4 Aa4 Sec (Dashed) Surface - GTW RKffSO lIP _ItL
Della L2 -",02
where LI• ~, L3• L4 are the coefficient that
Vertic,,:! Distance at
control. the motion. T Tip 0.00039"
H Tip 0.00024·
Since the L. coefficient produces a [Link] T Rool 0.00034'
H Root 0.00021"
in the normal direction at An '" 0, the setup must
Scale Factors
be altered to hold spiral and pressure angle at the X Mag 1.00
V Mag 1.00
mean point when this coefficient is used. Thel 60.00
DiP 5000.00
Fig. 7 schematically illu tratesthe change in Tip Concave Gear Surface DiITIArrow
Scule: Met!!1 un .00 I..
surface topology due toa second order change in
AI. It can be een thai metal is left on at the in ide Fi • 7- Topology :1P'.apbohecond order t:J change •.
and outside of both tooth flanks.
RtuUal Motion. Theismotion is a movement Program Test
821.001
of the tool along the instantaneous grind radiu Rcr(solid) Surface - GTW RKfT'80 1182 100 I
Program Test
between tile tool and the work gear. The effect is
Sec (Dashed) Surface - 'GTWRKlT801/P RR
to modify the surface [Link] [Link] oppo ire Della R2 .. 02
Rom
manner on the tooth flanks. Mela] is removed OIl Vern,a] Distance .1
T Tip . JJOO23 ,
one flank and metal is left Oil the other, unlike the H Tip ·.00064"
T Root -.00033··
case of conventionally formed teeth. At anytool H Root ·.00056"

phase angle po ition designated by Aa~the change Scale Factors

X Mag 1.00
in po ition of the tool relative to the work in the V Mag 1.00'
Tbet 60.00
instantasecu radial direction is given by I
eu 5000.00
ll.r = RI Au + R2Au2 + R3Ao:3 + R4Ao:4
'lip Convex Geur
where R., R2• R3, R4 are the coefficients that
control the motion.
Since the R. coefficient produces a velocity Program Test
821.001
in 'the normal direction at. An '" O. the etupmust Ref (solid) Surface - GTWRKf1'8011821 00 I
Ptogram TeSi
be altered 'to accommodate the velocity when
Sec (Dashed) Surface - GTW:RKff801IPARR
this coefficient is used. Delta R2· .02

fig. 8 schematically illustrate the change in Vertical Distance at

T Tip 0.00052"
urface topology, due toa econd order change in II Tip 0.00032"
T Root 0.00046'
Ar. It can be een that metal i removed 011 the H Root 0.00037"

convex side and teft on the concave side on each Scale Factors
X Mag 1.00
side of the tooth center ection, V Mag 1..00
Thet 60.00
Tangential Motion .. The tool can also be DiC 5000.00
Tip Concave Gear Surface Diff/Arrow
Scale: Metal on .00 I"
moved .in a direction perpendicular to the in-
stantaneou grind radius. The effect is to mod!ify
me surface topology in the opposite manner Fig. 8 - Topology graph of second order M change.
JULY/A,UGUST leu 211
 01'1 the tooth flanks. ]t has a effect similar to the
Progt" m Tes
S21.001 effect de cribed above for 8. change i1ll radial
II Ref (solid) Surface - GTWRI(fTS01l82 1.00I
Program Tesl motion, At any cutter pha e angle position
Sec (Dashed) Surface -GTWRKff801/S I designated by .1a, the change in po ition of the
Delta SI - .01
'ROO!
tool. relative to the work. in a direction perpen-
Vertical Di: lance 31
T Tip -.00013" dicular to the in tantaneous radial i .given by:
:H Tip -.00031 •
TRoD! -.0001.6"
Ii Root -.OOO2H"
.18 =
S I .10: + S2 ll.a2 + S3ll.a3 + S4 0..0:4
Seal", Factors
where S I' S2' S3' S4 are the coefficients that
X Mag 1.00 control the motion.
V Mag 1.00
The! 60.00 Fig. 9 schematically illustrates the change
Dif 5000.00
Tip
in surface topology due to a second order
change in ll.s.
Program lest
8'21.001 Tooth Contact Analysis (TeA)
Ref (solid) Surface - GTWRKlT80J/821 00 I,
Program Tesl A .econd illustration of the five motion
Sec (Dashed) Surface - G'fWRKfI'IIO liS I freedoms is shown using TeA. TiQ aid in the
Delta S I - .01
comparison, the same job is used, and the effect
Vertical Distance at
t Tip 0.00023" of second order change. ofthe arne magnitude
H Tip 0.000 17"
T'Rool 0.[10023" are evaluated. Fig. W is the baseline TCA. It
H Root 0.00019"
repre ent a conjugate gear set with only length-
Scale Factor
X Mag 1.00 wise mismatch. Fig. 11 is the TeA of the radial
V Mag 1.00
Thel 60.00 Tip Concave Gear Surface DifrlArrow tilt change. Fig. 12i5 the tangentialliit change.
Oil' 5000.00 Scale: Metal on .00 I"
Fig. J3, i the axial motion. Fig ..14 is the radial
motion. Fig. I5is the tangential motion.
change.
Duplex Enhancement
Tooth Contact Analysis - Finishing The face milling duplex process is success-
Gear - Convex H H Gear - Concave fully used in many application ; in particular
IOn fine pitch jobs and jobs where the contact
II :iifii1j1 pattern is,enhanced by lapping. When grinding
the final, finishing operation, the de ired
contact pattern length needed for rna t auto-
motive applications cannot be achieved eas-
ily. Often a contact pattern. where the contact
length varies from top to bottom, called a
diamond pattern, results. This ha been a factor
in limiting the success of grindingas the final
operation ·in. the manufacture of hypoid gear
ets for land applications. Enhancement of the
Fig. 10 - Baselin. TeA. duplex process by exploiting the new motion
freedoms hould make grinding more attrac-
Tooth Contact Analysis -Finishing
tive asa final fini hing proce s becau. e of the
Gear - Convex H " Gear- Concave
ability to develop contact. patterns witha wide
range of characteristics.
TypicaUy, gear sets used in [Link] and

I I'ABLE 1- BLANK nATA SMALL AUTOMOTIVE

-I
Inch MM
[Link]/Module 6.154 4.]3
Pilch Diameter 7.638 [Link]
I lumber of Teeth
1.300
12/47
Face Width 33.02
Offset 1.250 3l!.7S
Spiral Angle 48°

Fig. 11 - TeA or, eeead order lI.a: change.

Cutter Radius 3.000
I 76.20

22 GEAR TECHNOLOGY
 truck applications cover the range from two to Tomh Contact Analy .j , - finishing
seven diametral pitch. Three set within thi Gear - Convex H H Gear - Concave I
range were designed u. ing the taadard duplex
method plus the new motions. Blank data for
the ets is given in Table I-
Smail Automotive - When the diametraI
pitch is six or higher, (he need for added flared
cup motions can be questioned. Fig. [6 is a
TeA comparison of the duplex job de igned
withoul added motion on the .Ieft and with
added motions on the right.. The job were
de igned with very Iittlerransmlsslen motion
variation .. Both designs are imilar. Sub tan- Fig. 12 - TCA of second order A~ change.
tial pattern length wa. obtained without the
introduction of flared cup motions. However. Tooth Contact Analy is - Finishing
even for this case, greater pattern length was I Gear - Convex H H Gear - Concave I
achieved 011 center, while limiting the length
when the contact moves to the inside or out-
ide of the blank. The long center contact i
beneficial. when gear noise isa concern.
Large.Automotiv8 -or coarser diametral
pitches, the benefits that can be achieved are
more easi1y seen. Fig. 17 isa TeA compari on
of the largeautomotive set designed without
added motion on the left and with added
motions on the right. The tandard develop- Fig ..13 - TCA of second order aJ change.
ment is reasonably acceptable, and enhance-
ment. of the contact pattem as a result of the Tooth Contact Analysis - finishing
Gear - Convex H H Gear - Concave
lapping could re ult in a setthat would be very
acceptable relative to noise quality. However,
__ I
if the development i . for a final finish grind,
the contact pattern has more than the desired
lengthwise mismatch.
Lengthening the pattern. by conventional
duplex methods would result in an unaccept-
able diamond pattern.
Theintroduction of additional motions 'can
ub tantially increa e the on-centerpanern lengtb.
while controlling the diamond condition. AI 0, Fig. 14 - TCA ,of eeond order 8I'change ..
note that pattern length at the toe and heel is held
to a reasonable lengthto maintain adjustability. Tooth Contact Analysl» - Finishing
Gear-Convex H H Gear -oncave
Large Truck - For diametral pitches in the
rangeoftwo, the benefits are dramatic. Fig. 18

LARGE AUTOMOTIVE TRUCK

I -
Inch
-
MM Inch I I MM
I
11.178
10.500
11/41
I
6.08
266.70 I!:!;~I 7/37
] 1.41
422.28

1.600 4.0.64 2.000 50.80

1.500 38.10 1.750 44.45
480 47°
3.750 95.25 6.000 152.40
Fig ..15 - TCA (If second order As change.
 is a TCA comparison of tile large truck job
tooth Contact Analysis - Finishing Tooth Contact Analysis - Finishing without added motions on the left and with
I Gear-Convex H HOe!!r-Concave r I Gear-s-Convex H H Gear-Concave I
added motions on the right. Pattern length on the
standard development is short. Typical bias-in
!~

at the toe and bias-out atthe heel contact

pattern exsst, Any furthes attempt to lengthen
the pattern by conventional means [Link]
severe diamond problems, Note that more trans-
mission motion variation was introduced into

X1
i I
this design on the assumption that the set is
more highly loaded and. as a re ult, requires
/\(\
i \ I
increased adjustability,
With the added motions, the lengthwi e pat-
tern was increased at the central position, the
diamond patterns, at the toe and heel were COD-
Baseline flared Motions
trolled, and the lengthwise mismatch at the toe
and heel were held ..
Fig. 16 - TeA of compadson 'of small [Link] daplex [Link].
Summation
The article presented a theoretical de cription
Tooth Contact Analysis - Finishing Tooth Contact Analy is - Finishing of the freedoms available on full CNC generators
I Gear - Convex. H H Gem - Concave I I Gear-Convex H H Gear-Concave [
and their application to the flared cup gear grinding

~I ~' process. Surface topology and TCA were III ed to

graphically define the effects of motion variation on
the tooth surfaces, Finally, the application of the

DIU'I LAi!J I LE ] motion freedoms to enhance the flared cup duplex

process was demonstrated using TCA.

-,' I In automotive and truckapplications the amount

1/f17rr n,,' 117\71

7\{\i K!\
and distribution ofrhe mismatch between mating
surfaces has a critical effect on sound quality,

TTIf\(\
2\)\ ret;. K]
J.
12(2\
I
Theoretically, the flared cup process combined
with a full eNC hypoid generator offers the mo-
tion freedoms that provide the necessary mis-

Baseline Flared Motions match control lor the duplex process allowing
both pinion and gear members to be finished

Fig. 17 - TeA [Link] of lar,gc automotive duplex design. ground in one operation I.

Referenoes:
Tootll Contact [Link] - Finishing Tooth Contact Analysis - Fmishing
[ Gear-Convex.H H [Link] I I. "Flaring Cup Grind Formate Bevel and Hy:poid GeMS."
] Gear - Convex H .H Gear - Concave I
Gleason Publication ETI349l1. January. 1990.
2. Krenzer. Theodere J. "Face Milling or Face Hobbing;
Choosing the Best Process." .AGM.A Fall Technical Meet-
ing, Toronto, Ontario, October, 1990..
3. Dodd, Harry D. and K. V. Kumar. "TechnologicalPunda-
mentals of Bevel Gear Fini h Grinding." ASUE Journal of
Gear Manujaclurirl:g, November/December, 1985.
4. Krenzer, Theodore J . .& Kent D. Yunker. "Universal
Bevel and Hypoid Generator." 4th Biennial International
Machine Tool Technology Conference, Chicago. Il., Sep-
tember.1988.
5. RUler. Meriwether L, Jr. "Second Order Surface Gen-
eration." Journal of Industrial Mathematics Society, Vol.
23,Patl 2, 1973.

Acknowledgement:
Originally presented at the A QUA Pal! Technical Meeting.
Baseline Flared Motions
199 J. Reprinted with permission. The opinions. statements.
and conclusion's presented are those ojthe aUl/lClr and ill no
[Link]. 18,- TeA compar:1 on of large truek duplex de slgn, way represent the position or opinion of AGMA.
24 GEAF! TECHNOL~QV --