DESIGN OF A

HIGH-SPEED
CYLINDRICAL
GRINDING MACHINE
 ^11 Mi
iiHinii liiilliMI lm In lil II
Il lil Ml
lil III II II Hl II i; il i||||n| iliii II
L i i In II il II II u u Il UI IM II

o
CO •-'
fSi o
OD o
o *-n

DESIGN OF A
HIGH-SPEED CYLINDRICAL GRINDING MACHINE

PROEFONTWERP

TER VERKRIJGING VAN DE GRAAD VAN DOCTOR IN DE

TECHNISCHE WETENSCHAPPEN AAN DE TECHNISCHE
HOGESCHOOL DELFT, OP GEZAG VAN DE RECTOR
MAGNIFICUS PROF. DR. IR. H. VAN BEKKUM, VOOR EEN
COMMISSIE, AANGEWEZEN DOOR HET COLLEGE VAN
DEKANEN, TE VERDEDIGEN OP DONDERDAG
10 JUNI 1976 TE 16.00 UUR

DOOR

TONNY STORM

werktuigkundig ingenieur

geboren te Ede

J^Zl^^ /^-v;

BIBLIOTHEEK TU Delft
P 1798 1547

C 538280
Dit proefschrift is goedgekeurd door de promotor
PROF.IR. D.A.A. KOOLEN

.
CONTENTS
Page

L i s t of s3nTibols 1
Summary *+
1. Introduction 9
2. Basic starting points for the design 10
2.1. The product 10
2.2. The process 10
2.3. The machine 10
3. Basic calculations 12
3.1. Grinding power and forces 12
3.2. Speeds of grinding wheel and workpiece 13
3.3. The grinding wheel 17
4. The wheel head 19
4.1. The grinding spindle 19
H.2. The spindle drive 21
5. The hydraulic motor 22
5.1. Basic concept 22
5.2. Geometrical design 27
5.3. The manufacture 29
5. M-. Conclus ion 35
6. The construction of the machine 36
6.1. General design 36
6.2. Deformation of the machine structure 45
6.2.1. Method of calculation 45
6.2.2. Execution of the calculations 46
6.3. Design of the work heads 61
7. Drive of wheel head, slide and workpiece 62
8. Apparatus for dressing of the grinding wheel 69
8.1. Wheel wear and dressing conditions 69
8.2. Design of the dressing apparatus 71
9. The control of the grinding machine 73
9.1. Control of the grinding process 73
9.2, Control of the grinding wheel position 74
10. The power supply of the grinding machine 78
11, Secondary aspects of the design 81
11,1. Safety 81
11,1. Coolant-supply 82
12. Drawings of the machine 84
13. Conclusion 105
Appendix A Deformation of the grinding spindle 106
Appendix B Calculation of the hydrostatic bearings 109
Appendix C Contact of rotor circumference 113
Appendix D Leakage and mechanical losses in the
hydraulic motor 114
References 117
Curriculum Vitae 120
LIST OF SYMBOLS

Symbol Unit
2
area m

a depth of cut mm

a difference of maximum and minimum

rotor radius mm

length of spindle part between front

bearing and grinding wheel mm

infeed of dressing wheel

mm/rev
angle between the axes of grinding wheel degree
and workpiece

b width of the rotors of hydraulic motor mm

b length of spindle part between mm

spindle bearings

b width of bearing gap mm

b width of grinding wheel mm

stiffness N/mm

specific heat W/kg °C

w
D spindle diameter mm

d diameter of grinding wheel mm

s
d diameter of workpiece mm
w
d diameter of work head spindle mm
ws
AT temperature rise °C

e excentricity mm

e efficiency

dynamic viscosity kg/m.s

tangential dressing force N

dt
F' specific tangential dressing force N/mm
dt
radial dressing force N
dr
F' specific radial dressing force N/mm
dr
 Symbol Unit

tangential grinding force N

normal grinding force N

displacement of a point P caused mm

by deformation

angle, polar co-ordinate degree

angle between plane of the grinding wheel degree

and the workpiece axis

G grinding ratio

Y angle between radius vector and tangent degree

h height of gap mm
2
I moment of inertia kgm
i transmission ratio

L effective length of hydrostatic bearing mm

length of travel in x-direction mm

length of travel in y-direction mm

y
M torque Nm

M torque related to inertia forces Nm

torque at dressing wheel Nm

dw
torque related to grinding forces Nm

mass kg

number

r.p.m. of dressing wheel rev./min

dw
r.p.m. of grinding wheel rev./min

r.p.m. of workpiece rev,/min

w
angular velocity rad/s

angular acceleration rad/s

P grinding power kW

P' specific grinding power W/mm /s

P dressing power kW
dw

- 2
Symbol Unit

P pump power kW

friction power kW
w
P pressure N/m^

Q oil flow m /s

R radius, polar co-ordinate mm

P density kg/m

s feed mm/rev.

t acceleration time s

feed rate mm/s

speed of dressing head mm/s

dr
infeed rate mm/s

relative speed of bearing surfaces m/s

peripheral speed of the grinding wheel m/s

peripheral speed of the workpiece m/s

w specific wheel wear mm /mm.s

metal removal rate mm /s

specific metal removal rate mm 3/

/mm.s

- 3-
SUMMARY

The thesis describes the design of a high speed cylindrical grinding machine.
From research in a number of institutes it appeared that an increase of the
wheel speed offers the possibility to obtain a high metal removal rate at a
good surface quality and a high geometrical accuracy of the workpiece. Be-
sides the investigations concerning the cut-off grinding process at the Uni-
versity of Delft showed that specific metal removal rates up to 1600 mm3/mm.s
at a wheel speed of 100 m/s can be realized.
Starting from these facts a high speed cylindrical grinding machine is de-
signed capable of machining shafts with a maximum length of 750 mm and a
maximum diameter of 250 mm at a metal removal rate of 20.000 mm3/s and a
wheel speed of 150 m/s. The grinding process will be realized according to
the figure that shows the relative movements of the grinding wheel and the
workpiece.

U
Sahernatic view of workpieoe
and grinding wheel.

The first part of the design concerns the drive of the grinding wheel. It is
decided to drive the grinding spindle directly without any transmission. This
demand leads to the design of a special hydraulic motor with a maximum power
of 150 kW at a speed of 10.000 r.p.m.
The second part of the design applies to the machine structure. This structure
supports the grinding spindle and the workpiece. The main requirement for the
structure concerns the maximum deformation that is allowed with regard to the
accuracy of the workpiece. In order to calculate the deformation of the
structure a "finite element" program is applied.
The third part is related to the drives of the slide, the wheel head and the
workpiece and the control of the machine. The drive motors that are applied
are electrical stepping motors, which results in a relative simple control
system. This control system achieves in the first place the movements of the
grinding wheel relative to the workpiece according to the workpiece geometry.
Secondly the control system provides the correct conditions for the grinding
process.
Finally some secondary aspects of the design are discussed. These aspects con-
cern in the first place the dressing of the grinding wheel, which will be

4 -
realized by means of a diamond wheel. Secondly the safety of the machine in
the event of a wheel bursting is considered.
At last the supply of coolant both to the grinding zone and the dressing zone
is regarded.
The final design is shown in the drawings that form part of this thesis.
SAMENVATTING

Het proefontwerp heeft betrekking op een rondslijpmachine voor hoge snelheden.

Uit onderzoek in een aantal instituten is gebleken dat een verhoging van de
slijpsnelheid een grote verspaningscapaciteit mogelijk maakt bij een goede
oppervlaktekwaliteit en een grote nauwkeurigheid van de werkstukgeometrie.
Bovendien is door onderzoek van het doorslijpproces aan de Technische Hoge-
school in Delft aangetoond dat een specifieke verspaningscapaciteit van
1500 mm3/mm.s bij een slijpsnelheid van 100 m/s gerealiseerd kan worden. Uit-
gaande van deze feiten wordt een rondslijpmachine ontworpen die in staat is
assen met een maximale lengte van 750 mm en een maximale diameter van 250 mm
te bewerken bij een verspaningscapaciteit van 20.000 mm /s en een slijpsnel-
heid van 150 m/s. Het slijpproces zal worden gerealiseerd volgens onderstaande
figuur, waarin de relatieve bewegingen van de slijpschijf ten opzichte van het
werkstuk zijn aangegeven.

u
Schematische voorstelling van
werkstuk en slijpschijf.

Het eerste deel van het ontwerp betreft de aandrijving van de slijpsteen. De
slijpspil zal direkt worden aangedreven zonder overbrenging. Deze eis is aan-
leiding tot het ontwerp van een speciale hydromotor met een maximum vermogen
van 150 kW bij een maximale snelheid van 10.000 omw./min.
Het tweede deel van het ontwerp heeft betrekking op het gestel van de machine.
Dit machinegestel vormt de ondersteuning van slijpspil en werkstuk. De voor-
naamste eis die aan dit deel van de machine wordt gesteld betreft de maximale
vervorming ten gevolge van de slijpkrachten die met betrekking tot de nauw-
keurigheid van het werkstuk is toegestaan. Voor de berekening van deze vervor-
mingen wordt gebruik gemaakt van de "eindige elementen methode".
Het derde deel wordt gevormd door het ontwerp van de aandrijvingen van de
slede, de slijpkop en het werkstuk en de besturing van de machine. De aan-
drijfmotoren die worden toegepast zijn electrische stappenmotoren, waardoor
het besturingssysteem betrekkelijk eenvoudig wordt. Dit besturingssysteem
bewerkstelligt in de eerste plaats de bewegingen van de slijpschijf ten op-
zichte van het werkstuk overeenkomstig de geometrie van het werkstuk. In de
tweede plaats worden door het besturingssysteem de juiste procesomstandig-
heden tot stand gebracht.
Het laatste deel behandelt enkele nevenaspecten van het ontwerp. In de eerste

- 6 -
plaats betreft dit het afdraaien van de slijpschijf, waarvoor een diamant-
schijf wordt toegepast. Vervolgens wordt aandacht besteed aan de veiligheid
van de machine in het geval van een slijpschijfexplosie.
Tenslotte komt de toevoer van koelmiddel, zowel bij het slijpen als bij het
afdraaien, aan de orde.
Het ontwerp van de slijpmachine is weergegeven in een aantal tekeningen.
- 8
INTRODUCTION

Grinding was probably the first metal cutting process in the world. The abi-
lity of abrasive grains to remove particles of metal has been used during
many centuries and for a long time it was the only possible way to remove me-
tal. Since the use of tool steel, highspeed steel, carbide and aluminium oxyde
for metal cutting the application of grinding is mainly restricted to finish-
ing operations and the machining of hard materials. The main reason for this
is the fact that the use of the materials mentioned above made it possible to
realize an optimal tool geometry and, on account of the large dimensions of
the cutting edge,a high metal removal rate. Besides the cutting process using
tools of wellknown shape and dimensions is better predictable and also more
reliable than the use of a grinding wheel with an unknown number of cutting
edges with unknown shapes and unknown cutting conditions.
However, during the last ten years there has been a large increase in re-
search concerning grinding all over the world. One of the main aspects of
this research was the effect of an increase of the wheel speed on both the
metal removal rate and the surface quality. From this research it appeared
that an increase of the wheel speed results in a substantial decrease of the
grinding forces and besides in an improvement of the surface quality [l, 2,
3, 4 ] . This decrease of the cutting forces offered the possibility to raise
the metal removal rate to very high values. This way of grinding became known
as abrasive machining. The application of abrasive machining or high-speed
grinding in industry followed, mainly in mass production like the motor-car
and ballbearing industry. In these industries many products are manufactured
by forging followed by grinding. The turning process can be eliminated in
many cases. At this moment wheel speeds of 80 m/s and even higher are used
in industry resulting in a specific metal removal rate of about 50 mm /mm.s,
while in laboratories much higher values have been reached.
Besides the development described above there has been another application of
the grinding process that is of great interest, namely cut-off grinding. This
process has often been considered as different from other grinding processes,
but apart from the wheel v/idth no fundamental distinction can be proved.
Cut-off grinding has been subject of research in many research institutes
[5, 6] and also in the laboratory for machine tool design of the University
of Technology in Delft, Netherlands. In this laboratory a special cut-off
grinding machine has been designed and built [v]. The results of the research
carried out on this machine have considerably contributed to the design of
the grinding machine described in this thesis. One of the main results con-
cerned the high metal removal rate that was reached on this cut-off grinding
machine. At a wheel speed of 100 m/s a specific metal removal ratf^of
1500 mm3/mm.s has been realized.
The development described above leads to the conclusion that in the near fu-
ture grinding will become a process for high-metal removal and will compete
with first-operation processes such as turning and milling. This prediction
was also part of the outcome of a survey commissioned by the Science Research
Council in England [8, 9 ] . This survey showed besides that it will be ne-
cessary to use greater inputs of power than those used at present, thereby
leading to the development of more rigid machine structures to accommodate
high powers without sacrificing the work surface and geometrical accuracies.
Another conclusion of this report is that external cylindrical grinding is
expected to be the most important area of grinding. However, the design of
a machine capable of realizing an optimal grinding process is a prerequisite.
Starting from the above this thesis describes the design of a high-speed cy-
lindrical grinding machine.

- 9 -
BASIC STARTING POINTS FOR THE DESIGN

The product

The design of a machine tool has to be based on the manufacture of a well-

defined type of product. If the design concerns a cylindrical grinding ma-
chine at least the maximum diameter and the maximum length of the work-
pieces have to be known. The grinding machine described in this thesis
shall be suited to grind workpieces with a maximum diameter of 250 mm and
a maximum length of 7 50 mm. Although this choice is rather arbitrary the
area of application has to be of such importance that it justifies the de-
velopment of a totally new type of grinding machine. In the range of di-
mensions mentioned above many products can be found. Typical examples are
machine tool spindles, shafts of electric motors and gearboxes and so on.
Generally these products are manufactured from cylindrical bars or from
semi-manufactured products obtained by forging.

The process

The basic idea for the design is to provide a grinding machine capable of
removing metal at a rate comparable with that of modern turning machines.
So, if compared with a heavy-duty lathe with a maximum power of about
50 kW, using carbide or aluminium oxyde tool material, the maximum metal
removal rate has to be about 20.000 mm3/s. Just like the choice of the type
of product that will be manufactured the determination of the metal removal
rate cannot be based on accurate data concerning a well-known production of
certain workpieces. In consequence the metal removal rate will be chosen
according to the data that may be expected. It has already been assumed
that the workpieces that will be machined are mainly shafts of different
types. These products have usually in common that metal removal represents
an important part of the cost of manufacturing. So, in order to reduce this
cost the rate of metal removal should be high. For this reason the maximum
metal removal rate shall be 20.000 m m / s according to a high production
turning machine for the same type of product.
Until now the maximum value of the specific metal removal rate that has
been achieved in grinding is about 1600 mm /mm.s, while wheelspeeds of
125 m/s have been applied. The fast development in grinding makes it like-
ly that within a few years higher values will be possible. In consequence
the design of the grinding machine will be based on a specific metal re-
moval rate of 2000 mm3/mm.s at a wheelspeed Vg = 150 m/s. This means that
under these conditions the effective wheelwidth is 10 mm.

The machine

Starting from the data concerning both the kind of workpiece and the grind-
ing process it has to be decided in what way the process will be realized.
The two major points are the shape of the grinding wheel and the movement
of the grinding wheel relative to the workpiece. As mentioned before the
width of the wheel is small compared with the length of the workpiece.
This means that the grinding wheel has to move both in radial and axial
direction of the workpiece. In this way the wheel has to be able to gene-
rate at least cylindrical surfaces and planes perpendicular to the center
line of the workpiece. So the grinding wheel needs two surfaces touching
the workpiece in two straight lines that are perpendicular to each other.
This requirement is met by two conical surfaces with axes that coincide
and of which the sum of the vertical angles is 180° (see figure 1). The value

U
FIGURE 1.
Schematic view of workpiece and
grinding wheel.

of the angle a between the axis of the cones and the center line of the work-
piece has to be determined, based on the following considerations. Consequent
on the high speed of the grinding wheel the internal stress in the wheel is
very high. For this reason it is important to avoid an increase of this stress
caused by the grinding process. This means that the conditions concerning the
grinding process have to be such as to eliminate forces perpendicular to the
plane of the grinding wheel. As the stiffness of the wheel in this direction
is small the elimination of these forces results besides in a smaller wheel
deformation and hence in a higher accuracy of the workpiece. Based on these
facts the value of a has to be determined such as to cause the smallest de-
formation of the workpiece by the grinding forces. As this deformation is
mainly determined by that component of the grinding force that is perpendicu-
lar to the surface of the workpiece the value of a has to be chosen as near as
possible to 90° in order to reduce this force. However, it will appear from
the design that it is not possible to realize a value of a greater than about
60°. Another point that is of importance to the design of. the grinding machine
is the maximum wheel diameter that will be applied. On the one hand there are
a number of factors that tend to a small wheel diameter. In the first place
the stiffness of the system "grinding wheel - machine - workpiece" increases
with a decrease of the wheel diameter. Secondly the kinetic energy of a ro-
tating grinding wheel is smaller, at the same circumferential speed, when
using a smaller wheel diameter. This fact and the wheel diameter itself sim-
plify an effective protection against the consequences of a possible explosion
of the grinding wheel. On the other hand a large diameter of the wheel results
in a longer wheel life. Besides the problems of the design of a proper drive
for the grinding wheel at high speeds become more difficult at a smaller dia-
meter. At last it is very important to be sure that a grinding wheel according
to the demands can be manufactured. Based on these considerations a maximum of
500 mm shall be applied. A grinding wheel of these dimensions has already been
applied in industry at a wheel speed of 125 m/s and this application got the
approval of the German D.S.A. The minimum diameter depends on the dimensions
of the flanges that hold the grinding wheel and shall be 300 mm. This means
that the speed of rotation of the grinding wheel ng varies from 57 30 to
9550 r.p.m.

- 11
BASIC CALCULATIONS

Grinding power and forces

It has been mentioned before that cut-off grinding is the only grinding pro-
cess until now where a specific metal removal rate exceeding 1000 mm3/mm.s
has been realized. As this grinding process and the grinding process that
will be applied on the machine that is designed differ substantially from
conventional grinding processes, even in the case of so called "high-speed
grinding", it may lead to errors if the results of these processes should
be extrapolated and applied to the design without further consideration.
For this reason primarily the data concerning the cut-off grinding process
will be used for the design calculations of the grinding machine. In the
laboratory for machine tool design of the University of Technology in
Delft an extensive research concerning cut-off grinding has been carried
out [7, 10]. From this research and research in other laboratories [5, 6,
11] the data are derived.
At first it is necessary to know the power that is required for a metal re-
moval rate of 20.000 mm3/s. From figure 2 it appears that at high metal re-
moval rates the specific power P' varies between 6 and 7.5 W/mm3/s for
different materials. The values of the specific power, shown in this figure,

P W/mm^/s
* IOJ

9.

T- 1 r-^ 1 1 1 1 r
200 400 600
800 1000 1200 1400 1600 1800 2000z' mm /s mm

FIGURE 2.
Specific power P' at cut-off grinding.

concern the power input at the grinding spindle. This means that these
values include losses due to friction in the spindle bearings and air
friction at the grinding wheel surface. As the research concerning the
grinding power was carried out at a wheel speed Vg = 100 m/s and specific
metal removal rates smaller than 1600 mm3/mm.s it is necessary to determine
what the specific power is expected to be at 150 m/s and a specific metal
removal rate of 2000 mm3/mm.s On the one side the power will increase at
higher wheel speeds due to an increase of the friction. On the other side
the grinding forces and hence the grinding power decrease at higher wheel
speeds as a result of a decrease of the chip thickness at each grain of the
wheel. Based on these considerations the specific power P' at a wheel speed
vs = 150 m/s and a specific metal removal of 2000 mm3/mm.s is assumed to
be 7.5 W/mm3/s at maximum for the materials that will be ground on this ma-
chine. In this value only the air friction at the wheel surface is included.
As the maximum metal removal rate amounts to 20.000 mm3/s the maximum power
for this process Pmax = 150 kW. From this value it follows that the maximum
grinding force tangential to the wheel circumference

max 1000 N.
t max

From the research mentioned above it appeared that the grinding force per-
pendicular to the active wheel surface F^ is about three times the force in
tangential direction F-j-, So Fn max = 3000 N. In chapter 2.3 it is noticed
that the direction of this force is perpendicular to the axis of rotation
of the grinding wheel.

Speeds of grinding wheel and workpiece

The movement of the grinding wheel with regard to the workpiece can be
separated in a rotation of the workpiece and a linear movement of the
grinding wheel parallel to the center line of the workpiece or a movement
of the workpiece parallel to its center line. The ratio of these two speeds
determines the direction of the force Fj^ as will be shown in the following.

FIGURE 3.
Schematic representation of
workpiece and grinding wheel.

From figure 3 it appears that the force F^^ is perpendicular to the axis
of rotation of the grinding wheel if:

F • sind F • cose (3.1)

na nr

- 13 -
 FIGURE 4.
Workpiece, showing the areas of contact
of the grinding wheel.

In figure 4 the areas of contact of workpiece and grinding wheel are shown.
From this figure it appears that there are two different situations concerning
the intersection of the surface of the workpiece and the surface of the grind-
ing wheel. Which situation will appear depends on the dimensions of the work-
piece and the grinding wheel and on the relation between the depth of cut a
and the feed s. The relation between these data can be calculated as follows:
At first it is assumed that the forces perpendicular to the wheel surface are
proportional to the metal removal rate achieved by this wheel surface. This
means that the equation (3.1) can be written a s :

z 'Sine z -cosct) (3.2)

a r

In this equation z^ and T.-^ are the metal removal rates achieved at the areas
Aa and k-^. Corresponding to figure 4-1 the values of z^ and Zp can be calcu-
lated from

z = 0.5-a'l'v +0.5»a'P*v, (3.3)

a a >
z = (a«s-0.5*a'p) (3.4)

In these equations v^ and v^^^ represent the speed of the grinding wheel parallel
to the centerline of the workpiece and the peripheral speed of the workpiece;
1 and p represent geometrical data of the linearized representation of the
areas of contact, on which the equations (3.3) and (3.4) are based. As

TT.d
w w
in which d^ is the diameter of the workpiece, the equation (3.2) changes into:

0.5'I's
a«v (— + 0.5»p) •tg(i)=a«v (s-0.5'p)
W TT'd w
from which:
_ 0.5.p(l+tg(|))
(3.5)
" ^_ 0.5-l>tg(^
TT ' d

- 14 -
As, at the chosen value of cj) = 30°, it appears that in all cases the value of
p exceeds the value of s, it is not possible that the workpiece and the grind-
ing wheel intersect according to figure 4-1 and at the same time the grinding
process meets the equation (3.2).
If the intersection of grinding wheel and workpiece corresponds to figure 4-II
the metal removal rates z^ and z^ can be calculated from:

:(a-l - 0.5-q«l) • V + (a- 0.5'S-q) • V (3.6)

w
0.5-q-s-v (3.7)

In these equations 1 and q represent geometrical data of the linearized repre-

sentation of the areas of contact as shown in figure 4-II. The equations (3.6)
and (3.7) are derived from this representation. In the same way as before these
equations lead to:

a = 0.5-q'(l + • ) (3.8)
ltg4)
+ tg(j

From this equation it appears that at a value of (}> = 30 and a workpiece dia-
meter and a grinding wheel diameter within the limits assumed before the in-
tersection of the workpiece and the grinding wheel corresponds to the repre-
sentation of figure 4-II. This means that at any value of a there is a value
of s at which the forces perpendicular to the plane of the grinding wheel are
eliminated. The ratio of a and s depends on the value of a, the diameter of
the workpiece and the diameter of the grinding wheel as shown in figure 5.

a mm

FIGURE 5.
The ratio a/s at various grinding conditions.

- 15 -
 From calculations concerning the geometry of the intersection of the grinding
wheel and the workpiece it appears that the value of q in figure 4-II can be
approximated by

q = 0.74'a

Besides, as the value of v^ is very small if compared with the value of v^,
the equations (3.6) and (3.7) can be approximated by

z = 0.63'a.s.v (3.9)
a w
and
z = 0.37.a'S.v (3.10)
r w
The values of z and z can also be calculated from:
a r
z = a-z' (3.11)
a a
and
z=s'z' (3.12)
r ^
The maximum value of v^ will be necessary at the maximum metal removal rate and
at the same time at the maximum value of z'^ and z'^^. From the ratio of a and
s and the ratio of z^ and Zp it appears that z'^ is always less than z'p. This
means that only the value of z'^ can reach its maximum of 2000 mm3/mm.s.
As the maximum metal removal rate z^g^^ - 20000 mm3/s:
z = 0.37 • 20000 = 7400 mm~^/s
r max
and hence:

s = z /z' = 3.7 mm
max r r
The corresponding value of a depends on the diameter of the workpiece and the
diameter of the grinding wheel. At the maximum workpiece diameter of 250 mm
and the minimum grinding wheel diameter of 300 mm a = 9.5 mm. At small work-
piece diameters it is not possible to reach the maximum metal removal rate as
the value of a exceeds half the diameter of the workpiece. So the maximum value
of a that will be applied shall be limited to 16 mm. The corresponding value
of s at a workpiece diameter of 50 mm and a grinding wheel diameter of 500 mm
is 1.2 mm. Consequently the metal removal rate decreases proportionally to s.
From the above it follows that at a workpiece diameter of 250 mm and a grinding
wheel diameter of 300 mm the equation (3.10) leads to:

7400 = 0.37 • 9.5 • 3.7 • v

w max
from which
V = 569 mm/s
w max

At a workpiece diameter of 50 mm and a grinding wheel diameter of 500 mm

z = 7400 . 1 ^ = 2400 mm3/s

r 3.7

so:

- 16 -
 2400 = 0.37 16 1.2

from which

V = 3 38 mm/s
w
Assuming that workpiece diameters smaller than 50 mm will not be taken into
consideration the maximum speed of rotation of the workpiece

130 r.p.m.
w max
The minimum value of n is determined by the desired speed v,, at the maxi-
w . . '^ w
mum workpiece diameter and the possibility to apply specific metal removal
rates smaller than 2000 mm3/mm.s.
If the minimum value of z' that will be applied is 1000 mm3/mm.s the speed
v^^ at a workpiece diameter of 250 mm will be 285 mm/s. So:
21 r.p.m.
w mm
From the speed of rotation of the workpiece and the value of s the relative
speed of the grinding wheel with regard to the workpiece in axial direction
Vg can be calculated. It is obvious that v^ reaches its maximum value at the
smallest workpiece diameter that allows the maximum value of s. From figure
5 and the calculations before it follows that this diameter is about 100 mm.
This means that

w max max 6.7 mm/s

a max TT • 100

In order to position the grinding wheel with regard to the workpiece in a

short time a "rapid travel" speed v^ j^ = 50 mm/s will be realized.
Besides a movement in axial direction will be necessary in order to move
the grinding wheel or the workpiece in a direction perpendicular to the
center line of the workpiece. Although the grinding process allows an in-
feed of 16 mm/rev., corresponding with the depth of cut a in the fore-
going, it has not much sense to apply such high values. So a maximium in-
feed rate Vf jj^^x ~ ^ mm/s has been chosen. The "rapid travel" speed in this
direction will be v, , = 25 mm/s.

The grinding wheel

In chapter 2 it has already been assumed that the maximum diameter of the
grinding wheel will be 500 mm and the minimum diameter 300 mm. Besides the
diameter the width of the grinding wheel is of interest. From figure 6 it
can be seen that the width of the grinding wheel can be devided in two
parts b , and b ^ that are in relation with the value of a and s. The
si s2

FIGURE 6.
The width of cut at the
grinding wheel.

- 17 -
 minimum value of b , has to exceed at least the value a^ax * costj) and bgn rriin
has to be greater than s^^^ • sinifi. The choice of the wheel width is based on
the following considerations. If the value of bg2 equals exactly the value
of Sjjjgj^ • sincj) the wear of the grinding wheel will cause an error in the
workpiece diameter. There are two possibilities to avoid this error. At first
the wheel wear can be eliminated by adapting the position of the grinding
wheel in such a way that the distance between the workpiece and the wheel
surface is kept constant. It may be necessary to dress the grinding wheel
continuously or intermittend in order to restore the geometry of the grinding
wheel. Secondly the value of bg2 can be chosen two or more times the value
°f ^max ' sin(}). In this case the error caused by the wear of the grinding
wheel can be corrected by that part of the grinding wheel that did not
participate in the grinding process before. However, after some time the
total width of the grinding wheel will be involved in the process and the
wear of the grinding wheel will also cause an error in the diameter of the
workpiece. In consequence the grinding wheel needs to be dressed and its
position has to be corrected. The first method requires a continuous mea-
suring of the workpiece diameter in order to obtain a reasonable accuracy.
The second method, however, results in an increase of the wheel width and
hence an increase of the damage that can be caused by an explosion of the
grinding wheel. Based on these considerations the following values are
chosen: bg^^ = 14 mm and b „ = 5 mm. So the total wheel width bg = 20 mm.

18 -
THE WHEEL HEAD

The central part of the grinding machine is the wheel head comprising the
grinding wheel, the spindle, the bearings and the drive. From the fore-
going the dimensions of the grinding wheel, the required speeds and power
and the forces at the wheel circumference are known. The first point to
determine is in what way the grinding wheel, the grinding spindle and the
bearings will be set up. It is most obvious to mount the grinding wheel at
the end of a rotating spindle and to arrange the drive at the other end of
the spindle. Other solutions, such as the application of bearings at both
sides of the grinding wheel or a fixed spindle and a bearing in the center
of the grinding wheel are not considered because they do not allow a prac-
tical and fast change of the grinding wheel and besides the space between
the grinding wheel and the workpiece is too small to accommodate any bear-
ing construction. Both the design of the grinding spindle and the design
of the spindle drive will be discussed in the following.

The grinding spindle

The most important factor that determines the design of the grinding
spindle concerns the deformation, caused by the grinding forces, that is
allowable with regard to the workpiece accuracy. It is assumed that the
error in the diameter of the workpiece caused by the deformation of the
grinding machine, including the spindle and the bearings, shall not exceed
a value of 0.025 mm at the maximum grinding forces. This means that at an
average workpiece diameter of 125 mm and an average length of 500 mm the
error in the diameter of the workpiece, including the deformation of the
workpiece itself, corresponds to the tolerance according to an ISO-tole-
rance quality 6. In this consideration the radial deformation of the
grinding wheel is not taken into account. This deformation and the errors
in the positioning of the workpiece with regard to the grinding wheel may
result in the possibility to reach an ISO-tolerance quality 7 at maximum
grinding power, with the exception of small workpiece diameters. As it is
not possible to determine an optimal distribution of the deformations for
all the machine parts it is assumed that the deformation of the grinding
spindle and the spindle bearings shall not cause an error in the diameter
of the workpiece that exceeds 0.01 mm. Besides the error in axial direct-
ion of the workpiece shall be smaller than 0.01 mm. Starting from the
above it is necessary to determine what factors are of importance with
regard to the deformation of the grinding spindle. In the first place the
forces that cause a deformation that is important with regard to the
workpiece accuracy have to be known. In figure 7 the grinding spindle is

c ..'' ^
FIGURE 7.
Schematic representation
the grinding spindle.
of

- 19 -
 drawn schematically. From this figure it appears that only the deformations in
X- and y-direction have an influence on the workpiece accuracy. A deformation
in z-direction has only a negligible influence on the dimensions of the work-
piece. This means that only the force F^ has to be taken into account for the
calculations of the spindle data. The second important factor is the grinding
spindle itself. The deformations depend on the diameter of the spindle D, the
distance of the bearings b and the point of application of the force F^, de-
termined by the distance a from the front bearing to the center of the grind-
ing wheel. At last the deformation of the spindle bearings forms part of the
total deformation. In order to know this deformation it is necessary to choose
the type of bearings that will be applied. It follows from a rough calculation
that the application of ball- or roller bearings is not possible because of
the high speed of the spindle. So the only solution is to apply hydrodynamic
or hydrostatic bearings. As the thickness of the oil film in hydrodynamic
bearings at the desired stiffness of the bearings is very small the generation
of heat in the bearings may cause severe problems. For this reason hydrostatic
bearings will be applied, using one pump for each pocket, in order to reduce
the heat generation as much as possible.
Starting from the above the design data both for the spindle and the bearings
will be calculated. The main requirement that has to be met concerns the maxi-
mum deformation of the spindle and the bearings as a result of the normal
grinding force F^^. This requirement means that in the point P of the grinding
wheel (see figure 7) the displacement in y-direction fy caused by the maximum
force Fn max ^^ '^°"t allowed to exceed a value of 5 ym. At the same time the
displacement in x-direction fj^ has to be smaller than 10 ym. Besides this re-
quirement there are some other factors that are of interest for the design of
the spindle. At first the diameter of the spindle determines not only the de-
formation of the spindle but also the heat generation in the bearings. Second-
ly the position of the bearings relative to the spindle has a great influence
on the stiffness of the total system. At last the demands for maximum stiffness
of the bearing and minimum heat generation are contradictory. It is obvious
that it is not possible to design a grinding spindle that meets all the de-
mands mentioned above. For this reason a choice will be made based on a calcu-
lation of the influence of the various parameters on the stiffness of the
spindle, the temperature rise in the bearings and the power needed for the
oil supply to the bearings. At first it is assumed that the temperature rise
in the bearings shall not exceed 15° C. At this value cooling can be realized
without expensive and complicated cooling systems. Secondly the design of the
wheel head does not allow a distance between the centers of the grinding wheel
and the front bearing smaller than 100 mm. It appears that at a stiffness of
the front bearing Cy = 15 • 10^ N/mm, a stiffness of the rear bearing
C;^ = 7.5 • 10^ N/mm, a spindle diameter D = 120 mm and a distance between the
bearings b = 250 to 325 mm the deformation of the spindle does not exceed the
values mentioned before. The calculations on which these values are based are
written down in Appendix A. These calculations are rather simplified, as the
exact geometrical data of the spindle are not known at this stage. When these
data are known a more precise calculation will be carried out. In Appendix B
the calculation of the data for the hydrostatic bearings is carried out. Be-
sides the data mentioned before the design is based on the following consider-
ations, At first the minimum value of the oil viscosity is determined by the
fact that standard hydraulic components will be applied. This means that the
oil viscosity is at least 5 cSt. Secondly the total area of the oil film is
kept as small as possible in order to reduce the friction losses. At last the
length of the front bearing is mainly determined by the demand to reduce the
distance between the centers of the grinding wheel and the front bearing. The
main results from this calculation are as follows:

- 20 -
 Front bearing Rear bearing

Bearing diameter D = 120 mm D„ = 80 mm

^ -6 A
Oilfilm thickness h =30-10 mm -6
o h 30-10 mm
o
Bearing length Le = 90 mm Le, = 80 mm
^ -5 3 A
-5 3,
Total oil flow Q =16-10 m/s ^A = 12 • 10 m/s
^ 5 2 18,5 • 10^ N/m^
Oil pressure P = 1 6 , 4 - 1 0 N/m ^A =
V

The geometrical data concerning the detailed design of the bearings are
described in Appendix B,

The spindle drive

From chapter 3.1 it appears that the drive of the grinding wheel has to
meet the following requirements. At first the power needed for the grinding
process has to be 150 kW over a speed range from 5730 to 9550 r.p.m. This
means that the maximum torque at the grinding spindle comes to 250 Nm. Se-
condly the speed of the spindle has to be controlled continuously in order
to keep the speed at the wheel circumference at a constant value of 150 m/s.
The most common way to drive machine tool spindles is the use of electric
motors in many cases together with mechanical drives. An important dis-
advantage of these drives is the fact that both the weight and the dimen-
sions are of such a kind that practical application in this case is almost
impossible. The only way to keep the dimensions of the drive within accept-
able limits is the application of a hydraulic drive, as the ratio of power
to weight is much higher than ever can be obtained at electric motors.
Looking at the hydraulic motors that are available it appears that it is
not possible to get a hydraulic motor with a power of 150 kW and a maximum
speed of 9550 r.p.m. This would mean that it will be necessary to apply a
mechanical transmission between the motor and the grinding spindle. This
mechanical transmission, however, presents severe difficulties. At first
the transmission causes a static load on the spindle and hence an unde-
sirable deformation. Another objection is the fact that it is very diffi-
cult to avoid vibrations when using mechanical transmissions at high speeds,
These vibrations cause a disturbance of the surface quality of the workpiece
and in some cases inadmissible forces in the grinding wheel. Based on these
considerations it has been decided to design a new hydraulic motor that is
suited to drive the grinding spindle directly at the required speed and
the required power. The design of this hydraulic motor will be described
in the following chapter,

- 21 -
THE HYDRAULIC MOTOR

Basic concept

It is obvious that the principle of the motor has to be based on the method
of positive-displacement, as a fixed relation between speed and flow rate,
independent of load, is essential for the drive. Starting from this it has
to be determined in what way the displacement of oil from inlet to outlet
will be transformed in a rotation of the shaft of the motor. At first reci-
procating pistons or even vanes should be avoided. At the high speeds that
are applied these movements may cause problems as a result of high inertia-
forces or may cause inadmissable vibrations. This means that only rotating
parts should be used, preferably at a constant speed of rotation. In order
to achieve a rotary movement of the shaft of the motor as a result of a
displacement of a certain oil-volume there has to be at least one chamber
of which the volume changes as a result of a rotation of the motorshaft,
One possibility to realize this change of volume is to apply one rotor that
rotates on an axis that is also rotating, in a non-cylindrical housing. On
this application the well-known Wankel-coiT±>ustion-engine is based. Because
of the complexity and the difficulties concerning the balancing of the ro-
tor this solution is not considered. The other possibility is to use at
least two non-cylindrical rotors, each of which is rotating in a cylin-
drical boring, in such a way that the external surfaces of the rotors to-
gether with the internal surfaces of the housing form at least two cham-
bers of which the volume changes as a result of a rotation of the rotors.
The design of the motor will be based on this solution. Figure 8 shows a

FIGURE 8.
Schematic representation of a hydraulic
motor with two rotors.

cross-section of this system. Both rotors are supposed to be disc-shaped

and to rotate between two planes perpendicular to the axes of rotation of
the rotors. From the figure it appears that the rotors separate the inlet-
chamber from the outlet-chamber. This means that the circumference of each
rotor has to touch the cylindrical boring of the housing in v/hich it is
rotating at least in one place. Besides the circumferences of the rotors
have to touch eadi other continuously. These requirements and the demand that
there has to be a linear relationship between the change of the volume of
the inlet- and outlet chamber and the rotation of the rotors determine the
shape of the rotors. It appears to be very complicated to translate these
requirements into mathematical relations and to derive from these relations
the correct shape of the rotors. For this reason a rotor shape will be cho-
sen and calculations will be carried out in order to check if the require-
ments mentioned above can be met,
The shape of the rotor can be described by

R = RQ - X (5.1)

where R is the radius of the rotor, RQ the maximum value of R and x a func-
tion of the angle (J) as shown in figure 8, If a second rotor is suited to
act together with this rotor, according to the requirements described be-
fore, and if the shape of this rotor meets the equation

R' = R^' - y (5,2)

it is obvious that there has to be a relation between x and y in those points

where the rotor circumferences touch each other. If it is assumed that these
points lie in a plane through the axes of the rotors, the sum of x and y
has to be constant. The first demand for the motor concerns the relation
between the speed and the oil flow of the motor. This relation has to be
linear, which means that at constant speeds of rotation w and o)' of the
rotors the oil flow Q has to be constant. From figure 8 the following equa-
tion can be derived:

Q = Ü) • b(R^ - 1 • x) • x + üj' • b(RQ' - 5 • y) • y (5,3)

where b is the width of the rotors,

As the sum of x and y equals a constant value a, this equation can be written
as follows:

2
(o) . Ro - w' - Ro + Lü' - a) • X - i (w + 0)') - X = constant (5.4)

As the above equation is not true for each value of x the demands for the
motor cannot be met by this system. For this reason a system with two inlet-
respectively two outlet chambers of which the total oil flow is constant at
constant rotor speed will be considered. This system can be realized by
using three rotors, one rotor being the central rotor acting together with
two identical secondary rotors.

- 23 -
In figure 9 a system with three rotors is shown. As the value of R in the re-
lation between the polar co-ordinates of the rotorcircumference has to vary from

/
/

// /
• / /

.' /

FIGURE 9.
Hydraulic motor with three rotors.

a minimum to a maximum it is obvious to choose a sinusoidal variation. The most

simple relation is represented by:

R R a sin (n ) (5.5:

If the circumference of the central rotor meets this equation the secondary
rotors have to meet the following relations:

R' R ' a-cos 2 (n' • (f' ) (5.6)

o
and:
R" = R ' - a-cos (n' -(j)' ' ) (5.7)

If the secondary rotors touch the central rotor in lines at the angles (J) and
(f) + a the position of the lines of contact on the secondary rotors is de-
termined by an angle <^' that equals — • <^ and an angle ({>' ' n ((f+ct).
n' n'
In this case it is also assumed that the lines of contact lie in the planes
through the axes of each two rotors that act together. If the oil flow through
the motor is calculated, based on a value of
2TI_
" = 2 Ï Ï ^ "' n
where i may e q u a l a n y whole nuTnber:

Q = cü-b(Ro-5-a-cos2(n-(j)))-a-cos (n-(ti) +
w.b(RQ-5-a-sin2(n-(t> ) ) - a - s i n ( n ' ( { ) ) +
uj' . b ( R o ' - 5 a - c o s 2 ( n ' -<}>' ) ) - a . c o s 2 ( n ' •<\ • ) +
u>' • b ( R o ' - 5 a - s i n 2 ( n ' -(J)' ) ) - a - s i n 2 ( n ' •(] ' )

at constant rotor speeds w and w', or:

24
 2 4
O = b'((jj-a-R +ü)'-a-R '-i-wa -cos (n-*)
^ o o
2 . 4 2 4
-^•w-a -sin (n-(|) )-2-w'-a -cos (n'-(t>')
-l-oo'-a^-sin'^(n'-(I)')) (5,8)

From this equation it appears that the oil flow Q is not constant. However, the
system with three rotors offers the possibility to correct the shape of the
rotors in such a way that a constant oil flow is obtained. In the first place
it is possible to choose an other value of a, but in order to avoid a restrict-
ion of the oil flow to the rotors it is preferable to choose a value of a = Ti,
This means that the axes of all three rotors lie in one plane. The second solu-
tion is to apply a correction x to the radius R and a correction x' to the
radii R' and R'' as shown in the following equations:
2
R = R -a-sin (n-(t))-x
o
R' = R '-a.cos^(n'-è')+x'
o
R" = R '-a.cos^(n' -(j)' ' )+x'
o
If the equation 5,8 is changed according to the relations above the demand for
a constant oil flow Q leads to:

X = I" .(/(l+sin2(2-n-(j)))-l)

and;
x'= |-(/(l+sin^(2-n'.<|)')) + l)

and the condition:

CÜ-R =0)' -(R '+a)

o o
This last condition means that the difference of the circumferential speeds of
the rotors is very small. The equations 5,5, 5,6 and 5,7 change into:

R = R - ^ . (2-sin^(n-(t.)+/(l+sin^(2.n-(l)))-l) (5,9)
o I
R' = R '- ^ , (2-cos^(n'-({)')-/(l+sin^(2-n'-*'))-l) (5.10)
o ,2
R " = R t- .| , (2.cos^(n'•())" )-/(l+sin^(2-n'.(!)" ))-l) (5,11)
o 2
The oil flow through the motor can be expressed by the following equation:
2 2
Q = a)-b(R -a- •^) + cj'-b(R '-at^) (5.12)
- o ^ o z
The second demand for the motor concerns the fact that the rotor circumferences
have to touch each other. This means that the tangent planes in the line of
contact have to coincide. In figure 10 the circumferences of two rotors are
shown. In two corresponding points P and P' the angles y and Y ' between the
tangents in these points and the radii OP and 0'P' have to be identical,

- 25 -
 FIGURE 10.
Tangent lines in corresponding
points of the rotor circumferences,

As: R - |-(2-sin (n-(})) + /(l+sin^(2-n-())) )-l)

R
tgY= dR/d(t) 2sin(n-<))) •n«cos(n-(t))
- f(4.sin(n.*).n.cos(n.<l,) + 2( itsin2(2.n.<D)) )

and:
R Q ' - |.(2-cos(n'•(()') - /(l+sin2(2-n'.*'))'-!
tgY' = dR'/d(t)' a.,, , . ... , • / , ,,^, 2-sin.(n' -())' )-n-cos(n' •(()' )^
- 2(H-cos(n'.*')-n'.sin(n'-r)+ 2(ltsin2(2:n'-*')) ^

it appears that this condition is not met. However, by choosing the ratio of
R Q and n equal to the ratio of Ro'+a and n' the difference between y and y'
becomes very small, especially at values of a that are small compared with the
value of R Q and R Q ' . In the calculations concerning the final design of the
motor, in appendix C, is shown that an increase of the distance between the
rotor axes of only 0.003 mm is sufficient to provide a clearance between the
rotor circumferences. Manufacturing tolerances make it necessary to apply an
increase of the distance between the rotor axes that is at least 0,01 mm. So
practically the rotor circumferences touch each other continuously if the con-
ditions mentioned above are met,
At last it is necessary that the rotor circumferences touch the cylindrical
borings in which they are rotating. From figure 9 it appears that the central
rotor has to touch the boring at least in two places, while the secondary
rotors have to touch the boring at least in one place. In the lines of contact
of the rotorcircumferences and the borings the value of n-(J) respectively
n''(J)', respectively n'•())'' equals TT or a multiple of TT , In these lines the
value of y respectively y' and y'' equals JL, so the tangent planes to the
rotor circumferences and to the cylindrical borings coincide. From the above
it appears that the minimum value of n is 1.5, resulting in three lines on the
rotor circumference where the radius R shows a maximum value. In two of these
lines the rotor circumference touches the boring of the housing. The minimum
value of n' equals 1. It is decided to choose these values for the rotors of
the motor and that because of the following. In order to keep the value of
n-<j) equal to n'-cj)' and equal to n'-(t)''-n-a it is necessary to use an accurate
gear transmission as the shape of the rotors, even at high values of n and n',

- 26 -
 is not suited for this purpose. As the influence of errors in the gears be-
comes smaller at small values of n, respectively n', it is advisable to choose
these values as small as possible. Even in this case the gears have to be very
accurate. The final design of the motor is schematically drawn in figure 11.

FIGURE 11.
Schematic representation of the
hydraulic motor.

This figure shows that, as there is always a difference in the pressure in

the inlet chambers and the outlet chambers, this pressure difference results
in varying forces perpendicular to the rotor axes. In order to balance these
forces on both sides of the main rotors a system of compensating rotors is
arranged. These rotors have the same shape as the main rotors and are mounted
in the same position as the main rotors. Only the value of a in the relations
5.9, 5.10 and 5.11 is smaller. Besides the width of these rotors is half the
width of the main rotors. If the inlet and outlet chambers have positions
opposite to that of the main rotors, the forces applied to these rotors make
balance with the forces applied to the main rotors. Based on the design de-
scribed above the calculations concerning the geometry of the motor will be
carried out, according to the demands calculated before.

5.2. Geometrical design

The basic data for the design of the motor are the values of the torque and
the speed at which this torque has to be realized. From the chapter 3 it
follows that the maximum power at the grinding wheel Pjnax has to be 150 kW,
at a speed range from 5730 to 9550 r.p.m. Consequently the maximum torque at
the grinding wheel Mjj,ax has to be 250 Nm at a speed of 5730 r.p.m. Similar to
the derivation of the equation 5.12 the torque at the hydraulic motor can be
expressed by:

- 27 -
 2 , 2
M = p-b-(R -a- ^ ) + — .p-b(R '-a+ •!-) (5.13)
o 2 O) o 2
where p is the pressure of the oil. As <})' = —j- • <p the same ratio applies to
cü and co , so — = —p - 1.5.
ü) n'
In the first place it is necessary to determine the maximum value of the
pressure p. It is obvious that at high values of p the dimensions of the motor,
mainly determined by R Q , R Q ' and b can be small. Although a high pressure re-
sults in a higher leakage, the leakage gaps are smaller at smaller dimensions
of the motor. So it is decided to apply the maximum pressure at which
standard hydraulic components, like pumps and valves, can be used. This
pressure is at this moment 35,10° N/m^,
It is clear that a certain part of the power that is needed for the drive gets
lost by both leakage of oil in the motor and friction in the motor and the
bearings of the spindle. As the real values of both the volumetric efficiency
of the motor and the total mechanical efficiency of the spindle drive are not
known at this stage it is assumed that the mechanical losses at a speed of
5730 r.p,m. are 10 per cent of the power supplied at maximum pressure and that
the volumetric efficiency at these conditions comes to 95 per cent. So the
maximum torque that is applied by the oil pressure on the rotors:

M z i£ , M = 278 Nm
r max 9 max
As w R = a)'-(R '+a), the equation 5,13 changes into:
o o ^
M = p-b-a(2-R -1,25a) (5.14)
r o
This equation shows that it is possible to choose different values of b, a and
RQ that lead to the torque that is required for the motor. However, in the
first place a value of a that is small compared with the value of RQ reduces
the difference of the circumferential speeds of the rotors. Secondly a small
value of R O makes it impossible to create a supply opening of satisfactory
dimensions. At last the value of a cannot be too small, as this may lead to
undesired hydrodynamic effects in the oil gap between the rotor circumference
and the housing. Based on these considerations the values of a, RQ and b will
be chosen,
As described before a system of compensating rotors will be applied in order
to balance the hydrostatic forces on the rotor axes. The value of a for these
rotors will be a^ = 1 mm while the value of a for the main rotors will be
ajjj = 4 mm. As the center distance of the main rotors has to be the same as
the center distance of the compensating rotors:
R + R' = R + R'
om om oc oc
The values of R and R ' are chosen as follows:
o o
R = 39,0 mm - 22 mm
om ^o m
R = 37,2 mm = 2 3 . 8 mm
oc ^o c
The above results in the following equation:

M = p-b-{2(R -a -R -a )-l,25-(a^ -a^ )}

r om m oc c m c
or:
M = 218,85-p-b
r ^

- 28 -
As n =35.10 N/m and Mr = 278 Nm:
* max max
b = 36 mm

Starting from the dimensions of the rotors the gear transmission between
the three rotor shafts can be calculated. The data that are prescribed by
the dimensions of the rotors are the center distance of the gears and the
ratio of the transmission. If a hardened steel with a tensile strength of
at least 750 N/mm^ and a Brinell hardness greater than 5000 N/mm2 is used
a module of 1.25 and a width of 15 mm meet the demands both with regard
to the bending stress in the teeth and the normal pressure on the face
of the teeth. Finally it is necessary to apply bearings to support the
rotor shafts. As the forces on the rotor shafts only consist of the forces
caused by the gear transmission it is clear that the bearing forces are
very low. This means that the dimensions of the bearings are mainly de-
termined by the geometry of the other parts of the motor, like the rotor
shafts and the gears. The choice of the bearing type is determined by the
properties at high speeds and the accuracy that can be realized. For this
reason ball-bearings will be applied. In order to reduce the play of the
bearings two bearings are mounted and a small preload is applied. In the
drawing 1 the final design of the motor is shown. From this drawing it can
be seen that the motor-housing is composed of nine parallel plates that
are held together by accurate bars that also locate the position of each
plate very accurately. The plates 3, 5 and 7 contain the borings in which
the rotors revolve, while the plates 4 and 6 separate the compensating
rotor systems from the main rotor system. In order to prevent any unba-
lance in the rotating parts of the motor no fixing arrangements like
splined shafts, keys or screw thread are applied. Both the rotors and the
gears are mounted on the shafts by means of glueing. The type of glue that
is used is an anaerobic product that hardens in very small gaps at the
absence of air. This way of mounting parts together reduces the cost of
manufacturing considerably.

The manufacture

It is clear that the rotors are the parts that are most complicated with
regard to the manufacture. There are two possibilities to obtain the cor-
rect shape of the rotor circumference. In the first place the rotors may be
machined on a numerically controlled milling machine, either with a so
called x-y control or a R-(|) control. However, there are very few machines
available with sufficient accuracy. A second method is to make use of the
typical shape of the rotors. As derived before the shape of the rotor is
determined by

R = R^ - I" (2-sin^(n-(t)) + /(l+sin^(2-n-(|)) )-l)

This equation can be written as follows:

R = R + J ' cos(2-n-(|)) - ^ • /(2-cos^(2-n-cj)) (5.15)

- 29 -
 -^.co3(2.n.<D) -?-.\/2-cos2(2.n.(I))
2 2

FIGURE 12.
Mechanisms that generate the shape
of the rotors.

The parts of the equation 5.15 that depend on the value of (J) can be generated
by the mechanisms that are shown in figure 1 2 . If these mechanisms are con-
nected with the table drive of a milling machine it is possible to machine
the rotors with a high accuracy. However, this system only effects the posi-
tion of the center of the cutter relative to the workpiece according to the
equation 5.15. So a correction has to be carried out in order to obtain the
correct shape of the rotor. The way in which this correction is effected is
shown in figure 13. In each point of the circumference of the rotor the tan-
gent to the cutter has to
coincide with the tangent
to the rotor circumference
This demand can be met by
CORRECTED a rotation of the rotor on
POSITION the point in question that
equals -^ - y, where y re-
presents the angle between
the radius vector and the
tangent in that point.
CUTTER The value of this angle
follows from:
ROTOR
dR/d(j)
FIGURE 13. tg(y- Y) R
Correction of the position of the
cutter relative to the workpiece.

In order to simplify the system that has to realize this rotation the value
TT
of — - y IS approximated by the value of the angle 3 that meets the equation:
1.2 'n-a'sin(2'n-(())
tgB = (5.16)
R

The total system for the manufacture is shown schematically in figure 14

- 30 -
DRAWING 1: Hydraulic motor.
31 -
 FIGURE 14.
Schematic representation of a system
for the manufacture of the rotors.

This figure shows also the way in which the angle g is adjusted. From the
foregoing it follows that, in order to machine the circumference of the ro-
tor, three movements of the rotor relative to the cutter are necessary. At
first a rotation of the rotor on its own axis with a speed w^, where the va-
lue of ojjp depends on the cutting conditions. Secondly there is a linear move-
ment according to the equation 5.15, based on a rotation of the cranks of the
mechanism with a speed 2-n-a), At last a rotation of the rotor on a line on
the rotor circumference is applied, also based on a rotation of a crank me-
chanism with a speed 2-n-üj. The ratio of the speeds of rotation has to be
maintained exactly. In the laboratory of the University of Technology, where
the hydraulic motor has been built, the system described above has been
applied. In this case electric stepping motors were used for the drive of
the crank mechanisms. With these motors the ratio of the drives is determined
by the ratio of the frequencies of the pulses supplied to the stepping motors,
The correct adjustment of the frequency ratio can be realized very simple,
Figure 15 shows the manufacturing set-up as it has been used on a small
milling machine in the laboratory.

- 33 -
 FIGURE 15.
Set-up for the manufacture of
the rotors.

As mentioned before both the rotors and the gears are mounted by means of
glueing. Figures 16a and b show two typical stages of the assembly of the
motor. In figure 17 the motor is shown after assembling.

FIGURE 16.
Two stages of the assembly of the motor.

34 -
 FIGURE 17.
Hydraulic motor according to the final design.

5,4. Conclusion

The hydraulic motor described before is in a stage of development. This

means that there is a number of aspects to be regarded before an optimum
design can be realized. In the scope of this thesis it is not possible
to cover all these points. For this reason the foregoing describes the
basic aspects of the design. The calculations concerning the leakage flows
and the friction losses in the motor are carried out in the Appendix D,
Summarizing the description of the motor the following data determine the
design of this motor:

Maximum speed n = 9550 r,p,m.

max
Maximum torque M = 250 N,m
max
Displacement per revolution 5-10"^ m3

Maximum pressure Ti - 35-10^ N/m2

'max
Weight G = 120 kg

In the chapter describing the total power supply for the grinding machine
the oil supply system for the hydraulic motor will be discussed,

35
 6. THE CONSTRUCTION OF THE MACHINE

6.1. General design

The basis for the design of the grinding machine is the movement of the
grinding wheel relative to the workpiece. This movement can be devided in
three parts: a rotation of the workpiece, a movement parallel to the axis
of the workpiece and a movement perpendicular to the axis of the workpiece.
Both linear movements can be carried out either by the grinding wheel or
by the workpiece. It is decided to effect these movements by the wheel head
and that based on the following considerations. The deformations of the
bed and the slide of the machine depend strongly on the dimensions of these
parts. Specially the parts that carry out movements along guideways have a
considerable influence on the deformation of the machine. The length of a
slide, carrying the workpiece and the work heads that support the workpiece,
has to be much greater than the length of a slide that carries the wheel
head. Besides a smaller slide reduces also the dimensions of the complete
machine structure. The above concerns both movements of the grinding wheel
relative to the workpiece. So, in order to reduce the deformation of the
structure, both movements will be carried out by the wheel head. In fi-
gure 18 the workpiece and the grinding wheel with the grinding spindle are
drawn schematically. In
this figure x and y re-
present the directions
according to which the
wheel head is moving.
From this figure it ap-
pears that it is pos-
sible to achieve a move-
ment in y-direction by
composing a translation
in y'-direction, parallel
to the center line of the
grinding spindle, and a
movement in x-direction.
This way of moving the
wheel head has the fol-
lowing advantages. In the
first place the drive of
the wheel head in this
direction will not meet
FIGURE 18.
with any force caused by
Grinding wheel and workpiece the grinding process, as
drawn schematically. the forces perpendicular
to the grinding wheel
are eliminated as de-
scribed in chapter 3.
This means that the influence of the drive with regard to deformations
caused by the grinding forces is also eliminated. Secondly the construction
of the wheel head can be more simple as the center lines of the grinding
spindle and the guideways are parallel to each other. Besides the direct-
ions of the movements of the wheel head it is necessary to know the length
of the travel of the wheel head in each direction. The length of the
travel in y-direction is determined by the dimensions of the grinding
wheel and the diameter of the workpiece. From figure 18 it can be derived
that the length of the travel in y-direction

36 -
 d (d - d . )
w max . . „„„ s max s mm . /c i \
1 = + s m 30° • + 1 (6,1)

where 1 is a displacement of the wheel head that is necessary to handle the

workpiece during loading and unloading with sufficient safety. From the above
equation it follows that the travel in y'-direction

1 , =(175 + 1) - cos 30° (6.2)

y'
The travel in x-direction is determined by the length of the workpiece and
the value of 1 ,. In this case also an extra length 1' is added. So

1 = 750+1 , • sin 30° + 1' (6.3)

X y'
Based on the equations 6.2 and 5.3 the following values are chosen:

1 , = 225 mm and 1 = 1075 mm

_X! -2
These dimensions, together with the dimensions of the workpiece, the grinding
wheel, the grinding spindle and the spindle drive determine the design of the
machine structure and the slides. However, before working out this design it
is necessary to consider a number of aspects that have an important influence
on the design.
In the first place it has to be decided whether the material for the machine
will be cast iron or steel plates. It is difficult to make a decision based
on both technical and economical aspects as this decision requires an extend-
ed knowledge about the production of these machines, specially with regard to
the total number of machines to be manufactured per year, the number of types
and so on. So only technical aspects are taken into account. The advantages
of cast iron concern in the first place the freedom in the design of the
machine parts. Secondly the damping of cast iron is better than the damping
of steel plates; however, a construction of welded steel plates offers good
possibilities to apply sufficient damping to the complete structure. The
disadvantages of cast iron concern in the first place the relatively low mo-
dulus of elasticity. This results in an increase of the weight of the ma-
chine parts in order to obtain the same stiffness as when applying welded
steel plates. Secondly the flexibility of the design is very small. Even a
little change of a part of the machine entails great expense. Based on the
above considerations it is decided to apply welded steel plates for the ma-
chine structure, the slide, the wheel head and the work heads,
Another aspect concerns the guideways for the slide and the wheel head. As it
will be necessary in some cases to move the wheel head and the slide with low
velocities a stick-slip effect may appear. In order to avoid this effect
guideways with rolling elements or hydrostatic guideways can be applied. As
the damping in roller or ball guideways is very poor and as there is already
a hydraulic system for the spindle bearings and the spindle drive, in which a
third system can be incorporated, hydrostatic guideways will be used,
A third point is the fact that there will not be any possibility to inspect
the grinding process visually. In the first place it has no sense to look
at a phenomenon that does not give any information about relevant data con-
cerning the process. Secondly the safety of the machine can be increased by
avoiding openings or windows for inspection.
Although there are some other points that are of interest for the final de-
sign, the above data give sufficient information to set up a preliminary de-
sign, The drawings 2, 3 and 4 show the front view, the top view and a cross
section of the machine. The first point to notice is the triangular cross

- 37 -
 section of both the bed and the slide. This shape has the following advan-
tages, First a triangular cross section of both the bed and the slide makes
it possible to obtain a high stiffness of the structure at relatively small
dimensions, not only with regard to the area of the cross section of the
machine, but also regarding the plate thickness. Secondly, as can be seen
from the drawings, it is possible to realize small dimensions of that part
of the machine that is most influenced by the forces between the grinding
wheel and the workpiece. Particularly the distance between the guideways
and the place where the grinding forces apply is chosen as small as possible,
Another point of interest is the box-shaped construction of bed, slide,
wheel head and work heads. From litterature [l2] it is known that the ap-
plication of closed box-shaped structures offer a high stiffness. In order
to increase the torsional stiffness of the bed and the slide diagonally
placed plates are applied as shown in the drawings. The drawings show also
that the two work heads of the machine are identical, with the exception of
the motor that drives the workpiece. Although in many applications the work
head that carries the drive is much heavier and more rigid than the other
work head in this case the forces that determine the accuracy of the machine
are the same for both heads. So the requirements concerning the stiffness
of the heads are also identical.
It is clear that the above considerations do not produce the evidence that
the design as shown in the drawings is the only possibility to meet the
demands that are made. The judgment about the shape of the complete machine
or the shape of parts of the machine depends highly on the view of the de-
signer. However, each design has to meet the requirements that are de-
termined by the process and by other factors like the demands made by the
manufacturing methods for the machine parts. Some of these requirements have
already been the starting point for the preliminary design, like the di-
mensions of grinding wheel and workpiece, the length of the travel of the
wheel head both in y- and x-direction and so on. Other requirements concern
points like the supply of coolant, the dressing of the grinding wheel, the
control of the machine and provisions with regard to the safety. These re-
quirements will be discussed in separate chapters. At last there are re-
quirements of which the question whether they are met or not can only be
answered after calculations concerning a finished design. Should these cal-
culations show that the requirements are not met the design has to be
changed and the calculations in question have to be carried out again. In
this case these requirements concern the deformation of the complete struc-
ture caused by the grinding forces. The calculation procedure concerning
this point will be described in the following section of this chapter. In
the last section of this chapter the design of the work heads will be con-
sidered.

38 -
DRAWING 2: Front view of the grinding machine. 39
DRAWING 3: Side view of the grinding machine.

41 -
DRAWING 4: Top view of the grinding machine. 43
6.2. Deformation of the machine structure

6.2.1. Method of calculation

The deformation of the grinding machine as a result of the grinding forces

will be regarded as far as this deformation causes a change of the re-
lative position of the grinding wheel and the workpiece. According to the
considerations in chapter 4.1 it is assumed that the deformation of the
machine structure shall not cause an error in the diameter of the work-
piece greater than 0.01 mm. The error in the axial direction, caused by
this deformation, is equally not allowed to exceed 0.01 mm. The deform-
ations of the grinding spindle and the work head spindles are not included
in these values. First the method according to which the calculations
will be carried out has to be determined. These calculations have to take
into account both deformations in the plane of the plates of the struc-
ture and deformations perpendicular to this plane. As besides diagonally
placed plates are applied it is obvious to choose the so called "finite
element" method to calculate the relevant deformations. This method can
be described as follows [13]. A machine part can be devided into a number
of elements of a relative simple shape c.q. bars, squares, triangles and
so on. The elements are connected one to another at their corners, call-
ed nodal points. The basis for the calculations is given by the relations
between the forces in the nodal points of an element and the displace-
ments of the nodal points caused by these forces. From these relations
for each element the relations between external loads and deformations
of relevant points of a machine part can be calculated. It is obvious
that the solution of a large number of equations with a large nuii±ier of
unknowns leads to matrix solutions that can only be solved by using com-
puters. As a consequence computer programs have been developed in order
to allow for the solutions of these calculations, One of these programs,
called ASKA •) is available at the Computing Centre of the University of
Technology in Delft [l4] and this program has been used to calculate the
relevant deformations of the structure of the grinding machine. In this
thesis only the input and the output of this program will be considered,
the program itself will be regarded as a calculation tool of which only
the directions for use are of importance. The parts of the program input
that are important for the user are the "ASKA Processor Control step",
the "Topological Description" and the "Data Input", The"ASKA Processor
Control step" comprises the instructions that take care of the way and
the order in which the data have to be processed in order to obtain the
desired results. The "Topological Description" describes the elements
in which the structure is divided and numbers the nodal points. Besides
in this part the degrees of freedom that are restricted are mentioned.
Finally the "Data" input comprises the co-ordinates of the nodal points,
the material properties, the loads and the prescribed displacements.
The output of the program depends on the instructions in the "Processor
Control step" and contains in the first place the numerical results of
the calculations. Besides information concerning errors in the program

•) "Automatic System for Kinematic Analysis", developed by the

"Institut fur Statik und Dynamik der Luft- und Raumfahrtkon-
struktionen" in Stuttgart, W.-Germany.

- 45 -
or the data, the degrees of freedom in each nodal point, data concern-
ing the structure of the matrices and so on may be printed out. The fol-
lowing section describes the calculation of the deformation of the struc-
ture of the grinding machine.

Execution of the calculations

First of all the type of element that will be used has to be chosen. As
different types of elements in one structure lead to more complicated
calculations one type of element will be applied for the complete struc-
ture. As the plates of the structure are loaded both in a direction in
the plane of the plates and perpendicular to this plane the element has
to accommodate this load. Besides the shape of the element has to allow
for a division that approximates the real structure as accurate as pos-
sible, On account of this the so called TRIB 3 element is chosen. This
element is a triangular plate element that takes six degrees of free-
dom into account in each nodal point. The only disadvantage of this ele-
ment is the fact that there is no stiffness against a rotation on an
axis perpendicular to the plane of the element. However, a stiffness in
this direction can be taken from adjacent elements or, in case this is
not possible, a rotation in this direction can be suppressed in the
program without affecting the accuracy too much. In order to keep the
calculations surveyable the main structure of the machine is devided
into substructures. These substructures are the bed, the slide, the
wheel head and both work heads. Each of these substructures is connect-
ed to one or more of the other substructures by means of a fixed con-
nection like bolts or by means of movable connections like guideways,
The fixed connection shall be approximated by connecting the nodal
points of both substructures in that area where the connection of the
real structure is arranged. The two parts of each guideway are regarded
as being a part of the substructure on which they are mounted. On the
two parts of each guideway nodal points are chosen and connected in
such a way that a movement in the desired direction is allowed. This
means that in each nodal point of this connection only one or two de-
grees of freedom are connected. The deformation of the oil film of the
hydrostatic guideways is neglected as this deformation is relatively
small compared with the deformation of other parts of the structure.
Finally the drive elements for the slide and the wheel head have to be
regarded. These drive elements may be cylinders and pistons or screw
drives. In this stage of the design it is assumed that screw drives
will be used both for the drive of the wheel head and the slide. In the
first section of this chapter it has already been mentioned that the
drive of the wheel head is not affected by the grinding forces. As a
consequence this drive has not to be taken into account in the calcu-
lations concerning the deformations of the structure. However, the
drive of the slide has to be considered in the calculations. For this
reason the drive of the slide is represented by a substructure with
the same elastic behaviour as a real drive screw. The ends of this
substructure are connected to the bed and the slide in a number of no-
dal points that represent the real mountings. Although the wheel head
drive is not affected by the grinding forces it will be necessary to
connect the wheel head to the slide in at least two corresponding nodal
points of the guideways. In figure 19 the main structure of the grind-
ing machine is shown, while figure 20 shows the composing structures,
Starting from the above the relations between the loads and the dis-
placements in relevant points of each substructure are calculated. These
FIGURE 19.
Main structure of the grinding machine.

47
 FIGURE 20.
Substructures, representing the wheel head, the
work head, the slide, the bed and the drive screw.

- 48 -
relations are laid down in so called stiffness matrices. After these stiff-
ness matrices have been drawn up the substructures are connected into the
main structure and from the values of the external loads the displacements
of the points of the structure in question can be determined,
First of all the topological description and the data input of each sub-
structure have to be considered. The first substructure, representing the
wheel head, is shown in figure 21. In the figure the division of the plates
of the structure into the TRIB 3 elements is indicated. In this substructure
two parts are of major interest. First the nodal points in which the grind-
ing spindle is connected to the wheel head, the points 1 to 8 and 52 to 59.
In these points the load, caused by the grinding forces, is applied to the
structure. Secondly the nodal points in which the wheel head is connected
to the slide by means of guideways, the points 9, 24, 39, 62, 11, 26, 41 and
54, It is obvious that, with regard to the deformations of relevant points
of the main structure, only the forces and the displacements in these points
of the substructure are of importance. The data input concerns in the first
place the co-ordinates of the nodal points of the structure. The co-ordinate
system for this substructure is chosen as follows: the x-direction is paral-
lel to the line from point 14 to point 16, the y-direction is parallel to the
line from point 14 to point 67 and the z-direction is parallel to the line
from point 14 to point 20, It will appear from the description of the other
substructures that only the substructure of the wheel head is related to a
co-ordinate system that differs from the co-ordinate system in which the
other substructures are described. The reason for this is the fact that the
calculation of the co-ordinates of the nodal points is more simple if the
plates of the substructures are parallel either to the x-y-plane, or to the
x-z-plane, or to the y-z-plane. In this respect the position of the wheel
head differs from the position of the other substructures. However, an im-
portant consequence is that in the data output the values of the displace-
ments are also related to this co-ordinate system. The second part of the
data concerns the thickness of the plates of the substructure. In first in-
stance a thickness of 25 mm is chosen for all the substructures, with the
exception of the guideways as the dimensions of these parts are known and
fixed. From the results of the calculations it will appear for which sub-
structures or parts of substructures other values have to be chosen in order
to meet the requirements mentioned before. Finally the data concerning the
material of the substructure have to be given, namely the Young's modulus
and the Poisson's ratio, in this case for steel.
The second substructure represents one of the two work heads and is shown
in figure 22. In this substructure also two parts are of interest. First
the nodal points where the work head spindle is connected to the work head,
the points 1, 5, 2, 3, 9, 4 and 16 and the points 52, 53, 41, 54, 55, 45 and
56. Secondly the points where the work head is connected to the bed of the
grinding machine, the points 13, 25, 37, 39, 15, 27, 39 and 51. The co-ordi-
nate system for this substructure is chosen as follows: the x-axis is paral-
lel to the line from point 46 to point 10, the y-axis is parallel to the
line from point 46 to point 44 and the z-axis is parallel to the line from
point 51 to point 46. Both the values of the plate thickness and the data
of the material are the same as those for the first substructure.
The third substructure represents the bed of the machine and is shown in
figure 23. In this substructure a large number of nodal points are of inte-
rest, not only with regard to the calculations, but also with regard to a
general view concerning the deformation of this substructure. In the first
place the nodal points in which the slide is connected to the bed, the
points 14, 31, 48, 65, 82, 99, 116, 133, 150, 167 and 16, 33, 50, 67, 84,
101, 118, 135, 152 and 169 are taken into account. Although the slide is

- 49
» so n

•fc lit 3 Tf «0 3J S3

V» » 4* 67

A! Z3 3 aT A « 83

Substructure 1, the wheel head.

FIGURE 22.
Substructure 2, respectively
6, the work head.
FIGURE 23.
Substructure 3, the bed.

- 51
connected to the bed in only ten nodal points it is possible to consider dif-
ferent positions of the slide by drawing more points of the guideways into
the calculations. Besides the deformation of the complete guideways shows
clearly the deformation of the bed as a whole. The same considerations con-
cern the nodal points where the work heads are connected to the bed. In this
case the points 4, 21, 38, 55, 72, 89, 106, 123, 140, 157, 5, 22, 39, 56, 73,
90, 107, 124, 141 and 158 are drawn into the calculations. In the points 7,
24, 8 and 25 the end of the screw that drives the slide is connected. Finally
the complete structure will be supported in three points, represented by the
nodal points 35 and 52, 79 and 96 and 103 and 120. The co-ordinate system
for this substructure shows the same directions of the axes as the co-ordinate
system of the second substructure, besides the material and the plate thick-
ness are the same,
In figure 24 the fourth substructure, representing the slide of the machine
is shown. In this substructure also the nodal points in which the slide is
connected to the other structures have to be taken into account. The nodal
points of the guideway between the slide and the bed are the points 15, 34,
53, 72, 91, 3, 22, 41, 60 and 79, those of the guideway between the bed and
the wheel head are the points 16, IP, 35, 37, 39, 55, 57, 74, 76 and 93, In
the points 89, 70, 58 and 78 the screw that drives the slide is connected,
Both the orientation of the co-ordinate system and the material are the same
as in the previous substructure, just as the thickness of the plates.
The fifth substructure, representing the drive screw, is shown in figure 25.
This drive screw is connected to the bed in the points 67, 68, 72 and 73 and
to the slide in the points 4, 5,9 and 10. From the assumption that this
substructure has to represent a screw with a diameter of 40 mm follow the
dimensions of the plates. The material is steel with the same properties as
the material for the other substructures.
The sixth substructure represents the second work head and is identical
with the first work head. The only difference concerns the load applied to
this substructure, as this work head is provided with the work head drive.
Based on the calculations of the stiffness matrices of the substructures
the deformations of the main structure are calculated. The program for these
calculations also comprises a topological description of the structure. In
this case this description generates nodal points in such a way that all the
significant points of the composing substructures are indicated as nodal
points of the main structure and are renunÜDered. As the co-ordinates of the
nodal points of the main structure are determined by the relative position
of the substructures it is necessary to determine what position will be de-
cisive for the calculations. As it is not significant to choose the most
unfavourable conditions the calculations shall concern a workpiece diameter
of 150 mm, a length of the workpiece of 750 mm and a grinding wheel dia-
meter of 400 mm, while the position of the grinding wheel is in the middle
of the workpiece. In the topological description the points in which one
or more degrees of freedom are suppressed are mentioned. This is necessary
in order to obtain a correct support of the structure. In this case, where
a support in three points is applied, in one point three degrees of freedom
have to be suppressed, in the second point two degrees of freedom and in
the third point one degree of freedom. In figure 26 the main structure with
the significant nodal points is shown.
The data input of this program concerns the loads applied to the nodal
points of the structure. These loads are based on the grinding forces at
maximum grinding power. As the workpiece, the grinding spindle and the
work head spindles are not included in the calculations, these forces have
to be replaced by the loads applied to the structure.

- 53 -
 FIGURE 24.
Substructure 4, the slide.

FIGURE 25.
Substructure 5, the drive screw.

- 54 -
FIGURE 26.
Main structure with significant nodal points.

55
 In figure 27 a top view of the grinding machine is shown schematically. In
this figure the loads are indicated, while the places where the grinding
spindle and the work head spindles are connected to the structure are shown
as points. Actually these points represent a number of nodal points. In
figure 28 the division of the load in each of these connections is shown,
From these data the deformations according to the six degrees of freedom
are calculated. The results of these calculations are obtained as a list
of numerical values of the deformations in each nodal point of the struc-
ture, In order to get a better insight in the deformation of the structure
as a whole and in the contribution of each substructure to this deformation
the deformations are plotted in the figures 29 and 30, The displacements
of the center lines of the grinding spindle and the work head spindles
are shown in figure 31, In these figures the deformation of the drive screw
is eliminated, as this deformation exceeds the deformations of other sub-
structures to such an extent that the figures do not show the deformation
in the substructures clearly. The deformation of the screw is 0,007 mm in
x-direction. The total deformation of the structure results in an error
in the diameter of the workpiece of 0.0045 mm and an error in axial di-
rection of the workpiece of 0.0105 mm. From these values it can be con-
cluded that a reduction of the thickness of the plates of the various sub-
structures is possible. At the same time the diameter of the drive screw
has to increase. Based on the deformations of the substructures as shown
in figure 29 and figure 30 the following values of the plate thicknesses
are chosen:

wheel head - 18 mm
work heads - 12.5 mm
bed - 16 mm
slide - 18 mm
drive screw - 16 mm (corresponding with a
diameter of 50 mm)

According to these values the calculations are carried out again. The re-
sult of these calculations is shown in the figures 32, 33 and 34. In these
figures the deformation of the drive screw is eliminated for the same
reason as in the previous figures. The value of this deformation is re-
duced to 0.0054 mm. The total deformation of the structure is now increased
in that way that an error in the diameter of the workpiece of 0,0075 mm ap-
pears, while the error in axial direction of the workpiece becomes
0,0107 mm at the grinding conditions mentioned before. If compared with the
demands it may be possible to reduce the stiffness of the structure further
specially with regard to the deformation in radial direction of the work-
piece, However, the deformation in axial direction is not allowed to in-
crease, Although it may be possible, by extended calculations, to obtain
a more optimum distribution of the deformations the cost of these calcu-
lations do not balance the savings that can be obtained. So it is decided
to apply the plate thicknesses mentioned above. If compared with the
original thickness of 25 mm the weight of the structure is reduced with
35 per cent. Not only the reduction of the cost of the machine but also
the saving in the material itself justifies the calculations carried out,
The complete programs and the data output of the computer calculations
are not incorporated in this thesis but they are available at the Univer-
sity of Technology in Delft,

56 -
WOKKHEM)

WHEELHEW P"»

FIGURE 27.
External loads, applied to the structure.

FIGURE 28.
Division of the loads in the connections
of the spindles to the structure.

- 57 -
 32 M 34 36

» I —

I ! .^ .^ •z^z

15 17 18

zr

10 4
X

I'/im

FIGURE 29.
Deformation of the bed in four
significant lines.

' -r-
J-

8 . I'

FIGURE 30.
Deformation of the slide at the
connections of the guideways.

58
 y^"
l^m

FIGURE 31.
Relative displacement of the grinding wheel
and the workpiece.

1 T
37 38

1—r zzi
26

1 :

13 17
HE;
irz

10

FIGURE 32. 1>tm

Deformation of the bed after changing 1>im

the plate thickness.

- 59 -
 X

FIGURE 33.
Deformation of the slide after changing
the plate thickness.

Ujim

^ym

FIGURE 34.
Relative displacement of the grinding wheel
and the workpiece after changing the
plate thicknesses of the structure.

- 60 -
Design of the work heads

In addition to the machine structure and the wheel head the work heads and
particularly the work head spindles determine the accuracy of the position
of the grinding wheel relative to the workpiece. According to the con-
siderations in chapter 4,1 the errors in the workpiece dimensions both in
axial and radial direction, caused by the deformations of the work head
spindles and the bearings, are not allowed to exceed a value of 0,005 mm
at maximum grinding forces. It has already been mentioned that both work
heads are almost identical. The difference concerns the drive of the
workpiece and the axial bearing of one of the spindles. Based on the same
reasons as mentioned with regard to the guideways of the slide and the
wheel head hydrostatic bearings will be applied for both work head spindles
An important advantage of these bearings is the possibility to move the
spindle in axial direction without special arrangements. As one of the
spindles has to move in axial direction in order to fix the workpiece
there is no need for any special design of the work head in question. Be-
sides the oil supply for the bearings can be used to apply the clamping
force for the workpiece. As the load and the speed at the work head
spindles are not such as to make it necessary to design special bearings
standard hydrostatic bearings will be applied. The dimensions of the
spindles and the bearings follow from calculations similar to those car-
ried out at the design of the grinding spindle. The results of these cal-
culations lead to the following data:

spindle diameter d = 105 mm

ws
bearing distance b = 250 - 350 mm
14
radial stiffness front bearing C = 150 - 10 N/mm

radial stiffness rear bearing C = 50 - 10^ N/mm

stiffness axial bearing C = 50 • 10^ N/mm

AA

- 61 -
7. DRIVE OF WHEEL HEAD, SLIDE AND WORKPIECE

The design of the drive system for the wheel head, the slide and the work-
piece is determined by the following requirements. First the grinding wheel
has to move according to the geometry of the workpiece in a direction pa-
rallel to the center line of the workpiece and in a direction perpendicular
to this center line. Secondly this movement and the rotation of the work-
piece have to be realized at a speed that is determined by the required
metal removal rate. Besides the ratio of the linear speed of the grinding
wheel and the circumferential speed of the workpiece has to meet the re-
quirements concerning the grinding forces as described in chapter 3. The
way in which the drive system will be designed depends on the types of pro-
ducts that have to be machined and on the number of products of each type,
As the products that will be machined, according to the considerations in
chapter 2,1, are generally manufactured in medium to small batches, the
system that generates the path of the grinding wheel relative to the work-
piece has to be flexible. This means that the information concerning the
geometry of the workpiece and the speeds of the wheel head, the slide and
the workpiece has to be provided in such a way that a change of this in-
formation related to a new workpiece can be achieved as simple as possible.
As besides during the grinding process corrections have to be carried out
in order to compensate the wear of the grinding wheel and as movements in
a direction perpendicular to the center line of the workpiece are realized
by movements of both the wheel head and the slide it is obvious to apply
a numerical control for the grinding machine. This means that the inform-
ation concerning the geometry of the workpiece, the speeds of the wheel head,
the slide and the workpiece are supplied to the control of the machine in
numerical form. During the process the speed of the workpiece is adapted in
such a way that the resulting force at the grinding wheel is perpendicular
to the center line of the grinding spindle while the position errors of the
grinding wheel caused by the wheel wear are corrected continuously. These
two control systems will be discussed in the next chapters,
Starting from the above considerations it has to be decided what type of
drive will be applied. First the drive of the workpiece will be discussed.
This drive has to achieve a rotation of the workpiece at a speed that va-
ries from 21 to 130 r.p.m. and a maximum torque of 125 Nm. The maximum power
for the workpiece drive is 3.4 kW. These data follow from the calculations
in chapter 3.2. The requirements for the drive can be met by both electric-
al and hydraulic motors. The possibility to incorporate a hydraulic drive
in the existing hydraulic system leads to the application of a hydraulic
motor controlled by an electro-hydraulic servo valve. The main requirement
for the motor concerns the stability at low speed while the stiffness of
the drive, mainly determined by the volume of oil under pressure of the
motor is a second important factor. On account of these demands a so called
roll-vane motor is chosen. An advantage is the fact that these motors are
available as a complete drive system, including the servo valve and the
tacho-generator that are built together with the motor,
Secondly the drives for the wheel head and the slide have to be designed,
These drives have to move the grinding wheel according to a well defined
path at a well defined speed. The drives for the wheel head and the slide
can be arranged in different ways. First it is possible to apply linear
actuators like hydraulic or pneumatic cylinders. Although these drives
are simple the control systems for the position and the speed are compli-
cated, specially as linear position measuring units are necessary as feed-
back elements, which increases the cost considerably. A second possibility
is the application of a screw drive with an electrical or a hydraulic mo-
tor, In this case rotating position measuring units can be used which re-

62 -
suits in a reduction of the cost. However, the accuracy of the position
measurement depends on the accuracy of the drive screw. Third it is possible
to apply stepping-motors and screw drives. In this case both the position and
the speed of the slides are controlled by controlling the number and the fre-
quency of the pulses supplied to the stepping motors. The accuracy depends
also on the accuracy of the drive screws. As the cost of the drive and the
control with electrical stepping motors is less than the cost of the other
drives mentioned above and as the risk of damage, caused by dirt from the
grinding process is relatively small because of the absence of precise mea-
suring units, it is decided to apply electrical stepping motors both for the
drive of the slide and the wheel head. Besides the application of electrical
stepping motors has another advantage, namely the fact that the complete
data handling from the input to the drive motors is in digital form. Be-
cause of this parts of the control system can be simplified, c,q, the part
that controls the movement of the grinding wheel in a direction perpendicular
to the center line of the workpiece. This movement has to be realized by
moving both the slide and the wheel head at a speed ratio that equals 2, This
ratio can be achieved by a fixed frequency ratio for the stepping motors
that drive the wheel head and the slide,
The choice of the power of the stepping motors and the design of the drives
depend on the load applied to the drive both as a result of the grinding
forces and as a result of the inertia forces caused by acceleration and de-
celeration of the slides. As the grinding forces in the direction of the
movement of the wheel head are eliminated the drive of the wheel head is only
determined by the inertia forces. The stepping motors that are considered
for the drives generally show a step angle of 1.8° and a maximum step fre-
quency of 5000 steps per second. By means of a so called half step drive
it is possible to drive the motors at a step angle of 0,9°,
First the drive of the slide will be considered. The grinding process requires
a maximum speed of the slide of 6,7 mm/s, while a rapid travel speed of
50 mm/s has to be realized. As this speed has to correspond with a maximum
step frequency of 5000 steps per second each step represents a displacement
of 0,01 mm. The accuracy that has to be obtained allows a smallest increment
of the positioning system of this value. In order to reduce the speed of the
rotating parts of the drive system, and in consequence the inertia forces,
a step of 0,9° is chosen. This means that one revolution of the stepping mo-
tor corresponds with a displacement of the slide of 4 mm. The screws that
will be applied are so called "roller screws". These screws consist of a
threaded screw shaft, an internally threaded nut and a number of threaded
rollers. The rollers are positioned in the space between the screw shaft and
the nut and mesh with the threads of both. The stiffness of these screws is
very high because of the large nunÜDer of contact points between screw, rollers
and nut. The diameter of the screw has already been determined, based on re-
quirements concerning the stiffness of the machine and will be 50 mm. Al-
though it is possible to choose a pitch of 4 mm for this screw it is decided
to apply a transmission with a ratio 1 : 3 and a pitch of 12 mm in order to
reduce the effect of the inertia of the screw. The transmission between the
motor shaft and the screw is realized by means of a toothed belt. These
belts provide an accurate and stiff transmission together with a good damp-
ing, specially with regard to the stepping effect of the motors at low speeds.
It is assumed that the diameters of the gears are 36 mm respectively 108 mm
at a width of 25 mm. After calculation of the power for the drives the di-
mensions of the transmission will be determined and a correction will be
carried out if necessary. The torque M required at the motor shaft can be de-
vided in two parts, one part M^, necessary to meet the grinding forces, and
another part M^ required to accelerate the slide. In practice the value of
Mg reaches its maximum value when accelerations of the slide are negligible
and also the maximum acceleration is required at a moment when the grinding

- 63 -
 process has not started. So the maximum value of M is determined by either
the maximum value of Mg or the maximum value of M^, The value of Mg can be
calculated from the following equation:
F^ - cos30° • s
M = : (7.1)

in which Fj^ - cos30° is that component of the grinding force that affects
the drive of the slide, s is the pitch of the screw and i the ratio of the
transmission between the motor and the screw. As the maximum value of F^
•n
equals 3000 N, the maximum value of Mg becomes:

M = 1,65 Nm
g max

The value of M follows from the equation:

M = I . 0) (7,2)
a
in which I represents the total moment of inertia of the slide, the screw,
the gears and the motor, determined at the motor shaft, while oi represents
the angular acceleration of the motor shaft. The moment of inertia I can be
calculated from:
I, + Ig s
I + I + + m -( )2 (7,3)
m o '2

where I^^^ is the moment of inertia of the motor, IQ and I. the moments of
inertia of the gears, Ig the moment of inertia of the screw and m the mass
of the slide including the wheel head. As the data of the motor are not
known a value of Ijj, = 1 - 10~3 kgm^ is chosen. If necessary a correction is
carried out when the exact data of the motor are determined. The moment of
inertia of a cylindrical part is calculated from:

I = 0,125 - m - d^ (7,4)
c
in which m is the mass of the part and d the diameter. The values of I , I,
J T o' 1
and I are:
s
I = 0.125 - 0.0712 - 0.035^ = 1.15 - 10~^ kgm^
o '^
I = 0.125 - 0,54 • 0,108^ = 9,4 • lo"^ kgm^

I = 0,125 - 15.3 • 0.05^ = 4.8 - 10~^ kgm^

^°= 3 . 5 - 74 . 10-3
I = 1 • 10 +1,15-10 + —
3^

+ 1200 - ('^ • ^°"')2 kgn,2

or:
I = 2,13 - 10 "^ kgm^-

64 -
The value of (i) is determined by the time or the distance in which the maximum
speed has to be reached. In this case the distance is assumed to be 4 mm at
a constant acceleration. This distance corresponds with a rotation of the mo-
tor shaft of 2 • TT radians. As this value has to be reached after t seconds:

2 • TT = 0,5 . (1) • t^

As the value of w after t seconds has to be 25 • TT rad/s the value of t

equals 0.16 s. This leads to an angular acceleration (I) = 500 rad/s^ and
from this value to:
-3
M 2.13 10 500 = 1.06 Nm,

If the efficiency of the drive is assumed to be 75 per cent the above calcu-
lations lead to the choice of a stepping motor with a maximum torque of
2,2 Nm at a speed range from 0 to 700 steps per second and a maximum torque
of 1,4 Nm over the whole speed range. A stepping motor that meets these re-
quirements is e,g, a "Slo-Syn" motor type M 112 - FD25, of which the relation
of torque and speed is shown in figure 35, The moment of inertia of this

6,
TORQUE
Nm
5.

I 1 1 1

500 1000 1500 2000 2500 3000 3500 4000 4 5 0 0 5000

SPEED (STEPS PER SEC)

FIGURE 35.
Relation of torque and speed of a
"Slo-Syn" stepping motor type M 112-FD 25.

-4 2
motor equals 8 • 10 kgm , A correction of the calculations according to this
value does not lead to a smaller motor.
According to the manufacturer's instructions the type of belt and the dimen-
sions of the gears are determined. It appears that a toothed belt with a
pitch of 5 mm and a width of 25 mm and gears with 16 teeth respectively 48
teeth meet the requirements based on the calculations above. Although the
dimensions of the gears are smaller than assumed before the decrease of the
total moment of inertia does not lead to the application of a smaller motor.
The design of the drive of the wheel head is based on the following consider-
ations. The grinding process requires a speed of the wheel head in a direction
perpendicular to the center line of the workpiece of 5 mm/s, while a rapid
travel speed of 25 mm/s will be applied. As this speed has to correspond with

- 65 -
 the maximum step frequency of 5000 steps per second the smallest increment
of the displacement is 0.005 mm. However, it appears that a smallest incre-
ment of at most 0.001 mm has to be achieved and that because of the follow-
ing. During grinding the grinding wheel wears and this wear causes an error
in the dimensions of the workpiece. In order to correct this wear the po-
sition of the wheel head has to be adapted to the new dimensions of the
grinding wheel. If the correction steps are too large marks or even mea-
surable steps on the workpiece surface will appear, A smallest step of
0,001 mm seems to be acceptable. As it is not possible to meet these re-
quirements with one stepping motor a second stepping motor is applied,
While the first motor drives the screw the second motor drives the nut.
This means that the absolute displacement of the wheel head depends on the
relative angular displacement of the screw and the nut. It has already been
mentioned that a movement of the wheel head perpendicular to the center line
of the workpiece is achieved by moving both the wheel head and the slide ac-
cording to a fixed ratio that is determined by the angle between the center
lines of the grinding spindle and the workpiece. As this angle equals 60°
a displacement of the slide of x mm and a displacement of the wheel head of
2 - x mm, relative to the slide, result in an absolute displacement of the
wheel head of x • /3 mm. As the input data for the drive system are related
to the absolute displacement of the wheel head a conversion of these data
to the numbers of pulses supplied to the stepping motors is necessary. In
order to realize an absolute speed of the wheel head of 25 mm/s the speed of
the wheel head relative to the slide has to be 28.87 mm/s. As it is not very
important to determine the rapid travel speed accurately the drive system
is designed as follows. One motor is driving the screw in such a way that
one step corresponds with a relative displacement of the wheel head of
0.005 mm. The second motor drives the nut in such a way that one step of
this motor corresponds with a displacement of the wheelhead relative to the
slide of 0.001 mm. If both stepping motors act together the rapid travel
speed becomes 30 mm/s. It is decided to drive both stepping motors at the
same frequency, while a correction of the position of the wheel head is
realized by supplying pulses to the stepping motor that drives the nut. The
power of the stepping motors is calculated as follows. During acceleration
of the wheel head in a direction perpendicular to the center line of the
workpiece the ratio of the accelerations of both the slide and the wheel head
relative to the slide is 1 : 2. In order to reduce the acceleration and de-
celeration time as much as possible the acceleration of the wheel head will
be twice the maximum acceleration of the slide. The acceleration of the
wheel head is composed of the two parts, one part achieved by the motor
that drives the screw, the other part by the motor that drives the nut. As
the acceleration of the slide ag = 0.3125 m/s^ the acceleration of the
wheel head will be a„ = 0.625 m/s2. The motor that drives the screw has to
accelerate the wheel head at _ . 0.625 = 0.52 m/s'^. In order to avoid a large
6
transmission ratio between the second stepping motor and the nut a pitch of
3 mm is chosen for the screw. At a step angle of 0.9° the ratio of the trans-
mission between the first motor and the screw equals 1.5. Based on the same
considerations as mentioned at the design of the drive of the slide a trans-
mission by means of a toothed belt is applied. The calculations of the moment
of inertia of the drive are based on the following assumptions. The diameter
of the screw will be 2 5 mm. Although the load allows a smaller diameter
this value has been chosen on account of secondary demands like mounting fa-
cilities for gears and bearings. The moment of inertia of the stepping motor
is assumed to be 2 - 10"^ kgm^. As the mass of the wheel head, including the
grinding wheel and the motor is 500 kg, the value of the total moment of
inertia I' at the shaft of the motor comes to 3 - 10~^ kgm^. The angular
acceleration follows from the value of the linear acceleration of the

- 66 -
wheel head and equals 1633 rad/s . The torque required to realize this acceler-
ation Ma' = 1633 • 3 - 10""^ = 0,49 Nm, A "Slo-Syn" stepping motor, type
M 093-FC 14 meets this requirement. The inertia moments of both the motor and
the transmission gears do not exceed the assumed values,
The motor that drives the nut has to achieve an acceleration of the wheel head
of i • a;^ = 0,1 m/s2. As one step of the motor corresponds with a displace-
6
ment of the wheel head of 0,001 mm and as the pitch of the screw equals 3 mm
the transmission ratio between the motor and the nut has to be 7,5 at a step
angle of 0,9°. The moment of inertia of the motor is assumed to be
7-10 kgm while the nut is regarded as a cylinder with a diam.eter of 50 mm
and a length of 100 mm, A calculation according to the equation 7.3 shows that
the total moment of inertia at the motor shaft I'' = 1,2 • 10""^ kgm^. As the
value of the angular acceleration is also 16 33 rad/s2 for this motor, the re-
quired torque M^'' = 0,2 Nm, A "Slo-Syn" stepping motor type M 063-FC 06
meets the requirements for this drive.
In the survey the relevant data of the drives of the work head spindle, the
slide and the wheel head are given. The arrangement of these drives is shown
in figure 36,

Work head spindle

Motor: Hydraulic "roll-vane" motor controlled by

electro-hydraulic servo valve
Speed: 20 - 130 r.p.m.
Maximum torque: 12 5 Nm

Slide

Motor: Electric stepping motor with step angle 0.9°

Speed: 0 - 5000 steps per second
Maximum torque: 2.2 Nm at 0 - 700 steps per second
1.4 Nm at 0 - 5000 steps per second
Screw: roller screw, 50 mm diameter and pitch 12 mm
Speed ratio motor screw: 3 : 1
Smallest increment: 0,01 mm
Speed of slide: 0 - 5 0 mm/s

Wheel head

Motor: Electric stepping motor with step angle 0,9°

Speed: 0 - 5000 steps per second
Maximum torque: 0,7 Nm at 0 - 5000 steps per second
Screw: roller screw, 25 mm diameter and pitch 3 mm
Speed ratio motor screw: 15- 1
Smallest increment: 0,005 mm
Speed of wheel head: 0 - 2 5 mm/s

b . Drive_of_the_nut
Motor: Electric stepping motor with step angle 0,9^
Speed: 0 - 5000 steps per second
Maximum torque: 0.2 Nm at 0 - 5000 steps per second
Speed ratio motor nut; 7.5 : 1
Smallest increment: 0.001 mm
Speed of wheel head: 0,5 mm/s
The control of the drives and the way in which the position corrections are
carried out are described in jii'.'^'^er 9.

- 67 -
 \n]}(/n/////////7h
MB •

SLIDE

Mwp WORKHEAD

FIGURE 36.
Drives of the slide, the wheel head
and the workpiece.

68 -
APPARATUS FOR DRESSING THE GRINDING WHEEL

Wheel wear and dressing conditions

The grinding process as described in the chapters 2 and 3 involves a con-

siderable wear of the grinding wheel. At high metal removal rates the
grinding ratio G may even reach a value of 4 [lO]. This means that during
the grinding process the position of the grinding wheel has to be corrected
in order to eliminate the errors in the dimensions of the workpiece caused
by the wheel wear. Besides it will be necessary to correct the shape of
the grinding wheel as the wear is not equally distributed over the wheel
surface. This correction will be carried out by dressing the grinding wheel
with a diamond dressing wheel. The reason why a diamond wheel is applied
is the fact that stationary dressing tools need an interruption of the
grinding process. This interruption, together with the time needed for
the dressing operation itself, results in a considerable increase of the
machining time. Although it is not known how the dressing process at cir-
cumferential speeds of the grinding wheel up to 150 m/s will proceed, the
research carried out on this subject [l5, 16, 17] gives an indication how
to choose the dressing conditions. This research did concern in the first
place the influence of the speed of the diamond wheel with regard to the
speed of the grinding wheel. It appeared that at high values of the rela-
tive speed between the diamond wheel and the grinding wheel the influence
of the directions of the circumferential speeds of the diamond wheel and
the grinding wheel is not significant. However, at high relative speeds
the roughness of the dressed wheel surface decreases, which results in a
smaller value of the metal removal rate than can be realized. In order to
reduce this effect it is possible to choose a wheel with a smaller number
of active cutting edges, which means a smaller diamond concentration,
Based on these facts a speed at the circumference of the diamond wheel
adjustable from 10 to 50 m/s is chosen. In the area of contact the di-
rections of the speeds at the circumference of both wheels are the same,
resulting in a relative speed of 100 - 140 m/s. In order to determine the
depth of cut of the diamond wheel a^ in mm per revolution of the grinding
wheel it is necessary to consider the wear of the wheel. In figure 37 the

FIGURE 37.
Progress of the wear of the grinding wheel.

- 69
 progress of the wear at the surface of the grinding wheel is shown. This wear
starts during the first revolution of the workpiece at the edge of the wheel
and finally generates a shape of the wheel as shown in stage d. If only a
cylindrical surface has to be machined the wear at the part A of the grind-
ing wheel is important. In this case the correction can be little as the
total wear is distributed over a wider part of the grinding wheel. However,
the dressing conditions will be based on the requirement that the original
shape of the grinding wheel has to be maintained continuously. In chapter 3
it has been determined that at the wheel surface Ap the maximum value of the
specific metal removal rate z'ujax = 2000 mm3/mm,s. This means that at this
part of the grinding wheel the maximum wear per mm width of the wheel
rl max ~ ^ max/'2' ^^ "^^^ minimum value of G comes to 4 the maximum value of
= z'
the specific wear w'jj,^j^ = 500 mm /mm.s. As the circumferential speed of the
grinding wheel Vg = 150 m/s the depth of cut of the diamond wheel a^j has to
be at least w'^^^/v^.
luax o So the infeed of the diamond wheel has to be such as
to realize a value of a^ max - 0.0033 mm. In order to know the power that is
required for the drive of the diamond wheel it is necessary to determine the
tangential force at the circumference of the diamond wheel. From research
concerning dressing with diamond wheels [l8] only data concerning radial
forces are known with sufficient reliability. However, the relation between
radial and tangential forces has been determined at single point diamond
dressing tools [l9]. As the shape of these diamonds does not differ very much
from the shape of the diamonds applied in diamond wheels this relation will
be used for the calculation of the tangential forces at the circumference of
the diamond wheel. Also in this case an extrapolation has to be carried out
as data concerning the process described in this thesis are not known. Based
on the results of the research mentioned above it is assumed that the
minimum value of the specific tangential force F'^^-f- = 2 N/mm at a specific
radial force F' ^-p = 10 N/mm. In figure 38 the position of the diamond wheel
relative to the grinding wheel
is shown. From this figure it
appears that at the maximum
depth of cut of the diamond
wheel a,d max at the surface
Ap, the depth of cut at the
surface A^ a'^ ^^^ -
a^ njaj^/tg30O = 0,0057 mm. As
both the radial and tangential
dressing forces are proportional
to the depth of cut the values
of these forces come to
dta 3,4 N/mm and
F'(;jpa ~ 1"^ N/mm. As, according
to chapter 3, the width of the
wheel surface Ap equals 12 mm
and the width of the surface
FIGURE 38. Aa equals 16 mm the value of the
Grinding wheel and diamond resulting tangential force
wheel during dressing. Felt = 78 N and the value of
the resulting radial force
Fdr = 295 N,
It is obvious that the above calculations are related to the most unfavour-
able situation. In cases where the amount of grinding wheel material that
has to be removed reaches the values mentioned above it is advisable to
apply a smaller v/heel width. Besides at high metal removal rates a substantial
part of the grinding wheel material is removed by the grinding process itself.
Nevertheless the design of the dressing apparatus will be based on the forces

- 70 -
and speeds calculated above, in order to be sure that extreme conditions
can be realized if necessary. A disadvantage of the method of continuous
dressing is the fact that the drive of the grinding wheel has to provide
also the power for the dressing process. This power P^j can be expressed by
the relation Pd = T^-^ • Vg. However, the tangential dressing force F(jt
seldom reaches a value that exceeds 25 per cent of its maximum because of
the fact that in extreme situations a smaller wheel width will be applied
and the fact that in most cases the dressing process has to correct only
a part of the surface of the grinding wheel. The maximum power for the
drive of the diamond wheel P;^;^ = F^j-j- jjjax • v^ max» i^ which v^j represents
the circumferential speed of the diamond wheel. So P^w - ^ ^^' I^^ order
to keep the speed of rotation at a low value and on the other hand to ob-
tain a longer life of the diamond wheel the diameter of this wheel will be
200 mm. This means that the drive of the diamond wheel has to provide a
maximum speed of 4800 r.p.m. at a maximum torque M.^^^ = 7.8 Nm,

Design of the dressing apparatus

From the first section of this chapter it appeared that the diamond wheel
has to be driven at a speed of 4800 r.p.m. A direct drive without a trans-
mission is preferred in order to avoid vibrations of the transmission at
the high speeds that are applied and to reduce the mechanical losses of the
total drive. As the dressing apparatus has to be mounted on the wheel head
at the circumference of the grinding wheel it is important that both the
weight and the dimensions of this unit are as small as possible. For this
reason and because of the existence of a hydraulic system in which this
drive can be incorporated a hydraulic motor is chosen. One of the very few
motors that can meet the requirements is the "Vickers" axial piston motor,
type MF 3911-25. This motor provides a torque of 12 Nm at a maximum speed
of 5400 r.p.m. and an oil flow of 54 1/mm. Although the control of the
speed of the diamond wheel does not make high demands upon the components
of the control system, an electro-hydraulic servo valve will be applied and
that by the following reason. Each diamond wheel will show errors in round-
ness. These errors may be transferred to the grinding wheel. If the ratio
of the speeds of rotation of the grinding wheel and the diamond wheel is a
whole nuii±>er the surface of the grinding wheel will show wavings that lead
to vibrations during grinding. In order to avoid this fenomenen a control
of the speed of the diamond wheel is necessary as the speed of rotation of
the grinding wheel varies between 5700 and 10.000 r.p.m. It is obvious
that the accuracy of the bearings that support the spindle to which the
diamond wheel is mounted has to meet requirements of the same kind of those
related to the bearings of the grinding spindle and the work head spindles.
Based on this fact and on the existence of a hydraulic system suited to
supply the oil for the bearings hydrostatic bearings will be applied. In
this case standard bearings meet the requirements both concerning accuracy
and stiffness.
The dressing head has to move in a radial direction relative to the grind-
ing wheel. In this case also hydrostatic guideways are applied in order
to obtain an accurate movement of the diamond wheel and particularly to
provide maximum damping with regard to vibrations caused by the grinding
process itself or by unbalance of the rotating parts, specially the grind-
ing wheel. As the dressing process is carried out at that speed of the
grinding wheel at which the grinding process takes place and as the dress-
ing apparatus is mounted to the wheel head the position of the dressing head
relative to the grinding wheel does not influence the roundness errors
caused by unbalance forces. For this reason the dressing apparatus is
mounted opposite to the workpiece in order to reduce the deformation of

- 71 -
 the grinding spindle caused by the normal grinding force Fn- The movement
of the dressing head is realized by means of a screw driven by an electric
stepping motor. It is obvious to apply this type of motor as this drive
has to be incorporated in the system that determines the position of the
grinding wheel relative to the workpiece. As the maximum depth of cut of
the diamond wheel ad max - 0.0033 mm the maximum speed of the dressing
head v^p jjiax " ^d max ' ^s max/^*^ ~ 0.525 mm/s. According to the smallest
increment of the displacement of the wheel head the smallest increment
of the displacement of the dressing head is also fixed at 0.001 mm. This
means that at 400 steps per revolution of the stepping motor and a screw
pitch of 1 mm a transmission between stepping motor and screw with a
ratio 2'5 has to be applied. As in this case accelerations and deceler-
ations are not of interest the power of the stepping motor is determined
by the radial dressing force F^p. The torque of the motor M^ follows from
the following relation:
F •s
M, = ^^ (8.1)
d

in which s represents the pitch of the screw and i the ratio of the trans-
mission between motor and screw. So M, = 0.02 Nm. In this drive a "Slo-Syn
motor type M061 - FC02 will be applied.
The complete control of the grinding machine, including the control of the
dressing apparatus, will be described in the following chapter.

72 -
THE CONTROL OF THE GRINDING MACHINE

The control of the grinding machine can be divided in two parts, one part
related to the grinding process itself, the second part related to the
generation of the geometry of the workpiece.

Control of the grinding process

The control of the grinding process concerns in the first place the con-
trol of the circumferential speed of the grinding wheel. This speed is de-
termined by the oil flow to the hydraulic motor that drives the grinding
spindle and by the diameter of the grinding wheel. In figure 39 the sim-
plified block diagram of this control is shown. The value of the diameter

servo- motor
O^ pump
valve

tacho
generator

FIGURE 39.
Simplified block diagram of the speed control
of the grinding wheel.

of the grinding wheel dg is derived from the control of the dressing appa-
ratus as will be shown in the discussion of this part of the control.
Secondly the control of the grinding process concerns the elimination of
grinding forces perpendicular to the plane of the grinding wheel. In fi-
gure 5 in chapter 3 the relation between the depth of cut a and the feed s
at different values of the diameters of the grinding wheel and the work-
piece is shown. At the values of a and s that meet this relation the
grinding forces perpendicular to the plane of the grinding wheel are elimi-
nated. As the value of a is determined by the dimensions of the unmachined
workpiece and the final geometry that has to be realized the only control-
able parameter is the feed s. The value of s depends on the speed of ro-
tation of the workpiece n^^ and the speed of the wheelhead in axial direct-
ion of the workpiece va- As a change of the value of n^ does not affect
the value of the metal removal rate it is decided to apply a control of
the speed of rotation of the workpiece in order to keep the direction of
the resulting grinding force perpendicular to the center line of the
workpiece. It is obvious that it is necessary to apply a device that de-
termines whether the requirement concerning the direction of the grinding
force is met. As this requirement means that the load in axial direction
of the grinding spindle equals zero this device can be a transducer that
measures the load applied to the axial bearing of the grinding spindle.
Because of the application of hydrostatic bearings this load can be de-
termined from the difference in pressure in the opposite chanÜDers of the
hydrostatic bearing. As the only function of this axial bearing concerns

- 73
the location of the grinding spindle in axial direction the pressure in
the oil chambers can be very little. For this reason the oil that flows
out of the radial front bearing is led to the chambers of the axial bearin
and from there through oil gaps coTnbined with two adjustable restrictors
as shown schematically in figure 40, The restrictors are applied to adjust

FIGURE 40.
Combined radial and
axial spindle bearing.

the pressure in the chambers of the axial bearings more accurately by

choosing the correct dimensions of the bearing gap. As the oil supply to
the radial bearings is realized by means of one pump for each chairÜDer the
properties of these bearings are hardly affected by these axial bearings.
In figure 41 a simplified block diagram of the control of the direction of

workpiece process data

control process
An drive s +As

pressure axial
transducer bearing axial grinding
AP
force

FIGURE 41.
Simplified block diagram of the
control of the grinding forces.

the grinding forces is shown.

Control of the grinding wheel position

The geometry of the workpiece is determined by the path of the grinding

wheel relative to the workpiece and by the shape of the grinding wheel.
In chapter 2 this shape of the grinding wheel has already been determined,
The relative positions of the workpiece, the grinding wheel and the
dressing wheel are shown schematically in figure 42.
 FIGURE 42.
Relative positions of grinding wheel,
dressing wheel and workpiece.

From t h i s figure it a p p e a r s that three parts o f the surface o f t h e g r i n d -

ing w h e e l a r e o f major interest with regard t o the geometry o f t h e w o r k -
p i e c e . In t h e first place the position o f part A determines t h e diameter
of t h e w o r k p i e c e . Secondly t h e part B determines t h e dimensions of t h e
w o r k p i e c e in a x i a l direction and finally the part C determines the shape
of t h e change o f a cylindrical part o f the workpiece into a plane p e r -
p e n d i c u l a r t o the center line of this w o r k p i e c e . According t o this it is
possible t o consider three c a s e s :
First t h e grinding wheel is machining a cylindrical part of t h e w o r k p i e c e ,
In t h i s case only the position o f part A o f t h e grinding w h e e l is of i n -
t e r e s t . According to figure 4 2 this position is determined b y t h e value
of y/^. This value can b e expressed b y :

(H sin(() (9.1)
^d^

in which y represents the y-co-ordinate of the center of the grinding

wheel 0, H the center distance of the grinding wheel and the dressing
wheel and R^ the radius of the dressing wheel as indicated in figure 42.
Secondly the grinding wheel is machining a plane perpendicular to the
center line of the workpiece. This means that only the position of part B
of the grinding wheel surface is of interest. This position is determined
by the value of xg, which follows from:

(H - R^) coscj) (9.2)

^B 0

Thirdly the grinding wheel is machining the change of a cylindrical part

into a plane perpendicular to the center line of the workpiece. In this

- 75
 case the position of part C as determined by the co-ordinates xg and y^ is
important. The reason to consider these three cases separately is the
following. If a cylindrical part is machined the value of y^ has to be
kept constant and independant of the wear of the wheel or the dressing,
At the end of this chapter it will be shown that the part A will always
keep its correct shape by means of the dressing operation. This means that
a change of the center distance of the grinding wheel and the dressing
wheel AH has to be followed immediately by a displacement of the center 0
of the grinding wheel Ayo = AH • sincj). In order to keep this correction
procedure as simple as possible the displacement Ayg is realized by a
displacement Ayg/coscj) of the wheel head relative to the slide. Although this
leads to a displacement of the point 0 in the x-direction it is sufficient
to correct only the data concerning the dimensions in axial direction by
a displacement of the origin of the co-ordinate system, a so called "zero
shift". In the case of machining a plane perpendicular to the center line
of the workpiece the same method of correction is applied. This means that
the displacement of the point 0 in x-direction A X Q = AH • cos(f) is realized
by a displacement Axg/sincj) of the wheel head together with a zero shift
in y-direction. Finally the machining of the change of a cylindrical part
of a workpiece into a plane perpendicular to this cylinder needs a dressing
operation that provides a correct shape of the part C and a simultaneous
correction of the position of the point 0 in x- and y-direction. The con-
trol of the movement of the grinding wheel can be regarded as a control
of the movement of the center 0 of the grinding wheel if the correction
procedure is carried out properly. This means that it is possible to
start from a standard two-axes numerical control as it is widely used at
numerically controlled turning machines. As mentioned before the shape of
the grinding wheel is corrected continuously by means of a dressing opera-
tion. However, it is necessary to determine the value of the displacement
of the dressing wheel AH that is necessary to achieve the correct shape
of the parts A, B or C of the grinding wheel. This means that this shape
has to be determined by measuring. At the University of Aachen in Germany
a method for measuring wheel wear during grinding has been developed [20],
This method uses a pneumatic device as shown in figure 43. By measuring

"o

Pi FIGURE 43.
Pneumatic device for the measure-
ment of the wear of the grinding
wheel.
^y////-

the value of the air pressure p-]_ the value of the height of the gap be-
tween the nozzle and the wheel surface x can be determined with a maximum
error of about 0.001 mm. In figure 44 the positions of the measuring de-

76 -
 FIGURE 44.
Positions of the
measuring devices.

vice with regard to the grinding wheel are shown. It is obvious that ac-
cording to the parts of the grinding wheel that determine the shape of the
workpiece three measuring devices are applied. These devices are placed
quite near the place where the dressing wheel touches the grinding wheel.
The value of AH will be determined as follows. Depending on the shape of
the workpiece that is machined one of the measuring devices is considered.
First the dressing head is moved towards the grinding wheel until the gap
between the wheel surface and the nozzle reaches a certain value and this
value is kept constant during grinding. This means that the dressing head
has to be moved if the wear of the grinding wheel does affect the value of
the gap. In this way an accurate control of the geometry of the workpiece
can be achieved,

77 -
THE POWER SUPPLY OF THE GRINDING MACHINE

The power supply for the grinding machine concerns in the first place the
hydraulic system for the drives of the grinding spindle, the workpiece and
the dressing wheel and the oil supply for the hydrostatic bearings. Second-
ly an electric system is applied with regard to the electric motors that
drive the pumps and the power supply for the drive of the stepping motors.
This second part of the power supply is left out of consideration as only
simple a.c.-motor drives are involved. Besides the power supply for the
stepping motors is designed by the manufacturer and is included in the
control system for these motors. However, the complete hydraulic system
has to be designed based on the specific requirements of the machine. In
figure 45 this hydraulic system is shown schematically.
The first part of this system concerns the drive of the hydraulic motor
for the grinding spindle. The oil for this motor is supplied by a variable
displacement axial piston pump PVl controlled by a servo mechanism and an
electrohydraulic servo valve SVl. As this system is a so-called closed
system it is necessary to compensate the leakage flow of both the motor
and the pump by means of a boost pump P, This pump provides also a cooling
of the oil in this system by means of a flow of oil through the pump
housing and a cooler. The oil supply for the servo mechanism is achieved
by a pump PV3 that forms part of a second system. This part of the hy-
draulic system supplies also the oil for the drive of the work head, con-
trolled by the servo valve SV2, and the drive of the dressing wheel, con-
trolled by the servo valve SV3. Besides the oil for the hydrostatic bear-
ings of the work head spindles and the hydrostatic guideways of the wheel
head, the slide and the dressing head is supplied by this pump PV2, This
means that this pump has to maintain a constant pressure in this part of
the hydraulic system. The third part of the hydraulic system concerns the
oil supply for the hydrostatic bearings of the grinding spindle and the
dressing spindle. This part of the system has to be arranged separately as
for this application a very low viscosity of the oil is required with re-
gard to the heat generation at high spindle speeds. Two so-called multiple-
pump units PV4 and PV5 are applied in order to supply the oil to the eight
chambers of the bearings of the grinding spindle and to the bearings of
the dressing spindle. It is obvious that in all the systems a very effect-
ive filtering of the oil is necessary. Besides cooling systems will be
applied in order to remove the heat generated in the bearings, the pumps
and the motors. Based on the calculations of the grinding power, the
dressing power and the power for the hydrostatic bearings the total power
for the grinding machine will be about 225 kW, taking into account the
efficiency of the pumps and the electric motors,
 GUIDEWAYS OF
THE DRESSING HEAD MOTOR OF
GRINDING
MOTOR OF MOTOR OF SPINDLE
DRESSING HEAD WORK HEAD
11
GUIDEWAYS OF
THE WHEELHEAD

^--s
i — ^
»t
T<M GUIDEWAYS OF [JJ-NAA -<^
THE SLIDE
WN|7 5
<H -<XH ;;r~^^^-N;f Nk

L>-^^w>-A (?)-txy
I ^ J

cz>" TT

JL.
w
EfE
)i )
IZ
ll(
^ )
IZ
I : I -'j'
SV3 SV2 BEARING OF THE SVl
WORKHEAD SPINDLES

g>-£Xh

r\
WNl L.i

I j
¥^
Lti3
BEARINGS OF
DRESSING SPINDLEI

' I

FIGURE 45.
Hydraulic system of the gHnding machine.

- 79
11. SECONDARY ASPECTS OF THE DESIGN

11,1. Safety

One of the most important aspects of the application of high speed grind-
ing concerns the safety of the machine in the event of a wheel bursting.
In many institutes extended research has been carried out to develop
wheel guards that offer satisfactory protection against operator injury
and machine damage caused by fragments of a burst grinding wheel [21,
22, 23, 24],
The design of the grinding machine does already include a number of ad-
vantages with regard to the safety. First the machine will be closed com-
pletely when operating. Even inspection windows are avoided as these
windows do not provide any facility for inspection of the grinding pro-
cess. Secondly the plane of the grinding wheel is almost parallel to the
front of the machine which enables an effective protection against frag-
ments of a wheel leaving the wheel guard. Thirdly the width of the grind-
ing wheel is only 20 mm which means that the kinetic energy of the ro-
tating grinding wheel is relatively small.
Starting from these facts the following measures are taken. At first a
heavy wheel guard is applied as shown in figure 46. In order to absorb the

s
ly/yy/

/
/
/
/
1 \\ /
/
/
//
/
/
J Y /
/
\ /
\ /

FIGURE 46.
Mf X

Wheel guard with energy absorbing wedge.

energy of burst grinding wheels two plates forming a wedge-shaped opening,

between which fragments of a burst grinding wheel can be caught, are
applied. The plates are preloaded by means of diaphragm springs as shown
in the figure. At the front of the wheel guard there has to be an opening
in order to allow the grinding wheel to meet the workpiece. This opening
is determined by the dimensions of the largest workpiece. This means that
an additional protection from fragments which would escape from this open-
ing is required. As a consequence special movable guard plates are arranged
direct behind the workpiece. The guards have to be movable as from the

81
 frontside of the machine the workpiece has to be loaded, A second measure
that is related to the prevention against the bursting of a grinding wheel
concerns the shape of the grinding wheel. It has been calculated [25] and
measured [22] that a small inner diamieter of the grinding wheel reduces
the stresses in the grinding wheel caused by centrifugal forces. These
stresses reach a minimum value if no central hole is applied. So it is
decided to use grinding wheels as shown in figure 47 provided with a num-
ber of small holes arranged at a circle. The wheels are mounted by means

FIGURE 47.

I Grinding
central
wheel
hole.
without

of a number of bolts that keep the flanges together as shown in the figure.
It is obvious that a testprogram has to be carried out in order to investi-
gate the measures that are taken with regard to the safety of the operator
and the risk of damage of the machine,

1.2. Coolant-supply

A considerable part of the power supplied to the grinding process is trans-

formed into heat. As the heat generation increases at high grinding speeds
the problems related to an effective cooling are also increasing. One of
the main problems at high circumferential speeds is the presence of an air
film around the wheel which has to be penetrated before the coolant can be
introduced into the grinding zone. Research on this subject [25, 26, 27]
leads to the application of high pressure fluid delivery systems and
scrapers that could remove the air layer. However, in consequence of the
wheel wear these scrapers need to be adjusted frequently. For this reason
an air nozzle is applied providing an air jet tangentially along the wheel
circumference in a direction opposite to the wheel rotation as shown in
figure 48, This system has proved to be satisfactory at high grinding

FIGURE 48.
Coolant supply to
grinding zone.

82 -
speeds [27]. A second coolant supply is necessary at the dressing zone. In
this case a scraper plate will be applied as this plate can be mounted to
the dressing head and in this way follows the wheel circumference. A fa-
vourable aspect with regard to the cooling is the fact that at high metal
removal rates a substantial part of the heat is removed by the chips as it
has been investigated at the research concerning cut-off grinding at the
University of Delft. It is obvious that an effective filtering of the
grinding fluid is necessary in order to prevent strange particles entering
the grinding zone.

- 83 -
12. THE DRAWINGS OF THE GRINDING MACHINE

The design of the grinding machine based on the considerations and the
calculations in the preceding chapters is shown in a number of drawings.

Drawing A shows a section of the grinding spindle, including the bearings,

the motor and the grinding wheel. The flow of oil and coolant to the bear-
ings is shown schematically. The coupling between the motor and the
spindle is a standard type, although a modification is applied.

Drawing B shows three views of the dressing apparatus with the stepping
motor drive.

Drawing C shows a horizontal section of the wheel head. The wheel guard
covers both the grinding wheel and the dressing wheel. The dressing unit
is enclosed within a cover that is indicated by a -•-.- line.

Drawing D shows a front view of the wheel head. In order to show the de-
sign of the wheel head more clearly the wheel guard is left out partially.
At the circumference of the wheel guard a part of this guard can be re-
moved in order to mount the grinding wheel.

Drawing E' shows a vertical section of both work heads. The driven work head
is provided with a drive motor and an axial spindle bearing. The second
work head is constructed in such a way that the oil that leaves the rear
bearing is used to apply the axial clamping force to the workpiece.

Drawing F shows a cross section of the work heads and also the device that
clamps the work head to' the bed.

Drawing G shows the drives of the wheel head and the slide.

Drawing H shows the front view of the grinding machine. In this case the
cover of the machine is indicated by a -.-•- line.

Drawing I shows the side view of the machine.

Drawing J shows the top view of the machine.

84 -
 ^ ^

::3:;B" ' H

HYDRAULIC MOTOR

loomm
Vl/

DRAWING A: Section of the grinding spindle.

5 -
 mm^.

r^
K?
^Vl
• / / .

\ 777-

ifii

LAI

DRAWING B: The dressing apparatus. - 87

 100 m m

/
V

K>
-1:

DRAWING C: Horizontal section of the wheel head.

ioomm
'pvavi laaiicn. ai[q Jo naia quoa^ :a DNIMVHQ

lUUIOOI.
 100 m m

DRAWING E: Vertical section of the work heads.

- 93
DRAWING F: Cross section of the work head.

- 95
 R^^^

WHE EL H E A D DRIVE

J L

ioomm
SLIDE DRIVE

DRAWING G: The drives of the wheel head and the slide.

 500 m m

DRAWING H: Front view of the grinding machine.

- 99
 500 mm

DRAWING I: Side view of the grinding machine.

101 -
 soomm

DRAWING J: Top view of the grinding machine.

103 -
CONCLUSION

In this thesis the main aspects of a design of a high speed cylindrical

grinding machine are described. However, this design has to be regarded as
a first step of the development of a grinding machine suited for industrial
production.
In the first place a considerable part of the data concerning the grinding
process and the dressing operation can only be determined by realizing
these processes. These data will lead to a better understanding of the
criteria for a final design. Secondly a study of the economics of the pro-
cess has to be carried out. As it has been mentioned in the first chapter
the design is based on a certain type of product. So it is necessary to
know data concerning the various products of this type, the quantities of
each product, the quantity of metal that has to be removed and so on.
Based on these data and the data concerning the manufacture of the grinding
machine a correct choice about the main dimensions and the power of the
machine can be made.
The grinding machine, built according to the design described in this
thesis, offers the possibility to study both the process itself and the
economics of the process. This study may lead to a final design of a
grinding machine that provides both a high metal removal rate and a high
geometrical accuracy of the workpiece in one operation at reasonable cost.
 Appendix

A, Deformation of the grinding spindle

The basis for the calculation of the dimensions of the grinding spindle and
the stiffness of the spindle
bearings is the deformation
that is allowed with regard
to the accuracy of the work-
piece, In figure A,l the
forces applied to the grind-
ing wheel and the grinding
spindle are shown. As it
has been considered in chap-
ter 4.1 only the normal
grinding force F^^ is taken
into account. This force
causes the forces F^ and
Fy in the bearings of the
spindle. The values of these
forces depend on the dis-
tance between the bearings
b and the distance from the
front bearing.
FIGURE A.1.
Grinding and bearing forces
applied to the grinding spindle.

In figure A.2 the deformations of the grinding spindle and the spindle bear-
ings are shown. These de-
formations result in a dis-
placement of the point P
of the grinding wheel in x-
direction fp^ and a dis-
placement in y-direction
fpY, From the demands con-
cerning the accuracy of the
workpiece it follows:
f ^ < 0.01 mm and

V < 0,005 mm
The displacement of P can be
separated in two parts, one
part represented by dp and
dp', caused by the deform-
ation of the bearings d^ and
dy and one part represented
by gp and gp' , caused by the
deformation of the spindle,
The deformation of the grind-
ing wheel is not taken into
FIGURE A.2.
account in the calculations.
Deformation of the grinding The value of di and dp' can
spindle. be calculated as follows:

106 -
The displacements at the bearings are:
F^ Fy
d, = — and d,, = —
A c^ V Cy

where c^ and cy are the bearing stiffnesses. The linear displacement of the
point B of the spindle:

^B = (^A ^ dy) . ^ - d^
and the angular displacement in B:
__ £A_Liv
So the displacements in x- and y-direction of the point P caused by deform-
ations of the bearings are:

X "^A "^ '^V ^"^^

d = • R • sina + ( • (d +d.) - d ) • cosa (A.1.)
P b b " A A

a+b ^A + '^v
d y = ( - (d.,+d.) - d.) - sina - -^ • R • cosa (A.2.)
^ b b

where R represents the radius of the grinding wheel and a the angle between
the grinding spindle and the y-axis. The value of gp, caused by the deform-
ation of the spindle, is calculated as follows:
The linear displacement of the point B of the spindle:
F -b^.a F -a3
A n
+
'^B 3.E-I 3-E-I

and the angular displacement in B:

2 2
F -b"^ F -a
^1 A n
è = +
^ B 3-E.I 2-E-I
where E is the Young's modulus of the spindle material and I the moment of
inertia of the spindle. The displacements in x- and y-direction of the point
P, caused by deformation of the spindle, are:
9 9 9 3
F -b F -a F -b -a F -a
X / A n s „ . , A _n s /» -^ \
gpy = ^J^^
.A -^ ^TiTT^
n , • ^•^^^'^
. -" .^A 3.E-I ^n T^i^i)
. „ • cosa /(A.3.)
„,\
9 3 9 9
F -b -a F -a F -b F -a
From t h e e q u a t i o n s A . l t o A.4 i t follows:

f/ = F^ • R { ^ + (1+^) • ^ - ^ + ^ ^ ^ + - | — } - s i n a +
P n ,2 b b-c„ 3-E'I 2*E*I
b -c. " V„

2 3
T^ ra+b //„a, 1 a s a a-b a ,
+ F • {-— - ( ( 1 + - ) - —+ ) - + ^ „ ., + „ „ }-cosa
n b b e b'C;^ ^''-^A 3-E'I 3-E-I

(A.5.)

- 107 -
 2 3
^ V „ ra+b,,, a, 1 a . a a-b a ,
P n b b e b ^A A 3-E-I 3-j:,-I
a a 1 a-b a^
- F^-R-{-:^ + (l+r) - T-—- = ^rrr^ + ^r-r—} . cosa (A,6,)
^ b'^-Cfl •'-' b-cy 3-E-I 2-E-I

The value of a is chosen as small as possible and equals 100 mm, while the
demands concerning the accuracy of the workpiece are related to the maximum
grinding forces, so F^-j = 3000 N. By means of a computer program the optimum
bearing distance b is calculated at different values of the spindle diameter
D and the bearing stiffnesses c^^ and Cy. Based on these calculations the
following data are determined:

150 10^ N/mm

^V =
= 75 •10^ N/mm
^A
D = 120 mm

b = 250 - 325 mm

- 108 -
Calculation of the hydrostatic bearings

The grinding spindle is supported in two radial bearings and one axial bear-
ing. As there is no load in axial direction the axial bearing only has to fix
the position of the spindle. In the description of the control of the machine
it is shown that the axial bearing is used as a measuring device for the de-
tection of a load perpendicular to the plane of the grinding wheel. Conse-
quently the demands for this bearing are related to this control system and
will not be considered in this appendix. The design of the bearings is based
on the following data:

stiffness of the front bearing cy 15-10 N/m

maximum load at the front bearing Fy 4000 N
stiffness of the rear bearing c^ 7,5 • 10^ N/m
maximum load at the rear bearing FA 1000 N

The data concerning the stiffness of the bearings follow from the calcu-
lations of the spindle in the appendix A, From these calculations it also
follows that the spindle diameter has to be at least 120 mm. As the influence
of the diameter of the spindle at the rear bearing on the deformation of the
spindle at the grinding wheel is very small, a smaller diameter of the
spindle at the rear bearing can be chosen. One of the most important criteria
for the bearings concerns the heat generation in the bearings. For this rea-
son the oil-supply to the bearings will be realized with one pump for each
bearing chamber. In figure B.1. the hydrostatic bearing is shown schematic-
ally, A system with four bearing chambers is chosen in order to reduce the
influence of the direction of the load on the stiffness of the bearing that

FIGURE B.1.
Schematic view of a
hydrostatic bearing.

should appear at three chambers, On the other hand an increase of the number
of chambers increases the friction in the bearings 128], The calculations of
the bearings are based on the following relation between the oil flow Q
through a gap with a height h, a width b and a length 1, caused by a pressure
difference Ap:

109
where n is the viscosity of the oil. Besides the most unfavourable direction
of the load is chosen, this means the direction towards the gap between two
chambers. In each bearing chamber the oil flow that enters the bearing equals
the flow that leaves the bearing. In chamber 1:

Qo = Qi + Qi-2 + Q2-4 (B.2,)

where Qg is the oil flow from the pump to the bearing chamber, Qi the oil flow
that leaves the bearing chamber in axial direction and Q1-2 and Q1-4 the
oil flows that leave the bearing chamber in tangential direction to the cham-
bers 2 and 4. Because of the symmetry of the load Q1-2 = 0, According to the
equation B.1,:
TT
,2
Pi D 3
Qi = 12 - n • 1 - • (h - e - cos(t)) • d(j) (B.3.)
a 0

where h is the height of the gap between the spindle and the bearing at a
concentric position of the spindle, 1^ the length of the bearing gap in axial
direction, pi the pressure in chamber 1 and e the displacement of the spindle
caused by a force F. From the equation B.3. it follows:
Pl • D - a
(B.4,)
12 - n • la
with:
^ 3 T r ^ ^ 2 „^ 2 T T 2 3
a = h - — -3'h -e + 3 h ' e •— -— -e
o 2 o o 4 3
As:
(Pl - Vk) ' K • K^
Ql-4 = — e_
12 • n . 1t

in which Lg is the length of the bearing and l-j- the length of the gap that
separates the chambers
Pl - D - a (Pi - Pi+) • L^ • h^^
Qo = 12 - n . 1 " 12 - n . 1, ^'-'-^
a t
In the same way:
P4 • D • b (Pi - P4) • L • h
e o
Qo = 12 . n • 1 ' 12
"
- n
••
> 1^
^t
^ ^^-''^

with:
b = h --tSh -e + S h -e . - + - e^
o 2 o o 4 3

As the forces in the bearing have to be in equilibrium

F = (Pl - P4) • D • L (B.^.)

From the equations B,5., B.5, and B.7. it follows:

110
 1 h^ b + a b-a
F ( + - ) = Qo • (B.8.)
L-12-n'l 12'n'l^'D a-b a-b
e a t
From this equation the relation between the force F and the displacement e can
be calculated at various conditions both concerning the geometry of the bearing
and the flow and the pressure of the oil.
The second requirement for the bearings concerns the heat generation in the
bearings. This heat generation can be divided into two parts. One part con-
cerns the energy dissipated as a result of the pressure drop in the bearing
gaps. This power P , supplied by the pump, can be expressed by:
4 - Qo - PQ (B.9.)
Pp= 1
in which P Q equals the average value of the pressures in the four bearing
chambers and e represents the efficiency of the pumps. According to the
equation B.1.:
4 - Qo • 12 • n • 1
p 3
2 • TT . D - h -^
o

and from this equation it follows:

24 • n • Qo^ • 1
P = (B.IO,)
^ TT • D • h 3 , e
o

The second part of the heat generation is caused by the friction in the oil
film in the gaps of the bearing. This power P^^ can be expressed by:

V2
p =: F ri (B.11.)
w w •ho

In this equation F^ represents the area at which the friction appears and V
the speed at the circumference of the spindle. The equation B.11 can be written
as:
2 2
P = (2 • w - D - 1 + 4 . 1 (L - 1 )) • "^ 'y J (B.12.)
w a t e a 4 -h
o

in which w represents the speed of rotation of the spindle. The temperature

rise AT of the oil from the inlet of the bearing to the outlet can be calcu-
lated from
P +P
AT = —^ (B.13.)
4 • Qo • P • c^

in which p represents the density of the oil and c^ the specific heat of the
oil,
The choice of the bearing parameters is based on the following considerations.
At first the requirements concerning the stiffness of the bearing have to be
met. This means that hQ has to be small while the values of D, 1^, L^ and 1-^-
have to be large. On the other hand the requirements concerning the heat

- Ill -
 generation lead to small values of D, 1^, L and 1.^. while h^ has to be large,
Finally there are some limits like the minimum viscosity of the oil, the
available pumpunits and the dimensions of the bearings with regard to the
dimensions of the spindle. In order to find a compromise that results in a
practical design a computerprogram has been drafted. Based on the results of
the calculations according to this program the following data determine the
front- and the rearbearing of the grinding spindle:

Frontbearing Rearbearing
Qo = 4 . 1 0 " ^ m ^ / s Qo - 3.10"^ m3/s
h = 30.10~6 m ho = 30.10-6 m
o
1^ = 4.10-3 ^
h = 4.10-3 m
If = 8.10-3 m 1^ = 8.10-3 m
Le = 9 . 1 0 " ^ m
D = 12.10"^ m D 8.10 ^ m
4,35.10-3 kg/m.s. n 4,35.10~^ kg/m.s.
n
At these data and at a speed of 10.000 r.p.m. the following values of the
stiffness, the pump pressure, the pump and friction power and the temperature
rise of the oil are obtained:

Frontbearing Rearbearing

C^ = 1536.10^ N/m C^ = 7 6 8 . 1 0 ^ N/m

Po = 16.4 . 10^ N/m2 Po = 1 8 . 5 . 10^ N/r
P = 350 W P = 300 W
P
P^ = 3300 W Pw = 850 W
AT 13.8° C AT = 5.8"" C

112 -
C. Contact of the rotor circumferences

In chapter 5.1. it has been shown that the rotor circumferences do not touch
each other mathematically. This means that the rotor circumferences have
two lines in common. In figure C.1. this situation is shown. In this figure

FIGURE C.1.
Relative position of
the rotor circumferences.

the rotor circumferences have a point P at angles <() respectively ({)' in

common. If the rotor circumferences have a second point in common there must
be an overlap of the two rotors in a point Q quite near the point P. In or-
der to determine whether this overlap appears the following calculation is
carried out. At a very small angle 6 the value of R((f) + 6) is calculated
according to the equation 5.9 in chapter 5. From the value of R((f) + 6) the
value of the distance 0'Q and the value of the angle p are calculated. The
figure C.1. shows that an overlap of the rotor circumferences appears if the
value of R'(())' + p) exceeds the value of the distance 0'Q. By means of a
computer program these calculations are carried out at different values of
the angles (j) and (j)' and for different values of 5. From these calculations it
follows that the maximum width of the overlap, determined by the value of
x = R'(<)) t p ) - 1 does not exceed 0.0025 mm.

- 113 -
D. Leakage and mechanical losses in the hydraulic motor

D.1. Leakage

In the hydraulic motor leakage flows appear as a result of gaps between

the rotors mutually and between the rotors and the housing. First the
leakage at the place where the rotors touch each other will be calculated.
The flow of oil through a gap is determined by the dimensions of this gap,
the pressure drop from the inlet to the outlet of the gap and the visco-
sity of the oil. From the geometry of the rotors it follows that the gap
between the rotor circumferences is identical with a gap between two cy-
linders with diameters that equal the average diameters of the rotors as
far as the height of the gap is concerned. In figure D.1, this gap is

FIGURE D.1.
Gap between the
rotor circumferences.

shown. According to this figure the value of the height of the gap h can
be expressed by:

h = h . +R (1 - coscj)) + R' (1 - cos(})' ) D,l.

mm

As, according to the rotor geometry described in chapter 5, 2 • (J)' = 3

and 2 - R = 3 • R' the equation D,l. changes into:

2 3- 1
h = h . + R • (- - coscj D.2,
mm 3 - - cos - ^

It is assumed that the oil flow Q through the gap is determined by:
, 3
Q = 12 • n iP ÜJ - R - h D.3.
6cj)

in which ri represents the dynamic viscosity of the oil, p the pressure in

the oil film and w the angular velocity of the central rotor. The maximum
value of Q appears at u = 0 and the maximum pressure drop from the start
to the end of the gap Ap = 350 • 10^ N/m^. From the equation D3 it follows:

AP
D.4,
r+
12 d(f)

114 -
in which the values -3 and +3 determine the values of cj) at the start and
the end of the gap. The solution of the equation D.4. is obtained by means
of a computer program and leads to a leakage flow Q = 2.95 • 10"^ m3/s per
m width of the gap at a value of hj^^j^ = 2,10~2jnm and a viscosity of the
oil n = 0,087 Ns/m2. As the total width of the gap equals 0,144 m the leak-
age flow QLJ_ = 4.25 • 10-5 m3/s.
The second part of the leakage flow concerns the gap between the rotors
and the housing. As the leakage at the rotor circumferences is relatively
small, this part is neglected. The flow of oil QL2 ^"t the gaps between the
side planes of the rotors and the housing is determined by the equation:

Q, = "P • ^ • ^' D.5.

L2 12 • n • 1
In this equation b represents the width of the gap and 1 the length of the
gap. As the average length of the gap is assumed to be 1.5 mm at a total
width b = 1,65 m this part of the leakage flow QL2 " ^•'^^ • 10-^ m3/s at a
gap height h = 2.5 • 10-2 mm,a viscosity of the oil n = 0.087 N.s/m^ and
at a pressure drop Ap = 3.5 • 10"^ N/m^. The choice of the values of h and
1 depends also on the heat generation in the oil film. The total leakage
flow at maximum pressure Qj^ = 6 • 18 • 10"^ m3/s.

Mechanical losses

The mechanical losses in the hydraulic motor concern the losses caused by
friction in the bearings, the gear transmission and the oil films in the
motor. The first two parts of the losses depend on a number of unknown
factors like bearing pre-load, lubrication of the bearings and the gears,
geometrical accuracy and so on. In this calculation only the friction
losses in the oil films between the rotors and the housing are taken into
account. As the friction losses at the circumferences of the rotors are
very small if compared with the friction at the side planes of the rotors
these losses are neglected. The power P L generated in an oil film between
the surfaces that move relative to each other can be determined by the
following equation:
v2
p^ = A ^ . n . - D.6.

in which Ap represents the total area of the oil film and V the relative
velocity of the two surfaces in question. From the dimensions of the ro-
tors and the maximum speed of the rotors the value of P L can be calculated.
At the dimensions of the oil film mentioned in the first section of this
appendix the friction power

P, = 24,2 kW.
LI

- 115 -
 In the figure D.2. the total efficiency e of the hydraulic motor as far as
the losses calculated above are concerned is shown as a function of the speed
of the motor n at maximum power.

100. FIGURE D.2.

Efficiency of the
90. hydraulic motor.
>>
Ü 80.
c
e
'ü
70

60

50.

<^

y 5000 6000 7000 8000 9000

116
REFERENCES

[1] Ernst, W. ErhShte Schnittgeschwindigkeit beim Aussenrund-

Einstechschleifen und ihr Einfluss auf das
Schleifergebnis und die Wirtschaftlichkeit.
Aachen, Dissertation T.H., 1965.

[2] Giihring, K. Hochleistungsschleifen, eine Methode zur Leistungs-

steigerung der Schleifverfahren durch hohe Schnitt-
geschwindigkeit .
Aachen, Dissertation T.H., 1967.

[3] Voller, R. Sinnvolle Anwendung des Hochgeschwindigkeits-

schleifens.
Zeitschrift für industrielle Fertigung 64 (1974)
no. 5, p, 259 - 264,

[4] Stewart, I,J, Einfluss des Hochgeschwindigkeitsschleifens auf

den Schleifvorgang,
Fertigung 2 (1971) 2, p, 75 - 81,

[5] Shaw, M.C., Mechanics of the abrasive cut-off operation.

Farmer, D.A. and Paper no. 66-WA/Prod.-13
Nakayama, A. ASME, New York, 1966.

[6] Mueller, J.A. Abrasive Machining - Cut off.

ASTME paper MR 69 - 204.

[7] Storm, T. Adaptive control in cut-off grinding.

Proceedings of the 13th M.T.D.R.-Conference,
Birmingham, 1970.

[8] What the future holds for grinding,

Metalworking Production
118 (1974) p, 48 - 55,

[9] Research hots up to meet tomorrow's demands,

Metalworking Production
118 (1974) p. 56 - 59.

[lO] S.B. Baars Doorslijpmachine.

Internal report. University of Technology,
section Machine tools, Delft, 1969.

[11] Rüggeberg, T. High-speed and high-efficiency abrasive cutting.

Proceedings of the International Grinding Con-
ference, Pittsburg, Pennsylvania,
april 1972, p. 524 - 537.

117
[12] Kaminskayia, V.V. Calculation and research on machine tool
structures and foundations.
Proceedings of the 8th M.T.D.R.-Conference,
Manchester, 1967.

[13] Paulsen, W,C. Finite Element Stress Analysis.

Machine Design 43_ (1971),
nr. 24, p. 47 - 52
nr. 25, p. 146 - 150
nr. 26, p. 90 - 94.

[14] ASKA - User's Reference Manual.

I.S.D.-report nr. 73
Stuttgart, 1971.

[15 ] Pahlitzsch, G and Einfluss des Abrichtens mit diamantbestückten

Schmitt, R. Rollen auf die Feingestalt der Schleifscheiben-
Schneidflache.
Werkstattstechnik, 58 (1968) nr. 1, p. 1 - 8.

[16] Pahlitzsch, G. and Wirkung von Korngrosse und -Konzentration beim

Schmitt, R. Abrichten von Schleifscheiben mit diamantbestück-
ten Rollen.
Werkstattstechnik, 59 (1969) nr. 4, p. 158 - 161.

[17 ] Klemm, W. Untersuchungsergebnisse bei der Anwendung von

Diamant-Abrichtrollen.
Fertigungstechnik und Betrieb
19 (1969) nr. 11, p. 669 - 675.

[18] Scheidemann, H. Einfluss der durch Abrichten mit zylindrischen

und profilierten Diamantrollen erzeugten
Schleifscheiben-Schneidflache auf den Schleif-
vorgang.
Braunschweig, Dissertation, T.U., 1973.

[19] Frank, H. Das Abrichten von Schleifscheiben mit Diamanten

und der Einfluss auf das Schleifergebnis beim
Aussenrund-Einstechschleifen.
Aachen, Dissertation, T.H., 1967.

[20] Schreitmiiller, H.J. Pneumatisches Messverfahren zur Ermittlung des

and Dederichs, M. Schleifscheibenverschleisses.
Industrie-Anzeiger _93 (1971)
nr. 68 p. 1733 - 1734.

[21] Opitz, H. and Untersuchungen an Schutzhauben für Hoch-

Langbein, R. geschwindigkeitsschleifmaschinen.
Industrie-Anzeiger 95_ (1973)
nr. 17, p. 328 - 331.

118
[22] Krug, H, and Beeinflussung der Bruchgeschwindigkeit umlaufen-
Schoor, N. der Schleifscheiben durch die axialen Einspann-
krafte.
Industrie-Anzeiger 96 (1976)
nr, 65, p. 1463 - 146"7,

[23] Suzuki, I. and Schleifscheibenbruch wahrend der Rotation.

Wada, R, Werkstatt und Betrieb 105 (1072)
nr, 9, p, 643 - 647,

[24] Safety in High Speed Grinding.

Publication nr, M3071E of Cincinnati-Milacron,
England,

[25] Frank, H. Hochgeschwindigkeitsschleifen, Werkstückgüte

und Zerspanleistung.
Publication of Guilleaume-Werk, Bonn-Beuel,

[26] Kortendieck, W, Merkmale beim Schleifen mit erhöhter Umfangs-

geschwindigkeit,
Werkstatt und Betrieb 104 (1971)
nr. 3, p. 157 - 163.

[27] Grinding fluids can break the air barrier,

Metalworking Production 118 (1974) nr. 1
p. 58 - 73.

[28] Opitz, H. Aufbau und Auslegung hydrostatischer Lager und

Führungen und konstruktivp Gesichtspunkte bei der
Gestaltung von Spindellagerungen mit Walzlagern.
Bericht über die VDW-Konstrukteur-Arbeltstagung
am 21. und 22. Februar 1969.

- 119 -
 CURRICULUM VITAE

14 October 1939 born in Bennekom.

1951 - 1956 Marnix College, Ede.

1956 diploma H.B.S,-b.

1956 - 1964 University of Technology, Delft,

mechanical engineering.

1964 diploma werktuigkundig ingenieur.

1964 up to the present member of the scientific staff of the

University of Technology, Delft.

120
 STELLINGEN

1. Bij het vervaardigen van produkten dient aan spaanloze

bewerkingen zoveel mogelijk de voorkeur te worden ge-
geven boven verspanende bewerkingen.

2, De losse kop van draai- en slijpmachines dient h$t

werkstuk even nauwkeurig te ondersteunen als de
vaste kop.

3, Een weloverwogen materiaalkeuze zal het gebruik van

staal en gietijzer doen afnemen.

4, Alleen het onderzoek naar de toepasbaarheid van nume-

rieke besturing kan reeds tot een aanzienlijke produktie-
verbetering leiden.

5. Het onderwerpen van de z.g, "contacten met derden" aan

strenge normen zal de afstand tussen de Technische
Hogescholen en het bedrijfsleven vergroten.

6. De benaming "sportief pookje" bij auto's is een ver-

zachtende uitdrukking voor een minder goede oplossing
van een aandrijfprobleem.

7, Evenmin als een onderneming bij het vertrek van een

direkteur de produktie stopt of afremt dient het ver-
trek van een hoogleraar aanleiding te zijn het be-
treffende onderwijs te onderbreken.

T, Storm
Delft, 10 juni 1976