\C MSSBM0S THER0 eS CORPORAMON

oaa0 .. .....

(NASAoCRo-130225)' DEVELOPMENT OF A dc N73-22MOTOR WITH VIRTUALY ZERO POWEREDNAGNETIC BEARING 'lCambridge ThermionicCorpO.) 5 p HC $:6.00 CSCL 09E Unclas

< aOW p~~~~~~~~~~~~r' G/n 0 2 2 11 _

-r, I,

==3 0 7

i

I

.i

:V.

';i l

.- · '', .' I

. .1 .1

- , /' 2°

| ::S~~~~~~~~~~~

https://ntrs.nasa.gov/search.jsp?R=19730013425 2020-07-29T01:14:31+00:00Z

DEVELOPMENT OF A DC MOTOR WITH

VIRTUALLY ZERO POWERED

MAGNETIC BEARING

Prepared under Contract NAS5-21587

CAMBRIDGE THERMIONIC CORPORATION

Cambridge, Massachusetts

For Goddard Space Flight Center

NATIONAL AERONAUTICS AND SPACE ADMINISTRATION

(

TABLE OF CONTENTS

SECTION I --

VZP DEMONSTRATION

MODEL

SECTION II --

VZP CONTROL

TECHNIQUES

SECTION III --

TORQUE MOTOR

SECTION IV--

WORKING MODEL

SECTION V --

SUMMARY

PREFACE

1.1 Magnets

1.0 Magnet Tests

1.2 Virtually Zero Power Principle

1.3 General Difference Between Electromagnetic& Electromagnetic/Permanent Magnet Systems

1.4.0 Development of Breadboard Model of the VZPMagnetic Bearing System

1.4.1 New Technology

1.4.2 Mechanical Configuration

1.4.3 Displacement Sensing System

1.4.4 Rate Sensing System

1.4.5 Description of Behavior of Assembled System

1.4.6 3rd Servo of Centering System

1.4.7 Conclusions

2.0 Magnetic Bearing Motor and Motor DriveCircuitry

2.1 Magnetic Bearing Scanner Mount

2.2 Close Gap Bearing

2.3 "Bang-Bang" Magnetic Bearing AxialCentering Servo

2.4 Clamshells

2.5 Velocity Only Axial Control

2.6 Conclusion of Control Experimentation

3.0 Instability

3.1 Torque Motor

4.0 Electronic Controller

4.1 Automatic Starting

4.2 NASA Magnetic Bearing and MotorSpecifications

4.2.1 Supplementary Data

5.0 Recommendations and Conclusions

Page

i

1

3

4

9

10

11

11

12

12

12

14

14

16

16

18

20

21

24

33

35

36

37

37

38

40

41

I

PREFACE

Previous magnetic bearing - DC motor system delivered to NASA by

Cambridge Thermionic Corporation required 20-30 watts of continuous support

power. Although this system performed will the continuous support power was

a disadvantage for space applications.

The object of this contract was to develop a magnetic bearing - DC motor

with a minimum of power required for the support system. It was felt that

permanent magnets could be made to carrythe steady state bearing loads

while the electromagnets could be required only to counteract the transient

base disturbance inputs.

The period of this contract was to be nine months in length. However,

within three months the basic principal was reduced to practice. This principal

was called "Virtually Zero Power" (VZP) and was demonstrated in a working

device at NASA, Goddard Space Flight Center on 25 February 1971.

Following this demonstration, work proceeded to increase the radial

support capability, design and fabricate a drive motor and design and fabricate

the electronic drive circuitry.

Subsequent investigation showed that the radial stiffness could be increased

by armoring the permanent magnet with steel. This also required the use of

relatively narrow magnetic air gaps which put a greater burden on the

mechanical alignment and electronic servo system. A magnetic bearing using

steel armoring on the permanent magnets was designed and built. This device

was to be a Magnetic Bearing Scanner Mount (MBSM).

NASA provided a motor to be integrated into the magnetic bearing system.

This newly developed motor was considered to have qualities suitable for

scanner pointing applications.

The motor was integrated into the magnetic bearing system and the complete

device was delivered and demonstrated at NASA, Goddard Space Flight Center

on 19 September 1972.

ill

SECTION I - VZP DEMONSTRATION MODEL

1.0 MAGNETS

It became evident early in the development that there was need for a

feasibility study to determine if it would be technically possible to carry

all steady state bearing loads (axial, radial, and torsional) solely through

the use of permanent magnets (requiring zero power).

The NASA Engineer stressed the importance of utilizing a high magnetic

flux density (B) (through the use of permanent magnets), in the formula2

F - kB . It was felt that the required magnetic modulation, for stability

in the presence of disturbances could be obtained more efficiently in this

manner.

The selection of the most efficient geometrical configurations among

the large number of possibilities became a basic problem. It was early

concluded that there was insufficient information available to fix on a

particular design. It was therefore decided to: (1) Examine a number of

possible designs, (2) Make measurements, (3) Further investigate the use

of the new Samarium-Cobalt magnets. (See Figures 1 and 2)

In Figures 3 through 9 in the Appendix are graphs taken from data on

several configurations, including some with very small (1" x 1/4" discs 25

gram) Samarium-Cobalt magnets and more with very large (50 lb.) Alnico

magnets. In connection with the use of the large magnets, two extremes

in L/D ratios were examined. The large diameter tube (L/D = 1.337) is on

the edge of torsional instability while the smaller diameter tube (L/D = 2.675)

exhibits excellent torsional stability.

There is very obviously a discrete point on an L/D curve ratio where one

passes from stability to instability - for any given ring geometry and gap.

Although not a specified part of the project - since it would be sufficient to

employ an arbitrarily high L/D ratio.

Engineering data on Samarium-Cobalt magnets, as well as samples, were

provided by the Raytheon Company of Waltham, Massachusetts. Because of

the newness of these magnets and the consquent lack of understanding of their

behavior, some of the data concerning Force/Displacement attached to this

report must necessarily be regarded as preliminary. It has become evident, how-

ever, that one is dealing with a very unusual material which must be handled

by quite different methods than those normally employed with older "alnico"

type magnets. Specifically, it is evident that if is desired to obtain the very

high force levels possible with this light weight material, one must reduce the

"armoring" (by ordinary steels) to the shortest possible path lengths. The

Samarium-Cobalt magnet must be thought of as a large "air-gap" magnet.

A possible geometrical configuration began to evolve. It developed partly

from the absolute necessity to keep the pole piece lengths - in the use of

Samarium-Cobalt magnets - extremely short. Additionally, it is believed

essential to have very high speed (low inductance) servo operation, an attempt

has been made to utilize a linear ("ironless") motor in the high field provided

by the supporting permanent magnets. Such an approach does appear to have

merit, expecially due to its simplicity and apparent high order of theoretical

efficiency.

A test of such a system was conducted on the "magnetic gate" fixture. It

showed that a + 3 lbs. of force change (electric modulation) is provided through

the step-function application of 2.4 watts. (This is in a 24 lb. permanent

magnet total available force level (at zero gap)). Results were obtained at a

.050" air-gap. Due to flexural limitations of the test fixture, smaller air-gaps

could not be examined. The mechanical assembly was not rigid enough to

handle to higher force levels. It was apparent however, that percentage

modulation was tending to increase with gap reduction. Even so, such a 10%

modulation might prove sufficient for servo control of a complete system.

Because of the clearly demonstrated superior properties of the Samarium-Cobalt

materials, it was then proposed to abandon any further activity with Alnico

-2-

type magnets so as to be able to concentrate on the efficient use of the new

Samarium-Cobalt material. Subsequent emphasis was directed toward (1)

linear generators, (2) linear motors suitable for use with Samarium Cobalt and

"Virtual Zero Powered".

1.1 MAGNET TESTS

After selection of Samarium-Cobalt magnets, emphasis was next placed on

construction of apparatus for evaluation and test. Two new cup cores were

manufactured to contain the Samarium-Cobalt 1/4" x 1" magnets. Each magnet

of the mating pair was fitted with a 1/2" diameter (.025" square cross-section)

flux concentrating ring. On one of them there was built on, as an integral

part, a shaft extension (1.5" x .150") so that additional devices might be

installed on the end of the system. These new cup cores were not as thick as

the set previously made because there is no armoring other than the rings. A

miniature rate generator was constructed. A small linear motor coil was also

built to fit into the leakage flux path of the "rotor" cup core, while being

fixed to the "stator" core.

In addition to the above, a rigid stand was constructed to hold the stator

coils approximately six inches apart and to clamp them securely in this position.

(See Photographs 1 through 6.)

Quite clearly the elimination of the steel armoring to the Samarium-Cobalt

magnets resulted in a higher effective force level both axially and, to some

exent, radially as well. Axial measurements were reasonably straightforward.

(See Figures 10-13)

Shims were employed (.003") to fix the air-gap and the breakaway force

measured for each increment in displacement as shims were added from zero air-

gap to the maximum of interest. The dial of the Chatillion force gauge recorded

maximum force encountered and, since the breakaway point was precipitous,

the method yielded high accuracy. All elements had to be made very rigid to

preserve pole face alignment and prevent "cocking", in the face of high force

levels.

-3-

Measurement of radial forces was not quite so easily accomplished. In

fact, considerable effort was actually put into the design and construction

of the apparatus before a satisfactory method was developed. The difficulty

was in preserving accurate pole face alignment at small air gaps, in the

presence of the high axial pulling force. This method finally worked out is

illustrated in Figure II. The data from this test is shown in Figures 12 and 13.

It was found very important to have the "stabilizing" rod long and stiff

in order to preserve pole face alignment. However, when all had been put

in order, it was readily possible to measure radial forces down to .003" air-

gaps with excellent repeatability.

Some testing was also done with the idea of using one end of the magnetic

bearing as the velocity (linear) motor. The constituted a negative feedback

system and was attractive because of its simplicity. However, results were

inconclusive, largely because sufficient force could not be obtained from the

motor.' In the space available within the cup core only 56 turns of #28 wire

could be inserted and still allow for the necessary .025" radial "play". If

the scheme was to receive further attention, larger cup cores would have to

be employed. Tests did show that response time of the whole system was

sufficiently short, and this was the point of most interest at the time.

Also the miniature rate generator was attached to the shaft and was

tested. An unsatisfactory low output level was encountered, but again the

time constant appeared reasonably short. (This small unit was built with the

idea that it might later be placed within the shaft of the rotor).

1.2 VIRTUALLY ZERO POWER PRINCIPLE

Relatively early in the investigation, a rather far reaching decision was

made: namely, "to fabricate a complete system as quickly as possible to

prove out the overall theory of the Virtually Zero Power System". This was

done and the principle was proved and there was no longer any doubt whatso-

ever concerning the validity of the theory of "VZP".

-4-

To construct a complete operable system in a short period of time

required making numerous compromises. Previously built "cup cores" were

known from tests to be quite efficient. Their effective use, however,

entailed working at close air gaps (of the order of .005" - .010") where

dimensional tolerances are critical. However, to achieve a time table of

one month for completion of the apparatus made it impossible to consider

their use.

It was found that a triplet configuration of I" diameter "raw" Samarium-

Cobalt magnets yielded 2 lbs. of radial force at air gaps of .100" to .200".

To be sure, such a system could not be expected to give a steep radial force/

displacement curve, but if fairly large radial clearances were employed, it

appeared sufficiently high to allow construction of a workable demonstration -

if rotor weight was kept low. Unfortunately, it was not possible to secure

more than two Samarium-Cobalt magnets with large enough clearance holes

from the Raytheon Company in time. This meant that only a pair of doublets

were available.

The system was designed to separate the permanent magnet radial support

magnets by approximately 5". Two axial force coils were buil It to act

oppositely against a very light central disc of BFM steel. These force coils

had to develop sufficient force to enable starting the system from zero air

gap. It could be accomplished by application of 16 watts from the amplifier.

Next, a very simple displacement system was built. This system consists of

two photo-cells looking at two micro-miniature lamps. The rotor carries on

one end a ring, mounted on transparent plastic, which divides the light

equally to the two sensor cells when the central disc is halfway between the

two force coil cup cores. Such a simple system proved adequate. See

Figure 14. -

A light shield was made to avoid effects of stray light and surrounded

the whole sensor system. The rate generator employed makes use of one of

the rotor carried support magnets - a temporary but effective measure.

-5-

When the rotor is precisely in the central position (unstable) the rotor

"falls" to one side or the other to the stopped position (+ .025") in 8-17

seconds, depending on the amplifier gain. This free fall situation is

naturally dependent upon amplifier gain, P.M. air gap, etc., but in

general is a measure of rate generator effectiveness.

Upon assembly of the apparatus, it was immediately found that the

force coils vibrated as a tuning fork. This required that the free ends be

rigidly clamped, which was done - the clamp being additionally used to

support the force coil terminals.

By 2 February 1971, the system was operational. Some time was spent

in observing the behavior of the unit. It was soon found that:

I. The bearing could be operated with virtually zero power to the force

coils (Bearing Horizontal). Actually .05 watts.

2. That this situation also applied to Vertical Operation IF the amplifier

was re-balanced in this position.

3. That the system would self-start in any position. Starting required

approximately 16 watts peak and was completed in a measured 150

milliseconds.

It was found that initial balancing adjustments were critical, in that it

was easily possible unwittingly to balance one unbalanced part of the system

against another oppositely unbalanced part. When this was done, a totally

unstable system resulted! A procedure, outlined below, was then worked

out to avoid the problem:

I. Balance the amplifier with power stages shut off. This is done by

adjusting the amplifiers recommended bias potentiometer while viewing

the zero-center output meter.

2. Adjust the relative positions of the P.M.'s along the rotor shaft so that

the equa! force (unstable) position occurs with the central disc is pre-

cisely midway between the force coil cup cores.

3. Then physically move the fixed portion of the displacement sensor

system (with the force coils ON) to make the central disc arrive in

-6-

the previously determined central position. (Note that the disc's

axial motion follows the sensor motions, i.e. has the same sense).

4. Observe the zero-center meter. It should now read zero. If it does

not precisely conform to zero, then minor amplifier balance adjust-

ment may be made.

5. Note that the central axial position is that position at which the

rotor is equally likely to move left (or right!). With the displacement

sensor shut-off and this position is achieved, then the rotor will hold

this position for a considerable time - staying away from the destina-

tion pole (whichever) for as much as a minute. (On rare occasions

when all balance adjustments are nearly perfect.)

It was found that there was a small double zero position, but it

indicated further detailed examination as its elimination would permit a

closer approach to absolute zero power.

A very poor coast time was immediately noted. This was attributed to

the open field resulting from the use of "raw" P.M.'s reacting on the

axial shaft which penetrated two of them.

As experimental work continued, it was learned that with the displace-

ment sensor shut-off ("rate alone") that center position could be manually

maintained for extended periods. This was done by watching the zero-

center meter while manually adjusting the amplifier balance control.

Practice was required, after which it was possible to hold close to center

positions for a long as was desired. Such controlling required no appreci-

able skill since ample time was available for introduction of correctional

signals. This was due to the long time constant of the rate system as a

whole. The next logical step was to introduce the 3rd servo or centering

system. This was done on 3 February 1971 and proved highly effective.

It should be remembered that this system senses the unbalanced conditions

across the force coils (essentially the zero-center meter signal) and serves

to remove this unbalance by biasing the amplifier in the proper direction

by the proper amount.

-7-

Due to the low voltage level of the force coils (< 4V) it was not possible,

at that time, to introduce the desired dv/dt signal. In spite of this, it was

found possible to secure reasonably stiff servo response.

With the "3rd" servo in operation, it was immediately possible to secure

reasonably stiff servo response. With the "3rd" servo in operation, it was

immediately possible to see automatic compensation for applied axial loads.

To be sure, nothing like optimum operation was achieved, but, in principle,

the system was certainly a success. The rotor moved in opposition to axially

applied loads with total motion of + .020" before control was lost. Within

these limits, and with the axial loads not exceeding several pounds, the

system was stable and obviously kept the required power to low levels -

usually less than 1/2 that required to counter the loads without the system

in operation.

The concept of the Virtually Zero Power magnetic bearing system was thus

proved out. Clearly, much remained to be done before a sophisticated motor-

bearing could be fabricated. But it then became quite apparent that VZP

magnetic bearing systems were practical.

At this time, an additional interesting experiment was conducted. The

2nd (displacement) servo was disconnected (after starting) and reliance was

placed only on the 1st (rate) and 3rd (force balancing system).

In spite of the previously mentioned non-optimum characteristics of the

3rd servo system, operation was quite stable. The possibility of such a system

had been predicted before testing so it was particularly gratifying to observe

the actual validity of the idea.

In order to conserve overall power, it would be possible to disconnect the

#2 servo system whenever the rotor is axially centered. This could be done

(simply) by deriving power for the sensor lamps from un-differentiated output

of the amplifier. This power is essentially zero when the rotor is in the center

positions when starting is required, (thus the lamps are "On").

-8-

As a result of the work accomplished, during this early period it was

concluded that the concept of a Virtually Zero Powered magnetically

supported rotor was feasible. Further work was then pointed toward adding

two additional magnets, and increasing the overall performance of the #3

servo system, and to add a motor. See schematic in Figure 15.

1.3 GENERAL DIFFERENCES BETWEEN ELECTROMAGNETIC &

ELECTROMAGNETIC/PERMANENT MAGNET SYSTEMS

The following is a listing of the differences between electromagnetic

and electromagnetic/permanent magnet systems:

I. No thermal problems. Control coils operate cold (very close to

ambient.)

2. There is no mechanical strain accross the rotor and, therefore, a

lighter supporting structure is possible. (In all fairness, this could

also be done with the EM System).

3. Almost exactly 1/2 the number of electronic components are re-

quired. (The circuit is differential rather than push-pull).

4. There are no large steady state currents in electronic system -

they operate in standby condition, except for transients. (About

2 watts + unresolved #3 servo differential power of approximately

< 0.1 watts.)

5. It is readily possible to secure a satisfactory displacement signal

with I (one) micro-miniature lamp only. Multiple redundancy is

also possible.

6. Due to the "empty weight" being lower in the permanent magnet

system, the payload to empty weight ratio can be higher.

7. The configuration is more attractive from a dynamic standpoint since

outboard weights are very much smaller.

8. Cocking problems are reduced.

-9-

1.4.0 DEVELOPMENT OF BREADBOARD MODEL OF THE VZP MAGNETIC

BEARING SYSTEM

There then followed the following series of events:

I. Preparing for demonstration of the breadboard model unit.

2. Demonstration of the unit at Goddard Space Flight Center on 25 February

1971.

3. Collection of data on the unit.

In Figure 16 is a plot of amplifier performance which shows that there is

little, if any, "cross-over" distortion. It will be recalled that earlier mention

has been made of a "double-zero" position. It became evident that in view of

the linearity noted in the amplifier characteristics, it clearly could not be

attributed to the amplifier. Therefore, the "double-zero" condition needed to

be explained in some other way. It then was believed to be a fundamental

characteristic of the system as a whole.

It was also noted that the rotor must accelerate (axially) either to the

right or to the left from the neutral position as determined by the magnetic

fields of the permanent magnets. The function of the force coils is to induce

this acceleration to approach zero, as a limit.

Assume, for example, that the rotor is accelerated to the left but is very

close to the zero acceleration position. We now move the bias control so as

to start the acceleration through zero and beyond - to the right. One can

see that to counteract the right hand acceleration (if we wish to approach

zero again) the bias must now be moved through zero - to the left again.

There is, therefore, an unstable "flip/flop" condition existing about the

zero acceleration-zero power position. This is'fundamental to the system as

conceived at the time.

It became important to examine the magnitude of the unstable region.

Measurements showed that the actual axial motion, in the unit, was of the

order of plus or minus 0.0003" and that the power required to work against

this instability was about .05 watts.

-10-

It was then believed that the unstable region could be reduced in

magnitude by a number of means such as (I) increased displacement system

gain, (2) increased rate system gain, (3) reduced permanent magnet force

level - this being the least desirable. But it was thought that the "back-

lash", as in mechanical devices, could not easily be eliminated entirely -

but must be allowed for in bearing use. In some applications, for instance,

the system may be deliberately biased to one side so that the rotor cannot

accelerate through zero. This, however, would cost power. The Principal

Investigator, Mr. Joseph Lyman, completed his portion of the task assign-

ment at this point with the verification of the Virtually Zero Power principle

by practical demonstrations and test.

1.4.1 NEW TECHNOLOGY

A fundamental technological development confirming previous work of

the principal investigator was then reported. A stable Magnetic Suspension

system was fabricated and a fundamental basic electromagnetic principle

reduced to a practice, wherein a Magnetically Supported rotor utilizing

permanent magnet primary support had been made stably operational utilizing

Virtually Zero Power. The system is described as follows:

1.4.2 MECHANICAL CONFIGURATION

1. A triplet of 1" diameter "raw" Samarium-Cobalt magnets yielded >2 lbs.

of radial force at air gaps of .100" to .200". Such a system could not

be expected to give a steep radial force/displacement curve, but if

fairly large radial clearances are employed, it appeared sufficiently

high to allow construction of a workable demonstration - if rotor weight

was kept low. Unfortunately, it was not possible to secure more than

two Samarium-Cobalt magnets with large enough clearance holes from

the Raytheon Company in time. This meant that only a pair of doublets

were available. This proved to be sufficient, but provisions were made

to complete the triplet pairs as soon as the additional magnets were

received.

-II-

2. The system was designed to separate the P.M. radial support magnets by

approximately 5". Two axial force coils were built to act oppositely

against a very light central disc of BFM steel. These force coils had to

develop sufficient force to enable starting the system from zero air gap.

This could be accomplished by application of 16 watts from the amplifier.

1.4.3 DISPLACEMENT SENSING SYSTEM

1. A very simple displacement system was built. The system consisted of two

photo-cells looking at two micro-miniature lamps. The rotor carried on

one end a ring, mounted on transparent plastic, which divided the light

equally to the two sensor cells when the central disc was halfway between

the two force coil cup cores. This simple system proved adequate and is

diagrammed in Figure 14. A light shield was made to avoid effects of

stray light and surrounded the whole sensor system.

1.4.4 RATE SENSING SYSTEM

1. The rate generator employed made use of one of the rotor carried support

magnets - a temporary but effective measure. When the rotor is precisely

in the central position (unstable) the rotor "falls" to one side or the other

to the stopped position (+ .025") in from 8-17 seconds. This free fall

situation is naturally dependent upon amplifier gain, P.M. air gap, etc.,

but in general is a measure of rate generator effectiveness.

1.4.5 DESCRIPTION OF BEHAVIOR OF ASSEMBLED SYSTEM

1. When the system was operational, some time was spent in observing the

behavior of the unit. It was soon found that:

a. The bearing could be operated with zero detectable power to the

force coils (Bearing Horizontal). Actually .05 watts.

b. That this situation also applied to Vertical Operation IF the amplifier

was re-balanced in this position.

c. That the system would self-start in any position. Starting required

approximately 16 watts peak and was completed in a measured 150

mi ll iseconds.

-12-

2. It was found that initial balancing adjustments were critical, in that it

was easily possible unwittingly to balance on unbalanced part of the system

against another oppositely unbalanced part. When this was done, a totally

unstable system resulted. A procedure, outlined below, was then worked

out to avoid the problem:

a. Balance the amplifier with power stages shut off. This is done by

moving the amplifiers recommended bias potentiometer while viewing

the zero-center output meter.

b. Adjust the relative positions of the P.M.'s along the rotor shaft so that

the equal force (unstable) position occurs with the central disc is

precisely midway between the force coil cup cores.

c. Then physically move the fixed portion of the displacement sensor

system (with the force coils ON) to make the central disc arrive in the

previously determined central position. (Note that the disc's axial

motion follows the sensor motions, i.e. has the same sense.)

d. Observe the zero-center meter. It should now read zero. If it does

not precisely conform to zero, then minor amplifier balance adjustment

may be made.

e. Note that the central axial position is that position at which the rotor

is equally likely to move left or right. With the displacement sensor

shut-off and this position is achieved, then the rotor will hold this

position for a considerable time - staying away from the destination

pole (whichever) for as much as a minute. (On rare occasions when all

balance adjustments are nearly perfect.)

3. Also, it was found that there was a small double zero position. No serious

attempt had yet been made to uncover the cause of this situtation, but it

deserved further detailed examination as its elimination would permit a

closer approach to absolute zero power.

4. A very poor coast time was immediately noted. This was attribute to the

open field resulting from the use of "raw" P.M.'s reacting on the axial

shaft which penetrates two of them.

-13-

5. As experimental work continued, it was learned that with the displacement

sensor shut-off ("rate alone") that center position could be manually main-

tained for extended periods. This was done by watching the zero-center

meter while manually adjusting the amplifier balance control. A little

practice was required after which it was possible to hold close to center

positions for as long as was desired. Such controlling required no appreci- -

able skill since ample time was available for introduction of correctional

signals. This is due to the long time constant of the rate system as a whole.

1.4.6 3RD SERVO OR CENTERING SYSTEM

1. This system senses the unbalanced conditions across the force coils (essentially

the zero-center meter signal) and serves to remove this unbalance by biasing

the amplifier in the proper direction by the proper amount.

2. Due to the low voltage level of the force coils (<4V) it was not possible,

within the developmental period, to introduce the desired dv/dt signal. In

spite of this, it was found possible to secure reasonably stiff servo response.

3. With the "3rd" servo in operation, it was immediately possible to see automa-

tic compensation for applied axial loads.. To be sure, nothing like optimum

operation was achieved, but, in princip!e, the system was certainly a success.

The rotor moved in opposition to axially applied loads with a total motion of

+ .020" before control was lost. Within these limits, and with the axial loads

not exceeding several pounds, the system was stable and obviously kept the

required power to low levels - usually less than 1/2 that required to counter

the loads without the system in operation.

1.4.7 CONCLUSIONS

In effect, the concept of the Virtually Zero Power magnetic bearing system

had been proved out. Clearly, much remained to be done before a sophisticated

motor-bearing could be fabricated. But at this point it appeared that it could

certainly be done.

It has now been shown that permanent magnets can be made to carry the

steady state bearing loads while the electromagnets are only required to counter-

act the transient base disturbance inputs.

-14-

The servo system for this type of magnetic suspension is not unique in itself.

However, the combined system provides some unusual characteristics. The

first being the reduction of power consumption to a minimum. The second how-

ever, is that the rotor moves in opposition to axially applied loads so that the

permanent magnets continue to carry the load and maintain the power consump-

tion at a minimum.

-15-

SECTION II - VZP CONTROL TECHNIQUES

2.0 MAGNETIC BEARING MOTOR AND MOTOR DRIVE CIRCUITRY

Attention now turned to the drive motor for the magnetic bearing. The

rotor consisted of a 1/2" diameter Samarium-Cobalt magnet magnetized

across the diameter. The stationary motor coils were arranged around the

diameter of the rotor magnet at 90° intervals. This was done to provide a

two phase system. Hall Effect devices were installed (I per phase) in the

motor coils to provide non-contacting commutation.

In the motor drive circuitry current feedback was utilized to provide

maximum torque and efficiency. (See Figure 17 in Appendix.) Speed

control was obtained by use of a negative voltage feedback loop around

the amplifier. Top speed was ultimately controlled by the back EMF

generated inthe coils. See Parts List- Figure 18.

Pertinent motor data can be found in Figure 19 in the Appendix.

Subsequent improvements in the motor drive circuit components may

be seen in Figure 20 in the Appendix.

2.1 MAGNETIC BEARING SCANNER MOUNT

At this point in the performance of the contract, Cambion was requested

to submit a proposal for a modification to the contract to include a magnetic

bearing scanner mount, MBSM, using the techniques developed during the

first part of the contract.

It was recognized that the use of the magnetic bearing device as a

scanner made accurate speed regulation necessary. Such regulation is obtain-

able if the device is rotated by a synchronous motor. The average rotational

speed is then as constant as the frequency of the supply, and if the waveform

is sinusoidal the instantaneous torque is constant producing constant instantane-

ous speed. Figure 21 shows the schematic of a 2-phase, .8 Hz power supply.

It is made of a classical phase shift oscillator built arount Z1, an operational

amplifier. The frequency is set by C1 and resistance R1 and P1 in series and

C2 and the resistance in parallel R2 and P2. P14, R14 and T14 are used to

-16-

adjust the gain to give a good sine wave at a reasonable amplitude. T14 is

a thermistor which is traditionally used to stabilize the amplitude. R3, C3,

R20 and C20 provide the 90° phase shift with Z2 making up for the resultant

attenuation. P15 permits amplitude equalization of the two phases and P3 sets

the phase difference of 90° .

The motor is driven by the transistors Q1 through Q4 which are all 3 ampere

Darlington power transistors.

Use of larger (1" diameter) diametrically magnetized rare earth cobalt

magnets permitted a more potent motor and the larger diameter made it easier

to provide windings with less space unoccupied by copper.

All the motor drive circuitry, like the support, was powered from a + 12

volt source.

The prime purpose of the Magnetic Bearing Scanner Mount, MBSM, device

was to rotate a scanning mirror in a satellite.

It includes an integral motor which, like all the rotating elements, would

be supported by magnetic rather than mechanical bearings. This avoids the

problems that often affect mechanical bearings in the high vacuum of deep

space. The mirror was not included, although the mounting provisions were

provided.Specifications:

Mechanical

(I) Length (C.G. mirror to opposite end MBSM) 9 + 1/2"

(2) Diameter 5 + 1/2"

(3) Weight 7 + I lb.

(4) Potential speed 48 rpm

(5) Support Capability I (with load applied symetrically 5 lbs.to shaft not mirror case)

(6) Stiffness (as above under support capability) 350 lb/inch

(7) Support Capability 11 (with plan or load at 4 5°angle 5 oz.with respect to shaft mirror case)

(8) Mirror mountingprovision 1/4-28 NF thread

Electrica I

Input voltage: + 12 volts

Input power: Less than 8 watts

-17-

2.2 CLOSE GAP BEARING

An attempt was made to provide a bearing with magnet gaps of less than

.010" therefore increasing the radial stiffness. This close gap bearing was

connected to electronics used to run VZP magnetic bearings previously but

it was not possible to levitate the rotor. See Figure 22. (

A large number of experiments to determine the cause of this ensued.

The electromagnets were insulated from their aluminum plates, the plates

were slit, and the same experiment performed with one of the electromagnets

in the same coaxial position without any of the bearingstructure. This

latter test showed that the single pole at 500 Hz. was due to the structure

and it was found that it could be caused by simply inserting the force ring

aluminum partition in front of the rate coil. Slitting this partition radially

improved matters slightly - zero phase shift moved up to 600 Hz.

Finally, the partition was shielded with a piece of .014" transformer

silicon steel on each side. The two silicon steel discs being joined by

ferromagnetic screws through the partition, which also held the shields on.

This eliminated the problem and pushed the zero phase shift up to 10 KHz.

It was decided to widen the .008" air gaps to .018" to reduce the stiffness

and thus the band width requirement.

Even then, levitation was not achieved. The rate coil did not fit the

one inch diameter rate magnet too closely - there is an eighth inch

clearance on the diameter - but the assembly and disassembly of the bearing

had made it somewhat eccentric, and it was thought that we were troubled

by acoustic feedback via rate coil I bobbin/magnet contact. To prevent

this, the magnet was moved so that it came close to but did not enter the

bobbin. This sacrificed half the rate coil output. We also shifted to the

pulse mode of operation and achieved levitation. In this mode the output

of the rate coil is fed to a Schmitt trigger circuit whose trip points are

-18-

symetrical about zero voltage. When the preset velocity was reached in

one direction the electromagnet that was pulling the rotor toward it was

shut off and the other turned on. A circuit that accomplishes this is shown

in Figure 23.

There are various advantages and disadvantages to this arrangement.

One of the plus features is that the electromagnets are driven from the

highest voltage available thus insuring the fastest response time. With the

bearing levitated, it was possible to make some observations impossible

without levitation. The pulse mode of operation meant that the rotor was

constantly subject to reversals of the electromagnets and the frequency at

which this took place was of interest. One would expect that as the

Schmitt trigger trip points approach each other and also zero voltage the

frequency would increase. This did happen, and it was possible to main-

tain levitation from 250 Hz. to 1,500 Hz. with less power being drawn at

the higher frequency. A trip point of 100 or 200 millivolts was sufficient.

The actual output of the rate coil did not look just as one would have

expected from the theory of operation. Although it should not have matter-

ed which way one sent the current through the nonpolorized electromagnets,

operation was possible only with proper poling.

It appeared that there was a considerable direct transformer coupling

from electromagnet to rate coil . Whether this inhibited linear mode

operation more than pulse was not known.

With so many changes made to secure levitation, it would have been

necessary to analyze each variable at a time to see which was critical.

However, success with wider gap bearing led the development to return to

previously proven approaches.

During this period of investigation two other areas were explored. A

"Bang-Bang" type servo control was considered and "Clamshell" type

magnetic circuits were also considered. The results of these two con-

current investigations are detailed in the following two sections.

-19-

2.3 "BANG-BANG" MAGNETIC BEARING AXIAL CENTERING SERVO

In Figure 23 is shown a circuit used to axially center the rotor without

displacement measurement. The following explains the physics behind this

circuit.

In a passive radial bearing (although the use of this method is not restrict-

ed to this case) there is a null force position somewhere. But this no force

position is unstable because if the rotor approaches either magnetic pole it

is accelerated toward same.

Now assume that in some unspecified manner one applies a servoed force

field to the rotor such that the rotor motion is vectorially opposite to any

other force applied. To make this statement more concrete let us assume that

the radial support magnets providing the passive radial, and the unstable

axial force are temporarily removed. If one now pushes the still servoed rotor

with one's finger it pushes one's finger back. Imagine that the radial support

magnets are again active. The servoed force field will cause the rotor to flee

from the magnet poles rather than be drawn to same and come to rest where

the nonservoed force field iszero i.e. at the formerly unstable equilibrium

point.

One method of producing such a servoed force field is as follows:

Let the servoed force field be of the simple "bang-bang" type i.e. two

valued either +Fs or -Fs. When the rotor achieves a +V0 axial velocity let

the -F force be applied and maintained until it causes the rotor to reach

-V0, when the +Fs force is applied. This simple servo will produce the

desired result as demonstrated below.

Triggering the servo force from +Fs to -Fs when +V0 is achieved and vice

versa is equivalent to triggering whenever the kinetic energy of the rotor

reaches 1/2 mV02 . The successive triggers will occur when the various

forces acting on the rotor have made a change in its kinetic energy of mV0 2 .

-20-

/-

If

Sometimes the net force on the rotor is Fe+Fs and sometimes Fe-Fs, where

Fe is the external (perhaps finger) force. But the work done is (Fe+Fs) x i

in the first case and (Fe-Fs)(-x2) in the second. Where x i and x2 are the

displacements. But both the increments of work are equal being equal to

mV0 2 . Thus:

mVO0 2 = (Fe+Fs)xl = (Fe-Fs)(-x2 ) = (Fs-Fe)x2

If Fs>Fe, which is essential for proper operation, it is seen that x2 >xI.

But xl is the direction in which Fe would move the rotor in the absence of

Fs, thus the net actual movement per cycle, xl-x2 , is against Fe.Q.E.D.

The circuit shown in Figure 23 carries out this motif quite simply. It

happens to be the exact one in use when the NASA representative

visited us and the values etc. are by no means sacred but are put in to be

concrete. The blocking capacitor C is used to prevent strain on power

supplies and Ql, Q2 during experiments. It may be replaced by a direct

connection. The voltage follower ZI

is not essential, we have used

similar circuits with the rate coil going directly to Z2 . Z2 , a Schmitt

trigger, is the essential element, doing the triggering discussed above.

The diodes DI

and D2 are worthy of mention. They determine which

force coil gets the current, but they also permit the energy from one force

coil to enter the other when the switch is made and prevent inductive kicks

from sending currents into QI and Q2 then.

2.4 CLAMSHELLS

During the Magnetic Suspension work it was found to be of interest to

increase the magnetic forces by armoring Samarium-Cobalt magnets. When

it was found that NASA was interested in having such done, Cambion

utilized some of the work previously done to produce what was called the

"Clamshell" structure.

The largest Samarium-Cobalt magnets available at the time of this work

were hollow cylinders 1" O.D., 1/2" i.D. and 1/4" long available from

three suppliers. Subsequently, larger magnets were made available.

-21-

In addition, extensive work was done on a previous NASA Contract

Number NAS5-11585 showing that when utilizing fringing to make a radially

passive device, the ring width to air gap length could not profitably be de-

creased below 5, which restricts one practically to high gap permeances,

excluding leakage.

The clamshell structure did not maximize the air gap energy. It employed

the Samarium-Cobalt magnet to provide a fairly high remnant, Br, and was

also advantageous because it would stay magnetized with an adversely low

L/D ratio.

Ideally, the polarized material would be as close to the air gap as possible

with a small surface area (to prevent leakage) flux concentrator used to bring

the pole tips close to saturation. There are two things to mention about the

output area of the concentrator. One mentioned previously is that the ring

width should not be less that 5 times the air gap, and conversely it should not

be too large, say above 10 times. If the magnet had sufficient volume to

saturate an unduly large ring, say 20 times the air gap distance, one should

make two concentric rings etc.

The clamshell structure was selected as bringing the polarized material as

close to the air gap with as little leakage as possible considering that the

magnet was axially magnetized.

Figure 24 shows the first clamshell configuration tried. The high permeance

of the load meant the BD would approach Br, or about 8,000 gauss for Samarium-

Cobalt. Neglecting leakage, the area of the ring structure would be 1/3 that2 2

of the magnet to aim for 24,000 gauss. The area of the magnet is T'1/4(1 -1/2)

= 3/16 or 0.6 sq. inch. The first trial was made without thinning the ring from

.050 inches radially to .035 as shown - it being easier to remachine after trial

than making a new piece. Thus the ring area was 1.3 iti'x .05 = 0.2 sq. inch.

The single force ring partition shown mechanically grounded in Figure 24 was

clamped in a 3 jaw lathe check. A half inch diameter aluminum rod, 18" long,

had all the other parts slid on and affixed to one end of same. The other end

-22-

had a short piece of flexible wire joining it and the lathe tailstock. It was

then possible to axially pull the clamshell via the tailstock to nearly center

the ring, thus simulating the magnetic bearing situation.

The parts of Figure 24 were all designed so that the air gap should come

out zero. Then a piece of .014 inch transformer lamination was used to make

a washer and increase the air gap. This produced an air gap of about .008"

each side.

Force measurements were made by pushing on the clamshell with a

Chatillon force gauge in 4 directions and averaging the results.

As described, the clamshell took 2 pounds to push it beyond its maximum

travel. The latter occurred at about .050" in radial displacement.

Theclamshell was then cut down to conform to Figure 24 as drawn i.e. with

a .035 inch radial ring. This produced 3.5 pounds.

A single magnet in "opposition" as shown in Figure 25 was permitted to

stick itself to the outside of the clamshell. This did more good than had been

expected. The force rose to 5 pounds.

A second outboard magnet brought it up to. 7.5 pounds. The configuration

conformed exactly to Figure 25.

Two 2 inch diameter washers on the outside of the outboard magnets

brought the support capability up to 10 pounds. This is shown in Figure 26.

This seemed like a useful force and we turned to making a complete bearing

using this structure. We did, however, try three other experiments which may

be of interest.

First, since it was easy, we put in an additional washer making the air gaps

.014 instead of .007 inches. This always cut the force in half. This statement

applies to Figures 24 through 26.

Second, clamshells were made using Vanadium-Cobalt steel available at

the time. Although it is doubtful that saturation was reached in the cold rolled

steel used in the preceeding, it was felt that any increase in force using the

high B material would indicate that we had saturated the cold rolled. Actually,S

-23-

there was no improvement. However, while it is certain that the high field

alloy can go higher than cold rolled when both are correctly annealed we

annealed neither, and there doesn't seem to be any data available on the

performance of unannealed and fairly heavily machined high field material

versus similarly treated (or untreated) cold rolled.

The ring was then cut down from. 035 to .025 inches on the radius. This

reduced the radial support to 8 pounds. (See Figure 26) It should be noted

that this brought the ring/width/air gap ratio below 5.

2.5 VELOCITY ONLY AXIAL CONTROL

An important investigation was the attempt to correlate the theories re-

garding velocity only axial control with emperical observations. In particular,

operation of the suspension in the linear mode was repeatedly attempted. This

was done because it was felt that this mode may require less bandwidth and thus

permit one to operate stiffer bearings without increasing the frequency response

of the circuit elements - particularly the force electromagnets. Increasing the

bandwidth of the electromagnets means subdividing the solid silicon steel

(laminating) currently composing same, to reduce eddy currents. However,

such an approach is difficult and makes the structure less rugged. The results

are summarized in deltail in Section 8.1.

The non-linear or bang-bang mode was discussed previously. As mentioned,

suspension can definitely work while the whole system is oscillating by pure

electromagnetic feedback. In the simple non-self-oscillating theory one can

easily visualize the waveform that will be at the rate coil terminals. With

some simplifying assumptions (immediate appearance of the force in one electro-

magnet and disappearance in the other and assuming that the electromagnet

force is constant over the range of displacement permitted) the velocity will be

a symetrical triangular function of time as shown in Figure 27.

Figure 28 shows the observed wave shape. It is simply the triangular

function riding on the electromagnetically induced square wave as shown.

-24-

It is apparent that the device cannot function according to the simple theory

if the Schmitt trigger point is set below the electromagnetically induced

voltage - in this case about 100 mi ll ivolts.

But the suspension actually worked well below this voltage. The whole

system goes into oscillation at a high frequency e.g. 5KHz. The current

switches from one coil to another at this rate. Whether the mechanically

generated voltage phase modulates the 5 KHz carrier mentioned previously

to maintain axial control or whether the device works a la the simpler theory

is now known. This latter is possible because it may be that under these

conditions the mechanically generated voltage is large enough to swamp out

the electromagnetic induced EMF. It may also be a combination of these two

effects.

While the circuitry is similar to that earlier shown in a National Semi-

conductor, LM311H comparator, used in the Schmitt trigger in lieu of the

1/2 747 operational amplifier previously shown.

This was done because the 747 had a very low corner frequency (approximate-

ly 10 Hz) so that it would be stable as a voltage follower, without any external

capacitance for compensation. The schmitt trigger circuit used did need an

amplifier - a comparator is even more appropriate. The comparator needs no

internal roll off and the LM311 as used had a rise time of only 1.5 microseconds.

With the 747 the highest self-oscillating frequency was about two KHz.

Apparently its response and the natural parallel resonant frequency of the rate

coils (5 KHz) worked together to limit the self-oscillating frequency thus.

With the LM311, the oscillations take place at 5 KHz- the band-width of

the compartor being infinite in comparison.

The above discussion of the use of the 747 versus the LM311 is merely

reporting. No significant improvements in practical levitation were noted

using one versus the other. The use of the high speed comparator merely made

the experimental apparatus correspond more closely to the idealized theory.

-25-

Using the simple theory, not contemplating electromagnetically coupled

oscillation, one would expect that if the field direction of rate magnet, poling

of rate coil, choice of force coil connections were once made properly that

any of four different combinations of the above would be equivalent (the origi-

nal configuration plus two reversals of three polings). In addition, since the

force electromagnets were used in the soft iron attractive mode the direction

of current flow in same were immaterial.

Actually one particular combination of the variables gave optimum results.

In fact, it was not certain that we had a device which would work at all with

two combinations. We had not made a special effort to ascertain whether this

was possible. There was a minor exception to the above. One could reverse

the rate magnet and rate coil connections with impunity.

From the above it appeared certain that the direct mutual rate coil - force

coil inductance played an important role in practical suspension operation.

While, as mentioned heretofore, it was not known that the coupling and

sometimes oscillation of the circuit degrading the suspension performance, it

was suspected that this inhibited the linear mode of operation (described later

on) and it certainly did no good. This last statement was based on the following

reasoning:

If oscillation was helpful the necessary feedback could always be

provided by circuitry rather than by stray coupling. Furthermore,

with large electromagnetic feedback the percentage of electro-

mechanical voltage went way down.

Having decided that it would be beneficial to reduce the electromagnetic

feedback, some thought and simple experimentation was devoted to this end.

Actually what one wants to reduce is the ratio of stray electromagnetic

coupling to electromechanical. The actual values of voltage was immaterial

until one reduced the electromechanical voltage below the point where

amplifier noise (including long term drift) became important.

-26-

The resulting rate coil structure was not very efficient. The magnetic

field intensity, H, vector had components parallel to the rotational axis

and perpendicular (radial). If the magnet had a permeability more than

three we are assured by the textbooks' thumb rule that we may consider H

to be perpendicular to the magnet surface. (Apparently three is a good

approximation to infinity). The material has a permeability of about 1.05

so this simplification was not good. By using a test similar to the familiar

iron filing set up it has been concluded that the lines hug the cylindrical

surface of the magnet very closely. Thus a coil that would change its flux

linkages considerably with a small movement of the magnet would want to

hug the corner of the Samarium-Cobalt cylinder closely. Increasing the

O.D. of the coil would always further increase the voltage output of the

coil, but the increase would be small because of the weak field a little

way from the edge of the cylinder. The stray field pickup would however

rise more than linearly with the coil radius thus degrading the signal to

noise ratio.

Assuming that one continues to use a structure similar to the present,

ther rules are: (I) hug the corner as closely as possible - considering the

fact that the magnet moves (2) make the O.D. as small as possible and

(3) use as fine wire as possible. (2) and (3) are interdependent. Fine

wire permits more turns with lower O.D.

It should be noted that the circuitry need not be d-c coupled. This

is one of the is one of the advantages of the VZP principal. Previously

this fact was used to prevent the Schmitt trigger from causing an annoying

large steady current from flowing in the force electromagnet when during

experiments the Schmitt stayed locked one way.

One may capacitively couple the rate coil to the input circuitry, thus

eliminating the effect of offset drift. Similarly one may transformer

couple the rate coil to the input. These variations would enable one to get

-27-

along with less rate coil output voltage. A typical rate coil is 1.2"

I.D., 1.50" O.D. and axial length about 0.25". This does not hug

the magnet corner very well. As a positive by-product of this construc-

tion it is not useful to bring the magnet extremely close to the coil.

Little difference in output was noted between having the rate coil near

plane coincide with the magnet edge or 0.1" away. The 0.1" spacing

was convenient to prevent bumping.

It was best if the inevitable radial motion of the rate coil caused

no change in flux linkage and thus no spurious signal. Actually

lateral motion did generate a signal and the ratio of the axially genera-

ted signal to radially generated for equivalent sinusoidal excursions of

the same frequency provided a rejection figure of merit. Minimizing

this might invalidate the three rules set down previously. At the rota-

tional speeds we had employed the spurious signal never caused problems.

The axial component of the rate magnet flux was not useful in genera-

ting a rate output. Thus the magnet was used inefficiently. The rate

coil is to a dynamic loudspeaker as a generator is to a motor. Thus one

might be led to use a loudspeaker like structure. The dimensions of the

magnets practically available did not encourage one to attempt this.

The reasoning being that the addition of soft iron to the 1" O.D., 1/2"

I.D. and 1/4" long Samarium-Cobalt magnet would cause it to work into

too high a permeance to fully utilize the energy product. However

working into a permeance of 2 instead of I, only costs about 10% of the

energy product. Merely putting a large (2" diameter) soft iron washer

on the magnet surface away from the rate coil nearly eliminates the

reluctance of half of free space and approximates this. The increase of

flux density in the wire was thus roughly 2 x .9 = 1.8. Experiments

showed that the simple addition of such a washer almost doubled the rate

coil output with no complications.

-28-

Axial stability will be achieved if the levitated object is subject to

a servo system such that the object moves in the algebraically opposite

direction to any non-servo force. This omits the desirability of damping,

but with the velocity signal available damping is produced in the usual

way although, as will be seen, there is a switch of algebraic sign.

Take Newton's Law with the mass constant.

F = ma (I)

Let the servo force be, (Ka), proportional to the acceleration, (a),

and (K) be a positive quantity. Equation (I) becomes:

F + Ka = ma (2)

or---

F = (m - K)a (21)

If (K>m) the coefficient of (a) is negative and the object will acceler-

ate in a direction opposite to (F). It will achieve a velocity and displace-

ment that moves it against (F). The real mass, (m), has been given the

attributes of the negative mass, (K-m).

In the case at hand, if the levitated object approaches pole I, and

leaves pole 2, it is subject to a net force that without the servo would

cause it to soon collide with pole I. According to Equation 2 this would

be prevented by adding a force proportional to the acceleration to (F)

with (K) being positive. If the actual acceleration, (a), is directed away

from pole I, the (Ka) would appose (F) and the situation seems sensible.

This argument is circular in that it assumes what it is to prove, but it

does show that no inconsistency existed in this case. Conversely, the

equation (F + Ka = ma) with (K>m) cannot be satisfied for a positive (F)

and positive (a) since (Ka) alone exceed (ma). Thus (F) and (a) cannot

both have the same sign.

In a sense we have changed the real mass, (m), into the artificial

mass, (m-K), a negative quantity. As will be seen later, this connects

up with the non-linear or bang-bang case. It is easy to show that either

-29-

the bang-bang or linear all electronic system essentially turns positive

circuit elements into negative e.g. positive inductances into negative.

The practical servo system used must provide damping as well as the

function discussed above. This may be done by adding to (Ka) the

quantity (Dv) where (D) is a positive constant and (v) is the velocity.

F + Ka + Dv = ma (3)

In order to apply linear theory to this equation one must assume,

traditionally but somewhat inaccurately, that (F) is a linear function of

(X). Taking (x = 0) to be halfway between the magnetic poles, assume

that (F = +AX), with (A) a positive quantity. The positive coefficient

(A) corresponds to reality in that this will cause the pole collision we

are seeking to avoid. Writing (3) over, with this modification, using

Newtonian notation, and putting all the terms on the left side yields

(K -m)x+ Dx + AX = 0 (4)

or in transform notation

(K -m)s2 x+ Dsx + A = 0 (4)

To avoid positive exponentials in the solution, and thus bound (x),

this must be a Hurwitz polynomial. This requires that (K-m) and (D) be

of the same sign. (K-m) is positive and thus the choice of (D) as posi-

tive is correct. This is convenient as it allows the acceleration and

velocity signals to come from the same operational amplifier.

Figure 29 shows one of the circuits tried. Z3 and Z4

were used to

reduce the crossover distortion which would be produced if the output of

Z2 went directly to the bases of QI and Q2 . The negative feedback

divided the three diode drops that would trouble one without Z3 by a

factor equal to the open loop gain of Z3 .

ZI

is a voltage follower buffer that preventsthe rate coil inductance

from affecting the wave shaping elements connected to Z2 . Z2 and its

-30-

associated network did the real work. Its transfer function is:

e R0 =-I0

~~e. R C - I ~(R s + R5)e. oR s' o0 0

RoThe (RoCIs) term is the acceleration and the term the velocity.

The signs are proper. Selecting (Ro), (Ci) and (R2) permits one to adjust

the magnitudes of acceleration component and velocity component appear-

ing at the output of Z2 .

It is often desirable to test using straight velocity feedback only. Feel-

ing a strong viscous drag as the rotor is moved from pole to pole is reassuring

and it also permits one to phase things properly. With wrong phasing, the

rotor clicks from pole to pole more snappily with the electronics than with-

out because of the positive feedback. However when (K>m) the poling

should be reversed. This was rather annoying since when the poling was

reversed with (K<m) one really clicks rapidly from pole to pole. The

existence of a small (K) has reduced the effective mass but not made it

negative and the permanent magnet force knocks the rotor around like a

feather.

One way of looking at this whole matter is to note that the addition of

an acceleration term turns the free fall differential equation under an

attractive force proportional to (x), into the equation of damped simple

harmonic motion.

Presumably one can then take over all the mathematical paraphenalia

used with this classical stable second order system. However, the whole

notion of negative effective mass etc. is not very satisfying intuitively.

In trying to fully understand the operationof the VZP a few all electron-

ic models of the electromechanical system have been drawn. Newton's Law

with the mass constant is modeled by the primitive circuit shown in Figure 30.

-31-

diE = Lq = Ld i (6)

dt

with E = force

L = mass

q displacement

q = i - velocitydi

q = dt - acceleration

Figure 31 shows an all electronic analog that makes the real positive

inductance look like a negative inductance, (L-AM), (AM L). Here and

in what follows (A) is not the large number tending to infinity associated

with operational amplifiers. It is a finite number like 5 etc.

Figure 31 brings one into the currently popular area of active circuits,

where negative impedance convertors, gyrators etc. abound. Tying the

VZP to the powerful theorems of active circuit theory seemed to be the

best way to proceed, but this had just been started.

Figure 32 shows the analog of an electromechanical bang-bang. It

is a bang-bang negative impedance convertor. Note that, as with the

bang-bang VZP, no obvious differention of the current (velocity) was done.

The drop across RI

was analogous to velocity or current not its derivative,

unlike Figure 31, where the mutual inductance performed the differentia-

tion.

Figure 33 shows a circuit which was an imperfect, but interesting

analog of a suspension without any servo control. It was imperfect since

it lacked an inductive element, the mass analog, and only lead to a first

order equation not a second. But like the un-servoed bearing it had an

equilibrium point, which was unstable viz: (ei = eo

= 0). The slightest

departure from these conditions caused run away in one or the other direct-

ion.

Figure 34 shows a bang-bang mode of preventing this, and keeps

the charge (analog of displacement) on C in the vicinity of zero. A2

and its associated resistors form a Schmitt trigger symetrical around zero

volts input.

-32-

4

2.6 CONCLUSION OF CONTROL EXPERIMENTATION

The following conclusions were drawn from the work done on Velocity

only Axial Control:

I. It was still not clear whether operation in the linear mode was possible.

The equations show that it was, but the analogous all electronic equiva-

lent systems "looked" stange as will be seen.

2. While the axial control system can be made to operate just as described

previously, one more often obtains operation in a mode that is a modifi-

cation of same.

3. As far as simple user tests of the quality of suspension are concerned,

the modified mode was superior. By simple user tests one means ease of

hand starting of the bearing by pushing it off a pole and resistance to

the stopping of suspension by introducing outside forces axially and/or

radia I ly.

4. The modified mode was simply a mode in which the whole circuit was

normally oscillating by reason of the direct electromagnetic coupling

(transformer) between the force electromagnets and the rate coil. The

additional mechanical/electrical coupling via force electromagnets

acting to change the velocity of the axial motion was superimposed on

the above.

5. While it was possible to operate the suspension in accordance with the

theory set forth in previous sections this could only be done over a

small range of critical velocities (critical velocity is where the force

magnets are switched). Attempting to reduce this velocity by lowering

the Schmitt trigger voltage caused the Schmitt to be triggered

"immediately" by the square wave voltage induced in the rate coil

when the current started its approximately linear rise in the newly

"on" electromagnet. This did not prevent levitation. Setting this

velocity too high did prevent levitation, presumably by allowing the

armature too close to a pole.

-33-

6. The above was, as mentioned, caused by the fact that there was

considerable magnetic coupling between force coils and rate coil.

It was felt that this coupling may prevent operation in the linear

mode. This was because amplitude stable oscillations were only

possible in non-linear systems.

-34-

SECTION III - TORQUE MOTOR

3.0 INSTABILITY

Two basic problems remained in the magnetic bearing system. One was the

lack of mechanical rigidity and second was the inadequate frequency response

in the system.

The lack of mechanical rigidity caused vibrations to occur during high axial

force levels which destabilized the servo system.

The frequency response of the system was inadequate for the fast response

required by the narrow( + .007") magnetic gaps. This was because of the high

non-linear force levels in the axial direction.

At axial gaps of + .017" it was possible to operate the magnetic suspension

system with an additional one pound weight attached to the rotor. However,

stability was poor and power consumption was relatively high. The system was

very sensitive to outside disturbances which would disrupt its operation and it

was found that a soft mounting surface was necessary.

With the magnetic suspension system operating the following measurements

were made:

Radial Stiffness 125 lb./in.(each bearing)

Axial Support Capability 10 lbs.

Power Requirement + 13 VDC at 50 ma+ 11 VDC at 50 ma(2.4 watts total)

Time Constant 0.16 seconds(time for rotor to fall from centerto one pole - .017")

Rotor Weight 1.5 lbs.

A two phase DC motor as described earlier in this report was connected to the

magnetic bearing. The bearing was rotated up to 2600 RPM where rotor vibration

was encountered. Below 2600 RPM operation was very smooth.

-35-

To reduce the instability a change was made to the force coil configuration.

One of the two force coils was removed and a Samarium-Cobalt permanent

magnet armature was added to the rotor. What resulted was a force coil that

could push and pull. The flux path of the permanent magnet armature passed

through the force coil thus providing a bias level for the force coil to work on.

This method reduced the inherent time lag in the force coil thus increasing the

response of the magnetic suspension system. The system then operated in a more

stable manner. However, one drawback was that the additional permanent

magnets added a radial and axial destabilizing force thus adversly effecting the

axial and radial stiffness.

3.1 TORQUE MOTOR

The NASA representative recommended at this time that Cambion consider

incorporating into the magnetic bearing system a motor built for NASA by Sperry

Marine Systems Division. The motor was a three-phase ironless armature type

and was developed with specific characteristics in mind appropriate for magnetic-

ally supported systems.

A study was made of the possible integration and it was found that the motor

could be added to one end of the bearing with a minimum of changes.

Cambion designed and made the mounting fixtures for the motor and installed

the motor on the magnetic bearing. Bearing operation did not change except

for some increased sag (003") at the motor end when in a horizontal position.

A complete wiring diagram may be seen in Figure 35.

-36-

SECTION IV - WORKING MODEL

4.0 ELECTRONIC CONTROLLER

The electronic controller had evolved somewhat from the circuitry that had

been used on the first VZP model.

A schematic diagram of the final controller may be seen in Figure 35. The

position signal is generated by the six lamps and the six photocells. The velocity

signal is generated by a permanent magnet mounted on the rotor which induces a

voltage in the rate coil proportional to the axial velocity. The rate coil consists

of 50,000 turns of #50 AWG wire wound on a 1.2" diameter bobbin.

The velocity and/or position signals are amplified by an operation amplifier

(Burr-Brown 1506/15). The operational amplifier is operated open-loop and has

a gain of 104 dB. A balance control is included so that the output of the opera-

tional amplifier may be set to zero with zero input. The amplified signal then

goes to a complementary-symetry power amplifier. The power amplifier consists

of a bias network and two Motorola darlington silicon power transistors MJ4032

(PNP) and MJ 4035 (NPN). The signal from the power amplifier then goes to

the force coil which controls the position of the rotor. The force coil is made up

of 250 turns of #23 AWG wire with an inside diameter of 1.15", and outside

diamter of 1.5" and is .725" long.

A portion of the signal from the operational amplifier is fed to a RC integrator

and then back to the non-inverting input of the operation amplifier. This signal

is required for virtually zero power operation or, in other words, is necessary for

the force coil to position the rotor at the neutral force position.

4.1 AUTOMATIC STARTING

Attention now turned to automatic starting of the bearing upon application of

power. Because of the high axial forces present in the "clamshell" type magnetic

bearing, automatic starting would not take place. At an axial magnetic gap of

.017" the radial stiffness was 250 Ibs./in. (both bearings). However, for this

-37-

value of radial stiffness 4000 Ibs./in. of axial stiffness would be required to lift

the rotor off the magnetic pole faces. The servo system could only provide 500

Ibs./in. which was enough to control the rotor when near its center position but

not enough to lift if off the pole face.

Temporary mechanical stops were installed in the air gaps and it was found

that with a mechanical air gap of + .005" the bearing would start automatically.

However, this method was not completely dependable and was not pursued further

because of work needed to be done on the motor.

4.2 MAGNETIC BEARING STIFFNESS

The three-phase D.C. motor, supplied by NASA, was integrated into the

magnetic bearing assembly. The added weight of the motor did not permit

operation in the horizontal position because the bearing sag exceeded the radial

clearance of the motor.

Efforts were made to increase the stiffness of the magnetic bearing and, there-

fore, decrease the sag. These efforts were ineffective because of a bent shaft. The

1/2" diameter aluminum shaft was replaced and much better results were obtained.

However, rotor-stator contact still occurred in the horizontal position.

The magnetic gapswere decreased to increase the radial stiffness. On the

motor end the gap was reduced to + .007" and the opposite end was reduced to

+ .014". However, because of the destabilizing force of the permanent magnet

force coil the radial stiffness was not as high as expected. The + .014" gap

should have been 300 Ibs./in. radial stiffness but because of the P.M. force coil

being reduced to 75 Ibs./in., the opposite gap (+ .014") was 150 Ibs./in. as it

sould have been. This resulted in a total radial stiffness of 225 Ibs./in.

After conferring with the NASA representative, it was decided that the motor

air gap would have to be widened. The maximum radial clearance for the motor

as designed was + .0075". This occurred between the outside diameter of the

stator and the inside diameter of the rotor. Since this was the limiting clearance,

we decided to take .010" off the inside diameter of the rotor outer ring. Removing

this amount of material would increase the radial gap to + .0125".

-38-

Because of the construction of the rotor and the close proximity of the magnets,

the outer steel ring had to be removed by machining.

The new steel ring was constructed from carpenter BFM Iron to the new dimen-

sions. The new ring was installed on the rotor and the motor was assembled on the

magnetic bearing.

Although the motor gap was now + .0125", adjustment for the motor centering

was still difficult. Adjustment was completed and the magnetic bearing-motor was

hand carried to NASA-Goddard and demonstrated in operation on 19 September

1972.

On the following page is a list of the specifications for the magnetic bearing-

motor as delivered. (Performance specifications for the motor itself are not

included.) For a layout diagram of magnetic bearing and motor see Figure 36.

-39-

P

4.2 NASA MAGNETIC BEARING AND

Description

Length

Diameter

Weight (with motor)

Weight (less motor)

Potential Speed

Radial Support Capability I(symetrical load)

Stiffness (radial)

Maximum Force (radial)

Horizontal Support Capability II(with mirror)

Input Voltage

Input Power

Contract

9 + 1/2 inches

5 + 1/2 inches

7+1 lbs.

48 RPM

5 lbs.

350 Ibs./in.

5 oz.

+ 12 VDC

Less than 8 watts

Actual I

10 inches

4.5 inches

6.25 lbs.

4.15 lbs.

2200 RPM

6.5 lbs.

225 Ibs./in.

6.75 lbs.

9 oz.

+ 12 VDC

0.5 watts(2.4 watts inposition mode)

4.2.1 SUPPLEMENTARY DATA

Description

Magnetic Gap (motor end)

Magnetic Gap (opposite end)

Axial Stiffness

Axial Support Capability

Servo Time Constant

Sag (rotor horizontal in IG environment)

System Resonant Frequency(Operating)

Electromagnet Gap

Radial Stiffness (motor end)

Radial Stiffness (opposite end)

Actua I

+ .007 inches

+ .014 inches

3570 Ibs./in.

10 lbs./amp

4.8 seconds

.002 inches at motor

4384 Hz

.032 inches nominal

150 Ibs./in.

75 Ibs./in.

-40-

r end

MOTOR SPECIFICATIONS

5.0 RECOMMENDATIONS AND CONCLUSIONS

Previous to this contract magnetic bearings, in general, had three main

faults: 1. They required relatively large amount of power to operate, 2.

They contained considerable dead weight and 3. They were large in size as

compared to ball bearings.

The device developed under this contract plainly shows that continuous

power consumption can be reduced to extremely low levels - milliwatts.

Further investigation showed that the size and weight of a magnetic bearing

can be reduced by a ratio of better than 2 to 1.

By the addition of steel armoring to rare earth - cobalt permanent magnets,

high levels of radial and axial stiffness are achieved with very little expended

in weight and volumn. This then produces a magnetic bearing that can perform

jobs that were previously done by mechanical bearings but without the problems

of lubrication and friction.

The use of rare earth - cobalt permanent magent material in the force

driver also reduces the size and weight because of a reduction in steel and

copper required to do the same job. An added benefit is the increase in the

response time of the force driver because of the biasing effect of the permanent

magnet material.

Although the device developed under this contract shows a great improve-

ment in magnetic bearings it is only the beginning. Further research and

development can produce a new generation of magnetic bearing that will be

useful in both space and commercial applications.

-41-

APPENDIX

4,/-4

Figure I -

B-H Curve Showing Characteristics of Rare Earth-Cobalt, Platinum Cobalt, Indox VI

and Alnico V

.~~~~~~~~~~~~~~~~~~~~~. ~12000

1. RARE EARTH COBALT

2. PLATINUM COBALT 11000

3. INDOX Vl BB

4. ALNICO V

10000

kg B

45-O

36. 9000

27

' ~~~~~~~18 '~70009~~~~~~~~~ ~~~~8000

27 36 45 54, koe

7000

6000

5000

3000

2000

1000

000 8000 7000 6000 5000 4000 3000 2000 1000

BHc (oersteos) 661959-31-16-70

PR I NT ED

IN U.S.A.

IF:35

Figure 2

RAYTHEON COMPANYMicrowave and Power Tube Division

Waltham, Massachusetts

SAMARIUM COBALT MAGNETS

MAGNETIC PROPERTIES

Coercive force -

Intrinsic coercive force

Residual induction

Energy product, max

Curie temperature

Temperature coefficient

7, 500 - 9, 000

25, 000

7, 500 - 9, 000

20

850

0. 02

Oersted

Oersted

Gauss

MG Oe0°C

%per °C

PHYSICAL PROPERTIES

Specific gravity

Electrical resistivity

MECHANICAL PROPERTIES

Tensile strength

Compressive strength

Flexural strength

8 g/cc

5x10-4 ohm-cm

8,000 psi \

>> 10, 000 psi

12,000 psi.

___ Figure 312-280

I I I

J E l ,~~~~~~~~-,1 7

r-!L! r7 + ThT:T !

I I tI _i_ -TT iI

II lii fI I/ t I [_I l II __ :-_ -i_I~~~~~~~~~~~~~~~~~~~~~~~ I JJ I-/..

_._ i ~~~~~~ ~~~I I 11 __L q -,-- I i-JLIi i L I IT [---IFjT~'T M ~tFI~~~~~~ I I- : ,

~,:~_L~ ~ ~ ~II iir ,i_i__ I __

I I IiI I IJJI_ C'LW_ ~_ II i 't

.... I I

-- !- : ~J~--.L i ~~~~~~~~~~I F !i ,i I ~iI,

-I I

I4zl t + 11- , ~, -~:zz Ejq-100 ~~~~~~~~~~I Ii I

I I I -I-- .-

..... ,O& r0 ~-0 - 03 .0--- -- 1 :~-t,

.J Squares, i I tTIc n1 -Fro i m I I _Ar-i '- -

-,o !-_;-__ -' - '---_ ,\ ___ I I ~ : , I ~

!-LLLL I I__L_\ i Iid _k_±~_j_-i- J i '~TI i:i ____ I/ ___LL__ I __-;__-__,mTtT I-r-F-II'-, -- ' I .... - ....

I i I F-F' l'l i , I- \-I i fd- _a~~~~~~~i '_~LLL i,\~ ,\, ;~~~~~~~~~~~~~~~~~~~~I i i i ; i

- I &! M\~ i!!-\'?~¥:--r'~JI 1 I II I; Ii i-

I-

- ~

I ! ttI ! II I T-FI[ '\--\-- I-' !H_ ,,BC 1 il i__L -\-ii ,

-1 ldi____ TT~ I I Tq-T¥ I- li

IQ iiii I __________ i7 F T'p :% ,'_.LI_ ;_i_~; ~ !L! L_~~~~~~~~~~I_ I I IId] 5tq~- -jI

-t J! ]-Z~~~~~~~~~~~I I I-L pI I-it i'-L ~1Ti / · i ' I!! ~~~~~~~~~~~~~~I I i ~ I I I ~%-;¥-JF~

_ i3O -+ lI I~ I I I I\tf-\~''~

too f~iL lFF I I I it!t-i TTI\-~'4~ -- L-

''~"

10~~ ~ SqIe to th I Ic lJ ;I AI I i- I----II -F- ~!il~q

Figure 4. . . . . , | - l- eI, , . . .S. . . -

AtI : AJ AXlIAL-I 'i IJIF'OK~C ti

-I L I IM I 'E ._. -- iRSIII- I]:-'-, ,.~ - ,'--F-H-H-I Li-l-j |0 |S X [ r ;I I!| i

K.,! .DEC 1i , I I7. I LL1 t

ti1. l I t! . ]'

1[4,. A N. Iq7 /1i

i I I i IMEF heS~~~~~~~~~~~~~~~~~~1IL LI

f 4Ivvv1,-C4 F- fiII i Ii~~~~~~1

> _j T rF~~~ 1!liL| lL2 L f <I j;n1 1jg.t~ll-g'r!

q

S

4/1--IL-i

4 ! I

.11 I_ l

I.,iVJ

0

- i I i I i i i i I I I

I IIEO & I I LFAYFRH- I I,---, - II .. :~JJL I N-..-~JJ_ I'LL2'

I

%.

.030 .OS .02010 Squares to the Inch

r IcPpt Ir

T*- KT- -H -i:F T -F--i | i I1 t i.015 .Ol .OGS 0

r' - .NT /l.,,4FC'-

12-280-- l. . u

It

e,-. 3%ACl

z0

0~

L 2.ci0I'-

A-- I

I I I I I I I I I I I I I I I I , I I I I i l , i i t i;I1I1I I 1 [I I .. .. .. ...

I I ! !1 II

I i ' ! I I ! ] It

Ie 1I41

! ,!!l--i

1 .;

Figure 5!_t ! II0:.:: J I I:: LI t tI ! I / _ B I:.'..

FF -LF? . _ +-r!]_I- _Z~ LT ['-_I-- ...I-'~ IL ' Wir--,1~_ ,~Z-r T_ l -_ L -L' I - ii X Il 'Al- L7 DISP:A E EN'TZ. E RS 5_' ~ - I

L171L1L L I _t . .,i r~ --44j 774 AXIAL SPLACE MNT .RSU -t - I | --. -,- r -I , FORC_ Li - . ,, 1

14 f ! L | -~--- | 0 < LI .' -- -TA'.K-E_ MR--. !- I'N=_e; r-_M., . . ,.4N_t- t -H v' :)' E~fY~1=l I ' .~t~Jk r-- 'L ' -i-4--I - - t---l~ E ,---l- I: LLRL wmI V._..A:L._- jme oN".i_ i!

L _LLL_~_!__ L I '~L J~ I I L I I1 L.jII I ' !''I. :m I , i.__L~L~_~---.~-

_ I L I~~~~~~~~~~~~~~~~~

A I-, L4E-l-tkJ.9 "IF

LI-WLL.'_ *_ _ _._ _ __ ___ _ . . --

W I E 1 ! u;t-1 -- i u n -' i-t:--- L_ l__

-

.. r-..- L_ _; ' L i -- T-; -i:-t-I-- I+I I ! Tq-?'_F1 j i _ _ _ _ _ __II_ __-1 I I I I I_ _ _ _ i I

I' i I I - 1-i------ r -'- - i---i--- +-- -,-',2-r .- -- , -: I _ --- , -

I I , I , . , _

I --I-

,_ i ;- -

-IL , I I I I-,-'-; _ [-.i

- *- = --T - i--'---' ' ! tj, ''---.. r- --_ - -- :-I _ t II

ujj~I _-1- - H1 -'i I _-TK'--r-l- --_ li iL; i

I I, I

1i-1 IiI

Il -. -

.' - !1 -: -L

itlIl

t_:_I '. :_-I , !-'

t- I j

I

II

_/_

L!_t__ -- i-L!-I i; -r

I VTtII I _ I; __ I

F I ,I I _i

iLI

IRAYT I H

!.SA/ARIMAA M,_L IT~ ~%. I _. --U- - I I ---.I-- 1-t ; r I I - t-,-~ -L |

_! _1J ~ +--I I :_Ir + i i ! i i i I I Ii--.- ~~~tj- m'--r- ..-----,I T'_- ~ I _I_ L.' I- ?-H

I I

f-L i I I -. .. .......| -t t-- i _ -2 -! i-

| : : A ! I t ' I .

EONU:-iI UM--('r |--

I I L I t

. . I

- --ft-

Ii i1I~

HI L i q7|i± |+P 11 ! E' -. I+ I-!,-TI---,-t4-tJ---f ~_L~~~ _ , L' i l',',;i , , ,-rm-r-I

17FT7 Lilt> IL~~~~I q :7ii-%:;--!_cT_' , - 1. .-_i_;~_t_

I I '-j r -I o I

I iLLLL_ILJL-I__'_!_ f-f-.

0

i -- , , , I _ ,I _ I= i- -_- '-I... .:_: I '.

LI

ilI

l fi-:i(]Z_,~ . I .. L tt.

L I I -

I----- I - !I'O A ,-r I- IIEGo B^A LT-- _: ::--:i I .-- :, , :'

T_ _j--. _ F=I41

. , -I _. 7,

...---. .i-.--f I1_ '' :- -- I

I .i . _

,-77' 7

illii

I I I== -. I----I -, r=-; ,_ -1 I T,-, I I -

I I I-,

- !

-1_--i1I F

-I I I I

-i I I ' -,I 1 1 I

K I NU .POL; -E5 .. I_ .025 WIDg ., : .- ......,025 DEEP___ _r:__-i i ^ r 1- r r i

'; - - -'' - '' 1-'- - I-' I--; 1:

-I-I4 r L]li l II-.i-i- -.- --'-I 1 I

4-...1. ]__..' . L . -.-- ; ..... _ ... J--~- ~ .~. -'-~ I - 'I -I I -1-4- II--i- F-.-I- T- I F 1.A4 iV i -_ L LI I +-t-i -1 l:

_~-___-- __iI-tli±V 'I [I

_.-I -_j_ _01_h 1.-1II - -- ' --I j i '--~ IF- " -

47 "

4-4 - -I I-I-SCA ---- t-L LA -4R.-- 1_LLi_-S.CALF**_J ' _ ,OE'_OR; -Y '-i-l

4:

12-280M. ." II I

-!) 1;

IT T _'-I-i _. t : I I-I- i-- ! I -- i-,i-I- I _ I'-F1-I~'-II-F-F-F-F- i -T - I i _ ,:/

20

5sI

IL--LIT

]i4- FF ~jI IL~KF1Erq

clz

Ir 10

Q0

5

030 025 020 OS 010 0051/) Squares to th Inch tDIS PLAC .JE NT (INCHES)

,, I ., ; Iflmt.mill s t

,_,

I -- I

I-- I- ' t-:-F -- - -I , , I ',_ I

I i._ I r

J !

; i1

I i i ;I, /I , _ I- -I .=_". .. -- .Ii

, E

i -Fq-i%.

!. i: !

_ L1it.

L]

ad_ !I I

11i! I

I.0

L

--II

--Z-I Q-i

<

-_

1fT

L

! i

<

r-1

j _U

JL

R

_

.-I--a-

--I

, _

Q

O ~

-I

XI

IL i

iI I

t-6 ;H-r- __

rlj .=

-_-~:A

I._4

F;_ZG, I2

IF

-F--H--

_]:_i- L'l

tH(sknor) ?,80~C2

Ul

10 Squares to the Inch

12-280.1 2

-280

I{ "1 _

C_]

i. IL

Figure 6

0

l-,_ iz1i

_Z_-

X\Li 2"11\1

-- I|tF_-D

I

r _ri

I I;

I '

i.,

i-II

I- I.I-

F2

I-+

tL

:F I 2_( ~

i r7 c 1

r--t/' _C

_--

o 48

~0

A~~,.

o"

Il-

Il1

i'ii.

I 1, I

I- I I:

1,

f

IEl-,

i

Iiij---l-E

-�_w

�_I

.II

I _U

J I

7 i7r

1__. _/_i_7

7,X _

w-

=_ _! .1

."

II

i I -

l L

:- --

E*_, I~

_ _

II71'

-l--

-I ij i

Ilkl - ,__ I

;1_i -I _F;

r=,71

0

Figure 7

___ L2Jt i- 1i

7 I H

4. _ -i- tL - t _ T1 1 ! ' _ .l I 1 .--

l~~P_ _-E-i-.E|L+ tX- ¥1',_ +-_-"----t ~ i, | m-~ 0_.'-jil ;_L' ~--I ,._ _1.-%.l -,~ L__ ?..-7__l ~ ~ ~ ~ ~ ! 1llilI

~~~ | ~ ~ '"-;-q.-o Z! F - ....'I-; . --' --I! -! ....i: :_ , ,~ ~ [ -1r_ ,.

t 5 t ...............1 ""li ..... ql 'irt F ,S1' ..... r-'-':t ,~i-t- .,------! o.

j1 iVL.3c0i9 ' , /--; .'

- .....ALN I CO. MAGNET I -- '

.- , t ---; -FI.,'t,f -i . ... -1- --- --- K

I -W .IZI I I I- F T-1-I <, I T;--

I +.l F I K &111

a

0

uJccU1C.)

0u_

10 .Sqlares to~ te Ic10 ,Cqtiares to the llnch

I i

I i I I rI- I

. I I I I t I I I I I i I

i , i -, Ij _

' -- - :1- ' ---:- :--::--- .. ' --.-.--

L------ ; ' -l ,'4_i -

--

i t I ! I !.I| I I i I I i }I , 1 . 4i

I I I I I i

~D .030

L, .I -LE

IL i .' ., !J..-1 l I.... 1 77 1 = .......! ..... =L ...

AXI AL DISPLACEMENT VERSUS---- -: 1FOR CE " DIA. T"rU,) I ------. -' -;i.

--DATA-TAKfN-BY-'MR.-J-0S'EPH--iIYMAN ..*DEC Mq70 ... '- -

DRAW B __ PA U L _.. I_.IPSON__ t.W.tJAN 1q71 . - --. - 7 -

JijiiI- 'm;

, .27. .

:.-:-|x.- _?I} LI

--i--I -r -- '

-

-,---i r- i

*., / j ' ---- ---

I ..... .... -L

! -- -I, ... =-_______________ I _____ _ __-_I__I___ ___ -__ --___ _________

I ~~~~~~~I, . - t', - '/iI.,:.::- -h-j--.--I i. : ! I 1I . ' I I I .- . i- . -I -I r-- '

tam ,;3rid , ,G,-i 2 ise.'srI i ... |---' - ' " i

{LF;-LZL'_4, : i , - - I_ ii, - ,

ii'I i

-.I+_ I

-- , -- It t77 (7.77I - - _T 4_> iEl, KFPL77172 T ,. - - I

I j; ,,,,I I.i ____!_'

i ' , i i ~ j | i t- 17 ~ '-|-- ¥

i - -v_ ,i- - ---. '

'-71

. /

,--- i- /- Z't=r:-'__J-. \__ 7--t: -:i-r--'-E-r -!----

-t-L-t..... i- n:_.4_ .. _L -0- ]- - ~

'_L__.. -'V ---- -- :- -1- - ,-- .. T-i-i-- ~----"-1.070 .0o10 . I10w

) ~49

', I _--|.-___

I- _- , , ___

-- ?- -I -------i--I~ ~, . ............

4_- -. _ _!-_ -- ;4 I L._-- I-- -- i- -I

I ;i _-

-: - -

I-; i- .' 1 -I

i i l -i _ _

-1--- TI,

_ __ _ _ __ .__ I

-4

i CFT- -i ~J '/_ -'- 1 T -

- --... ,1, ' ;

-ll=-T-

l-'-i -. ... : -- i- -. .

.1 i_i_.. I 1_ :L .- i , i1

i i ,

fi'7!_ _

- -- i ,-

'-t-!--, -

./ 3

12-280

ALAI

I

i'i -'-

L! I ! 1-I . .

_ -F j _I-

35

30

1 I-| __1L1+.\______ -JI I _ ' It-I i-i-?- !-T'-;K'[FQ_,

I I II li I

. ! H-_II

It ~

I i .±27!_+LLl -1P I iS-_ TI- i'~-

:2 ., :..

- 1:-- .,-I.~ - 'i - I _i- fl ' ,!, , m", 'f -; '+ '--i .....'.h..'~-;\i',%]. it,Tlr '9: f

i. , 4-. q2 U IO

-.\---\_ ____ -.. _.

ao

1is

-i I

-±&± .'!= i I_-, . I ._ _.1 j i I_ '_

i LLI -iF-!.050

DISPLACEMENT (NCHES)

i ~ _

r-;- -;----L -

--- ,l _ -. r-;! 2- T''-

I , _ -i I. _t 1

. I -1__ , _! i

iJL

------ ! -. - _ i- I

\. .

-I ---- - -I_ i'!,.,[ -T-i _,N411 1 T:$ j �_H --f ; 1r- 1

I _i -I-; ~ /I ! -I

';_-i- \ -- t-j.4 .'_. -.!- -_..--- 5- - - I - - -'- -,

IZ

A1II __...... _n.. .. .

Li I _. _ .

I , .- 1_1 [1-

- -; -.; -_ 'i

_ .- .- \

I 1 I ', , I I I I I-IU1 .

, _Lt- I-FT- JIt! i _'

; I

-.I- -J

I ! ' _i!__

I

I I

_,7_LI.!; :. I : jij .. . j , _.I_l_i_L_l__i,

1-., , .1t, I I

Figure 8

7 -- i RADIAL DISPLACEMENT: VERSUS_1 !_r FORC1E .- (a" D-A.TU3EP) - 7-

L L 4 D _ -Ar1rA-KEN-BBY. 1MR. J-O-E.Pt YM_ N'MI ]_ _L 5 D- .EC -I7 +! IVHT.LL - F._!_l_-i i -,- - ' - RAWN Bi ,Y__PA:LMpSi-

-r _; 5 J Si hJ. i 117 l |XA 4 I- + 4 I L 4

F: F ;!]i lI _L I~'- ~ -~ ~-----"--' .:.

-I L _

-_ 1 ' i_ D0 _ _ E-~~~~~~~ F-v'ji OR~YEAF

[ ,, i i. ..; i ! : 1-;-[ J-!T-i, I i--i i tF :)fI~'~-,-.. !ŽjI

I-[; 'II.FII- I!IAI I L FL

!-]:-~ I--_F I

;L_ H i

r! 1 .41LIF_.___4

i- :. --.--. . .

I _ . _F_ _ t | -tF!_L_.LT1 _ _ i ,_!.

IF_ -

_._- -- _ F __I1._.___ I----_. !I

;F'

I_ , :_Y <:

IZi-.Z-i--' Ir i -.j 4 - - ----- r --- q-]------------

' ::!'?I j ! ' ---:L_:_i i .... Wi ; , , .l . , . _, _ - - ' , ' .- _ I, !_ _

J~ Lr I J! i_,L! IL 3.J i 1 , r r, = J ! ~ 1.~'r__4__

-

,~ j~-'-', t, , ii ii

iihi FFF -1 -F v-i - -

-12 i 1 = I ___-_I -'-:_ -- ':- -ftl , -- : K -t-- ii-t-- ----L --- -- :;.. , '~.L |-- ,':- ,-~-' +!:'.- --_ .____:...-_LT- ..._L_::_,l:"--:--r_-'-. _=: -'~-IZ~ T-'-:-- --t---::-L1-i:_-:T __=E ,_:_ - iiII JIiI ! _: __ I I I | ', _ ~ i I_ ' + ' -iS I : '- 1 4', |; ,F ..._: .... ......... ...

~~~~T-i- r-t-IT-: -:-':":- _+ --....-- ---]- +-. I .-r,._.....m n~~~~~~~~~~~~~~- -:- ........- r2T-: -I--:-'-:--:- ,.....

-i- -' -i_-1-I . . .t- -- l- .I .j i .. -'---.-;--"t---:ir:-

-r-_IFIF ~F I __i__ __ _ _ s _.___ .. II_._ -' -- ,__._

.

. ...

- , ,LL : 1_-___:

HU~fL :-q -t . ... :

>~~~~~~~~~~~~~~~~~~hI- I:'L*It-iI I - - ---- I I F

_LLK~ Li-1_

l __ FIr T-l- -

-

-- i-- - --I-T-------I_

._l _rt . || .I I e '--f< ------ .......- l; I tI- I Il ., I I I

rI-:-.-'--l-'.~ -j -r_- -r - -4-------I---:----j-.--.i:_ _ t_ .... 1L f __=

I L i'LTtJ I I <

K I X F F ne i 1 d - -t I- Fz t r 1 / A ;- r -|-<7-| --i -|

Lb) I III FL~ ~ ~ ~~~~~~HF- III i14~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~t- '__ I:!_ _._______ ,'-j~-J--!-_________

_:__M F : .... I ....

X ~ ~ ~ ~ ~ ~ ~ ~ ~ - ,_ ; _' m'-[ ' n-F :-- ]-F- ;--, i--r-T- '--- - F --.- -r-j-T- r. ..... ,-....---[l. ...........................................t'.._!

_-~~~7

,~~~Fl ' I ,_ 1FF :~~I,

ii~~~~~ ~~ ~~~ J~ LLOLlNE - ''______7 - --- ___[- T

____T--___ I~-- ~ - - ~ ~- ~ - . .. a , _ ~____._--_ . L__ * ; _

I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I

t-J--I It 4I I-iI__--- ~ - I:I I_ _:-_--.~-,-1' ..I I~~~~~~~~~~~I '.Z":T1F- . . . . . . . , .i... . . -rT ...... 1F~i

I i I F ' F:ri ~ T-J-T-- I ' ': , ' * , --i F I Fi I !

-I [1-T~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-"" 7-~-~ '~, ' ' L~~_ '__. ::~ _::, .=::% "'~ / :-~-'~~lLL-~ ............ .....

JLL-' ~ ~Y--F~~-~-:':--~~:- _ -:,ZZ_-_ E__ -. LFIt:, F- -/ I-F-__ ._.I_._-__'_ITImiPr

¶I I LL " JJl I I II F LtlI I !~ ~ r F ,F II IFT~ Iq-qY l" _~ 'T]iq-]i- J-

i T~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

7{TV~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ' __T' '_..Tjj.Xjjjjjj' , ,1 ---- F ' JF . I ;-FrJI I I

__I_- : 'I~ I .. . ... ~k' /;1hA -q.,I--~""-T:-I-T::-- ---- t-':--

.) qures to the Inch3010 Squlares tol the Inlch, ,1 C 0 o 010

DlISP LACEMEN-r (INCHE S).O0W50. .030

12-280e.. r.1 , ^

-i-T

I [iL

lV

:20

a-1-

iii

c0OIL

I i I ! I i I I I I7 7 T ii , 7 I i i I , ' r ! i i I i 1

-~~~~~~~~~~ _I _IL_!- .......... >1_ 1,1 ':' L'zF'_~z 6''J

--LZ-I r-:" "rvli' S -:-"-J ?"' ', -"- I I--~-- -_. -I __-I-I ... , _ .- - I I-- - - I . I I

.a

I

I

I

I

i

F

__l

=I I

A1

I. -.

Figure 9I I I _iT -Fi

- ----

;gI-.t

.......... ! --, -t '= -.r L_.I. !-_< t 'I I I 1-. _- iT1 _1 --

ANGULAR DEV.IATION VERSUS ..-..FORCE ( ... TU.).... .... .........D-ATA-.A K'EyiMR.- J-0-SEPW-YM-AN-

-7_ D E CI9 70 h L - --H I-L------ DRAWN BY PAUJL.S1IM PS QN 7',--

-; J A N H q .'712:ti.'--ii , i ,{, -T ._-L T

'1q.Ill

_ -_i

il t tI Li i i'--'_ i F L _hl _1_J_

"; . .... : .... .~ _-'_- : ..... 'i";?" , ; -F!i'-l-r-iS-ti-J--.- ~l-l-'r-h-

i Fl' I T- i 1- !

--

-1 .- , L'

- I _ _i _ l--- :- i _' I

___t- i t mt I + - t- 1 ;ii . L -th -t-i--- __I LII-

t- -- t= !I ,,! ,:'i' :.l ,?-I'"-~.--I-t-<-'-:- :f- ... %-, :-~-- -,-.r-1- ,- -- ' T;" ,-i'-;

L._L _ ', -i [ ,; '.F ,_ .'[: -, .. ! i -- ; ', ___ _:_i- __ __.__ ,.

LII i I It /tl ~ i-'. i ' i i i It-F---CTr-r- m .- - FT-1---=-----t--F ---- -f%-:1-r-~-m.-- _ _,.

- -"F-- -- -

I !F

I i;

:. IA_ ... i ......_,: ..... . z-.I -. - - i---,-. .... .:-.'--'t . h.05 I --- --- .-

1--Ii

I -' i f T .pi-I iI-%,~-7--- -- -i I ._ t-11- .- Irl- =41l177 -, i- --~7 '-r

II

/H lFiii-_I !_I_ .F

_1 _441 -I,..

I

I L _, -~ ----- -- -, - --.- I- F- ._--F--H- I --- A 1 --.- ,-

I - I PIVOT

-_-_--.

II -247S --_ =E 2,,675. _

-I-HT

r-- , I I F7' . ' i I'A-LN I CO"--MA G NET--_: _- - -'- i --

-H,-I -T

-FI I-I

I . .

F I

_- ii_ 'II-j .77TT

I I-- l

I. , l .i .,I,

I -_ 1_ 11_ __ _

I -____I _ _ t:_lLti-_

I, ,

I I i I

1--H--H-H ?H-Fi-H~-H p -, ! . I j !-I- L__ I; _

I- I;

; iLL I I I

FI - F It _

I I IF 1FFII!I F! iF l r!L !LJ i

IZIIZLLL IIi -, i | H -w

i I - I I I

! /

/-F, -

! I I

1 , , .~

I i I '- -

I , - FL VIi-:_ - -I-'r'

tI ! I iiITi- I F-

,1,- 4 It FTTL --ilt I 1-l i t ! X i

IF! L IFI

! Ir-- .-

-T

I

--- i--. r-- - ,_ :r-, = - -I I- i-: '-?!-

- --- ! t-I-I

I___: . .. ! IL_ . _ . I - _ _ _ - . . - -- _

1 ' .... ,---r -r?- If' ... ! :_ --- _ l

~-tiIohz I iil.! ! ! I I ~' t ! i i i l'i I I t i I I I i ~ t I ~ i ; : ~ , I ~t I -t- ! t-r I t i I r

I 1 i t ._J! , I ~ -_L , i i t ! 'Ii T-7-T-- F I I ,--|-7--Iq-X,~~~~~~~~~~~ ___L_ Z~E__ 2i E_;~.1'_ E . T..-ri:!r- ~, ~. :---~.':-. t ';N[': ~ "'"'-'II '----':-+ "--' Fl

|T l;t h~ttW i.K 'H _ 1-r L, JtIFP___IF_;. _! _ i _

rl : , _1 ._ L--I !-il............_.i|_,'..Lt-i L

H~~~~~ I 1!

... .... -F----4-:::' ... ... I...--H F. 4 / 'J1'L.IL ............,

7~~ __ ___ ___ __~F~~hT1i~F IF ____ -i~~~~--.~---- ...... --,- I -F 4.i ___ L I-~I-~-LI.L 'i-i--_-i~i -~

1S ~ ~ ~ ~ ~ ~ ~~~~' ,'''''''!'' ''I. r ! 5I

51

12-280X 2-280

I[

4IC'0z

-.

'U

u2c¢o0IL

10 Sqas to Ohe Inch11 Squlamzs to tlle Inch

I. U 10° 5e0 5Oa 0 106sO 15" 2.0o

ANGULAR DEVIATION (DEGREES)

.i-i-

I -1-f 1-11 I I

-4

iI

- I

--I

-1I

-1I

I

-1

I]

i

L

j

-1

t

! , I ; I 1 . i . - I I

IiLpl� I �jl I -, I�_,E I- . . - . -__ - '0113,.1 .!, CO � �- 'r� �--T-__ 1 1 ; t I '! -, 7,.!I- iv l-dEd . ,iohJ, t I

.Clar

-t--l_9-r-

L---Li___

-:_-E-i-'mT-7'

I I

-LIt! i .1 . .-

r0I -m'---i'-~ r7--ri I IF i-i . I'

I :! _

-- I[EZF..'z- Ft

I1:;T

PHOTOGRAPH 1

1-,

/

^ ^ -jS

SAMARIUM COBALT MAGNETS IN CUP CORES AND LINEAR MOTOR COIL (1" GRID)

PHOTOGRAPH 2

•

/

/ - **<_ ,

MAGNET IN CUP CORE WITH FLUX CONCENTRATING RING

52

PHOTOGRAPH 3 FT""* --" r???i»«Wf?W5W ,^f'*W*'- - ' , '"« ,-tW, " " , ' . f f l ^ J |

SMALL RATE GENERATOR

PHOTOGRAPH 4 ff)T^y»!IWWWPffS*?.WW "'. 'Try T'W*,^.--'"jfiyPWiV. gT^Hg-^^rv?—^!!' q ^ ? '" ! , , ^y- -^ 'V* ! •.•p-'P^*—>WWt .>. - , -ny.»-—7-T^JI

r*

• -

%*.

/ *

\ *%S

SMALL RATE GENERATOR WITH SAMARIUM COBALT MAGNET

53 £

PHOTOGRAPH 5

I

!

| f - • / • *

^ i«s»ai*'*

r f̂ !

. , . \ . -- ,«.^vV*** :^ &*&£ L.

AXIAL TEST SET-UP

-*

PHOTOGRAPH 6

"

« *

-I

]

.wawrHto. • .•»» . -r isyr^^.-^--^^.-- <&&

TEST FIXTURE

.̂ 4

Figure 109~~~~~''L'. :I4,~: ,,_, , ,U,0 k

)22 +A t4 L -| 0.. [ rTN t -L I Il L rl -' ^ I-- - -'~-Up- -o _________I I

zIt 112 IL It LIu l_ ._t . -! _-L -- !l- 1 ---- -t --r-b-V-; I--- -tl. o o t!sis ~~~~IL2L -i-i"---~ h: R-J i .-- '.

'i"- H -- H_-F'+ I -- --==--_i F'-- =+---.-F_+- L

Ir _I ; t 1---!- - -- 1----- -i--- -- IK---Kesio1j4 o - ---ide -t Bus in t up.ovlIWI _[_30 Shims

_L JL' _~_I1.lJL, l I,===== LE===F i -ZZ._fTe-_'_- ; ...-LI-i-i-er.'e ._ -_~ PPIII JLI p 3- hTrjK Al. Rod

___. -v .lLII __.,_.} .. _ H- tixi . . .i. : .:

aiG=ion u. -ForA Bronze Wire' -g - 1 .... R ' -- -i -- I . . . . l--T-, - i I --~ 'C? !n-ping'

' ! --t-_ t - -' _.

.... Il *;LL _. ...---- , , iT-i Cup Cores,

-- }-i- Plforc . .. C " ARo.lam in Stand --

7 -K~~~ - Bench-T-

I- -I V17 I , : I .... L

I-_L--N_ LIGae' 1 ."-if 'I J oI I re -- _ 'i ' : -

_

i~ ~~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ WH r

I11~~ ' ....

zM ! I: '_ - -- '---Pi . -...

-Q -_ I __L.____I _~~~~~~~ _ ___t __I__[__ _ __ ._ i I . -- . ._ ^| |_s_, ,_ ,

S . ! rjk . _|, . !._| .... . +_ .1 -:- - ---- -r --- I [1 m , , !~~~~~~~~~~1I I -. _.,LnL ~ ~ ~ ____ ___ ,__ 'l4Rsr ini I~o Usd - I- l---

-~~~~~~~: -?:'io A I~:} I I L B--neC°- 2I..

I ~~~~ I I i~~~~~~j 1 I ~~~- 7--- --- .

i ._13.a~ c Clar nc -. ' - ' r-

._~~ ~ _ .

FC' 7r I

11)s ~.. n /srestoh lc55

ho~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ;

HI~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~,

l-td-i AXIAL-TESTand- _i

ES-FIXTUREi- ,,-L J-

10~~~~~~~~~~~~~~~~~~L ', :: to (fi _.-I__--Inc

,0 Njuare to th IncI

T -) r-FI7vv4X:i-[7-vrm TF

- I,-f= =I .... I_V_

; Fttj .iA-L74++ A-Al

AXIA L DIS e LAC E NTi VERSU5-

--,: ...... _. -_:_*- ._1_._ _; - . ............... .- ]....... I_.|--,-i-'_FORCE i iA l : .....4 .....----

-DATA rAfiK -N-BTY MRTJos P H.Y

-0 RA WN BYift.AU LZ!L- P:SO WI _ ,_ . j_ - -t - I- i I -I r

, d j , ~_ i¶7 ,I LuI I - l i/ !-.I I

I _1 __ _1 , _ 1 |t__L__

' i_ ___J-_,_i_..L _ LL_ i-_- ------ _

_-L_|LL . , ,I---.. I !-iI'-- -I I-1-? -f-:I

'f 'i' ~ , tlli__

2-1:I:

'A-i+'T!-r F-F! _L<iF Tit- t-tt C W-E- ~ Z'TZ[~E-FB r .-~:-4-`-j.:`..:~.~~ J`~`-:;i�-.W-.--,:;L-T-W-i.- 7LJ_Llll |- _

ALJl 2-L--' ,L---- S - ',--- _._.--_ .... -_ ... t.J...' ; - "--,-:,,i- ... &:-¢~.-'Ci-i i-'.,b ~ t--P,~. 5t^Pu -; .... :-- c 8-:

I I [T I +- ' i , ' I I - 1,l l 2.i-:¢. -. --- l

I1-; h- .. I d-hLi ~i lIi_ L/TiLJ ' I !. ,,ml,,',1 ','___ _ _ ____ ____ I H I ~ ~ L Ll_44

/tw F-i-;- q l-:Fi-H- -- ?* ifi= F- -F: 1*;; -R------i :Fli- - T - -- -

= - - ; -

'-q~-'iFTq-'--T ...T-- .. - .......L].... -'I.......... i.......: :....... : ' ..... 1...' ...

-71~ -- F-[Y:-IF' Y~-:'~--F.T-- [--[ - '- -F'-i%:- -T -~ %' - 7-2

I ':-'--:!F.%. __ _

T'~-t-=FFZE :-i:_.rLF ZLFTL:-,:-_t;~-i_:I_-~'-~-r-.r:IZ_%Y~~~~~i_-t-I~~zTZ Iz[~T?_ Zj% -_LT:~Z; Tr:F~.. l ! i-M-~ ~ ~ Li-~I

, -I::_.:. .,__-. I :~__

,ii-r I--- --r- -II I -:-

- _ -,--- - - - - 7 - ---r--- _

-p__. _-[---=-.- --:--I ! ! I I . L I j !,,-

_-'- -------- i---- -4 I i-I'- - -t----'-i-' ._-_-_'L . .r- .' - __LZ__

-:... --, I - -L- I;-~ ; , It ~ I _! :- i II . -:_', ! 1 Iil

-ITC'F~rJ, ' i---i J :----:7dT:-'- '-F!-:- T-: T7 !___-_F'--,'-i -- " ;

~. f X--l LE1 IILL -!-t ,_',_T-_I. -I- , I I I __i_'--_--[ L i -,I -t--- i:

1 FI I -I I- -I I 4I -t F I-/,-i-[-

so::'-_L iFX .i'!: -! ! [: t '-::-- -- :2-_-EE~_Z|- F- ZZi-iZ_ :T_ --S

li 1 I I i i I .I ' I; i , I

I I' I/II t t !i ;iz LIK ± L

1-------/

i '%I_ 7--?

ftl-r -

L ;TZ !I. 5O0; . . . T _ _ _j

Li ! .-- ?I--:] -T-F- ~LriILf-L]F F|i_ ---F r -FIT?,-Z~"_.?..! +r-i N |l ~:-:-: ~-! ~~-1 FT-i- _--:'--T--.. -- -r

_I.'_ ____

,2"-i I i j Ii I I | T i IIi t L ' I : 1 ' ;

,L]L :[ vi- r i--;-- --.t- -- _ i -- :--L. I I _

L 4i +KL!iti7 K* i . .i I iI I IW I i i ; '_-!-'_

__ _I_ _1

i _l ! ii-r,. ' T-FIT!'-T:-'

II, i. . 4

* i I ,

I i 1

i - asa_ __ II1-- -1---4/. - - I

I 1 r

I,-

I ,1 ·

... ]-= --1---tN-=:f,t----- wFM- IQON+-4-',-. ' IOr-- ;-'--N-

. _, . . .. ..

* .....................FM . .. ._7 _A..._

aTFRAYTHEON -' - I---I ~_L4~ S:'A' 'AR-I VM,-C66-A-/Ti '_:'-_:

: MAG N T:-!-

I I II I

. .' . 1 , j

-, _ ,-- I-' ' - -. . ... -A-t-: ;

~~~.6

-.I : 1T I I; - . I,-- - ,-I - -! I_ !,+ 1. .. ' L .

_.i4 _ 4i t .1 _ j_

4 -- -- - 1FT:i : ! ,I Ii

i-I

.-r- --- ,-1 . __. I

,._:_-:-- t_

!:-41]

-- : I.1 ,. * I.

.- i----,- - =T -r -.-

.. _.: .: _ .....]-4J _ .- I -' ..'..

___~~~~ --...... L_. ....

I ILJ I t fil '. I t i I J FZ L, ]:_~_fZFZ I -_r] _F_iZEI -L~ ,TE

I __"--ri'-':---. I - -- I- - _- ' _ 7 -

I

Figure 12I. I. . 12 .I_12-280

--..e , I .

--- i-F-l --- T- ; ... .: '-

7w

'R I6N S -,S 0O. D.1:..B0aiŽ IROR_ -'--.oE a5'I R .oZ5Y:--

'7 ,-;.1..0-:'_-_

. I~~~~~~~~~~~~~~~II;1-' , 'I'. Ti-L-.-.-1--1--t~~~~~~~~~[- ;--t7i7-1

' 1--i WF' -; f' I It I ,

30

IU

az0

0.

0Ia,.

20

i| i L

Wi-' iII-I : L'ti-]-I-

I~~~JL I I~ I ! J% ! I ,L H -±±LL-.'LL{±-L-LL VJ4-Lj.p018 .Os .OI2 .001 .00& .003 0

10 Squ,.r. to.t ,. T c, DISPLACEMENT (t O ES)

r-. I I

__j-1

-1

;,,P l',i-l ,~¢' -,,L :q It |J-i;['.

:----:: ::::-::F: i- :---_; -_--ii- r-F

_ I

I ! I

t- ._ . I

__ = II ;_

s"7 -/i/ ' i ~ !TT1 ,7i fi: I 7

:_ -:i-: 1 i I -

------4 r: :::-- -:---- I

_.: :1I!:;! .I!i l

I I ..1'! ! , ' i

r I 1-1:LLttrI

"A6 A

$2-280-.... u1 I .

Ia

E-~q'w

0:.

i

lulU

0it,

.03010 Sqtuares to the Ih'nch

Figure 13

RADIAL DISPLACE ENT :VERSUS.FORCE . L V - ,-4- --D-ArA7 EWB YtIRTJ 0 SP WY-m AI8 JAN 1' 'WI K |-F ''+7-DRAWN BY PAULSI-Sl1PSON -YF

I P 1?

-- !---,-T --r--F-r-i ....

--.... = ... ': -\:-i--T-4 i- -'---, \ ,27 -t-"- :_- '__ ';. V

.010 0

DISPLACEMENT (INCHES)

__ _ __L -~__ -- 7 . ,

. t 1 . ._ - . , -- j

J_ _... .5 - ---- __----

~~__ d____

F- ---_ . j _

r-- -.-- -- _ __

_R I NGS ,5 O. D._SFrM I RON .025 :X .025_,

- i ... -...... j ., -...

-T. -RArrHEON - ! !- !J

'T ' [;--I -: r SAMARIUM COBALT;- ;-i. ;-i-M- - .AG ET - I . '-- ..

.., ......... ; ...... , ... MAGET ... : ..... ......... _

.57

Figure 14

20oa-I

o E

I-oOf

-4QjJ

,

0)

C:

0-

fn

LJ

3zI--Cc2<

CL

lC

D

U-Lii

z05-(X0 1.

0-O

f

U

.N>-_J

-J4::D

-Jn IC

4C001-t-0121

?-

..

?,o.

(;

.58

Figure 15uJ ;C

r~ _

wo

I

oo

-~

z

/1

' tt3

' 3 ~ 7/./'

qz

G

~ ~~;

5£

z z -- A

;?'

~"z<

' Q

>L

4 b

<

>J

·of M

~~~~~~~~~~~~~~~~~~~~~~CL

~_

__

z

} c0

-o ,

,U-.

in~

~~

~~

~~

~~

~~

~~

~~

~~

~~

~~

~~

~i

.~~

~~

~~

~~

~~

~~

~~

~~

~~

~~

~~

~~

~~

~~

~~

~~

~~

~~

~(.

CL

O

<

v >

I --

0I :

-/' X

I

H I,9.

C;|

X-

B-

di~

~~

~~

~~

~~

ri

+_zL

.

5 9,

~

PHOTOGRAPH 7 | . « . I . I - i i i i i M u u . i w c ' * • " ' ' -i-IAi.—»'»P'H»-J^i-«..J«'l^i<.".»Jlti»^i»lJ»i:J» 1 . JM)li>iJ»nim,-l, Hi mn»Jt»H>l|i,» , ' • • - " ) ' f«

VIRTUALLY ZERO POWER MAGNETIC BEARING IN OPERATION

PHOTOGRAPH 8

VIRTUALLY ZERO POWER MAGNETIC BEARING WITH FORCE COIL CLAMP AND DISPLACEMENT SENSOR (LEFT END) .

60

PHOTOGRAPH 9 Pf^tM^ft!' * : ^'r^M

I II t )

jfc'F *<**'l

VIRTUALLY ZERO POWER MAGNETIC BEARING IN OPERATION VERTICALLY

PHOTOGRAPH 10

I! ! I I I

^ VIRTUALLY ZERO POWER MAGNETIC BEARING

IN OPERATION AT CAMBION

«1

Figure 16

-h om, .+_OUTJPUT VOLTS>EASURED BETWE£M ,_-- ! ! _! .I I I 0 IV

X .1lO.ITLS-_B AND C._,ON. SCHIEMATIC''-rli-11 I , · ·I | -t ̂ _! --- -1

__ IjNPU_ iI LLIVO0LTSPM ASURE1DO_, L----- 1,LIiE/TWNPQ'lW1, A_ AND COM^I ' It.J ! i I' ' ! M A T ; ] I 2 f IL

I!,O IS.H EAT I IAre M.: - ! -I -_ , -L i

k!lIP_,UT. IeASUReD WITH KEITHLEYi-i--F F-Ii- -' :;I i:-i-

OUTPUT. VERSUS I IPUT' - -. 1.i-V.I RTUAL- ZERO. POWER. BSEARING

_XM)Cfffflet- -7 -----lLlL DATA- TAKEN-R. JOSEPH LYMAN -i_

76 MAR/H -I ql -1----i-lF --I-''-'1'--''-tT4---I I . I I,

bRAWN BY. PAUL_- SI PS 1O.,_25 . 4ARCH4. 7I -I - ,- .i .,--i I . . I I~~- I~

Fr;-i-!I4-I -

ZZ -ZZ~ OB-_HIMC 'RO`VO LTT.7AIMETERTi_ i__i___ _ . .....- !. .. i -

_ _ O PUr--MEASURED- W. ITH KEITHL-:EY '--F-C E THI-IIONCCOHOk'--~I -4-,-, .rT -.. .... . -.... .. - - --1 !-_- - - $ Cco445 cncrrd- Aveue 1-

... L ~__L. . ! ..... -.'I. R -

,, ,~--- Lt-I -- I- -'-- -i ' L-- _- _-_- .t- - -'-| -- _

_-i- ...... _ _ , _ ! ..... __-__ -_- _ .... !_ - -- --t-t-- 1-- ' -- F

-

:

-- ' "

_

-

FT-%q L--I~--.-:t EX-i+l _!................. I:-'----'--l' 'L------- -~~-,- '-: ----: .... ~---...I.... .... ,...... ..... t , ti , -r-- t i ! , m - -L | - ~ - i - S t r , ,t '-~ i_ .'

F7~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I

-- 7T-L4_LJI-1,_!L__I -n- ! | ___ !: ___ i _LL_ | ' t _ | - 1 1_'___ - t tI_~__ _l )_1_1 |_j _iLI L

I I t+ ~~~~~~~~~~~~~~~~~_ _ 4-I

- __I__ tI\-I i I ! t , ' 1 I.J|1 -F--j j -Fi-|-7-,- -'--- .I i -7-~-[1L' i -.,~I~5-;-'-

t~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-[-i-i-- ri~~lW~ -F _"l ' ~_

| F... ~ !t-Wt-fit k-i |-,-<: , _ -~ -h X--iT____ _. ___-

______ i t t f L l i - | | < L - ! | 1 1 ! L __ L I+ ;

I! __ I F'* ii-F- i II i _ l_!_1_._-f4-- K-i-W !K-,-lii I-i--I i-

-I F

f~i L ~ 7\--l--'FT il-I 1 Jl.LI-++7

| | f | ! ' i f

t ~ i

i-

_Xt4 -1-t I I I

i i -I- !+X7 + W ! | i r \ET 'r F-------r-ii F F F I ---

-F~~rF r !--1-!-1 L744S~~~~~~i 1 l-, r l3-y? i- -FF ¥"-"q-i -' T~ t j it r t ~~~~~~~~~~~~~~~~~~~~~! ' T i 7 ! : |1 ! W -; l=Si- _ - -ri_>S t_ _ , _ __ _ t_1_-r_- -- l__-_: -_ -I_-. ._

4~~~~~~~~~~~~~~~~~~~~~~~i

I I &!_ II I i _!

F !_i

_LLi ,L ~ i. I L_ L:.I ; i i _;-___FF__; F ,L ! _I ' ! ! !-! _L ! . _ !_ J I 1 I11I--TTW

-F,-r-i-7

2.0

I FFi -I I L

IIIII F1 !- IIL- i I

/,

:-Ji- F :7'_

T; .~ i | ' ! - II

I F..

I . ~1 - 7

!- l I F li I!---i'I:': , 1 j .....: -IF, F.F

I !/:i -r ii_ !T !:i- :1- ; [rr--t--?-r-A- -

1,00

I- I - i-I . : 1 '; '

I ! I

r I92..

F'.

- "iT - -I-:'--

F1 ilIEI l

I - .- -! -- -_.-.-.-- ;-.-

-_--- -ET~TI i

Il7l7 --r- -1

r-i-:l-l--r--

I I

ii --. .- -- I

'i-IF-?-iF----]

--- -L-i-

F=7 I

10 Squa:res to thc InchMILLIVOLTS INPUT

12-280

I

-i I-

1tL!id

_ / I 1 '.-- _ '_i I 4_I

. 1 _ -- :- -: -:/-1----

_!! !

·\ ' . ....

I IT

-!- ~-' ---\

_ ._ . _'. _.t

.' . . I

_-i-1- .-

f --! _1_! _111T :tI (! 1I I

.62 2-

! -i l-

I ..

: ·

..:

.'~~ ~ -- '

; .

_ -

.7

; ,

.

-t\}

.

' ';

.. +

,

, ..

., .

--

.

; t

., ,

i

., -

i

.,

,

* ,

X,

-'

f X,

..

.

..

.. ..

..

..

..

..

..

..

h ,

..

..

..

.,

.

£)

-.

.

I -. :

-;

7:L

,

L

.' .' .

-:. ...,.

,.. ; , .

©"

~.---

.4 --.

"'

'; '

.0-

-.+

.

-''

U'*'1

+: .:/

z0Iii0 L

O

a:

.4 CD

7 F..

I''I Z.a

-J5.@'it41

". I4

_. '

.,.

.L47

a

Z

+M

U)-

'

:U)

I- O.

,0

· ·

'v ,' +II

b. U

:t)~Ld '

IL .) 1

0-£ Pw-z

lU: ~

'

-U u W

4 4.

J.*

.+

I >bJU)

Zag

* 0

z V

101

us

w

i.V -I

ii' 'tj 2

fI M

_o'w_ 0

vl

.,

(-..)*f

cc

I: Ii

CFs

'

C :

|0:::~

~~

~~

~0

cr- w

>:ft

~~~~~~~o-

-r0 S~ ~6x 1f g

" * ,a: t~~e te tS .-l

: -, ;9

11l f- M

E

-.

:

.t -

t I

bi-~ 1

-IC

1C\J

.E

()

I-0I..

.. W

I1'8

.,,i 1,C

WL

P

9\ o

... .

., ,

.................. .

0

_ --f _

-',

.--.

;

... ;' '-S'' ,' ,'' ' --' ..... ;A

.

..

, * '

' .

.

..- ;

.,

..

-}

-1

--

l ''

I 'S

-1- ':I

..

s -

I' 1-

'

1 .-. I'.

1- 1'--

I t

r:-#- t

1 1_

1 1

e::£

I(p

¢LL '4-'I.

PiC

.I

LA2IIT N

OIS

JA

: I~

1

·,.%,W

N

OI S

]IA3H

~

l:

.K

ti

rI(-LL @

. -0

zM4-iu0*00IL.aL.IJCU~0

0VI P.

z U vI-

~I".'.!,L

o ·-

.

1-

I

..

I ;

I I

a I

I

!

1

.I

Id

It

r -:

:~~~

I.I

.j

I .

v

, f

7. .

.-.

I .

.x

I .:

.I

.~~~~~.

I

.

''

IE

...

,I..'

.,

.-, *

-.f

I

.: 4

.,

.

.,

)

:.

.,

, 5

.

:

I'', ~'.

..

..

..

: .

' !

..

.

..

/ .

*: .

.I

I

._

II4

I ... -

-S t

... .

--I

.' 'I '.

w

.

..

.I

.':

::-

-I

cc; ., .

.I

:

I .

--.

III -

iQY QTQTNCQTY N ODE PART OR NOMENCLATURE OR_TQ TT _ IDEN IDENT NO. DECRI

I-! _- .: -1 A ;R1 ___ TTON&. _MPTFTER. HT(HL.... ... - -POWER. ANALOG DEVICES #402_1 - -_ - --

1 H1 HALL DEVICE, SIEMENS #SV566 __~~~~~~~~~~~~ , , , , l _ . . i ,I , __

___ 1 L1 ' MOTOR ARMATURE COIL_ }1 i _ iiiiiirL ii·

-_ ___ R1 _ _

_ .. _ ... ._ 1 R2 5K /2WATT 10% POTENTIOMETER

_ ' L[ R3 .470 1/2 WATT i0% CARBON . .

- -. _. __' __ ._ . _ _ _ _,_- - - R4 .,_ . ..R4 ...._ 470 /Z2 WATT 10% CARBON

R5 470Q 1/2 WATT 10% CARBON

-_ - - - R6 _ _ 470 1/2 WATT 10% CARB ON

- - -. - R7 .__ ____ _2.5K 1/2 WATT .0% POTE)T' EZ

~~~~~~~~~~~~~~~~~._., ., . .,. . _

R8 25KQ 1/2 WATT 10% POTENTIOMETE

)~~___I 12 2 WATT 10% CARBON...

_ _- _____ ________ ~ S 47no ? 2 WATT In0% RARRnN

Rli 470Q 2 WATT 10% CARBON

Ri2 10KQ 1/2 WATT 10% CARBON

R13__ _ 10KQ2 1/2 WATT 10% CARBON.

-. - __ ___ ___ _ _

Q1__ ____ANSO!_ RNT . POW.TR M_,T]non

} Q2 . .rRANSISTOR POW/ER MJ900

. - _ _ -.I-__________________ ______________________________ .. i . _ ._,, , ,

_. _SK.E._E.T..m.

LM ,FIuRE, 8

_____ _ _____ SCL .HE ...___._,.....SCL-....- ... .._=' .....__.__,__.... ·................ .~. . . ._ ~~~~H.T . _V ,. '...

I

Figure 19

MOTOR DATA

1. O.D. Sam. Co. Magnet

2. I.D. Iron Collar (Return Path)

3. Mean Dia. Air Gap

4. Mean Field at Above

5. Magnet Length

6. Wire Size

7. Turns/Coil (There are 2 coils/phase, 4 per motor)

8. Active Wire Length/Phase

9. Back EMF @1200 RPM

10. Inactive Wire Length/Phase (Est.)

11. Resistance/Phase

12. Peak Torque/Ampere (1 Phase)

13. Average Torque for 1 Ampere

in Each Phase

14. Average Torque @0.05 Amp/Phase

1.5 Cm.

3.0 Cm.

2.25 Cm.

2,000 Gauss

1.5 Cm.

#36(25 Circular Mils)

600 Turns

3600 Cm.

10 Volts

5000 Cm.

160 Ohms

26 Oz. In.

26 Oz. Inc.

1.3 Oz. In.

. f5

Figure 20

u-C.

rw

(frVI

rY0--0

0

a C

)

CO L

LNJ

O6

U.)

Ci-DIN

0l-LU0!j~-v

; I

LeIa:-CLiU

J0:

IY

I -II

Figure 21

V)~

~~

~~

~~

~~

~~

I

> J

0n

CL

CH

ct,

l .v\A

~ \/\

/' .

, U

\J ff

X,

Coo~0

Page intentionally left blank

$g4iz717L(uXzuFigure 23

RATE COIL

= 100 KS2,1/2W,10%

= 10 Kf,1/2W,10%

= 100Q ,1/2W,10%

;= 1000L-,1W,10%

= 2,200f ,1/2W,10%

= 0.1fL,5W,10%

= 1/2 747 OP. AMP.

= 6A, 50V DIODE

= 16A, NPN DARLINGTON

= 16A, PNP DARLINGTON

= 25 MFD, 20V, NON POL ELECTROLYTIC

CHECKED CHECKED

VERIFIED APPROVED CAMBRIDGE THERM IONIC CORPORATI ON445 Concord Avenue Cambridge 38, Massachusetts, U.S.A.

DRAWN SCALE

D9DATE

CAMBRIDGE THERMIONIC CORPORATION Printed ,n USA~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ii

Q2

Q1

FC1

R7RC

R1

R2

R3

R4=R6

R5

R7=R8

Z1=Z2

D1=D2

Q1

Q2

C

(.) CAMBRIDGE THERMIONIC CORPORATION Printed in USA

(ibL7JLBU®U9®JFigure 24

FRAME

Figure 25

.035

Figure 26

075

-. 25C

RADIAL FORCE

_m FERROMAGNETIC

MATERIAL

CHECKED CHECKED

CAMIBRIDGE THERM IONIC CORPORATIONVERIFIED APPROVED

445 Concord Avenue Cambridge 38, Massachusetts, U.S.A. I................. _DRAWN SCALE

70DATE

DIS E. .. .............. i..___ .. , .......-..--------...... .CAMBRIDGE THERMIONIC CORPORATION

Printed in USA

I

I

TIME

Figure 27

TIMEMILLISECONDS

-- -4 MILLISECONDS

Figure 28

CAMBRIDGE THERMION-IC CORPORATIONCambridgo 38, Massachusetts, U.S.A.

CAMBRIDGE THERMIONIC CORPORATION

0 w

>0X~

H -

H HU> C

H EjX4 re>a 0

>

0~

H0U

0.2

0.1

-0.1

-0.2

445 Concord Avenue

( Z

Printed In USA

( v 7L I3U®Zk Figure 29

11'

ill

CHECKED CHECKED

VERIFIED APPROVED CAMBRIDGE THERMIONIC CORPORATION445 Concord Avenue Cambridge 38, Massachusetts, U.S.A. l

__________ ~~~~~~72_~~ . . ._ EON DRAWTN SCA LE

o _ _~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

1

7

I

CAMBRIDGE THERMIONIC CORPORATION Printed in USA

Figure 30.

Figure 31

L

Figure 32

L

R]

CHECKED CHECKED

CAMBRIDGE THERMIONIC CORPORATIONVERIFIED APPROVED

445 Concord Avenue Cambridge 38, Massachusetts, U.S.A. ALE.......................................... )RAWN | SCALE

.73DATE

© ABIG -HRIOI - DRD-----rined---

( hdIXIND@G~m( z [;~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~XBLKDL~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~j .~~~~~~~~~~~~~~

I

I

(3 CAMBRIDGE THERMIONIC CORPORATION Printed in USA

I

IE

CAMBRIDOGE THEORIONIC CORPORATION445 Concord Avenue Cambridgo 38, Massachusetts, U.S.A.

CAMBRIOGE THERMIONIC CORPORATION

( iL J7zJB3UŽ7®

Figure 33

Figure 34

Pnnted in USA

o.,

(-"X

to--%

'-/--_

ax~

~~

~~

WIL'la:.co_ON

II(t

H~~~~~~~~~~~~~~~t~~~

Q OQ_

Li

42-2CL

if +

1

i ",I'

Figure 35

1IAi

00IPL

ZI

c'J4.f

,. ' d

O

IQ 2

Q_/3

II

Q!a0IL0U-JQ

-cOtLId 0iv-1

LU 02)'

cO-

TH

Page intentionally left blank