A Comparison of ISO

4156/ANSI B92.2M,-,1980
With Older
Imperial Standards
J'ohn Finney
Spline Gauges Limited
,(Division ,ofIM&IM Pr,ec,ision Systems, Dayto,n. OHI)I
Picca,dliUy KilngsburyWarwickshire
England
The purpose of this article is to discuss ISO
and to compare it
tandards till in use .. In our
experience designing and manufacturing spline
gauges and other spline measuring or holding
4156/ANSI

B92.2M-.1980

with other, older

Although ISO 41.56 was published in 1.981

and is widely used in Europe, we find that it is
not u ed as frequentlya . it hould be in the

United States and United Kingd.om for a number of reasons. Onei

that many engineers do

devices for spHned component manufacturer

not seem. to be aware of it; another

throughout

because it is a metric module standard, the cir-

the wcrid, we are constantly

sur-

prised that so many standards have been producedcovering

whai is quite a small

Many of the standards aeemternational

ubject.
tan-

is that

cular pitch is different from that of imperialstandard-based

pitch splines. This

diametral

fact. can necessitate buying new cutters; how-

dards; others are company standards, which are

usually based on international
standards ..
Almost all have simileritiesi jhat is, they all.

ever, this should not present

deal with splines that have involute flanks of

metral pitch) cutter. Also, DP cutters have to

where new project

cutter is no more expensive

30,37..50 or 45 pressure angle and are for the

be periedically replaced,

most part flank-fitting

metric

diameter-fining.

or occasionally

major-

plines?

designed

any problem

are involved, as a module

than a DP (diawhy not change to

SO'

Furthermore,

spline gauge

to a module standard

are no more

'expensive than DP gauges. These arguments.

of course, cannot apply

[0

componentsthat

have been made for many years, andtherefore

Space Width
of Internal
Spline

have to be interchangeable.

In our opinion, the main reason that ISO

4156 is not used more :inthe e countries is lack
of awareness. We hope the following i informative

10

those who have not heard of it and

en COli rages tho e who are aware of it to u e .it.

We feel Ibis isnecessary
Tooth
Thickness
of External
Spline

because quality con-

trol the e days is of the utmost importance.

This standard (unlike the previous

standard
guarantee

) has many cia

imperial

es of fits that will

sliding, clearance

or interference.

Also, in genera], components manufactured

to

this standard are of a higher quality. astighter

22

GEAR

TECHNOLOGY

tolerances

have to be maintained

to achieve

Japan, among others ..At a meeting of all mem-

these better fits. Section 8.1 of the standard is

bers in Paris, two draft proposals were put for-

of paramount importance to component manu-

ward for consideration:

facturers.

based on DIN 5480, and the other was a com-

This paragraph

has wide-reaching

One proposal

was

by AFNOR

implications, as :it means that closer tolerances

pletely

have to be maintained to obtain the desired fits,

Briti h Standards ..The basis of the tolerances

meaning that qualitycontrol

during manufac-

forthis draft was taken from ISO 286; i.e., tol-

ture and gauging win have to be increa ed, but

erances for plain parts with an additional toler-

that is inevitable anyway in today's competi-

ance for variation allowance.

new draft proposal

and

After discussion, the new draft was adopted

tive global markets ..

ISO 4156/ANSI 892.2M-1980Its Origins

and contributions

from alii member organiza-

tions were invited ..The result of the collaborain the

tion i ISO 4[56. 111is standard also has been

United Kingdom that with metrication in mind,

published in the United Kingdom as BS 6186.

a new standard for involute splines was

required, Up to this time and indeed still today,

in France as E22-144/E22-] 45, and in the

United States as ANSI B92.1-1980. It is used
extensively in Europe and is gradually becom-

During

the ] 970s it was realized

the most widely

11

ed standards were BS 3550-

1963 and ANSI B92.] -1970. Both of these

ing more and more popular

standards

United States, where transition from imperial

are. of course, imperial

standards

based on OF. The DF is the number of teeth

per inch of pitch-circle

diameter and as such

provides a standard series of tooth sizes on various pitch-circle diameters ..The proposed standard,

of course,

European

would

standards,

have

to be, like

metric-mod

ule-based,

throughout

the

to metric is taking place.

Explanation of Variation Allowance

& Effective Size
To enable ISO 4156 to be compared with
the older standards.

it is nece sary to under-

stand the concept of effective fit. The effective

Similar to DP. the module is the ratio of the

tooth thickness of the space width of a compo-

pitch circle in millimeters

nent is the most important element because it is

to the number of

a standard series of tooth

this size that determines whether the fit is good

sizes. It was not atisfactory simply to convert

or bad. It is not only the actual tooth. thickness

DF to module because the pitch-circle diame-

or space width that determines the resultant fit

ters and tooth proportions would still be imper-

between an external and internal spline. This is

ial-based and would therefore give no degree

because, unlike plain pans. splines have multi-

of interchangeability

ple engaging surfaces. Each of these surfaces is

teeth and provides

AFNOR
authority)

with metric components.

(the French

national

and DIN were consulted

standards
to see if

subject

to

spacing/index

an

error

of

or parallelism,

either

profile.

which must be

there were any standards in existence on which

allowed for jf the parts areto fit. This variation

the new standard could be based. The existing

allowance is determined by allocating to each

French standard E22-141 was considered to be

of the elements a permissible

out of date, as it made no reference to' variation

sum of these tolerances is then referred to as an

allowance (di cus ed later). The German. stan-

error allowance e or variation allowance

dard DIN 5480 seemed to be the only other

tolerance. The

A.

It is generally accepted in most standards

possibility, since it did take into account varia-

thai. because it is unlikely that. these errors win

,John Fi,nnev

tion allowance. The result of this consultation

occur simultaneously

is design manager with

Spline Gauges Limited,
a subsidary of M & M
Precision Systems.

was the realization that more than a new metric

amounts

standard was required. What was needed was

allowance

an international

sum of the positive

standard

that could be used

To achieve this. working groups were established at the various standards organizations;
the working groups consisted
from

the United

Germany, Scandinavia.

on the same spline,

consisting

a variation

of a percentage
profHeer:mr.

of the

the index

error and the parallelism error for the length 'of

throughout the world ..

tives

and in their maximum

of representa-

Kingdom,

France,

the United States and

engagement is allowed.
To achieve at suitable fit. the nominal tooth
thickness of the external spline and the nominal space width of the internal spline usually
are equal to one half of the circular pitch at the

Dayton, OH. He has

more than rwo decades
of experience in 1M
design of splint! gauges,
master gears and splin
measuring instrumenu.
HI! is a member of the
Shafts for Machiner
Committee of th British
Standards Institution.

SEPTEMBER/OCTOBER

1994

23,

pitch-circle

diameter.

allocateda

The internal

spline is

space width greater than nominal.

A machining

m.i

tolerance

arrive ata maximum

determined

and minimum

to.

actual

values of total index, twice the positive profile

are used to calculate the

and lead variation

variation allowance; i.e., 0.6 (0.038+ 0.010+

0.0]0)

= 0.035

(Due to. rounding

of various

space width, and the variation allowance Ai. is

elemental errors the value is stated as .0.0.38 in.

then subtracted from each value to. arriveat a

the ANSI

tandard.)

maximum and minimum effective space width.

For Classes 4, 6 and 7. the machining toler-

Similarly, the external spline is allocated a

ances and allowable errors for a Clas . 5 pline

tooth thickness

that is less than nominal.

are multiplied by the following factors:

machining tolerance is used to arrive at a maxi-

Class 4= x 0.71

mum and minimum actual tooth thickness, and

Class 6= x ] ..4

the vaeiarionallowancei
tooth thickness.

Class 7= x 2.00.

added to each value

toarrive at a maximum and minimum effective

Table

1 shows a typical fit

condtrion between an external spline and an

internal one.

Because the minimum effective space width

of the internal

and the maximum

effective

tooth thickness of the external spline are equal,

a fit is always guaranteed .. To ensure this, the

ANSI B92.1-1970

GO plug gauges and the GO ring gauge have a

This standard is probably the most widely

used standardtoday.

As can be seen from

tooth thickness/space
(with a tolerance

widthequal

to nominal.

and wear allowed),

which

Go. plug gauge and one

Table 2, it shows four classes of fit. The maxi-

means that one single

mum effective tooth thickness of the external

GO ring gauge can be used for checking

splineand

fnur classes of fit. The NooO gauges have to

the minimum effective space width

of the internal spline are equal to nominal. The

machining

tolerance

and effective

clearance

be unique to each class of fit.

The only disadvantage of this system is that.

are different for each class of fit, resulting in

while a fit is guaranteed (provided the spline is

Class 4 being the closest fit and Class 7 being

accepted by the gauges), the type of fit is not.

the loosest

For example,

According

to this standard,

theallowable

if you were designing compo-

nents requirmgan extremely loose fit, Class 7

errors for a Class 5 spline (the most widely

tolerances would he used, However, :inthe case

used) are

of a poorly made component

with spline

maximum metal condition and with excessive

Total Index Variation

profile, index and lead error, it is possible for

Profile Variaticne
Lead Variation

0.038

+0.005/0.010

= 0.0]0

(for a spline

length of 25)
A described previously, i1 is not accepted

the gauges to accept the spline, providing, of

course, nominal. size .is not exceeded, The fit
obtained from this condition would bea tight
fit rather than the desired loose fit.

BS 35501..1960

that aU elements will occur in the same posi-

In this standard, the tooth proportions

similar

elemental

and lead errors)

or

are

to those in ANSI B92, 1-1970. The

allowable
Space
Width
Internal
Spline

on

For 30 teeih, 24/48 Df, 30 PA

0

T.able 1: eomparingl'oor

all

error

(profile,

are identical

spacing

to those of a

Class 5 spline in ANSI. The only pressure

angle referred

to. is 30. There are just two

classes of fit.. The internal spline

one tolerance

Nominal
1.661

band (identical

ANSI spline). For Clas

have only

to a CIa s 5

2, both extemaland

internal splines have the same tooth thicknes

I

and space width dimensions as an ANS[ Class

TOOlh
Thickness
of Extemal
Spline

5 spline. For Class 1 splines, the machining

tolerance and variation allowance are identical

to. Class 2. However, a clearance or deviation

allowance Cv is introduced, thus ensuring 111

24

GE AR He

H NOL.OGY

degree

of looseness

female

splines.

between

the male and

The clearance

Table 1. As shown. the tooth thickness dimen-

sions of the external splines are displaced or

adjusted by the clearance
concluded that. while

as

Table 3 - Comparing Two Classes of Fit in 88.3550.1963 30T, 24/48 DP, 30 PA

is shown in

Cv. It also can be

Class I

Class 2

Space
Width of
Internal
Spline

1.732

::J

3S50 has fewer fits, a

Class 1 spline does ensure clearance.

Nominal

DJN5480

1.661
1.623_

DfN 5480 has been in existence for many

years and is predominantly

used in Germany,

but also is used in many other countries ..It is a

metric module standard and was therefore con-

of External
Spline
III

for ISO 4156. Its popularity is

sidered a basis

due to its 1) being a metric standard, and 2)

offering

::JI

Tooth
Thickness

a vast diversity

* Min. possible

B92 ..1-1970 and BS 3550 do not. In the past, if

to achieve

fit. DIN

a certain

S480 was the only metric standard available to

enable you to achieve this (the only other metric standards

that were available

tion allowance and therefore were unsuitable),

consideredlas

the outside

diameter

ensure

i.e., DIN considers

that the

series of whole numbers. To

this and to enable common

Class II H/h

Clas 9 H/h

Class 7' H/Il

ominal

2.045

Tooth
Thickness
of External

--'1
udI-

I.
".01
I

Spline

m=O.080 A= 0.045 m =0.040'-=0.023

I.Actual

.=0.020 A= 0.012 m =0.014 A =0.00

I

Effective

Corrected

tion basis difficult; Having said this, these days

with modern

x circular pitch) to occur always on the pitch-

measuring machines. which have eliminated a

circle diameter, but sometimes on a corrected

great deal of human error, it is now possible to

pitch circle, either larger 'or smaller. This cor-

manufacture components that are accurate and

rection

to the

consistent, so perhaps there is nothing wrong

sizes are

in setting limits that were at one time consid-

does not cause any problem

as all corrected

specified in the standard.

computerized

machinery

and

ered impossible. Table 4 shows a small selec-

To calculate the variarjonallowance,

of the tooth thickness machining

60,%

tolerance is

tion of fits from this tandard.

ISO 4156/ANSI B92.2M-l980

used. This means that profile, spacing and lead

This new standard also allows engineers to

errors are not specified .. This can be a disad-

select various types of fits; i.e., sliding, close

vantage if a splined component manufacturer

sliding, clearance

wishes to check these errors on conventional or

the splines are not outside-diameter-based,

computerized

tooth profiles are not corrected, and the resul-

universal

gear

checking

and interference.

Because
the

machines. In the writer's opinion, DIN 5480, is

tant outside diameters are not whole numbers

an excellent

and, consequently,

standard,

Class 6 H/Il

Spline

possible for the nominal tooth thickness (0..5

but the vast choice of

fits must cause confusion

Effective

Internal

cutting

sary to correct the tooth profile; i..

e., it is not

designer,

Space
Width of

it is neces-

tools to be used in manufacture,

spline

Actual

Table 4 Comparing HIli Fits in DIN 5480 30T, 1.0 Mod.. 30PA

is

outside diameter of a spline should conform

toa preferential

effective clearance = 0.038

A. = 0.037

2.090*

a basis rather than the pitch-

circle diameter;

m = 0.033

were older

standards that did not lake into account variaIn this standard,

A = 0.037

of fits that ANSI

you wanted to produce a metric spline, andit

was necessary

= 0.033

10

the inexperienced,

do not conform to a pre-

ferred series. This, however, does not present

In addition, the machining tolerances for some

any problem, because a suitable tooth number

of the closer fits are extremely

can be selected for a particular application,

variation allowances consequently

which must make manufacturing

tight and the

minuscule,
on a produc-

The nominal tooth thickness/space

width is

calculated in tile same way as ANSI H92.1-1970

SEPTEMBERIOCTOBER

199'

25,

rvrod.,

Class 6 HJh

Class 7 :H1b

allowance A should becalculated

Class '5 H~h

It rates that the variation

manufacturer.

Table 5 . Compariag Four Clru;.ws or fl. Hlh

in I. 041561 NS! B92.lM - 1980.30T. 1!0
30 PA
Clas 4 HJh

using the fol-

lowing expression:

Space
Widlhof

lnternal
Spline

The expression

is different bam that u ed in

ANSI B92.]-]970

ominal

1.571

and BS 3550-]960,

which

states that variation allowance .is calculated as

follows:

Toolh

;l,. = 0.6 [lndex Variation + 2 (Posiuve Profile Variution] +

Thickness

of External
Spline

Lead Variat ion]

m .. 0.090"

.. 0.051 m =0.054 ./..= 0.034 m =0.033)..=0.024

m = 0.019'Ji.= 0.016

The resul! of the expres ion specified in the

newer standard is a value thai is lessthan that

Effective

of the old standard, as can be seen fromthe

following example. A. 30-looth. 1.0 module,
30 PA Class 5 spline is used.
HIe

[-IJd

e Fp
3.55 ..[l + '9
Fp in mierometers = 3.55 ~47.124 + 9
Fp in micmmeters e 33.370
Fpin millimeters :::0.033

Total Index Variarion

Fp in mierorneters

E basic
" 0.5 1tIn'""I--'I ......--1~
S, basic

."

where the length of arc L :: 0.5 x Pilch Circle

ircumference.

Zero

With effective
interference

Without
minimum
effective

Total Profile Variation = FI

Flin micrometers = 2.5 Ij>f + 1.6
F/in micrometers '" 2.5(1.375) + 16
Ff in micrometers = 19.4375
Ffin millimeters = 0.0]9

Witb effective
looseness

where Tolerance Factor Ij>f= m + 0.0125 mZ.

and BS 3550; i.e.. 0.5 x circular pitch. There
are four basic classes of fit, referred to as 4H/h,
SHih, 6H/h and 7H/h .. The suffix. H refer

Total Lead Varianon

to

Fin Micrometers

(as

F in Micrometers

= I{E

described in ISO 286 limits and fits for plain

F :in Micrometerse

10

parts).

F i.n Millimeters = 0.0.0

the internal. spline and h to theexternal

Iff

+5

+5

Table 5 shows the four HIb fits and can be

compared wi,tll the four ANSI fits in Table 2.
As can be seen, the fits obtained for Cia ses 4,
5 and 6 in ISO 4156 are significantly

closer

where g isthe spline length in millimeters.

Sub titutingthe

above values into the new

expression in ISO 4156:

than those from the older A_NS] standard (particularly Classes 4 and 5). The machining tol-

= 0.6 J

(0.033)2 + (0.019)2 +(0.0 I0)2 = 0.024.

erances for both standards are approximately

equal.consequently,

the closer fits obtained in

Referring to Table 2. we ee thai the allowable

ISO 4156 are due to a reduction in. the varia-

errors of index, profile and lead for an ANSI

lion allowance

A.

B92.1-1970 spline are reasonably comparable to

As slated earlier, Section 8. ~ of the tandard

has wide-ranging

26

GEAR

TECHNOlOGV

implications

for the spline

the above (it must. of course, be appreciated that

the above value for

FI i

total and should there-

fore be interpreted as
The variation

0.0095).

allowance.

however,

for an

ISO 4156 spline is less til an that for an ANSI

B92. I -1970

spline;

consequently,

when

machining splines to the later standard, greater

care and more rigorous quality control have to
be applied,
The above demonstrates that by using ISO
4156, it is possible to maintain closer fits than
those obtained from the older imperial standards.

Furtb.er Fit Variations

By using ISO 4156. it is possible to achieve
fits other than those shown in Table 5; i.e., the
NOOOP\u~

Hih fits. This is made possible by introducing

further fit classes to the external spline while

Gauge Tolerance

maintaining the H fits on the internal spline (it

able by spline manufacturers,

is beneficial to adjust external splines. as these

major diameter easier to control than the tooth

are usually

thickness.

machined,

while many internal

who find the

Comparison of Gauging Prneedure

plinesare produced by broaching).

These additional fits are designated f, e, d,

All. the standards that have been used for

js and k, and are obtained by applying a devia-

comparison have a section that clearly defines

tion allowance

the method of tolerancing

(Cv) to the shaft-tooth

thick-

for gauging. Note

ness. The deviation allowances to guarantee

the following points upon which all standards

clearance for a 30-tooth, 1.0' module spline are

are in agreement:

as below:

An internal spline should at the very [east

= --"0.020

be checked with

e = -0.040

L a full-form GO plug gauge designed

d = --"0.065

check "min. effective" space width,

The clearance becomes progressively

from

f to. d

[Q

larger

and ensures loosene s. If interfer-

enee is required, it is possible to employ fits k

2. a sector-type NOGO plug gauge designed

to check "max. actual." space width,
3. an additional

NOGO plug

composite

or js- Fit k guarantees interference, and with js

gauge to check "max. effective"

it is most probable that interference will exist.

(this is very occasionally required if it is neces-

Spline tooth thickness values for the two fits

sary to restrict the effective size).

can be determined

by referring

to Table

6,

that when using

fits /. e or d, the shaft major diameter

and

minor diameter also should be adjusted by a

value equal to the deviation allowance divided
by the tangent of the pressure angle.

major-diameter-fitting

standards
splines

an advantage in maintaining
this instance,

2. a sector-type NOOO ring gauge designed to check "min. actual" tooth thickness,
composite

NOGO

ring

gauge designed to check "min. effective" tooth

also contained

thickness where it is necessary to restrict the

effective size.

concentricity;

in

the spline flanks were used as

The posiuon of the gauge tolerance

and

wear allowances is the place where the various

standards do. net agree. As can be seen from

The fact that ISO 4156 does not contain

major-fitting

check "max. effective" tooth thickness,

which could be

drivers only,

greater

1. a full-form GO ring gauge designed to.

3 .. an additional

ISO 4156 specifies only side-fitting splines.

The older imperial

spline should at the very least

be checked with

which also. shows the clearance fits.

It should be remembered

Anexternal

space w:idth

splines should be offset by the

accuracies

achievable

by using the

Table 7. ANSI B92.1-1970

places the gauge

tolerance within the part tolerance,

BS 3550

puts the gauge tolerance outside the pan toler-

mN 5480 and ISO 4156, the

dose tooth thickness tolerances recommended

ance, and in

in this standard. This, however, may be debar-

gauge tolerances are bilateral .

SEPT~MBER/OCToeER

'994

27'