AD-A096 095 PERKIN-ELMER CORP NORWALK CONN F/S 9/5

ELECTROMAGNETIC ACTUATORS.(U)JAN 81 S ADELMAN F29601-79-C-0054

UNCLASSIFIED PE-14490 AFWL-TR-79-190 L

EEIIIIEEEIIIE

'---n- n ---

AFWL-TR-79-1 90 AFWL-TR-79-1 90

ELECTROMAGNETIC ACTUATORS

S. Adelman

Perkin-Elmer CorporationMain Avenue

-- Norwalk, CT 06856

January 1981

* Final Report

Approved for public release; distribution unlimited.

ID T'

AIR FORCE WEAPONS LABORATORYAir Force Systems Command D

'LU Kirtland Air Force Base, NM 87117

ii-

AFWL-TR-79-1 90

This final report was prepared by the Air Force Weapons Laboratory, KirtlandAir Force Base, New Mexico, under Job Order ILIR7913. Dr John R. Kenemuth (ARAA)was the Laboratory Project Officer-in-Charge.

When US Government drawings, specifications, or other data are used for anypurpose other than a definitely related Government procurement operation, theGovernment thereby incurs no responsibility nor any obligation whatsoever, andthe fact that the Government may have formulated, furnished, or in any waysupplied the said drawings, specifications, or other data, is not to be regardedby implication or otherwise, as in any manner licensing the holder or any otherperson or corporation, or conveying any rights or permission to manufacture, use,or sell any patented invention that may in any way be related thereto.

This report has been authored by a contractor of the United StatesGovernment. Accordingly, the United States Government retains a nonexclusive,royalty-free license to publish or reproduce the material contained herein, orallow others to do so, for the United States Government purposes.

This report has been reviewed by the Public Affairs Office and is releasableto the National Technical Information Service (NTIS). At NTIS, it will beavailable to the general public, including foreign nations.

This technical report has been revicwed and is approved for publication.

R. KENEMUTH, PhDProject Officer

FOR THE DIRECTOR

JAMES D. DILLOW DAVID W. SEEGMILLERLt Colonel, USAF Colonel, USAFChief, Advanced Beam Control Branch Chief, Advanced Laser Tech Division

DO NOT RETURN THIS COPY. RETAIN OR DESTROY.

UNCLASSIFIEDSECURITY CLASSIFICATION OF THIS PAGE (WIhen Data Enred)

READ INSTRUCTIONSREPORT DOCUMENTATION PAGE BEFORE COMPLETING FORMI. REPORT NUMBER 2. GOVT ACCESSION No. 3. RECIPIENT'S CATALOG NUMBER

AFWL-TR-79-1 90 c' 2&. TITLE (wnd S.15,tle) S TYPE OF REPORT & RE $4100 COVERED

ELECTROMAGNETIC ACTUATORS Final Report6 PERFORMING ORG. REPOR' NUMBER

, 14490 ,-

7. AUTHOR(a) S. CONTRACT OR GRANT NUMBSR'v)

S. Adelman F29601-79-C-0054 /IV'

9. PERFORMING ORGANIZATION NAME AND ADDRESS 10. PROGRAM ELEMENT PROJECT TASK

Perkin-Elmer Corpnor trr-. AREA d WORK UNIT NUMBERS

Main Avenue,,orwalk CT 06856 61101F & 63605F/ILIR7913I1. CONTROLLING OFFICE NAME AND ADDRESS 12 REPORT DATE

Air Force Weapons Laboratory (ARAA) January 1981Kirtland Air Force Base, NM 87117 13 NUMBER CT PAGES

4814 MONITORING AGENCY NAME & ADDRESSfII dillerent from C-ntroiitlv Office) IS. SECURITY CLASS -dt (li sopor

UNCLASSIFIED

1. DECL ASSIFICATION DOWNGRADINCGSCHEDULE

16. DISTRIBUTION STATEMENT (of this Report)

Approved for public release; distribution unlimited.

17 DISTRIBUTION ST ATEMENT (of the absrhrv- entered in Block 20, if diff.re,, from Report)

l6 SUPPLEMENTARY NOTES

'9 KEY WORD ICon.. '-o se , side a, n e e y end irIentify block , nuber

ActuatorDeformable Mirror

20 ABSTRACT 'Contlni, o re Op. ,de It n,...., , V nd IdeotI h, block n,,rrhe,

An electromagnetic actuator was configured, designed and constructed, as adirect replacement for the PZT devices now being used in deformable mirrors.The subject actuator has a stroke of 41 .5 A at 10.6 Pm, and is complete with amatching amplifier so that it can readily interface into existing equipment. Amagnetic circuit of Hyperco-50 iron, two BeCu flexures, and a tungsten push rodare used to construct an actuator with the required characteristics. Twoamplifiers, one a linear device and the other a switching amplifier, ar also

?Over)DD jAN 1473 UNCLASSIFIED

SECURITY CLASSIFICATION OF T.IS PAGE Il.,, D81E F,,-,.d,

UNCLASSIFIEUSECURITY CLASSIFICATION OF THIS PAGE(W47, Data Entered)

20. ABSTRACT (Continued)

described. During the program, the device was assembled and tested, andwhen several structural shortcomings were observed, the push rod threadswere reworked to provide additional integrity.

The actuator test results are described including test results on bothamplifiers. These test results show that the device performance was as perinitial predictions until the device was installed in the test cell and thefull output force was placed on the push rod. Performance then deterioratedas the threads degraded.

Test results after the thread rework indicated that the device is capableof stated performance.

Accr~siln For

NTIS CrA&IDTIC TABUnannouncedJustirtoation .. .. .

By_ - -

Dist ributicrij

Availbility Co6s I-- Avail ano/or

MA peci

,,,9 4981 J':D

UNCLASSIFIEDSECURITY CLASSIFICATION OF THIS PAGEMM.. Dat a eIe,

TABLE OF CONTENTs

Page

Optimization Study 5

1.1 Statement of Problem 5

1.2 Characteristics of Electromagnetic Actuators 5

1.3 Design Considerations 0

1.3-1 Magnetic Circuit Characteristics 7

1.3-2 Magnetic Circuit Design 9

1.3-3 Spring/Flexure Design

1.3-4 Coil/Drive Amplifier ]1

1.3-5 Push Rod Anaiysis/Design 13

1.4 Device Nonlinear it ies I

1.5 Thermal Considerations 17

..6 Peak Displacement Considerations 18

1.7 Actuator Resonances & Natural Frequency 18

11 Fabrication 19

2.1 Machining Processes 19

2.2 Coil 20

2.3 Assembly 20

2.4 Electronics 20

2.4-1 Class B Linear Amplifier 20

2.4-2 Switching Ampl ifier 2)1

2.5 Special Tool ing 25

III lest Procedure and Results 10

3.1 Mirror Cell Tests 0

3.2 Actuator Force Test 1O

3.3 Flectrical Test ing ii

3.4 Temperature Tests 2

3.5 Performance Tests in AFWL. Test Cell 12

3.6 'rests at AFWI. 2

Appendix A 37

LIST OF FIGURES

I,"i gur',:,t,

I SChematic of El;.-t rumdqn1t i,o A(-l at,,r I

2 Cross Section of Electromaqnetic Actuator 1

3 Drive Amplifier Simplified Schematic 13

4 Electromagnetic Actuator Conceptual Dr.-winq 14

5 Mechanical Model of Push Rod 15

6 Push Rod Allowable Load and Natural Frequency 16

7 Major Actuator Components

8 Assembled Electromagnetic Actuator

Electromagnetic Actuator in Test C]ll

I0 Schematic of Class B Linear Amplifie.y ..4

I Breadboard Class B Linear Amplifiei .,

12 Schematic of Switching Amplifier

13 Breadboard Switching Amplifier

14 Special Tools

15 Electrical Actuator Model (I

16 Static Input/Output Amplifier CharacerritrC:. 33

17 Dynamic Responses for Class B Amplifier 34

18 Frequency Responses for Both Drive Am li.tior! 35

19 Static Displacement/Current After Thread Rework 36

Al Magnetic Tractive Force Actuator Layout (G?-l)3H(A 38

A2 Post-Center (6R7-2064) 39

A3 Cap (h87-2065) 40

A4 Case (687-206C,) 41

A5 Post - Threaded ' 37,-7 42

A6 Nut - Difffrnt lal ,TA 1 ;t ((,e:-.l, , 43

A7 Arm~ir re itR7f~ 44

A9 .trr i:i, (6H7-2071 46

A10 Base (1-i-2!K22 47

All Ad j - tush Rod ((47-2" 3) 48

Conversion factors for U.S. customary

to metric (SI) units of measurement.

(Symbols of SI units given in parentheses)

To convert from to ultiply by

angstrom maters (i) 1.000 000 I E -10atmosphere (normal) kilo pascal (kPi) 1.013 25 X E +2

bar kilo pascal (kP&) 1.000 000 X E +2

barn "tter, (m') 1.000 O00 X E -28

British thermal unit(thermocheical) Joule (J) 1.064 3.0 1 E 43

calorie (thermchemical) Joule () 4.184 000

Cal (thermochemical)/cm' waga joule/a' (PU/M') 4.184 000 x E -2

curie gigs becquerel (Gq)* 3.700 000 X E #1

degree (angle) radian (rad) 1.745 329 X E -2

degree Fahrenheit degree kelvin (K) tK -(t.F * 459.67)/1.8

electron volt joule (J) 1.602 19 1 E -19

erg joule (J) 1.000 000 X £ -7erg/second watt (W) 1.000 000 X E -7

foot meter (m) 3.048 000 X E -1

foot-pound-force joule (J) 1.3SS 818

gallon (U.S. liquid) mter' (a') 3.785 412 X E -3inch meter () 2.540 000 X E -2

jerk joule (J) 1.000 000 X E +9joule/kilogram (J/kg)(radition

dose absorbed) &My IGY)* 1.0o0 Okilotons terajoules 4.183

kip (1000 lbf) newton (N) 4.448 222 X E 43

kip/inch' (ksi) kilo Pascal (kPa) 6.894 757 X E 43ktap nawton-sernd/m' (N-s/a') 1.000 000 X E +2

Micron meter (i) 1.000 000 X E -6

il meter (.) 2.540 000 X -5

mile (international) Meter (i) 1.609 344 X E +3

ounce kilogram (kg) 2.834 952 X C -2

pound-force (lbf avoirdupois) newton (N) 4.448 222

pound-force inch nootot-meter (N..) 1.129 848 X E -1pound-force/inch nevto/mter (N/ ) 1.751 268 1 E *2

pound-force/foot kilo pascal (kPe) 4.768 026 X E -2

pound-force/inch' (psi) kilo Pascal (kp) 6.894 757

pound-mass (lb. avoirdupois) kilogram (kg) 4.535 924 X E -1

pound-iss-foot2 kilogram-mater' (kg.a') 4.214 011 X C -2(iomnt of inertial

pound-iss/foot' hilogrlat/'ter' (kg/&') 1.602 846 X 1 *1red (radiation dose absorbed) Gray (6y)" 1.000 000 X E -2

roentgen coulowi/kilogram (C/kg) 2.$79 760 1 E -4

shake second (S) 1.00 000 K E -8slug kilogram (kg) 1.459 390 X [ 41tort (mm Hg. 0* C) kilo pascal (Pe) 1.333 22 K E -1

*The becquerel (Bq) is the SI unit of radioactivity; I Bq = 1 event/s.

**The (;ray ((;y) is the SI n it of absorbed radiation.

A more complete listing of conversions may be found in "Standard for Metric

Practice,", American Society for Testing and Materials, ASTME-380-76C.

3/4

I. OPT1IMIZAI ION STIUDY

1.1 STAt EMENi' (1 I'ROBI'M

In a tylric-11 deformable mi rror aqplicat ion, the iffirror surface f igurt, is

controlled by a servo loop, and must he diistorted or aldJUSt d in respon1Se tl a fn

external -ommand. The actuaLtor must gk-neral I y move the surface of the ni rrr11

on the order oif two to three wavel engtils. Force requirements are en1 thcL ordt r

of 10 to 100 pounds. Resolut ion re~quirents vary with tht. specific appIi ca-

t ion, with af resolution of several mill iontlis of an inch censldercd to I' qitt

T[he electromagnetic actuator should II, he i'eponsivc to tlitse rtehi l rert,,

and p rovide low hys tuercsis, low power con Stimpt ion, and the potenti jul tor ai

longer stroke, if requ ired.

1.2 CHARACTERISTICS OF El ECTROMAGNE' IC ACTUATORS

It has been suggested that at uat-rs can he const rue t cd using magn.' ic

tractivye force as thfe generat ing meehan i m. These t ra t iv xeorceU ae tt, tr- ai

be eoni figured to be ('apab I u o f hfigh f oree s ov'er reason bh e st rokes . ite t'( it-

nology which is utilized in these devica- has been relined over a nuimber ,

years, and is well understood. The tCehnia I1 issues which aire mive)Vd aire .11;

related tO several new magnietiit: materials and to th lfabri iation uil hand injg -j

these materials, as well as- to the faibricat ion ot Vt'r\ prec( i SeL 1 bry'l I i um- oppe1) r

f cxu res.

Bas iea I lI time dev i c wi leli i s he j fig < it i gil'c'c inlc'r t I i i - p) og r'aiir., is a a

1)1 a cup InagneL2t a s sem lt an.b id am a irmai iLtur i lauc r c a t sp,-iid mi r ii g ml lin i mi.

charac ter i ;t ic's of Lhe devict ewi I I heL di't Ltril i1' bct hi t 'ost1, rue t iou ;iul 1 i'

propert iit's of the miaterial . ohiel iar#' to h1 bk it iI Iit'edl

A sceit iC of the, devi t' is shc wii ii I igu rc' I . Ih,' d rivye , )I I inii c

magict it' flux: in the> 'c cti t.r p)ole 1c it(I, ocd toI 11 f lox t I s tilr'ut~-1 thc' irmift irt,

and ret urns thml;:h the r im. When majgne(t it I I i >: is lnucd a( rasag, i ,ip, h

two pleat'-s of MetalI a re attLr~is tcd. 'Il is t roi-t lyve flrcc i t civel b:

A ' x Ill

w r, - I ' rc c ill pmlicds

A si ir I 14c , art -. I Jn i

fI flix dens it V in I fi'si

and, the flux densitv is a function of gap length, and is given by:

B 1.26 NI (2)

R

where: N = No. of turns on Coil

I = drive current in amperes

R = gap length in cm

The deice is round, with a thin outer rim and a round (enter pole I)iece.

This form factor has been chosen to ensure that the electromagnetic tract ion

motor will be compatible with existing hardware.

1. DESI(;N CONSIDERATIONS

fhe performance of a traction motor can be calculated using the two basic

magnetic circuit design Equations I and 2. Conversely, a device can be con-

figured using the same basic equations. In an actual design, there are severa!1

additional factors which must be considered in order to account for non-ideal

behavior, and for material anomalies. The basic equations assume that all flux

in the magnetic circuit passes straight through the air gap, through the armature,

and through the rim and center pole. 'T'his is approximately true for ver small

air gaps, where the reluctance of the gap is low, and fringing of the field in

the gap is small. As the initial gap increases, because of longer stroke require-

ments, the reluctance of the gap increases correspondingly, and part of the flux

tends to leak between the center post and the outside rim. In addition, fringing

begins to occur, and the flux lines do not cross the gap perpendicular to the

surface of the gap. Because of these phenomena, it is necessary to apply a

correction factor which is a function of the device size and geometry. The

determination of some appropriate correction factors has been made by employing

a vector magnetic field simulation to calculate exact device performance, and

comparing, this performancc to calculated idteal performance and to exptrimcnetal

results. 1his simulation was done under a prevJouos pr gram.

In order to complete the design of tie tract ion motor actuator, ti( correction

tactor is estimated from the device geometry, and entered into the basic equat ions

as ' force mult ipl ier. The maximum flux density, which is a function of tht

material used In the magnet Ic c i rc it i t , and tie correctecd pea k f, rc e ar omb i ned

to determine the necessarv tractive area. The total forc is the sim ol tht,

force exerted by the center pole and by the outer rim. For a tiven total area,

maximum force Is realized when the center post area is exactly equal to teie rim

base Armature

/ V

Coil RestrainingSprings

Figure 1 . SchematiL of e t(c L romagn, t i c Act uator

area. Next, the coil and drive requirements are determined by using Equation 2.

As a rule, the more wire that is used, tIl 1owe.r the overall power dissipti t io

will be. However, as the amount of wire is increased, the size of the act uator

increases. The f in;lI (urrent reqluir ements aitrc dt crai 1 ned from tht- rnmax imum f I:-,

density, number t)I turns, and the ititi al gap.

1.3-I MCNETIC (I RCUIT CHL\RACTER ISTI CS

Si llet' I lIt . "r,c 1be. t w-eeln the1 ;lrmajtur t and ft e 1 ,isc, i port' I a ( ,1: 1 tpr 3

sqiire ,f the it dtvns it,., it is nece!;sarv t ii m. .imiz t .n h Iux dtcn i i w.. i. ,

(',lan bi generate'd icroess tik- gaip. There ai rc :I I Il s which ;art, formnilatcd t l,., l

high i ux densit is lxd maintain high pirmtea j ies. TH). prpcrt is 11 tI, ,,

al l,,s are generally .stahlished not oilI': h thci.r ompisit ion .ind mnlulH i, 1111,t

but by the' app] i,,t ion ,I a xt(,t Irr.t cn)" prre n Lc ss whi h1 *;p! I it.d

a ft Pr tht' maitk r a I i sm;, Iliited t o shape. As g generaIl ru I., the. b.t t . I lit

magnetic properties of an illoy, the hardor it is to machine and tread. -

tionally, tlh : alloys generally have poor structural properties, and the .onfi-

guration usual ly must (ontuin A support t ructure. A riumb.r( of the syccj .a] 1;

magnetic alloys which are manufactured have been evaluated at Perkin-Elmer.

When the various desired properties - saturation flux density, permeability,

magnetic hysteresis, and machinability - are traded off, the choice is c -r.o,.-J

down considerably. For instance, several of the alloys are formulated tc 1u

free machining, but can support only about 12,000 gauss. Several alloys hav

very high permeabilities, but are also Limited by a low flux density caabi i

Por this application, the design requireincits of high force and small \',l

di, Late d very high flux density, and since the gap is significant, ii., . -

bility is not critical, since the total path reluctance is determined by the gap.

The alloy which has been chosen as most suitable for this application is

ilyperco-50, a soft iron manufactured bj Carpenter Steel Corporation. H1-ypts r

is available in billts, and can be machined into the des,;irod folm factor. P'.

techniques for machining and subsequently heat treating this mater]al in oruer

to obtain the desired properties have been intensively studied at Yerkin-Elmer,

and a number of magnetic devices have previously been manufactured on a proto-

type basis.

In order to obtain the optimum magnetic properties, the material must

first be machined to form, and the unit assembled. The machining is performed

at precisely controlled feed rates and speeds. In its raw form the material

tends to be brittle, and care must be taken to ensure that the tool does not*

eave the material after a cut has been started. Additionally, only ccrtain

lubricating oils can be used for tht machining process, and the material must

be degreased immediately after machining. Th( material is then heat treated

for a precise time interval in a hydrogen oven at 1825 'P, fol lowed hy a

-ontrolled cooling cycle. Atter the material has cooled completely, the

listor tion introduced by the heat treati ng is measured, and the cric"ail 'i -

fac:, are ground back to shape. The ma t J, }rope rtief h,'.e bo,, esti, i:,.

by th, thermal history, and very high eomletatiorCs a meci aniual : Stres. "i

jisturb these pr .Jrties.

)it experience wit h thiia' alo,!; i n di.,it, -; th t al I Ieat ;Zci() lo ss or.

to, rl alized a.-ous the gap. h'lle devir-,, ,es eiqn l,] sum'im ; a flux ,isity ". , :,

vausi, iid thore are several othel matr1 ills whici .A t' " , e of th ' i

performance. These materials -oild be used if on alt Irl iv. Were I e-(i-1.

1 3- 2 MAGN TI IC C I RCJ 11 1 1)1S I (;N

A L ' rct Li(I on ,, ()r whtich 'act s a to bas i, pramtuI'rs c-an ho ('onfi1gur. 1i ito

follows,* when' )k 101iiL ft g iven by FEquAL ion I

F AB 10

I'he devi ce hais been constructed from Ca rpenteor SteelI llvpt'rc -,50, aind ai

flux density of at 1least 20 k iiogauss shoulId ho' achlie vableo withb proper bocat

t r2aLt ing. The detr ivat ion of parameters w il I]SLM 20ui'N k ilogauss as ai vc r

Sal e nulmbLo r . Thu toutal gap area reoq i rodJ i s thu i' t arc

(11.8) (100)1% 400 2. 79 5culnW 0 . 4 33 i

i'Lt- total tract ive air.'a is noinia i y divided equal] V hotWLc'n tho tu '] f.

I)I po and the rim. The cou21ttor pOSt j wI haxV' a 1/8 inc~h h)IL u for thc pi Ih .d,

and the OD is given by:

ol) (18)2 0.433

-+(4

OD 0. 54 i t.

Since the r im out or di amot or will bc 1 .0 inch, the inner di ameter is

given b%,:

I 0_

IDb 0.85 in).

Al Ilowing tor a 1 5; t I i>I nkg. is indicated bv prov l ins a i muilat ion

roan its, I ho 1;per shotll 1 c iho hit 17.Ihin r l) thL' oulttor w~ill willI at~l'-

I r on i .0( inc n t , 1.15 iricht and Lw l I he a .o' appro r ;iat( 1 v'. Hit' i no.-r po(stIi; I t. Ir t r-m' 0.55)! inli to 0.60 ill( 11.

Another t '.idett t 0 hi' Iliado i S I ('II I', I I. As t It- I ttIgt h1 i s i Ilk ,';is.d , t Ill,

I ix It- WI g 14 is l iTi(rot-is dA a; I, nrd i n g I%-. %,(, x r, as t lit,' I ong"t 11 is' inliroa.'iI,-

9 rt i' WI Lt ,,i n ho wounI 1d , Id i i ( io w t-r d i .s i t ioIu i oit rc i ,cs d ict I' ,- t ItoI

ttI tiftoittt oif -,qpt'r il,'.is.)s j n lit laI in'.'oat igal iol whiH it is' ptt I.

inltiit i%". idi It("; thaIt -il opt iMUrn lui I is hott '4'tt two Wtd tilto' il, hoC.-.

Ill';is nCuSt iY It ittl i; thulk ristIlt oI a (pit it tat ix',' 1"'-1 tit t 1( 1:11'j.1 s~~ It in"

9

and additional analysis is called for in this area. This analysis, however,

is beyond the scope of this program. A length of three inches was chosen.

In addition, there are several geometry refinements which appear to be

appropriate and w!,ich were incorporated. First, the corners were rounded

instead of left sharp. Second, the armature was cut out so that the face of

the armature matched the base profile at the gap. A sketch of the cross

section of the device is shown in figure 2.

/) |percu - S0

./ I , - 50

Coil Area

Figure 2. Cross Section of Electromagnetic Actuator

t0

II

1.3-3 SPRING/FEXURE DESIGN

In addit ion to tile magnetic circuit, tie actuator incorporates a spring

conf igurat ion wh ich has two ma in purposes. The f i rst is to support the

armature or moving port ion of the actuator to the base. The second is to

maintain a precisely controlled spring constaht . Again, machinabilitv is an

issue, and the spring lmitirial must have ai v.rv low mechanical hysteresis wiith-

in tile p recis ion elastic limit. Tensile, st rcugth should be high, and Y,,ung's

modulus very predictable. Natural Iy, the maiteriinl s bo non-magnetic su bs

not to interfere with the operation of ilhe magnetic circuit.

The only material which can seriously be considered is beryl I ium-cOpper.

A study of the various app] icabl allos indicates hat KBI : Iloy 172 oI irs

the required properties. Alloy 172 colIes in two forms disignated by mannfac-

turers as Condition A and Condit ion H. Cond it ion A is annlaled and st,f t,

permitting bending to shape. Condition i is part izally hardened and stlitblh

for machining operations which do not rCq ii 'r bending. Both types have, t he

same Young's modulus when lieat treated, but thie Condition Ht materi al cures to

a higher tensile strength.

The machining and het treIt ln, 01 those ] laOVs are well understood. 'lie

materiai is irst machined to the desired form, heat treated in a fixture, and

tile heat treatment is checked by measuring tilt hardness of the mater il . Mea-

sured spring constants generally check with calculated values verv well.

1.3-4 COIIL/DRIVEI AMPI FI ER

In order to generate tilt magnetic flux, the unit conlains a drive coil

wound around the center pole. This coil is interfaced with a power amplifier.

The designs for both tilit coil and the iamplifier are, very closely related. The

total drive requirement is determi ned by the wire area and the current dtensitv.

The drivi is specit iiid in iumpri-tkurns. 1i1' (oil can hi' configuroed ci ther

frm a I l.tr llumbr ,I tt-ns ,f fine wire, or Irom ai lcsscr number of turns

LI! ILclvier Wile. As the numbir of turns increases, the rturrilit decreuses,

bhlt tilt, indu lalice ,f the t'oil incrtises is tlt, square of tlii' number oi turns.

Ihis stibsequtenl increses the voltagie rt-quiiement of tit, amp Iifi r.

he i mplil ier should be ai trut currtnt amplifier, that is, it should

driv e tilt, coil with a tonst art currtnt pri,porti mal to tile amplifier input

vol t ge. Hie limitatilln ll il. amplifier dtii n is tihit suit;ble power t llns-

* Karwicki Bervl ium, In,

II

istors perform best at output currents less than 8 amperes. Since the driv

requirements for this unit are not severe, it seems that a 3-ampere coil will

meet all the salient performance requirements and will be reasonable to manu-

fac ture.

There are two ways to realize a device with suitable characteristics.

The first way is to drive the actuator coil with an ordinary feedback voltage

amplifier, and to ground the coil through a small, stable resistor. Tile re-

sistor voltage is now proportional to the coil current, and this voltage can

be used to derive negative feedback for gain stabilization. This technique

has the advantage that the gain is extremely stable (as stable as the feedback

riistor), and the disadvantage of requiring a feedback loop which may tend

to be oscillatory as the coil inductance varies due to non-linearities.

A second way to construct a current amplifier is to use a single power

transistor with a large emitter resistor in the output stage, and place the

actuator coil in series with the collector. The collector current is approx-

imately equal to the emitter current, which is determined by the base voltage.

That is, the impedance looking into the base is equal to the transistor '- times

the emitter resistance. This can be seen to be considerably simpler than the

first approach and is the technique which is generally used in TV flyback amp-

lifier circuits.

Figure 3 is a simplified schematic of an amplifier capable of supplying

from 0 to 6 amperes to an electromagnetic actuator. The amplifier controls

the current through the actuator, and allows the voltage across the actuator

to change by tOV as necessary. The amplifier gain will be adjusted so that

an input range of - 10 volts to + 10 volts will correspond to zero to full de-

flection.

Transistors Q1 and Q2 are connected in such a manner that IA = xIE' where

0.995< ut< 1.000. (refer to Figure 3). Thus, to a very good approximation,

IA (A1 V IN + A V BIAS) (6)

Diode DI and Zener diode )2 protect QI and Q2 from damage if an input that

causes IA to decrease rapidly Is applied to the amplifier.

Q2 and D2 can dissipate as much as 40 watts of power at full current outt-

put. A 2 in. x 4 In. x 6 in. convection-cooled heat sink is necessary to keep

Q2 and D2 within safe temperature limits. The remaining components art, mounted

12

(-n ai 3 ill. x ') in. circuit board. The protutype amplifier weiqlle- i

pound, and the goal for a final design would bo several ounces.

- I L

Figure, 1. hi iv. Ajinpic .z miif,.2hmt

1. 3- 5 PUSHi Rol) ANAIYS I: 'rJFS;N

The linkage between the maqnt cc actiiitor md~ the i iot ot

an axial leng(th adjustmei~t assembly, a pu,; iied, And( a "T I 1.71(i.

componlent-; flhe 1pmch rod is most c'ritical, wart-ari ir; nq orisi -rati, of~ 01),,"i

-- tural frejuen'r'-s and bivcki iq load.;. A pro1 Iimiriary sketch ot the i

device is shown in Figure 4.

1:11

Figure 4. Electromagnetic Actuator Conceptual Drawing

Three modes of vibration are possible: longitudinal, torsional, and

transverse (bending). The first two modes can be eliminated from this

analysis. The longitudinal vibration has a small amplitude with a high

natural frequency (over 10kHz), and torsional vibration does not siqnifi-

cantly affect performance. However, the transverse vibration occurs at a

natural frequency close enou;h to the operating frequency to require pro-

vision in the design to avoid resonance problems.

The push rod was considered as a "clamped-hinged" beam (Figure 5)

for which the following formula for tran:uverse vibration is 'Applicable.

14

- -- 'l l , J , . . . . ' I / lr .. .

f 15.4 1-

whe re

f 11= Natural frcquency

EI = Bending stiffness of the section

Length of the beam

Mass per unit lenq4tf

t Actuator

Transverse LaVibration

Figure 1_. Mcanical Motel-' (t .}o

When slender members are loaded in 'ormr)rf~ssien, they t ei,, to al;I. n

experie-nre significant lateral def I((ct i Cns th1a t tare fuc Ion 1 )II till- oATt I 1-

loads. Computat ions were performed us ing em;i iri cal co(l urnt 11k to: jO

valiues of push rod diameter and moduliv; ot e Ia:;tic it y. The ae n.e

diameter and modulus of elasticity Were aw1l jr'd in crn C It~t Mor 101 t nat 111 a I

frequency. The resulits of thoseo r-om;. uta t i s for *'' i k1~II(5 2

rods are presented in Fi gore? 6. Three mar I:1u u1r m! a.I.

ident ified; the rod motst be non -m-Agjne I i , t he natl oral r n.mu40 i. rth alo bermres'ewt 1ulodns X. i .u;

There are n* imerous (31 1 ena t i ys for tV, ii' sli roddriA.. ~ xm

brass; (E= 20 x F) lhf/iri. Irod with a diameter of 0.135 in. could he

em; lo'ed. tHowever ,it isacvntreo: for ih' rod to ill tiit,. r.a.

a design sl idy wil aIsic . k r""j'; ml an-1 1- 'T

cal se-tinn

/

Figure 6. Push Rod Allowable Load arid Natural Frequency

1.4 DEVICE NONLINEARITIES

One of the salient characteristics of this type of device is that the

force is proportional to the flux density squared, and therefore to current

squared. This is made even more severe when it is noted that the force is

also inversely proportional to the gap, and therefore, if the total stroke

is not insignificant with respect to the gap, the force may be proportional

to an even higher power of the drive coy rent.

'h,,I,, are several factor s which tozend to moderate the effects of thi -,

nonlinearity. In actual pra-ctice, the permeability of the magnetic mate'Ilal

decreases as the flux density increases, and it is possible to configure a

,init Mo that the permeability variation compensates for the squarinq factor

That is, the flux donsity becomes nroportional lo the square root of the

d1rive -urront in the operatinq regiqons of interest. Ii actual t racti (0

we have I lotted very near 1y linar for-e vs. :urrent c-haracter ist ic.s with

linear ;prlng. The, i :'c ri ', d,ue t,, a lail , :-;t rke sho uld riot r:xist,

sync, for th , Aevi-. L ir~q -,,l';ie r, , tr 'A ok- will be- emall -om a -,d t

the iri ial Ia -•I ',

A further c0'. a5d-~t ioni i s the inh i(ref,t inTstabi Ilit,. of the!,V

Fo r a i near sp ri rig/f lex ure constan It , a-.; hI f ur 1 en211 1 t isi rea ssI I l

pulls the arm. Lite -luser to the pole !;ti rust tire. T! 1i s s l t Of bot. hi. fe

analogous to moving a permanent maogne(t close-r arid clos ,;erI t ' 1" :iec f

At the critical position, the iron s-iddes 1'., pimp,, t o the, r-,i gnet -or

actuator, there is a crit ical vul Iue ()t -o tfit at wl:ii h t 1:- arlfo~t ire( is

pulled to the pole. In the solution; of the t wIot ns wii (-I (;ovterts

behavior- of the device, this is the pointt it '41 il-I. es t-;urret I: ~sito maintain a greater deflection. We hiv'e g: uti . iioil ut ot;t , !-!i -

this behavior as a function of coil 1io uritrt-s :~s I'fhc - ararr !-,It

initial ogap, device (geometry, and .lt-ivt'- citt,-tt .:s- t ui hrt - r,1!1!

which we have made, on devi ces vcry siini jar to t i A,-: I I- :I 'si t t-,

instabilIity occurs at a iefle,-t ion of iboiit I1,K ' fl-i: i Iu iil; I

possible to adjust the i,:.i t iaI qao) so) thur tiis h o:.onet~ut. wil toi,,

problem. A stop is instal led in ordeor to) revetit ut uttb''tt tort. ,

occur-ring and damaging the spring or tho so icil t

1.5 THERMAL CONSI i ERAT I~t

One )f the crit ical facto~rs tO o Oot ()1iie!(ed is th- fIfr r;taI e.xt s-

of the actuator due both to heat geti it el, i ri the, c's, I i n to s i 1! ....

in the magnetic circuit. Tie effect-; of t Is-iiml eci' ctito- mi tu L .:

the fact that the magntic circ:uit expsati'ls it, onle !i ret't t-sts, sIti 1; i

rod expands in the opposite di rection . Idleall y , the Inush roh isld .'

an expanision coeffic-ient equal to that. of lHypsrro-51o faoit, x 1 :. t1

'rhe one material t-hat standsi ott 'I'; lciIv il u high I 'Atisi the riqht e'xtItsi

c-oeffic:ient is L'rtylli tm. trfo~t sintely, 1ltry] Iisim -'unse1t 1.-r ou te

u-nt'i duet to tr'xi.- it% 10 'ti rat J[1 -s 1it it a r scsi too I w t fihli I-

it we-i !1(ot Sei - .5i j t, m.ii's' rt, t- pit*-t 't St titt5 e (it t'it

I !I-IoIse. . TIserr'for, ()I ' il mTII' Il sI t I' 1i -1' ,iut' OsS ui t!

,.h a a " i os- expatotenr 1'ef i- to . Bt'r I'I IitI sIt is :o 101de (i a-I

sui1table, for a l ater effort , afteor the- motre hart I ife-,y ''orsilerrut I oss, to%'I

bee#n resolved. (A seconid push rod was; fafrh'suted front iatios- e

th#e t ungsten i x cracked di, t o hr it t I(s'

1.6 LEAK DISPLACEMENT CONSIDERATIONS

The maximum allowable displacement is a function of the spring mec).a-

nism and the initial air qap between the base and the armature. For the

peak-to-peak motion of 31.811m, the spring is essentially linear, and the

devices of this type have been found to be unstable for displacements of

dbout 1,'2 of the initial gap. Since the power requirements will increase as

the squar- of the gap, it is desirable to keen the initial gap as small as-3

possible;. The gap will be set at about 0.00i in. = 7.62 x 10 cm. The gap

is initially adjusted to be uniform with a toolmaker's microscope, and can be

:t to within +i0*. The maximum displacement will be set by fastening metal

t stops in the gap.

1.7 ACTUATOR RESUNANCES AND NATURAL lPEPOUENCY

There are two mechanical resonances of interest. The first is !I,.

resonance due to the armature on the spring. This resonance is givcr by

f : I (8)

where K = spring constant (lb/ft)

M = mass (slugsl

For a typical spring on the device pictured in Figure 5, we have measured a

displacement of 0.0015 inch for a force of 83 pounds. They ofore:

83 x 12 6.64 xK - <.'.001(] 5 lb,/ft {). nOl 15

'lhe mass of the armature is

0.¢ n x . V1 b/in - i qX H

I. ;?x sli'(I'

ind: f '01), liz (11

Thi,- mav be lowerod sliqhtly by the mas!s of tie ush rod. 'The s'-ond r'-

nan ,e i; ,-aused by the transverse vi)b-,at ir of thr tiush if. The tpusL, ro!

will ).I in 'h diamotor, and will l, mlre, of a tungIsten allov with i.

- x . I ri ,...n, from Fiqur- ',, that the natiraI froqu.rvy ( : t,

t ow is qr',ater thai, .. 05 khiz. We 'to n(,t '-iv si ?(l Anly J 1 .i , I5 mr :: , .

H$'5( !1z r'srracllre limitat ion1 re iif-r'm('lt.

l~l

II . FABRICATION

2.1 MACHINING PROCESSES

The pieces of th actuator were fabric att,-d as per drawinrgs 687-2064 to

687-2073 which air inwluded as Appendix A. The machininq Procedures used ire

as follows:

I. Post-Center (687-2064)

The center post was turned to form from Carpenter Steel Hyperco-50.

2. Cap (687-2065)

The cap was turned and threaded from e)' stainless steel.

1. Case (687-2066)

The case is designed to fit over ti, base ar is turn: from .. $ st ,lin

steel. It- is fastened to the base with Lhoctite 15 retaini ,';mcn.

4. Post - Threaded (687-2()67)

The post is configured! to mate the actuator test cel. It, also, is

machined from 30 stainless steel, and the threads are turned. or. a lat,

It is fastened to the base with Loctite 35.

5. Nut - Differential Adjust (687-2068)

The differential nut is turned from 103 stainless steel. Tim, rO at

American Standard Threads are accompi ishod hy chanqinj the back qears '

the lathe.

6. Armature (687-2069)

The armatire is t rmied from hfyierco-5 ), and in r'ot 'id t mat A: t,

cent

, rmPt

7. Flex ire (',87-2r!7'))

The sprit:q flexure is machi ned fr,,m 8lit ' - A] o 1 72 , and subsequent lv

annealed.

8. S[.rir. ; (€:87-2W'1

The if I.rq is m ',- red ., f ro ,m fPeij - Alj 17jr 7 ,rd , ,Iaed,. khe-

ns 'was sr.l ctf -.xjrimeitally, aid is ,.)27 inr ht-,;.

9. base: )Os<-,, .!

The Las' i; titL) It )m 'art it 'r - '- 'r''-; art tie '1'lit' I I:.

w ,. ;,r , t 1 t t, I( - , ( r.

Trio , ' ,' . l ,, 'In "p'.'n 'r.. 1'- '11' .:' h

r, tr I wfre ma)'- :c. Iw ;', , . ; t.,-] ,' , ' w},n . ' , ''-,'' : :

;t,, ,rk ',' ,' th,, t' ,

2.2 COIL

The coil was wound trom #24 double formvar insulated wire on a delrin form

configured to match the actuator form factor. After several iterations, wt,

selected a coil witi, 410 turns total. The coil was potted, one layer at a

time, and pressed into the actuator. The fit is verv tight, and it cannot

be removed intact.

2.3 ASSEMBLY

The electromagnetic actuator was assembled in accordance with layout draw-

ing 687-10086 shown in Appendix A. First, the center pin was pr- s fitted

into the base. The cap and case were then degreased and secured to thC a rm1at turc

nd the base with loctite retaining compound 35. Both assemblies then required

holes to be drilled at the joint between the stainless stuel and the Hvperco-

50. This was done in a jig boring machine. The loctite joints were thei,

broken, by heating the assemblies in an oven to 800oF, and the bottom otas stl)eab ,

along with the armature, was sent to Carpentel" Stel for the proper condit i'ning

and heat treating, to generate opt imum magnetic propert ies. The cap and ise

were then refastened, and the 2-56 holes tapped in both.

After a one week settling period, during which the Hyperco-50 tended to

settle, both gap surfaces were ground flat, and the device was assembled. The

gap is set at 0.0042 inch by using a toolImaker's microscope , and the special

tool which is supplied. Three stops, which limit armature travel to 0.0012

inches were inserted into the cap at a later time.

The maj or actuator components are shown in Figure 7, and the assembled

ai tuator is shown in Figures 8 and 9.

2.4 ELECTRONICS

Tht. objectives of the amplifier dcsign phase were to construct an amplil itr

which could be used to driw. the actuator for proof of principal, and to, d.moni-

strate that a very small and efficient ampl it ier was teasiblec. The appr,,;o h

taken was to design and build a linear ampl if ier to use for test purpowss, aind

to design a switching ampl if ier as a paril lIcl efffort. The class BI linear amp-

1 if ter shumld have an et tic iency of about 50.', compared to hetter than 91 or

the swit ,hing dyv ii ,.

2.4-1 ClASS B LINEAR AMILIFIER

the ampl I fHer schemat ic is shown in Figure 1I. The de.sign e.mpi o ,rre.nt

feedback from resistors RIO (.2 ) and R4(IK',) back to the l,,w power l ,vcl 7'7

20

1

-ft

U

b

0 - - -,-." -.

CAR 1 2 3 4

~2

cc >0z

Or1L

cK)

244.

o)p-amp. (A1). The 747 nlitput Ir v ; I11- cascaded Dar I injt,, -! ra! istolI(.'N2.'liJ) ,mnu I., (A:. |,O i) . 'thu. Ilt er lit 1n t tr niu-O dii', '; the o'1 - jt

whilie zt-ner diode il .' ( IN i 127A) iu~~: volt ay ';u rA(ut cc! for t h'. col 1 i

junction of O). 'I"-e breadboard amplifier is illustrateo in Figure 11.

2.4-2 SWITCHING AMPLIFIER

To reduce size, weight, and power consumption a switching amplifier

design was undertaken. It is shown schematically in Figure 12.

The goal of high efficiency is achieved by operating the Unitrode PFcA,

Darlington power device in either full ON or full OFF mode, with the duty

cyle establishing the average power. The actuator's natural inductance

i te.rs the switching carrier current to an acceptable amount. The duty

cycle is controlled by comparing the command input or control voltage to a

triangle input. With the co0ntrol at zero, the offset potentiometer R2 is

adjusted until the comparator provides a ]0'. duty cycle square wave to the

outtut power device. As the control voltage increases positively the duty

cycle increases linearly until a maximum of nearly 100% or full on is

obtained. The switching amplifier breadboard is shown in Figure 13.

2.5 SPECIAL TOOLING

The following tools, illustrated in Figure 14, were fabricated as aids

in assembling, aligning and installing the actuator. The small wrench in

the lower center is used to loosen and tighten the nuts to adjust the gap.

The long thin device in the upper right engages the coarse and fine preload

adjustments, and the large tool on the left is used to turn the actuator

into the mirror.

a

4k

rrr- n UJ rNr-

C -j L ,-

1'4 4 <ij

i-c

L

CL

- t

ii -~

N ~4 A.

~. ~.Ft

*7 *~~-1

"S.,- I

AV

$'o.,.'~ ~

444

'r. jv 44V

1

41 S

I'Cr"

4'a;,

S It

- a --

lit. TEST PROCEDURE AND RESULTS

3.1 MIRROR CELl. TESTS

The first test performed was a displacement/force measurement in ordcr to

determine exactly how much force was necessary to displace the mirror. A billet

of material was machined, which fit into the test cell, and 1-is a recess whi i,

meshed with the nipple on the simulated mirror surface. A ca ibrated lorrt

gage was used to push on this billet, and the displacement at the billet wa-

measured with a dial indicator. It was determined that a force of 60 pund.,

deflected the test cell surface the required 1.2 x 10 - inches.

3.2 ACTUATOR FORCE TEST

The initial test on the assembled actuator included displace.1nt , and

force measurements. The maximum available force is determined by lirst -t , -

brating a relatively thick spring, by measuring its displacement due to a

known force, and the displacement as a function of current.

Our tests were first made without the push rod installed, and a 0.)78 inh

spring which had previously been made was used. This was the thickest BeCo

spring which we had available. The a,tuator was assembled and aligned with

gap of 0.0032 inches, and it was determined that the spring constant was 78,260(

lb/inch. Current was then applied, and the highest stable displacement obt, in--3

able was 1.55 x 10 inches, for a current of 1 ampere.

78,260 lb/lIn x 1.55 x 10- 3 in = 121 lb (L2

This implies that the device was capable of 120 lb. It wa.s n,,t readily poanlb%

to determine if larger forces oruld be realized, since a thicker spring was ii,,t

;jva lable. Also, this force is somewhat greater then we had anticipated, and

is certainly very adequate.

We then determined the static positional hysteresis of the device by

set tinug the exeitation at 500 mill i amperes, and varvin tlt, (urlirent so that w,

obtained this value of current by mnot,,riica lIv -pproia hing from ;,cr, or ,0c.

ampere. Using a B&S electronic dial india;ttr, we. est imat.d that t ' hvstcrtsis,

again without the push rod, was between 7 and 25 x 10 - 6 inches - or 0.18 to .64

micrometers. This est Imati, is of quest ionahle val iditv, si nce t het spring is

substantially thicker than the final apring, and it w;as wtrv (lt it tlt t,)

repeat the current accurately.

i0

With the assurancL. that the actuator had ample force capability, thc push

rod was then installed, and the mirror test cell was used to simulate the

actual mirror 3urface.

3.3 ELECTRICAL TESTING

Electrical tests were performed with both amplifiers with a dummy ],,,'Id.

The dummy load was establ ished from s inuso idal measurement of actuator irnpc-

dance characteristics. The resulting model is shown in Figure 15.

Re_

p0

R resistance due to copper losses (0.67 ohm)

1 = open circuit inductance (12.7 ml)

1, (I-K) = leakage inductanice (). J8 m1l)

Kl. = magnetizing inductance (12. m11)

= hvst vresIs inductan( e (.29 MH)

R = resi stance due to eddy ,rrent s (774 ohm)e

Figur. 15. El e trital A, t ;,tir ModcI

31

I4.

Static tests on both amplifiers were performed. DC actuator currenit

versus input voltage was measured and resulted in the plots in Figure 16. For

the switching ampl Iif it r it is no ted that one ampere of d.c . ecurren t prevailsa

with no input. Th s is due to the bias and ad just-munt for a ten percent duty

cycle. If the switching amplifier characteristic is translated down by one

ampere it Is obvious that this I invarity is inferior to the class B amp] ifihr.

The c lass B ampl if ier was tested dvniam ical t v with square wave and si no-

soidal input waveforms. The square wave tests were performed from 10Hz to, 1kHz

with peak currents ranging from 0.25 amp to 3.5 amperes. Some ot these results

are shown in the osc illoscope traces of Figure 17. Frequency responses for

bo th amplifiers are, shown in Figure 18.

3.4 rEMPERATURE TESTS

The ampl if icrwas environmentally t emperatulre t ested at maximum powcr atnd

frequency. The actuator current was set at a peak Current of I .- 7 amper. uisinig

the class B amplifier at a frequency of 150Hz. The current staved constant for

10 0 Finterval temperature rises up to 113 )11F (45')C). 1he maximum tenptratinrt, (,I

the ctutorcasereahed68oC while the aimp lifiir heat sink rose tr) 33"C ;it

room ambient (27 0C).

3.5 PERFORMANCL TESTS IN AFWL TEST CELL.

The push rod was then insta lied in place,* and the hot tom fl exure ad ju;t ed

to a slight preload . The actuator was t ightened into the test cell , and the

test,, were repeated. Full deflection oit 1 .2 mill linches was noted with a drive-

cutrren t of- 1 .5 amperes, and the hys te res is was co(nsistent with previous re sul ts.

Quan tit ati ye data, however.* was not taken diie to thlie fact that t lie pert rmnct

was noted to deteriorate shortlv after tihe device wais installed inl the teSt

c -I I . Anl examinat ion ot the de-vice showed that the threads on I li push rod

were, the mechanism for this deterioirat ion, and it is hel jeved that the higher

than ajnt it, ipated peak force Was a COnt r ihut inp ,ictor. When the f lexuires wvre

loiosened, the play inl th, push rod threads was st-ei t hioasandths of ;ii n i.

Inil i.illv', we were aole- to compensate for this play by increaising the pricload,

btit this finally resulteVd ill greater dete-riorat I On. F inallyI.v t he stroke wais

l imi ted to ;ibotit 1/3 of the spe( valuet'

i. h ll.S1 Al AFW.

lii satIctests worc repe~atod it AM ont 11 Sept umber, and a serie- ot

dvna mlI tests were run. At this time, the threads had deteriorated badly. Due

6

STATIC INFUT/OUTPUTAMPLIFER CHARACTEFRISTICS

5

W~4

L

SSWITCHING3 AMPLIFIER

zL f -

2 ~CLASS B AMPLIFIER

INPUT VOLTAGE (-V OLTS)0 .2 4 .6 .B L, 1.2 IA 1.6 1.s 2.0 2,2 2.A 2.6 2.- 3.0 3.2 3. 3U-

Figure 16. Static Intout<,)utput Aztplifi;r Characteristic-

33,

N

-. - r.

N.

£ C .4:C

t

_~ - 1 -1-

LLL -

LI- Ii'

__J < ! Lipo--

LiiL

-LL (.r

I -

0 1- - - - - - - - - - - - - - - - - - - - - - - -

0)--

to this deterioration, the stroke was limited to about 10 micrometer, . Thcese

data are available at AFWI,. The threads were reworked, and the diameter

increased subseque'it to this deterioration, and retested at Perkin Elmer.

This retest ineii :ates that a stroker greater than 14 micrometers IS Pos-I

sible. The results of this retest are shown in Figure 19.

X (.'))01')

15

14

12 S-cale 1

11 --

10 -

9 781

4

2

II

Scale 1 0.1 .2 .3 .4 .5 .6 .7 .8 .99 1.1 1.2 1.1 1.4 1.-)Scale 2 .2 .4 .6 .8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.H 1.0

F igure 19). Statirc D)i si, I dment /Cuhrrent Aft or Thread Power k

~ AI'iINMIlX A

Actuator Drawings

37

aa

off

o -w2

F 00

404

* 38

D.I-r.87- 2064-- 1 T

IP

U'

0 ~a

0 LA0

L)

40-

-A-A

39I

r-687-2065 -

InWaf

I 5 w a

0

r.:G.8 7- 2 OG~ - 9-,-E1

II L

414

r68 7-2067 -*--9~

1-0 0

04

o w

if-Ao5

so..z

U *~ I-I

Il rU

0 *

42 g

I Ix

z

00

43J

-X G87 20r.9----- --

c,

0 01lpA

0 KdLP; w

ulLU LO

Ii L

w (Ii

'-7

ir

44

II * ~~G87 2070 -Y1~!

0,2-fi 4,0

4 Ct z o~z

2u

1 L I. : , j 4

4-

0< 0 o. 2 4I - IJ'l '

0 .'-o

I "

45

r 687- 207!- ----. *YI

"I

zQn

bnI?,

a CL

(u

SiL mU

10~

44

Z! 6 87- 2 C72---I-'

U-)

10

II

L jr-I 0 &

~ ~A2 0 .

I w~L

oc 0047

I ~ ~ ~ E~W-Ga7- 2075i--YYY-:

oi

a~ I I -

LEI

2 :1; r-zf z3

l:2

ri

ii48

fibi ism