PRECISION IN MARINE-GEAR MANUFACTURE: THE MODERN APPROACH

7/26/2019 PRECISION IN MARINE-GEAR MANUFACTURE: THE MODERN APPROACH

1/10

153

Paper

19

PRECISION

IN

MARINE-GEAR MANUFACTURE:

THE MO DERN APPROACH

A. Hadcroft*

In previous years, a high standard of accuracy in hobbing machine correction and gear production was

achieved by labori6us and painstaking work. Today, accurate gears are the rule rather than the exception,

and manufacturing plant is usually fully utilized. Serious interruption to manufacturing schedules can no

longer be tolerated for maintenance of the established product quality. The paper seeks to examine some of

the more modem procedures and equipment, and to mention those areas in particular which require constant

vigilance and attention to ensure that a gear manufacturing plant is maintained at high standards of quality

and efficiency.

INTRODUCTION

AT THE 1958 CONFEREXCE, Timms 1)t gave a very able

review of the instruments and methods of measurement

available at that time for dealing with large turbine gears

and hobbing machines. Considerable progress has been

made since that date, and much of this has been reflected

in our experience at Manchester.

Gear cutting at the works at Trafford Park dates from

1916, and throughout these years the company has been

alive to the necessity for continual improvements in

accuracy. Newton in 1949 2) gave details of the elaborate

and effective means employed to correct the older machines

in the thirties and forties by means of cams to cancel out

transmission errors.

Having created satisfactory indexing by these methods,

reproduction of a new master wheel was possible, and

cams could then be dispensed with. It is of interest to note

that the first 8-ft worm wheel corrected in 1946 is still

today within the

A

grade quality of B.S. 1498:1954

specification for master worm wheels.

Th e painstaking effort essential to the old methods of

correction often resulted in a better standard of gear, but

the maintenance of this standard was costly. The lack of

structural rigidity in the machine tools, coupled with poor

foundations, gave little permanence to the alignments.

In

the late 1940s progress along the then established lines had

changed from development to a burden of perpetual

The M S . of this paper was received at the Institution on 20th April

I970 and accepted for publication on 12th June 1970. 43

Advanced Manufacturing Engineer E.E.-A.E. I . Turbine Gener-

ators Ltd TraffordPark Manchester.

t References are given in Appendix

19.1.

maintenance. At that time the manufacturers of large

hobbing machines in the U.K. were well behind those in

the U.S.A. in ability, both as regards accuracy and machine

design. The existing U.K. plant, most of it of 191418

vintage, was in urgent need

of

replacement. To continue

to renovate and, where possible, improve the transmission,

and at the same time accept the lack of rigidity inherent in

most existing machine tools, were now accepted as

uneconomic operations.

In

some instances attempts were

made to stiffen up existing machine structures, but in the

main this procedure was

o n l y

partially successful.

The issue in 1948 of

B.S.

1498, Gear hobbing machines

for turbine and similar drives followed in 1951 by B.S.

1807 covering turbine gears, comprised a challenge which

was taken up by two U.K. machine tool manufacturers.

These specifications also indicated the existence of a

potential market. At Trafford Park the potential marine

market and the definite interest of the machine tool

builders gave encouragement to a new start being made.

These conditions created an opportunity to install new

plant in a temperature-controlled environment, where

isolated foundations could be provided and rigidity

designed into the machine tools. Significant advances were

being made by gear hobbing and grinding machine manu-

facturers. These advances, coupled with developments in

electronics and metrology, marked the beginning of a new

era.

An expanding market would demand gear manufacture

in quantity, with little time for the ingenious but laborious

maintenance and correction methods of the previous

decades. The near-laboratory conditions under which

Ptoc

lnstn Mech Engrs

1969-70

Vol

184 Pt 3 0

at UNIV NEBRASKA LIBRARIES on October 16, 2015pcp.sagepub.comDownloaded from
http://pcp.sagepub.com/http://pcp.sagepub.com/http://pcp.sagepub.com/


2/10

154 A.HADCROFT

gears were then cut h ad to give way to work shop processes

using plant capable of economic rates of production.

Th e criteria by which the quality of the product is finally

judged are the contact marking between th e m ating gears

and their uniformity of relative motion. The former is a

function

of

the machine tool alignment, the feed screw

accuracy, the hob accuracy and its mounting, and strict

control of the ambient temperature.

It

is also controlled

by the post-hobbing process, whether

it

be shaving or

grinding. U niformity of relative rotation

is

a reflection of

the inherent inaccuracy in the kinematic link between the

work and the hob during gear cutting.

T o meet these conditions the gear shop in Manchester

was re built on a new site, and a fully automated tempera-

ture-control plant was installed, T h e shop Fig.

19.1)

accommodates the precision machine tools, together with

th e associated metrology equipm ent n a common area

144 ft long by 52 ft between th e crane stanchions. A 40-ton

crane is provided, which runs on a track extending along

the full length of the shop. T h e building is sited approxi-

mately north-south, and windows constructed from

sealed glass bricks are provided at a high level on all but

the south side, thus shielding the equipment from direct

sunlight. Separate air locks are provided for th e admission

of work and for pedestrian traffic.

TEMPERATURE CONTROL PLANT

I n preference to a plant h oused in separate cubicles, each

with its own control system, the machine tools and

metrology equipment are established in an open area

giving easy access for work handling an d supervision. Th e

temperature-control plant provides six changes of air per

hour and maintains temperatures within a total range of

2 degF at constant relative humidity.

Conditioned air is introduc ed throug h th e ceiling grills

and extracted at

floor

level. After fresh air has been add ed,

it

is cleaned by being passed throu gh rotary viscous filters

and cooled through direct-expansion coolers. The air is

Fig.

19.1.

Temperature-controlled gear shop

at

Manchester

Proc

lns tn Mech Engrs

1969-70

Vol

184 Pt 3



3/10

PRECISION IN MARINE-GEAR MANUFAC TURE: TH E MODERN APPROACH

155

t

Layout of complete plant gear shop.

Air flow diagram-gear shop.

Fig.

19.2a.

Temperature-control plant

LE Pre heat

Fresh

air

Schematic layout, temperature-control plant-gear shop. View on arrow

X

in Fig.

1 9 . 2 ~ ~ )

A Extraction fan.

B Control dampers.

C Filters.

D

Air conditioner.

E Plenum fan.

Fig. 19 2b Temperature-control plant

then passed through a preheater and air-conditioner, and

delivered by a plenum fan at the required wet and dry

bulb temperatures (Figs 1 9 . 2~ nd 19.26).

The plant has operated satisfactorily over a number

of

years, needing little attention other than the annual

inspection and overhaul.

MACHINE TOOL FOUNDATIONS

The gear manufacturing plant is housed in close proximity

to other heavy engineering activities. To preserve satis-

factory machine tool alignments under these conditions,

unusual foundations are required. Furthermore, distortion

in the machine tool scantlings under widely varying loads

must be avoided. To achieve this aim, the gear-cutting and

grinding machines are bolted to heavy fabricated steel bed-

plates that are themselves supported at three points on

resilient mountings (Fig. 19.3).

The bed-plates carry adjustable wedges on the top face

supporting the underside of the machine tool. The stiffness

of the fabrication is designed to permit adjustment of the

machine s internal alignment by the supporting wedges.

After final adjustment, the bed and machine are securely

bolted together to form a single unit that is free-standing

on three small areas. Before this type of foundation was

Proc

lnstn

Mech

Engrs

1969-70

Val

184 Pt 3 0



4/10

156

A.

HADCROFT

Support

oreas

Resilient mountings bolts

Fig. 19.3. Typical thr ee-point mounting

installed, the principle was proven on a pinion-hobbing

machine that was sited adjacent to a stanchion base sup-

porting a heavy crane track. Until the new foundation was

installed it was not possible to cut precision gears.

The position of the support areas for the foundation

minimizes the deflection caused by the varying work loads

and the changing position

of

the heavy column. Two

supports are placed directly beneath the work-table and

one under the column bed, the position of the latter giving

minimum bending moment at the extreme ends of the

column travel. The supports under the work-table being

directly beneath the area of changing load, no bending is

induced into the structure. Under these conditions the

whole machine is free to move in space but will remain,

together with the work-piece, as one integral unit.

The resilient mounting at each support area is hard

rubber, bonded between two steel plates

; ach is installed

in a manner that permits removal for inspection and

replacement. A common type of mounting was used for all

machines, but the size of the mounting and the type of

rubber varied accordingto the load. The natural frequency

of all the units is about

500

cycles/min. These mountings

and foundations have operated satisfactorily over the past

14 years; none has been replaced.

The continuity of satisfactory alignment is reflected in

the match between the products from the various machines.

Departure is indicated by the extended post-hobbing

work that is needed to attain a satisfactory match over the

tooth face.

The gear-shaving process has in some measure pro-

vided a method by which mismatch can be corrected, but

anything more than minor correction is unsatisfactory.

Furthermore, correction by gear shaving is something of

an art rather than a science, the time and cost for the work

being difficult to estimate. Hence, there is both a functional

and an economic advantage in these foundations; they

keep manufacturing times to

a

minimum by maintaining

machine tool alignment over long periods. In addition,

continuity in correct gear cutting provides reliable

knowledge of process times, which is essential in the

preparation of manufacturing schedules and cost estimates.

MACHINE TOOL ALIGNMENT

Whilst alignments, once set, are maintained over long

periods by the steel foundations, initial settings have to be

made and from time to time adjustments are required.

Rapid and reliable methods of checking and adjusting

alignments are needed to reduce to a minimum the outage

time of major machine tools. I n the late 1950s, research

work commenced on the development of an optical

replacement for the test pillar. By 1960

such an instru-

ment, later known as the reflecting Rodolite, had reached

a satisfactory stage in development. In its fully developed

state

it

included the mercury Rodolite; both versions have

been fully described by Dyson and Tillen 3).

These instruments will define a straight line within

0.0025 nm 0.0001 in) over 6 m 20 ft). When they are

being used, instead of the conventional test pillar, a hob-

bing machine can be prepared for examination in 5-6 h,

provided, of course, that all the supporting equipment is

available.

The two versions of the Rodolite are complementary.

They consist

of

two target gratings with a common sight-

ing head. The reflecting target is a 50 mm 2 in) diameter

grating ruled to 400 lines/in with a fixed reflector; the

Proc

lnstn

Mech Engrs

1969-70

Vol 184

Pt

3 0



5/10

157

PRECISION

IN

MARINE-GEAR MANUFACTURE: T HE MODERN

APPROACH

mercury target is similar, but

it

has a pool of mercury

reflector. Thus, the reflecting Rodolite will establish the

relation of a moving point relative to the axis of a rotating

table or chuck, while the mercury version will establish

the relation to a vertical axis.

It

is known that the use of a long outrigger from the hob

saddle to the table centre can result in readings which

contain exaggerated errors resulting from cross-wind in

the column shears. However, use of both targets in turn at

the centre of the hobbing machine table, sighted from the

same outrigger, will indicate the departure of the table

axis from the vertical. If the telescope is then transferred

to the hob position and sighted on the mercury target at

the foot of the column, a corresponding departure of hob

motion from a vertical line can be measured.

The reflecting instrument has a distinct advantage in

checking and setting horizontal pinion machine align-

ments, because the traditional pillar test was always

unsatisfactory when extended over an appreciable hob

travel.

Th e reliability and high accuracy of these instruments,

together with their rapid application, appreciably assists

in shortening the time required to make adjustments in

machine tool alignments.

FEED

SCREWS

Control and adjustment of feed screw accuracy are features

normally outside the capability of the gear manufacturer.

Nevertheless, at the present time reliable screws can be

obtained from firms specializing n this type

of

equipment.

Accurate examination n situ is readily carried out using

National Engineering Laboratory (N.E.L.) linear and

circular gratings. These tests show the cyclic error arising

from the thrust bearing in addition to the cyclic and linear

errors

of

the screw and nut assembly.

HOBS AND SHAVING CUTTERS

It has been recognized for many years that gear-cutting

hobs for turbine drives require a higher standard of

accuracy than that specified for industrial gearing.

Previously, hobs were manufactured by specialist firms to

the specification

MOY G 2

compiled by the National

Physical Laboratory. In

1959,

B.S.

2062,

covering gear

hobs for general purposes, was issued; this was extended

in 1960 to cover hobs for turbine drives. Whilst this

specification goes some way towards meeting the require-

ments, there are anomalies that require clarification. At

the present time these are overcome by the goodwill

existing between the hob manufacturer and the user.

The gear-shaving process now has common acceptance

as a method of refining the tooth profile and making minor

corrections to the helix angles. In use, the accuracy of the

shaving cutter profile is reflected in the tooth contact

marking of the shaved gears. The quality of these cutting

tools is specified by

B.S.

2007.However, in many instances

the ultimate refinement

of

the profile has to be defined on

Proc

lnstn

Mech

Engrs

1969-70

a cut and try basis to suit a particluar gear; the final

correction is assessed by the contact marking between the

mating gears.

GEAR PROFILE MEASUREMENT

The generation of a tooth profile cannot be adjusted whilst

cutting on a hobbing machine. This is not significant

because proven plant usually ensures continuity of satis-

factory products. With a gear-grinding machine this is not

so; adjustments can be made as the work proceeds, the

profile accuracy being dependent

on

the settings made by

the machine operator. Again, the profile accuracy is finally

judged by the contact marking between the mating gears;

but it is necessary to remove the gears from the machine to

make the check. Resetting is usually necessary for correc-

tion and final refinements. This time-consuming practice,

particularly when grinding gear wheels, amply demon-

strated the need for a portable profile-measuring instru-

ment of the autographic type that could be used on a

gear-grinding machine.

Development of involute measuring equipment was

started at the A.E.I. Research Laboratory at Aldermaston

in 1956, being aimed specifically at portable equipment.

The principle on which the development work proceeded

was the osculating circle at the pitch line. Resin casts were

made of the tooth profile; these were moved through an

arc of known radius past a measuring head that indicated

the departure of the profile from the true arc. The readings

were compared with calculated values. In the next phase

of development the cast was eliminated by arranging for

the measuring stylus to move over the gear profile

4).

In this manner a measurement was made of the devi-

ation of the profile in the normal plane from a circular arc

having a centre on the base cylinder and a radius equal to

the curvature of the theoretical profile at the pitch point.

The development instrument is a light but rigid tubular

frame that can be positioned normal to the profile by two

ball feet, locating in the tooth spaces, and a knife edge

resting on the outside diameter of the gear. A parallelogram

mechanism set to rotate a stylus through a predetermined

arc having a centre on the base cylinder (Fig. 19.4) is

driven up the tooth profile by a synchronous motor. The

stylus and the drive motor are electrically connected to a

recorder, which is arranged to give a trace of the profile

deviation from the arc. The full calculation for the

theoretical trace is complex, but this is normally done by

computer. Where the instrument is used over a particular

range of gearing, and for all practical purposes, a simplified

calculation is sufficient.

Before this instrument was accepted as a piece of work-

shop equipment, tests were made to prove its agreement

with other established profile-measuring machines. The

plotted readings from the Vinco portable ordinate

measuring machines were also compared with readings

from the profilometer; close agreement was found with all

the equipment. Finally, a suspension harness was provided

to support the instrument on the vertical face of a gear

V o l 1 8 4

Pt

3 0



6/10

158

A. HADCROFT

Involute

radius

of

curvature

of

involute at

the

pitch

ir le

4

Radius

o f arc equal to

C O M P A R I S O N OF

INVOLUTE

WIT

REFERENCE CIRCLE

THEORETICAL T R C E

y=calculoted

points o f profile

R o o t

I

Pitch

point

xDP

i Tip

ZXDP

Fig. 19.4. Trace

from

prof i lometer

wheel whilst mounted on a grinding machine, for which

purpose the instrument was originally devised. All the

development work was carried out using the laboratory

instrument, which was finally replaced by a new workshop

instrument (Fig. 19.5).

This work had established for the first time a portable

instrument that will autographically record the error

occurring in a gear tooth profile. By its use the profile can

be measured and corrected where necessary whilst the

gear is still mounted on a grinding machine. This facility

makes a considerable contribution towards reduced cost

and handling time, with an improvement in the overall

accuracy of the product. Th e instrument finds little use on

hobbed gears, other than when a finishing process such as

gear shaving is involved. I n this case it can be used as an

investigational instrument to help resolve problems, and

where necessary to provide inspection records that would

otherwise be unobtainable.

Similarly, in monitoring gear performance in service it

can be readily adapted to measure the tooth profiles of

installed gears. Very little preparation is needed other than

the removal of some of the covers from the upper part of

the gearcase. In this manner, a periodic inspection of the

tooth profiles can be carried out, n situ in a relatively

short time.

PITCH ACCURACY

In the early 1960s the N.E.L. was developing a circular

grating unit designed for the continuous measurement of

kinematic errors between rotating units. The equipment

was portable and capable of measuring angular phase errors

to an accuracyof

f0-5

econds

of

arc. I t was readily adapt-

able to the measurement of worm-worm wheel errors in

gear-hobbing machines and has since been fully developed

to

monitor the kinematic link between the hob and the work.

This equipment has been fully described by Smith and

McGregor 5 ) .

For correction work it was first applied to a 3.8 m

(150 in) hobbing machine at Manchester (Fig. 19.6) that

some years previously had been certified to conform to

B.S.

1498: 1954, Grade A standard. Th e machine accuracy

had deteriorated and the worm wheel errors as measured

from a spur test gear showed an accumulative pitch of

0.09

mm

0.0036

in) over a

4-2

m

166

in) span. The

predominant error was sinusoidal but there were some

superimposed short span errors. The one direction tested

indicated that the worm wheel axis needed to be moved

and some correction

of

the worm wheel teeth was

required.

Although the portable grating unit at the time was a

laboratory rig, it was decided, in conjunction with the

N.E.L.,

to use the equipment to monitor the correction

work. The hobbing machine has a built-in provision for

adjustment of the worm wheel with respect to the axis.

Its position is initially controlled by eight radial screws

and finally secured by bolts and dowels.

The first readings taken by the portable grating unit

confirmed those taken from the spur test gear; these are

shown in Fig. 19.7. Before dealing with the shorter span

errors the predominant sinusoidal error was adjusted by

moving the worm wheel centre. The main grating was

mounted above the table centre, with sufficient space under

it to allow an operator to work within the hollow table

journal. It was then possible, whilst the machine was

running, to shift the worm wheel centre and monitor the

result by the grating unit. Control was excellent and the

adjustments were positive; the results of the first move are

also shown in Fig. 19.7.

Proc lnstn Mech Engrs

1969-70

Vol

184

Pt

3 0



7/10

PRECISION IN MARINE-GEAR MANUFACTURE: THE MODERN APPROACH

Fig.

19.5.

Goulder-A.E.I. profilometer

Fig.

19.6.

Portable grating uni t mounted on 150-in machine

Proc lnstn Mech

Engrs

1969-70

59

V o l 1 8 4

Pt

3 0



8/10

160 A . HADCROFT

Machine: 150-in wheel Hobben.

Worm:

Wheel:

720 teeth.

0.500-in

pitch.

Fine

pitch

engaged. Single start.

Linear

pitch: 0.5025-0.4975

in.

Fig. 19.7. Errors recorded by po rtable grating uni t on 150-in hobbing machines before and after correcti on of table



9/10

PRECISION IN MARINE-GEAR MANUFACTURE:

THE MODERN APPROACH

161

In a comparatively short time the anticlockwise error

was reduced from 14 o 6 seconds of arc, and clockwise from

12

to

44

seconds of arc. Superimposed on the cumulative

error were a number of short span errors which required

correction at the tooth contact. Where correction could take

advantageofmetal on the flanks the procedure was no more

than a careful scraping at the right spot. This was defined by

the use of hard lacquer painted on the flanks, followed by a

period of running that readily showed the contact marking.

Where the error was due to metal shortage a study deter-

mined the best position for the worm wheel centre. This

was a compromise that would provide metal on the flanks

yet require minimum hand scraping to correct the overall

error that was consequential to a move of the axis.

The precise location of the teeth to be corrected was

determined by pieces of adhesive tape adhering to the

flanks of the worm wheel teeth. On going through the mesh

with the worm these teeth were shown by peaks on the

chart record. The tooth to be adjusted was easily found by

counting from these markers; hus, there was a positive

link between the record and the hardware.

The work continued for approximately eight weeks.

Much of this period was running time to create tooth-

contact marking, but there were a number of moves

of

the

worm-wheel centre to help diminish the hand-correction

work on the short-span errors. As the work continued and

the errors were reduced, the increasing cost for diminish-

ing return became evident. The work was stopped when

the fundamental component of the error had been reduced

to 2-7 seconds of arc in the clockwise direction, and

3 seconds of arc in the anticlockwise direction. These figures

brought the machine within the B.S.

1498: 1954,

Grade

A

standard. The final records are shown in Fig.

19.8.

This grating equipment also indicates the cyclic errors

due to the worm, the table drive gears, and the hob drive

gears. They are superimposed on the record of the funda-

mental component and distinguished by their various

frequencies. In this instance, however, no work was needed

on the other gear trains.

GEAR-GRINDING

MACHINES

The success with the portable grating unit, described

above, led to investigation of the equipment for use on

gear-grinding machines.

On some machines, such as the Maag SHS.150, the

accuracy of indexing

is

linked with the generation of

the involute profile. This is generated by the rotation of the

table combined with a translatory movement of the slide

carrying the table. This movement is derived from a lead

screw coupled to the table drive through a gear train. Th e

worm-worm wheel drive to the table can be measured in

much the same way as the hobbing machine. However, to

check the kinematic link that controls the generation, a

linear grating would be needed to monitor the table

translation and relate it to table rotation. The factor is

introduced into this relationship owing to the pitch of the

lead screw, which is 7r mm. The table is driven by a single

Proc lnstn

Mech

Engrs 1969-70

start worm and a 216-tooth worm wheel. Th e relationship

between the screw and table is expressed by

where L is the length of translatory movement, 0 the

rotation of the table in degrees,

P p

the pitch change gear

ratio, and

P

the number of teeth-change gear ratio.

It

was not possible to provide a calibrated linear grating

that was compatible with these conditions, therefore a

helium-neon laser unit was used in conjunction with the

grating unit. This work has been fully reported by

Smith

6).

The errors found in the machine were not significant

and no work has been carried out in the way of improve-

ments to the kinematic link between table rotation and

translation.

NITRIDED

GEARS

Considerations

of

economic production tend to favour

a

procedure that eliminates the grinding operation and

completes the gear before exposing it to the hardening

process. The nitriding process has proved to be satisfactory

in this respect and a large number of gears have been

manufactured in this way. Experience to date covers a

wide variety; at the large end of the range are pinions up

to 1.8 m 6 ft) in overall length and 2 ton in weight, and

wheels

1.3

m 50 in) in diameter and 35 cm

14

in) face

width. With conscientious stress relieving before finish

cutting, and correct matching of the elements before

nitriding, no significant change in contact marking has

been noticed.

To prove this, a sensitive method of measuring tooth

contact is needed; this can be satisfied by applying tool

makers marking dye to the teeth of one of the meshing

units, then driving one by the other. Some of the dye is

transferred; but, although this is of interest, the criterion

is the hard metallic marking that is evident on the dye-

covered tooth flanks.

A

careful study will determine the

precise areas of tooth contact to a tolerance of the order of

0.005 mm (2

x

in). This is readily proved by a con-

trolled change in alignment between the two gears.

The final marking is recorded by the use of transparent

adhesive tape, which, after application to the areas carry-

ing the marking dye, can be peeled away with the contact

pattern on the adhesive side of the tape. When fixed with

the adhesive side to stiff paper, a clear and permanent

record of the tooth contact is formed.

THE

SEVENTIES

Gear accuracy (customer requirements)

The accuracy standard currently accepted is tha t defined

by

B.S.

1807

in its various grades, but for naval installa-

tions certain tolerance bands are usually reduced to 60 per

cent.

At the present time there appears to be little demand

from the customer for anything better, and design has still

to justify existing standards. The real need is for the

V o l 1 8 4

Pt 3 0



10/10

162 A. HADCROFT

correlation of accuracy, loading, and noise parameters to

provide the design criteria for the machinery of the next

decade.

The revised B.S.

436

which is intended eventually to

replace B.S. 1807, is based

on

internationally accepted

tolerance bands. These have yet to be related to design

requirements in contrast to empirical figures that have

their origin in the most optimistic achievement of the gear

manufacturer.

Measuring instrumentation sophistication

The overall achievement in metrology in the last decade

has eliminated the need to cut test gears to prove the

accuracy of gear-generating machines. For large units this

was an expensive and time-consuming procedure that

recurred after each period of repair or modification. The

continuous monitoring of accuracy is currently carried out

by measuring the product rather than the machine tool.

Modern electronic equipment, such as that designed by

Hofler and the N.E.L., will readily measure the short span

and cumulative pitch errors whilst the gear remains

on

the

machine 7) 8). Unfortunately, when these measurements

are carried out on the hobbing machine they have no

reference to the axis of the installed gear. After removal

from the machine the meshing check serves to prove the

alignments and matching profiles.

Ultimately, it is hoped that some form of seismic or

grating unit will be developed to provide a slow-speed

check on the relative velocity of two meshing elements.

When applied to a meshing frame, this could then be

considered a composite check on the mating elements. By

their various frequencies the tooth contact, profile, pitch,

and axis alignment could all be identified from an auto-

graphic record. Since the source of the composite error

could arise in a number of machines, investigational work

could be required, using the equipment of the last decade,

which would be adequate for the purpose.

Hobbing

machine

users

Being concerned with gear-hobbing machines, the author

regrets that once again the U.K. has returned to the same

position as that which prevailed in the late 1940s-early 50s.

In Europe, other than in Western Germany, there is no

manufacturer of large high-grade turbine gear-hobbing

machines; nor, so far as is known at the time of writing, is

there an established manufacturer of this type of machine

tool in the U.S.A. In these circumstances the U.K. user is

forced to seek an expensive source of help and machine

tool supply outside the U.K., or once again resort to self-

help, as pre-1950. This time he is better equipped

technically and has better manufacturing facilities at his

disposal; but, unfortunately, he still lacks the design

experience of the machine tool builder with many machines

in service.

The machine tools already installed can be maintained

at their present accuracy standard and may even be

improved. Nevertheless, progress is not made by main-

taining that which exists; technical achievement comes

from the design and manufacture of new plant, and at the

moment we look in vain to

the U.K.

machine tool indusmy

for a lead in this respect.

Gear manufacturers

The author believes that the current need is for higher

production rates. I n general, it can be stated that accuracy

requirements can be met, but the metal removal rate of the

average large turbine gear hobber is usually less than

1

in3/min. In many instances the power available at the

main motor is

of

the order of 50hp, and an increase of

300

per cent

in

metal removal rate does not seem to be

unreasonable. Development in precision gear-cutting

tools and in cutting tool materials is needed; again, the

lead at the moment is being taken by firms outside the

U.K. The benefits which will accrue are higher produc-

tion rates and, above all, a reduction in capital investment

for plant expansion to meet heavier production

programmes.

CONCLUSIONS

No matter what care is taken to produce accurate gears,

and no matter what records are produced to support this,

the criterion is a satisfied customer.

Th e procedures and equipment described in the paper,

whilst not sufficient in themselves to satisfy this end, have

made a great contribution to the maintenance of quality

standatds at the authors firm in the context of a full

production schedule.

Future requirements are for a realistic assessment of

accuracy requirements and for redesign of machines and

tools to maintain this accuracy with faster rates of pro-

duction.

ACKNOWLEDGEMENT

Thanks are due to English Electric-A.E.I. Turbine

Generators Ltd for permission to publish the information

given in this paper.

AP P E NDI X

19.1

R E F E R E N C E S

I) TIMMS,.

2) NEWTON,. M .

3)

DYSON,

.

and TILLEN,

.

J.

4)

CROOK,

.

W.

and COLYER,

.

F.

Measurement of large gears, Proc.

Znt

Conf on

Gearing 1958, 269

Insm

Mech. Engrs, London).

On the accuracy of gear hobbing machine

tables, Proc. Znstn mech. Engrs 1948

161,

10.

The rodolite: a new optical

device, A.E.Z. Engng 1961 1 No. 10).

Improvements relating to

apparatus for measuring the profile

of

gear teeth, British

Patent Specification846404, 1958.

A

precision portable grating

unit, N . E . L . Repr 249 National Engineering Laboratory,

East Kilbride, Glasgow).

Some recent developments n gearing and machine

tool metrology, Sixth Int .

Round

Table Discussion 1967.

ully

automatic

merhods

of gear meaxurement

1968 14th May) British Gear Manufacturers Association).

Anew automatic

gear pitch comparator, N . E . L . Rept 1968 National Engin-

eering Laboratory, East Kilbride, Glasgow).

5) SMITH,P. and MCGREGOR,.

6) SMITH,

.

7) HOPKINS,

. R.

8 ) INGLIS,., RAFFERTY,.

S . and

SMITH,

P.

Ptoc

lnstn Mech Engrs

1969-70

Vol 184 Pr 3 0


PRECISION IN MARINE-GEAR MANUFACTURE: THE MODERN APPROACH

Documents