View
4
Download
0
Category
Preview:
Citation preview
HDL-CR-82-158.1
April 1983
Integrated Component Fluldic Servovalves andPosition Control Systems
by D. N. Wormley and K-M. Lee
Prepared by
Massachusetts Institute of TechnologyDinprtment of Mechanical EngineeringCambridge, MA 02139
Under contractDAAK21.79-C.0158
U.S. Army Electronics Researchand Development Command
Harry Diamond Laboratories
C-) • Adelphi, MD 20783
DTICA-o=.d f Pu rei- , . - aV.ut. -i . ELECTEjt
APR 2 a8N
04 27 018I q-°
--- S~~Uý W.7--- .
UNCL-ASSIFIEDSECURITY CLASSIFICATION OF THIS PAGE (WmnDota Ln.,r."c.
REPOT DCUMNTATON ACEREAD INSTRUCTIONSREPOT DCUMNTATON AGEBEFORE COMPLETING FORM
1. REPORT NUMBER bQINO. RECIPIENT'S CATALOG NUMBER
HDL-CR-82-158-11 9,006___________
4. TITLE (and Subtitle) S. TYPE OF REPORT 6 PFRIoD COVERED
Contractor ReportIntegrated Component Fluidic Servovalves and Position Final 1979-1982
Contrl SysemSS. PERFORMING ORG. REPORT NUMBER
7. AUjTHR(a) 8- CONTRACT OR GRANT NUMBER(a)
D. N. WormleyK-M Lee DAAK21 -79-C-01 58(HDL Contact: James Joyce)
9. PERFORMING ORGANIZATION NAME AND ADDRESS tO. PROGRAM ELLMEKNT. PROJECT. TASKAREA A WORK UNIT NUMBERS
Massachusetts Institute oil TechnologyPrgaEe:61,2ADeoartment of Mechanical Engineering PormEe .10.Cambridge, MA 02139 HDL Project: A44934
It. CONTROLLING OFFICE NAME AND ADDRESS 12. REPORT DATE
Harry Diamond Laboratories ApIl 19832800 Powder Mill Road 13 NUMBER OF PAGES
Adeiphi, MD 20783 841.MONITORING AGENCY NAME A AOORESS(II different from, Controlling Office) IS. SECURITY CLASS. (of timi report)
UNCLASSIFIED
* -,15s.. OECL ASSIFICATiON/DOWNGRADINGSCHEDULE
-C:": 16. DISTRIBUTION STATEMENT (a(f this Report)
Approved for public release; distribution unlimited.
17. DISTRIBUTION STATEMENT (of the abstract angered In Block 20, It different from Report)
IS. SUPPLEMENTARY NOTES
DROMS Code: 61111102.1-.44001 1DA Project: 1LI161102AH44
19. KEY WORDS (Continue an reverse side It necesem7' and Identify by block number)
-. .,Servovale Fluidir gain blockLaminai proportional amplifier Position control system
2ILAM CT ~eitilwe m rvers ebb If ,evwesay md idanitit r by block nwamhvr)
The operating characteristics of fluidic lamninar proportional amplifiers (LPA's) operating onhydraulic oil have been determined as a function of pressure and temperature. The usefuloperating range of these elements has been defined for application in multistage gain blocks andsumming amplifiers.
An operational servovalve constyucted from LPA's has been developed and coupled with a* f fluidic position feedback transducer, summing amplifier and ram to construct a closed loop posi-
DD I F 4'73 I E01ITION OF I NOV 611IS OBSOLETE1 UNCLASSIFIEDSECURITY CLASSIFICATION OF THIS PAGE (Whe1bn Data &ntrod)
UNCLASSIFIED- ~SkCURITY CLASSIFICATIOM OF THIS PAGE~~a Data SM1004
S.,
20.'pbstract (Cont'd)
tion control system. Static and dynamic experimental evaluation of the servosystem has shownthet its performance is comparable to that of a servo employing siectrohydraulic components.
This development effort has demonstrated the capability to develop high performance,closed loop servo components from standard, integrated component fluidic elements
2 UNCLASSIFIED
59CURITY CLASSIIFICATION OF' THIS PAGOE(Mo~hn Date Entered•)
...
CONTENTS
Pagei ~1.• INTRODUCT10N ........................ . . . .. . . . ........ 7 -
2. FLUIDIC INTEGRATED COMPONENTS. ....................... 11
2.1 Laminar Proportional Amplifier ......... 0.......... 11
2.2 Fluidic Channel Resistante.....o...................... 19
3. FLUIDIC GAIN BLOCK AND SERVOVALVE .......................... 21
3.1 Gain Block Configuration and Characteristics .......... 213.2 Fluidic Servovalve Configuration and Characteristics.. 273.3 Temperature Effects and Compensation .................. 333.4 Dynamic Response....*... ................. .. .. 39
4. FLUIDIC POSITION SERVO .................. *............... 47
4 . 1 F l u i d i c S u m m e r . . . . . . . . . . . . . . . . . . . . . . . 4 7
4.2 Mechanical-Fluidic Displacement Transducer............ 474.3 Fluidic Servovalve.. ............................... 514.4 Closed-Loop Fluidic Position Servo .................... 524 .5 I m p l e m e n t a t i o n . . . . *. *.. . . . . . . . . . . . . . .... . . . 5 3
5. SUMMARY AND CONCLUSIONS .................................. 57
NOMENCLATURE ..... I.. .......... ............... 73
DISTRIBUTION .................................................... 77
Aeossjon Fol---
DTIC TABuenamoUnoed
JUStiflostl~o
"_DI Itrlb.u_. t ton/
Avallabillt, Codes
",-.
S S 2 . .
2 * - -T
FIGURES
Page
1. Closed loop position servo and servovalve output characteristics. 8
2. Kinematic viscosity as a function of tewperature................. 10
3. Primary fluidic C format integrated components*........*......... 12
4. Silhouette for b - 0.5mm LPA in C format........................ 135
5. Comparison of theory and data of blocked load gain ............... 15
6. Experimental data illustrating the operating range of LPA ........ 18
7. Comparison of analytical and experimental nozzle resistance ...... 22
8. Gain block schematics........................................ .... 24
9. Gain block construction schematics ............................... 28
10. Experimental gain block output characteristics ................... 29
11. Comparison of computer-aid design prediction and experimentalgain block output characteristics ................................ 3 0
12. Servovalve schematic .... .*...*.. ....................... . 32
13. Effect of valve parameter a (on valve characteristics)........... 34
14. Servovalve construction schematics .................. ......... 35
15. Experimental data of servovalve output characteristics ........... 36
16. Comparison of experimental and analytical servovalve outputcharacteristics ................... o. o........ ... .............. ..set 37
17. Gain block compeitsated and uncompensated blocked-load pressureg...... ..... ....... 40
18. Fluidic servovalve compensated and uncompengated blocked-loadpressure gain .. o. e..... .... ..... . .. . .. . . .............. a. ..a;.9 41
19. Cain block and servovalve blocked-load pressure gain frequency.response........... .... . ......... ., 43
20. Gain block and servovalve no-load flow gatin frequency responses... 44
21. Comparison of blocked-load frequency response between inte-gre=ed component servovalve and breadboard configurationseIjvovalve...... .. ............ . . ...... ....... ;..,.; .. . 45
22. Comparison of no-load frequency response of integrated.coa-ponent servovalve, breadboard configuration and electrohydraulicservovalveo.o..... .... ...... ... *... . .. . .. . .. . . . . . ..... 46
23. Fluidic position servo block diagram... .e...... ............... 4 8
24. Mechanical-fluidic displacement transducer schematic...... ...... .49
25. Root locus analysis of fluidic position servo ................... 54
26. and W T of fluidic position servo ........................... 55n
27. Fluidic position servo construction schematic...................56
28. Fluidic summer schematic and static characteristics.............. 60
-"-
K Page29. Fluidic sumuer blocked-load frequency response ...................... 6130. Displacement transducer blocked-load static characteristics......... 62
31. Displacement transducer blocked-load frequency response.....*....... 63
32. Fluidic servovalve no-load static characteristics .......... .... 64
33. Fluidic servovalve no-load frequency response ........ 65
34. Comparison of step response between fluidic and commercialposition servo ........ . .. .......................... 66
35. Comparison of experimental and analytical step response of fluidicposition servo ....................... . ...... ........... 67
A-i Servo components descriptions ..................................... 72
TABLES
1. Hydraulic Oil Unvis J-43 specification ........ .. .......... 11
2. Characteristic dimensions of LPA's ................................. 16
3. Gain block configuration and incremental parameters ................. 28
4,. Fluidic position servo component configuration ........... .... 585. Values of parameters of fluidic position servo ...................... 59
A-1 Gain block stacking order and descriptions .......................... 69
A-2 Servovalve stacking order and descriptions .......................... 71
LN
5
.4"A
: 1. INTRODUCTION
Hydraulic control systems are wideiy used in applications where high force
levels, fast response and high power to weight ratios are required. Aerodynamic
* control surface actuators, machine tocl actuators, mobile equipment control sys-
tems and marine control systems frequently employ closed loop hydraulic control
systems. Important performance criteria for these systems include maximum force
and velocity capabilities, accur.cy, repeatability, reliability, maintainability
and cost.
The primary power modulation elements in high performance hydraulic systems
are servovalves. In position and velocity control systems servovalves are
utilized extensively as indicated in figure 1. Also shown in the figure are
typical pressure-flow output characteristics of commercial sliding spool servo-
valves. Servovalves with linear flow gains and high pressure gains are desired
to achieve accuracy and overcome actuator and load stiction.
Electrohydraulic servovalves consisting typically of a torque motor, a
first stage flapper nozzle or jet-pipe valve and a final stage sliding spool
valve are the dominant type of valves employed in high performance hydraulic
systems. In these valves the electromechanical interface and the sliding
mechanical elements contribute to valve cost, sensitivity to contamination and
sensitivity to failure due to radiation.
The high reliability, insensitivity to extreme environments and low cost
associated with no moving part fluidic elements and the potential for weight and
size reduction In comparison to conventional valves are attractive features for
servovalves. In addition, the implementation of a closed loop control system
employing only fluid and mechanical elements offers potential for reduction of
sensitivity to radiation and increased reliability with the elimination of
electro-mechanical interfaces. In systems where these &ttributes are important
and where the quiescent power drain associatec with open-center fluid valves can
be accommodated, fluidic servovalves and pure fluid-mechanical control systems
have high application potential.
The application potential of pure fluid servovalves and control systems
1R. Deadwyler, Two Stage Servovalve Development Using a First-Stage Fluidic
Amplifier, Harry Diamond Laboratories, HDL-TM-80-21 (July 1980).
7
-' . - . . . . . . -- .- ,--.-. --•- - - * - - -C . . . . . - . .
Actuator hposition
Set Processi LaPointI
Dislaowe,: Q
LL
Figue 1.Closd lop poiti~ seVoadsroalve otu
4.. cDarpcteristics.
00, Loa
8P
'4 4 -- 4 - . ..- 4.4. 4 -.. 4 . . ... , 4L
-4W- 4.. . . .. . . . . . . . .. -2 b..~
-,. motivated a study2 in which a pure fluid servoalve was constructed using laminar
* proportional amplifiers (LPA's) in a breadboard configuration. In the present
study, development of pure fluid servovalves and servosystems has continued.
Servovalves constructed from standard C format laminates 3 and interconnecting
elements have been developed to reduce packaging volume and weight, provide a
basis for standardization and to improve valve dynamic response. The static
and dynamic characteristics of the C format LPA elements individually and
integrated into the servovalve have been measured including performance sensi-
* tivity to temperature. Finally the servovalve has been employed in a closed
loop position control system which !.ncludes an actuator and fluidic position
transducer for evaluation.
Two aspects of the influence of temperature are particularly addressed
in this study, namely, the static characteristic performance and the laminar
operating range of the LPA as a function of supply conditions.
In all tests described in this report, hydraulic oil Univis J-&3 has
been used. The properties and specifications of the fluid are sumnarized
in table 1 and the kinematic viscosity as a function of temperature is shown
in figure.2 in the range from 20'F to 140*F. Univis J-43 changes in viscosity
over this typical temperature range. An exponential curve fit of the form:
S= u - X ( T - T o( I
whereV - kinematic viscosity at temperature T,
V - kinematic viscosity at reference temperature, T0 0
SX - viscosity - temperature coefficient,
may be used to approximate the kinematic viscosity of the hydraulic oil Univis
J-43 as a function of temperature
with X 0.02862 l/OC [0.0159 1/'F]
and v - 21.78 cSt. evaluated at 25*C [77*F]
The exponential curve of equation (1) is a good approximation to the fluid vis-
cosity as shown in figure 2.
D.N. Wormley, D. Lee, and K-M Lee, Development of a Fluidic, Hydraulic
Servovalve, HDL-CR--81-216-1, Harry DJamnnd Laboratories (February 19d1).
3M.F. Cycon and D.J. Schaffer, Design Guide for Laminar Flow Fluidic Ampli-fiers and Sensors, HDL-CR-28-288-1, Harry Diamond Laboratories (April 1982).
9
-.
LL
00
P 006--0
I I
.4.
10
b%..
TABLE 1 HYDRAULIC OIL UNIVIS J-43 SPECIFICATION
Specific gravity: 0.860?
Temperature ('C) Kinematic Viscosity (cSt.)
100 5.2
54 10.3
40 14.9
-18 102.7
-40 495.5
-54 2332.
2. FLUIDIC INTEGRATED COMPONENTS
Fluidic C format integrated components are basic elements in the servo-
valve. The pr'.mary elements are the LPA and the channel resistance which
are shown in tigure 3. The secondary elements are vents, exhausts, spacers,
transfers, base plate, input and valve manifolds. The primary elements are
standardized in design and manufacturing and are thus well documented and
have repeatable characteristics.
2.1 Laminar Proportional Amplifier
Laminar proportional arplifiers have been designed to operate in
the laminar flow regime. The detailed geometry of a typical HDL integrated
component laminar proportional amplifier with a summary of LPA characteris-
tic dimensions are illustrated in figure 4.
The LPA performance is influenced by the temperature of the op-
erating fluid through its influence o"? the fluid viscosity.
The analytical design procedures which predict the performance of
the laminar proportional amplifier based on the characteristic dimensions4
and the supply conditions have been discussed by Drzewiecki et al. The
4 T.M. Drzewiecki, D.N. Wormley and F.M. Manion, Computer Aided DesignProcedure for Laminar Fluidic Systems, Journal of Dynamic Systems, Measure-ment and Control, 97, Series G, No. 4 (December 1975).
- 11
(a) Iamiilar propontiona I awl) lIi fi (LPA)
(b) linear channel resistance
(c) nozzle or nonlinear resistance
Fig•ure 3. i'r imlr-' cy tL.jdic C format integrated components.
1?
B BtO
x/c/0
b p
b - supply nozzle width
- aspect ratio, h/bs
B control port minimum widthC
B 0 outlet port minimum width0
B - splitter widthB t downstream control edges spacing
X control port channel lengthcX0 outlet port channel length0
Xth supply nozzle throat length
X supply nozzle - splitter distancesp
Capitel.zed parameters are normalized to bsa
Figure 4. Silhouette for b 0.5 mm LPA in C Format3
13
LPA's with aspect ratios less than one are commonly used in the design of
multistage gain blocks. However, limited experimental data are available
for LPA's of aspect ratio less than one. The work described in this section
provides data and correlations for gain block and servovalve design using
LPA's in this range of aspect ratios.
The supply condition of an LPA may be characterized by the modi-
fied Reynolds number
NR- MR
R ) 2 (lxh
where b 2P
R V P
p - fluid density
P - supply pressure, gage8
T = nozzle aspect zatio, h/bs
Experimental data have been collected from three different C for-
mat LPA configurations (HDL 63020, HDL 72010 and HDL 61505) with aspect
ratios of a - 0.667, 0.55 and 0.333 and supply nozzle throat widths of
0.75 mm, 0.5 mm and 0.375 - respectively. The characteristic dimensions
are shown in Table 2. The LPA's have been tested under blocked-loadconditions over a temperature range of 5.5*C (42°F) to 48*C (118gF) and a
pressure range of 35 kPa (5 psi) to 11,032 kPa (1600 psi). In all tests,
the control bias pressures were adjusted to 5 percent of tl e supply pressure.
The analytical and experimentally measured blocked-load pressure
gain are plotted against the modified Reynolds number in figure 5. The
analytical model is based on the modified two-dimensional, incompressible,4
laminar jet deflection theory discussed by Drzewiecki et al. The experi-
mental gains were determined from the slope of the blocked-load characteris-
tics at the null position.
4 T.M. Drzewiecki, D.N. Wormley and F.M. Manion, Computer Aided DesignProcedure for Laminar Fluidic Systems, Journal of Dynamic Systems, Measure-ment and Control, 97, Series G, No. 4 (December 1975).
14
.......................................
0 en4- 4
'4.4J
41 0
0, 0 041 4
4444
NI 0
43 )
P0 %. C44
UT~b xnsszd POT PNOOT
o1
TABLE 2 CHARACTERISTIC DIMENSIONS OF LPA's
HDL design 63020 72010 61505
b (Mm) 0.75 0.5 0.375
"0.667 0.55 0.333
B 1.0 1.0 1.0
B3.08 4.0 4.733
x 8.875 13.312 21.c
B 1.2 1.25 1.330B 2.667 2.875 3.6
X 14.875 16.56 22.080
B 1.25 1.125 1.167t
B 0.55 0.5 0.533sp
Xth 1.257 1.25 1.25
* X 8.0 8.0 8.0sp
The experimental data closely follow the analytical predict,.on up
to N' " 120 beyond which transition-to-turbulence occurs and the theory
based on laminar flow fails. In the turbulent flow regime, the flow noise
K increases, offset increases and the gain decreases. Therefore, thetransition-to-turbulence establishes an upper design limit of operation for
the LPA. In the laminar flow regime, the blocked-load pressure gain in-
creases as the Reynolds number increases.
L.. A guideline to determine the point of transition from laminar
to turbulence for fluidic devices (when X - 8 to 10) has been discussed
by Drzewiecki et al. 5
CdNR
-200 (2).':. +
where Cd volumetric discharge coefficient, f(N•) or the expression may be
written as
5T.M. Drzewiecki and F.M. Manion, Fluerics 40: LJARS, The LaminarJet Angular Rate Sensor, HDL-TM-79-7, Harry Diamond Laboratories (December
16
S.. . . . . . .
C lr 200 (.. :. I = +Xt5(3)C R Itr l+X th
For Xsp 8 and X th 1.25, the point of transition from laminar to tur-
bulence occurs at N'. = 120.
The lover operating limit of the amplifier may be predicted from
the information on the control edge clearance. The amplifier ceases to
function if the jet cannot deflect. The distance from the control edge to
the jet edge, based on the assumption that the jet spreads linearly, is
given by Manion et al. 6
B 1Bv c (4)
whereBv - distance between control edge and jet edge
C9 = momentum flux coefficient.
With C9 - 1.32 C 2 and B - 0, the expression may be written as
47.5 B
d Ri• The average value of CdN' for the three amplifier conf igurat ion s is ap-
proximately equal to 10 and the point at which the LPA fails to functionoccurs at NR'-- 22.
The experimental data obtained for the blocked-load pressure
gain of the three different LPA configurations are plotted in figure 6.
The effect of the modified Reynolds number on the blocked-load pressure
gain may be sumarized by noting the four regions defined in the following
paragraphs.
In region (1), 0 < N' < 40, the blocked-load pressure gain is
less than 50 percent of its maximum value. In all three different con-
6F.M. Manion ana T.M. Drzewiecki, Analytical Design of Laminar Pro-
portional Amplifiers, Proceedings of the HDL Fluidic State-of-the-ArtSymposium, 1, Harry Diamond Laboratories (October 1974).
17
..................... .
0 0 kn Go 0 M t-
40 owe0 to4wi i.p4 kOO 0
%0 0 9.b 0
'Ur-
0@4 0 101
00
O41rS-d
4 "4".4 4J
r-
0 (Z 4.bu o
b" 4
181
figurations tested, the LPA gain is zero for NR< 20.
In region (2), 40 < NR_ 100, the typical LPA operating range is
bounded by two limits, a lower limit below which the LPA gain is too low
for use and an upper limit beyond which transition-to-turbulence occurs.
In region (3), 100 < NR' < 120, just before the analytical rre-
dicted point of transition-to-turbulence, the experimental data are
scattered. This uncertainty suggests that the point of transition-to-
turbulence is in this region.
In region (4), NR_ > 120, the blocked-load pressure gain decreases
and noise increases since the LPA's are operated beyond the transition-to-
turbulence.
2.2 Fluidic Channel Resistance
The channel resistor is one of the important elements in fluidic
circuits. In general, the channel resistance may be represented as a
laminar, fully-developed duct resistance. The channel resistance consists
of a linear portion due to the fully developed viscous dissipative flow in
the channel plus a nonlinear portion due to the entrance region pressure
Sdrop. A general expression for a channel resistance given by Drzewiecki 7
is12 U= + I + C + 0.4752q (6)
R 2bh~ l a bc
where
for 1 < a < 2, 0.35 < C < 0.5;c
for o > 2, C = 0.5,
x - channel length in the direction of flow,
b - average channel length,
b - minimum channel width,
Q volumetric flow through duct,
- channel aspect ratio, -h
c
7T.M. Drzewiecki, Fluerics 37. A General Planar Nozzle DischargeCoefficient Representation, HDL-TM-74-5, Harry Diamond Laboratories,"(1974).
19
= absolute fltid viscosity,
C = empirical constant.
For the case where the channel length is much longer than the en-
trance length, the nonlinear term may be neglected. An analytical study on
the hydrodynamic entrarce length for incompressible flow in rectangular
ducts has been performed by Han. The experiments to verify the analysis9of Han for aspect ralios of 5 and 2 have been performed by Sparrow et al..
The results of these studies to determine the entrance length may be sum-
marized as2
L- 0.055 c N foro a 5,c c(Jl7(4Oc )2Rc
Le U.127 N o for ac 2,
whe re
L - normalized entrance length, Ze/b
NR - channel Reynolds number, referred to b.c
For flow between parallel-plates, Schlichting10 has derived an expressionfor the entrance length as
L e
2 - 0.04 for b >> h.Oc2NcR
c
If the assumption of fully developed flow is Justified and the
entrance length can be neglected, the flow can be described by Poiseuille
flow between p&rallel plates and the expression for the resistance may be
stated as
R - 2 for b >> h. (7)c bh3
V88L.S. Han, Hydrodynamic Entrance Lengths for Incompressible Flow in
Rectangular Ducts, Journal of Applied Mechanics, 27, Trans. ASME (1960).9 E.M. Sparrow, C.W. Hixon and G. Shavitt, Experiments on LamAxar Flow
Development in Rectangular Ducts, Journal of Basic Engineering, Trans. ASME(March 1976).
10M. Schlichting, Boundary Layer Theory, McGraw-Hill, New York, New York,(1960).
20
.•" • " , " - - " : , • . i - • : . . . .: : ' - : .• . • , • •i' -. . - . ,* . -, / , . . .. . . ..-
-- 1 . .-.'. i7X -N,-. - - - -. - m
The standard nozzle, 5221A-20 shown in figure 3, has been used as
a channel resistor in the design of the gain block. Since the channel length
is much longer than the entrance length of the flow for a wide range of
Reynolds numbers, the quasi-fully-developed assumption is justified, and
equation (6) may be applied. Normalizing equation (6) to the linear portion,
the expression for resistance is
j-2 a +1 + 2) 1+ X+ 0.7 ( * _(- th C N' (8)
12 \b - ) d R
whereR
Im+
The nozzle resistance as a function of modified Reynolds numbers
has been determined experimentally. The range of temperatures and the
pressure drops across the nozzle are from 6C to 480C and from 5516kPa
to 10170 kPa respectively. The experimentally determined nozzle resis-
tance is compared with analytically predicted values using equation (8) as
a function of modified Reynolds numbers in figure 7. The experimental data
falls within 10 percent of the predictions in the range of Ný tested.
3. FLUIDIC GAIN BLOCK AND SERVOVALVE
The fluidic servovalve consists of a multi-stage LPA gain block and a
set of laminar flow resistors. The gain block is a basic power amplifier
while the resistors are used to provide feedback and summing functions.
3.1 Gain Block Configuration and Characteristics
The analytical design of the fluidic gain block to predict the
essential characteristics as a function of individual stage operating
Reynolds number, control bias pressure and the detailed geometry of the LPA
has been discussed by Manion et al. 4 ' 1 1
4 T.M. Drzewiecki, D.N. Wormley and F.M. Manion, Computer-Aided DesignProcedure for Laminar Fluidic Systems, Journal of Dynamic Systems, Measure-ment and Control, 97, Series G, No. 4 (December 1975).
11F.M. Manion and C. Non, Fluerics 33: Design and Staging of Laminar
Proportional Amplifiers, HDL-TR-1608, Harry Diamond Laboratories (September1972).
21
-' - W �Y�' -� ' � -. - F-. - - - - - . . -.---- -.--.------ �----I-�----..---'-----�---'--
-� - -- - - - �S. - - . - . ...-... . .-S. -.
I.
C
SciE
.1.4 5-S. *
0 0 N�5** .14 0dJ
Nem.5. - I.
�I �
N0 1
3 0
.- *64
K'0
1'¶5
I
0 4
4J -4�
N h.14* . 0
* Np5.-S -ISi. 1 0
ka I
0* �fl
S p 5
N o
� '.ouw�sTse�i rsuuwv *30CUOTUUWFQ
22.4
%��* � '5 5� '5% - -- -. . - -- - .
-- * . - .- ... - - -
The general design criteria of a gain block are as follows:
(1) Maximize the laminar operating range by matchingthe modified Reynolds numbers, of each stage:
"il -NM'2 - NR'3NRi =R2 = R3'
(2) Achieve 90' phase shift bandwidth requirement:
900 3 4x3 _ sp
21r2PdCdi p
(3) Maximize the blocked-load pressure gain:
K, K 'KK1 •1 2 3K
where K1 , K2 and K3 are blocked-load pressure gain of Ist 2 nd and 3 rd2 3
stages LPA respectively. R01 and R02 are output resistance of the let
and 2nd stages and R1 2 and Rt 3 are input resistance of the 2nd and 3rd
stages.
(4) Maximize the input-to-output resistance ratio:
R /Rii 03*
(5) Minimize the quiescent flow draw:
3Q8 i: £ Qsis
5i-l
at ud rdwhere Q (i 1,2,3) are supply flows of the 1st, 2 and 3 stages. An
iterative design procedure is generally required to achieve a design which
meets (if possible) the above design criteria.
The gain block shown in figure 8 has been designed using the
guidelines cited above.
The supply pressure of the three stage gain block is connected
directly to the final stage of the gain block. The first and second
23
'1A
I
L" ~Po
add
'i.r. - -.... -
3 Stage
P%* HDL 71505
S2nd stage
HDL 6302010
IL• . -, ....- • -.. I .Qs i• cd
p3
Figur'e B, Gain block schematic..
24
.~~ Stage-- . ~. - . * ~ . * * - . t * - -A m . . . . . . . . .
stages are supplied by reducing the main pressure through the pressure re-
ducers which are a number of nozzles in parallel.
The first and second stage supply pressures depend on the final
stage supply condition with respect to the fluid properties. The supplyth;ressure of i stage, P may be expressed as
P sP m 1a5 fb +i (9)
ni Rsi
where
a, no. of LPA's in parallel of ith stage,thni M no. of resistor nozzles in parallel of i stage,
thR * LPA supply resistance of i stage,
R a nozzle resistance of ith stage.ni
As the nozzle resistance, Rni, and the supply resistance, R.i are connected
in series, the flow through R and R are related by the continuityni si
equation as
- 2 (10)
Qsi ni
With equation (10), the ratio of the nozzle resistance, Ri, to the supply
resistance, Ri, may be expressed as
R 24 X [c 1 rH1"ni ni 1. _-ni-
:-" -- = " + n + C • Cd
Rs -2 [ -1B2 H nN Rini ni ni
S]~2 "i+ 0.95 • BiHi] " Cdi (11)
B niH ni n Cdi
where
0, O LPA aspect ratio,
"X U-i normalized nozzle length, Xin/bsi,
Bn a normalized nozzle average channel length, bn/bs,ni ni sit
Bni normalized nozzle throat width, bi/bi,
Hni - normalized nozzle height, hn/bait
25
-- - - - - - .... .... .. -- --
H H6
-ni Sn.
1 < < 2, 0.35 < c < 0.5; > 2, C = 0.5.B ini B n"S mu ~th I
Subscript i refers to the i stage.
The Reynolds number of the first and second stage LPA may be
. related to that of the final stage as
N b si Psi N(2NRi bsb aN-f (12)N1 7be
wherebs* LPA supply nozzle throat width, i th stage.bsi LPA supply nozzle throat width, i stage,bf 0 LPA supply nozzle throat width, final stage,
N Reynolds number of i stage,Ri.
N Reynolds number of final stage.Rf
" ~ As the first and second stage supply pressures a".. an implicit function of
the final stage Reynolds number, NRf* the first and second stage operating
Reynolds numbers depend only on the final stage Reynolds number. In the
following discussions, the gain block operating Reynolds number is referred
to the operating Reynolds nuaber of the final stage, N Rf* The first and
second stage Reynolds numbers can be related to the final stage by solving
_N• equations (9) through (12) simultaneously.K2In the report by Wormley et al. 2 a hyperbolic tangent curve has
been used to describe the nonlinear saturation characteristics of the gain
block. The output pressure/flow characteristics of the gain block may be
expressed as
QL Pcd Pod-l- tanh (13)
PLs Pcds ods
where
QL = output load flow,
Pod = amplifier output pressure differential,
"P cd amplifier input pressure differential,
2 D.N. Wormley, D. Lee, and K-H Lee, Development of a Fluidic, Hy-draulic Servovalve, HDL-CM-81-216-1, Harry Diamond Laboratories(February 1981).
26
S . .
f- - " I . . , •. . .ý .- , I - . '..- , ", "- - ,'- , .- ' .- ii..- .-. ,., - .i, -_ . , - - . . . . ' . , . , , - . . . ,
QLs = saturation output load flow,
Pods m saturation emplifier output pressure differential,
and where the saturation control pressure differential is defined as
P ous QLspcd - 0-- =~ L (14)c ds K K (4
V q
with the incremental amplifier static pressure gain K and flow gain, Kp q
defined as
K Kp - cd (15)
QL
LQpQ L
Kq a Pcd o (16)IPod =0
A three-stage fluidic gainblock with a single supply pressure
based on the standard packaging technique was cinstructed and tested. The
construction schematic of the gain block i's shown in figure 9. The
stacking order is listed in appendix A. The characteristic dimensions of
the amplifier laminates and the three stage amplifier parameters, measured
with Univis J-43 at a temperature of 27°C, are summarized in table 3. The
experimental output characteristics are displayed in figure 10. The com-
parison between the predicted and the experimentally measured characteris-
tics displayed in figure 11 shows that the analytical model with hyperbolic
tanh curve closely matches the experimental data.
3.2 Fluidic Servovalve Configuration and Characteristics
The conceptualization, analysis and design of a fluidic servovalve
constructed from the gainblock and fluidic resistance elements in a2
breadboard configuration are described by Wormley et al. . The flow feedback
resistance in the case in which the net flow feedback is equal to zero can
be eliminated. In order to minimize the loss of output flow due to feed-
2Drli Wornley, D. Lee, and K-M. Lee, Development of a Fluidic,
Hydraulic Servovalve, HDL-CR-81-216-1, Harry Diamond Laboratories
(February 1981).
27
" " "" , " . . . . . .' -• ' , .. .. • . , , .. .* z L, i .i ± ... . i ,*: . . -. . . . . . . . .-
input 'npt anolSignal rptMnfl
., """ " " . ........... a Stage Amplifier2n tge ~- ~ ]2 Stage Amplifier
NozzleF 3I Stage Amplifier
I st tegeI
SNozzle _ _ _., - . __ _ .
-- Base Plate
Valve Manifold
Figure 9. Gain block construction schematics.
TABLE 3 GAIN BLOCK CONFIGURATION AND INCREMENTAL PARAMETERSLAMINATE DESCRIPTION
Stage Design b (mm) a= h/b Number ofs Sections
1 HDL 63020 0.75 0.667 22 HDL 72010 0.5 0.55 3
" 3 HDL 61505 0.375 0.333 6
,_EXPERIMENTAL DATA
P 6895 kPa 1000 psiPods 2551 kPa 370 psi
-6 3QLs 8xW0 m /sec 0.5 cis
Pcds 6.72 kPa 35 in.H 0
K 277
q 8,6xi0-7 m/sec/kPa 0.35 cis/psi
R 5.80x10' 0 N-s/mi5 140 psi/cisa
,', 28
00
(U UM
* r41
N
o C4
10I 4, ,o4 = I"
41
l I1 II II# N N
0
04.V'
.400on in VA 0
I I p "s•""€ -
o,-
1
4
I2 9
L!
'I
4%*
:.: 29
F.
r-
441
tn 0
a*a
0. 0
u.~CLgi *!4440 . 0C 0
a oo~-r r-4.
0 14
P04
67...
I qC4 cLteso/ -mmm ol.-ii.94n
30
back and the complexity of the valve construction, the case of zero flow
feedback is chosen for the performance evaluation. In this particular case,
the resulting nonlinear normalized governing equation of the valve becomes
*LQ + - tanh + (17)Pods PPod
Ls ods [f ida dS
where
" Pidm maximum input pressure differential
Ixm P -Am. for R >>R (8R+ cds fp i.(8;.: + R
•* a
'Y RRi 1 1 l,(9
i- R K )Kp (19)fp +
a
R, M input resistance,
Rfp - pressure feedback resistance,
R a amplifier input deflection resistance.a
The schematic drawing of the servovalve is shown in figure 12.
The fluidic servovalve steady state blocked-load pressure gain and
no-load flow gain can be obtained from the linearized valve model as
aqL KG q _ ~--- (20)
dqs RiPod- 1+ +
Gs - PId - (21)dQLO ( R-y) +a
31
L P
G(_
- .•_ R '
pP od
'ý44
I Ra a
K Rfp
fpp
4 ~ id
Figure 12. Servovalve schematic,
32
.; o .. L- -r --.- ,,--- , -.. ..... •¶ r r r "''" "-V '• I • • " V 7W -' - - ' '
The steady state performance of the servovalve is characterized
by the parameters a and y. The valve parameter ot indicates the flow gain
linearity with respect to the maximum input pressure differential whereas
the parameter y provides information regarding the limit of servovalve
stability.
As shown in equation (21), the servovalve pressure gain is sen-
sitive to the valve parameter y. The fluidic servovalve has a maximum
pressure gain at y equal to one and has a negative pressure gain for y
greater than one. The valve is unstable when the valve pressure gain
curve has negative slope.
The effect of a variation in a is illustrated in figure 13. With
y - 1 to achieve maximum pressure gain, the servovalve flow gain becomes
more non-linear as a is increased.
An integrated component fluidic servovalve has been constructed
and tested. The gain block discussed in the preceeding section has been
used in the servovalve design. Two HDL 5196 capillaries are connected in
series to form an input resistor Ri. The pressure feedback resistor Rfp
consists of four parallel HDL 5026 capillaries in series, with a HDL 5027
capillary. The experimentally measured resistances at the temperature of
"27C (80.6*F) are::: =1010
R 4.42 x N-s/m 5 (105 psi/cis) and
Rfp= 7.52 x 10l2 N-s/mi5 (17850 psi/cis).
The servovalve construction schematic is shown in figure 14. The stacking
order is summarized in appendix A.
The experimental data obtained at a temperature of 270C and a
supply pressure of 6895 kPa (1000 psi) showing the output characteris-
tics are displayed in figure 15 and compared with the analytical prediction
in figure 16. Apart from the offset, the analytical prediction fits the
experimental characteristics well.
3.3 Temperature Effects and Compensation
For imcompressible flow, the working fluid temperature affects
the Reynolds number, NR - (b /v)/2Ps/P, through an influence on the fluid
33
1.0 II/Q~s C --.
L Ls a - 0.5
1.0
0.5
.4 od
4. "-1.0 -0.5 0 0.5 1.0 ods
;.'
QL/QLs
-10,-000. 1.02
a- 125
y
=1.0
________________ ~od
-1.0 -0.5 0 0.5 1.0 Pods
t• -
a ,, 2.S
"Y -1.0
,•.S
'•'
Pod
-'-1.0
-0.5 0
0.5 1.0 Pods
.4," Figure 13. Effect of valve parameter a on ValveCharacteristics.
* 34,p
4,E':• •,,',,'- -'','•," '-_'-1-
. ..2._
, ". . '", -.- •, . . . .. .J -• , -_.. ,.'- , ". _- " i . ,._ -. _. . .. • ._- . .i .... -
.. . - -
00 U,
'o - -4 F4 *q
4 04
414
35 - -
'..,, ..
, I,
A.-to
a!
4.dO
o (~4
36~
Yl-"
0 0
.494
04
a a60
100
9. 0
37
"viscosity. The effect of the change of operating Reynolds number on the
blocked-load pressure gain of the LPA has been discussed in section 2.1.
The stability of the fluidic servovalve is primarily deter-
mined by the valve parameter 1, which is a measure of the degree of
positive pressure feedback. The value of y can be increased by de-
creasing the pressure feedback resistance, or increasing the gain block
pressure gain, Kp, or the value of input resistance to the point where y
is greater than one and the valve fails neutrally. The additional flow of
the positive pressure feedback due to the increase in temperature tends to
decrease the valve stability. On the other hand, if the positive feedback
flow is reduced as a result of a decrease in temperature, the valve behaves
as an amplifier with reduced feedback. To compensate for the temperature
effects, the valve parameter y must be kept at a desired constant value in
the temperature range concerned. Since the ratio of two linear resistances
is independent of fluid viscosity and the factor 1/(1 + Ri/Ra) is not
significantly affected by temperature variations, the valve parameter y is
primarily dependent on the blocked-load pressure gain of the gain block,
K.p
The valve parameter a is proportional to the ratio of maximum
input pressure differential, P id' to the saturation control pressure
differential P cds The saturation control pressure differential, Pcds'
decreases as the temperature increases. The maximum input pressure dif-
ferential must be limited in order to keep the flow gain in a reasonable
linear range. However, the linearity of the servovalve flow gain is main-
tained at the expense of the maximum input pressure differential, Pidm"
Since the gain block characteristic performance is primarily
a function of final stage Reynolds number, this provides a means of ten-
perature compensation by maintaining a constant operating Reynolds number,1Rf. From the definition of the Reynolds number and the kinematic viscosity-
temperature relation as written in equation (1), we have
N R(T)Rv Ps(T) ~(T-T) P(T))T -
N(T) V Ps(T) e 0 T) (22)4Rf 0 s 0
38
If NR (T) = NR (T) - constant, thenf f
P (T)- e' 2X(T-To). (23)
SP(T 0 )
Hence, the supply pressure must be varied directly proportional to the
square of the fluid viscosity in order to maintain a constant Reynolds
number. To decrease the high pressure required at the low temperature op-
erating condition, the supply pressure is scheduled as a linear function
of temperature, i.e.
P (T) - P (T )
T-T (24)
0
The choice of A depends on the temperature range considered and the naxi-t
mum of safe supply pressure imposed by the hydraulic plant at the lowest
temperature of interest. As an example, consider the following,
0Reference temperature, T = 256C (77(F),
soReference pressure, Psa(To 0 6895 kPa (1000 psi),
Temperature range, 10C < T < 50*C, (50"F < T < 122*F),
Maximum supply pressure, P 11,032 kPa (1600 psi),
It follows X0 W 275.8 kPa/*C [22.2 psi/IF].
The experimental data for linearly compensated and uncompensated
gain block and servovalve blocked-load pressure gain as a function of
temperature are shown in figures 17 and 18. The temperature compensation
based on the supply pressure scheduling significantly reduces the
temperature sensitivity of both the gain block and servovalve pressure gain
and successfully extends the operating range of the servovalve beyond the
design temperature.
3.4 Dynamic Response
The fluidic servovalve dynamic model for small derivatives may be
expressed as
QL(s) G (s)Pd(s) - G (S)P (S) (25)L q9id qp od
39
4��%
*1
4.
*4%�
4,4..
EU
a.. * EU
I''S. U EU
6
4., #1 AI0 0
I0
* _
'1
Id
4.- 0
1*� SiI-p S
0(�%I -
U1W�� .TflUSSZd PWO'I-pS�po�
40
.4
I.. -a.,.I. - - - - . - .*.� ,. ....--.---.- . - - . -
I 0
*1 0
0
0 0
ON rd
M 02' >r. uo 04J w
r' 02 .~00
00
4 4i
00
uivq, axn~ssad po'q-pa(ooT
41
where G (S) K (S) (26)q -~Z (S) qiz a + a(S)
V..
G (s) G (S) [1 - y(s)]K (s). (27)qp pqp
K (S)
(a = -y [ + __a(S)(28
Zi(s) Kp(s)yG)(s) s)(S) (29)
Sfp [I + Zi_(S)_
S+ Za(S)
*, s Laplace operator.
,S., The input Zi(s) and feedback Zf(S) impedances consist of a resistance and
inertance. The input deflection impedance of the gain block is Za (s).
Two dynamic tests, flow frequency and pressure frequency response
tests, have been conducted point by point on both the gain block and servo-
valve at a temperature of 27*C. The experimentally measured pressure and
flow gain as a function of frequency are plotted in figures 19 and 20. The
test data show that the flow gain of the valve reaches 900 phase shift at
80 ,.z, the pressure gain reaches 900 phase shift at approximately 100 Hzand the experimentally measured pure delay time for both the pressure and
flow gain is 1.1 ms.
The comparisons of the frequency response between the breadboard
configuration and the integrated component fluidic servovalve are shown in
figures 21 and 22. The elimination of the feedback line capacitance which
is present in the breadboard configuration as described by Lee 12, leads to
improved response in the integrated component fluidic servovalve. The data
show that the pressure gain reaches 900 phase shift at 7 Hz for the bread-
1 2 D. Lee, The Analytical and Experimental Development of a Fluidic
Servovalve, Massachusetts Institute of Technology, Ph.D Thesis (April 1980).
42
I
14 01
3 0 0 00~
C3 0
100
00
14
04
59 J
43'~
beo
V4
re0
in
up Q090 3U v Q Jo ut
440
700 Breadboard
Con fiquration[-..
?00
0"0
So
40
00
10 so100,s.
oa
o\ o ..
UQ
F r q e n c y i n R
0
I0400
300
510 so 'ý1001z~q~ncyin H
Figure 21. Comparison of blocked-load
frequency
response between integrated
component
servovalve and breadboard
configuration
servovalve.
45
Znktoqrated5COnzent Servovaeve
B readboardConfiguration
- ~ .~ EectrohydraulicServiovalve
0
-10
5 10 ~Frequency in Hz5010
-00
Fiue22. Comp~arison of no-load freqwp-ncy response ofintegrated comrponent 4uervovalve, breadboard
-7.con figuration ade lectrohydraulic servovalve.
........
board configuration and at 100 Hz for the integrated component configuration.
figure 22 indicates that the flow gain reaches 90' phase shift at 60 Hz for
the breadboard configuration and at 80 Hz for the integrated component con-
figuration and electrohydraulic servovalve approximately. The comparison
shows that a si.gnificant decrease in phase shift has been achieved and the
dynamic performance of the fluidic servovalve is comparable to the electro-
hydraulic servovalve with standard packaging.
4. FLUIDIC POSITION SERVO
A closed-loop fluidic position servo has been constructed as shown in
figure 23. An integrated component fluidic servovalve, similar to that
discussed in the preceeding section, has been used as a power modulation
element. A fluidic summing amplifier is also used to perform signal pro-
cessing. In addition, a fluidic feedback transducer has been developed so
"that the mechanical displacement sensing is fedback in the form of a
fluidic signal. The elements in the control system are described in the
following paragraphs.
4.1 Fluidic Summer
The fluidic summer is shown in figure 23. For low frequency ap-
plications, the transfer function of the fluidic summer may be expressed
as-T aK (a) K e s (30)
where G
K - LPA (31)
2+-R
as
GIp,LPA LPA blocked-load pressure gain,R - Summer input resistance,
R = LPA deflection resistance,as
T - LPA pure time delay.
4.2 Mechanical-Fluidic Displacement Transducer
The mechanical-fluidic displacement transducer is essentially
a position feedback sensor in which the mechanical translational displace-
ment is transformed into a fluidic signal. The electrical equivalent and
the construction schematic are shown in figure 24. The displacement trans-
47
S.. ., .. ... ., . .. - . . . . . . , ,.. . . . . • . .. • .. . . , .i•...," " , .*' - -. . . . - " " " "-..
'P .idR- R
"Rfp Rfp
Ri- Y15 PS Ris R
S Pdid
!d~d Fluidic id Fluidic LHydraulic x mechanical
Signal Sumer-- Perova-e Ra-&Lad
Signal m&Iodisplacement
Feedbackran.da cer I
- pstImplementation
Rf
R
Figure 23. Fluidic position servo block diagram.
48'-4
-I.
xl 4J
.9.
?A "
41 wto'4
4J4
IQ.94
44 )
41
49.
4) Id
4Jy A
t o
41AU
4j~
0 0
500
•i ÷4- •-t- + o
50
S
ducer may be considered as a fluidic resistance bridge. It consists of two
pairs of fixed resistors and a pair of variable resistors.
The transducer sensitivity at blocked-load may be derived from
circuit analysis
P (s) P +2Kts) dt a t 1(2(=) -x Rf + sL (32)
v v
where
P - output pressure differential of transducer,
P " supply pressure of transducer,at
x - translational displacement,
x - half stoke,
RfRv- fixed and variable resistance, respectively,
Lf,Lv= fixed and variable inertance, respectively.
The resistance and inertance of the channel resistance are
R - !?Ax for b >> h,c bh 3
L - ' (33)Lc bh
and the time constant of the channel resistance, T, may be expressed as
I c h2
c 2r"T " -7-" 2"-7" (34)
c
If equal channel heights are chosen for both fixed and variable resistances,
the transfer function of the mechanical-fluidic displacement transducer may
be simplified from equations (32) and (34) as
'•'•K ( laP (35)
V
and the dynamics of the transducer may be neglected.
4.3 Fluidic Servovalve
The characteristic performance of the fluidic servovalve has been
discussed in section 3.2. In the application of the fluidic servovalve in
51
the position servo system, the output flow/pressure characteristics must
be designed to meet the particular ram and load requirements. The actuator
and load described as part of the position servo by Lee1 3 has been used in
this study so that a step response between the fluidic position servo and
a conventional electrohydraulic position servo may be compared directly.
From the characteristics of the fluidic sei-vovalve and the para-
meters of the actuator and load, the dimensionless group
"G d
A 2
r
and the time constant,
m < I secondd -
-may be calculated
where
G '(S) -G (s)/G (S)qp q p
d - damping coefficient
A r area of ram,rm - mass of load.
As a result of the high pressure gain and small load mass and friction, the
load dynamics for this particular system can be neglected and the dynamic
flow gain of the servovalve is the dominant valve performance parameter.
The experimentally determined flow gain from section 3.4 is
QL (s) G e vsG (s) L - se (36)
q Pid(s) (I + T q S)
where
G - steady state servovalve flow gain,qs
T v - pure time delay of the servovalve,v
T - first order time constant.q
4.4 Closed-Loop Fluidic Position Servo
With the dynamic characteristic performance of the fluidic summer,
fluidic feedback transducer and fluidic servovalve predicted in sections
52
L...... . •-
4.1, 4.2 and 4.3 respectively, the fluidic position servo can be represented
"*: by the block diagram shown in figure 25. The open loop transfer function of
pomition servo may be expressed as
Ke-TOGH(s) - (l+Tqs) (37)
where K KO a GS....K t 55 8 ,
K ,Ar
T " F + TV a
The dynamic performance of the closed-loop position servo may be expressed
in terms of two dimensionless parameters, namely, the normalized gain KT,
and the normalized characteristic time constant T/Tq and may be analyzed by
Root Locus analysis. Normalized gain for zero damping, which indicates the
limit of closed-loop stability, and normalized gain for critical damping,
which corresponds to a step response with no overshoot, are of particular
interest. KT corresponding to • - 0 and • - 1 are plotted as a function of
nT/q in figure 25. For simplicity and as a first-order guide in selecting
the combination of KT and T/Tq, the damping ratio & and normalized natural
frequency, wnT are plotted against the normalized KT with T/T as an q
parameter in figure 26.
4.5 Implementation
A fluidic position servo has been constructed and tested. The
construction is shown in figure 27. A flapper-nozzle valve with an elec-
trical torque motor has been used as a fluidic signal generator.
The gain block, used in the servo is a modified form of the gain
block described in section 3 in which the steady state flow gain has been
increased and the transport time delay has been reduced with no change in
blocked-load pressure gain at the design temperature of 25*C. The increase
of the no-load flow gain provides improved servo frequency response and
has been achieved by increasing the number of sections in parallel in the
final stage and by operating the servovalve at a higher supply pressure
whereas the decrease of the pure time delay is achieved by using LPA's with
smaller nozzle throat width (x S/b = 8) for first and second stages. Thesp5
53
%8
I--
I..K 0
�
s-Il ��Yl I-'
M�0 0
AI U
1wHU
Sr.b.c
(.4 H
-. * HU H
i- b4
U
U Hp1: Uo
* H
5' t
w
+ * H
0
IIH
fl-h
3 * C..'0
.5. 3
H0 H
0
H
o
A' * 0 * 0
H ,-4 o o o o
'.5 54
Pld s (1+T S)
1.6-TS 1 2
e TS-- +-'-(Ts)
1.4 1 12 T
n i+ 2
KtS1.2
/
,0.0
" •1.0 •
r 0.8, 5
". 0.6
0.4 , . .
0- .5,1.0,2.0
0 0.2 0.4 0.6 0.8 1.0 1.2Normalized Gain, Kt
omFigure 26. and W nT of fluidic position servo.
55
,,4
.. A
56
increase in first and second stage gain, resulting from the higher aspect
ratio, is designed to offset the additional pressure drop. Hence, K hasp
not been varied significantly. The same input and feedback resistors as
described in section 3.2 have been used to construct the servovalve.
The fluidic servo components are summarized in table 4. The
components have been tested individually for both static and dynamic per-
formance. The data are presented in figures 28 through 33 for the fluidic
summer, displacement transducer and servovalve respectively. As shown in
table 5, in which the essential servo component parameters are summarized,
the servovalve flow gain has been increased by a factor of 2.2 and the pure
time delay has been successfully reduced from 1.1 ms to 0.65 ms in compari-
:2 son to the value of section 3. The displacement transducer exhibits a
linear characteristic throughout the entire stroke tested in a blocked load
condition and significantly solves the mechanical-to-fluidic interface
problems encountered in the previous investigations. 1 3
The fluidic position servo response has been calculated based on
L responses to step inputs. The experimental data, in figure 34 show that the
fluidic position servo design with 5 percent overshoot exhibits performance
comparable to the commercial electrohydraulic position servo and a sig-
nificant improvement over the hydraulic position servo described by Lee
and Wormley.1 3
Figure 35 compares the experimental and analytical step responses
for the fluidic position servo. The preliminary analytical pure time delay
based on the sum of LPA transport time lags in both the fluid in summer and
servovalve is observed to be smaller than the measured time delay of the
servo. Since the dynamic responses of the components have been measured
individiaully, the additional time delay may be attributed to the inter-
connections between the components of the fluidic position servo.
5. SUMMARY AND CONCLUSIONS
The characteristic performance of HDL fluidic integrated components
essential for servovalve design have been evaluated as a function of supply
1 3D. Lee and D.N. Wormley, Multistage Hydraulic Summing and Signal
Processing Amplifiers and Fluidic Input Servovalve Development, HarryDiamond Laboratories, HDL-CR-76-233-1 (1976).
57
TABLE 4 FLUIDIC POSITION SERVO COMPONENT CONFIGURATION
FLUIDIC SUMMIER '
LPA design 81505 o = 1.666 b = 0.375 mms
= 90, G 10.5NR p,LPA
R 1.45x10 N-s/m5 350 psi/cisas
Input resistance, R1 6.863x10 9 N-s/m5 16.5 psi/cis
MECHANICAL-FLUIDIC DISPLACEMENT TRANSDUCER
10 5Variable resistance, R 2.O3xl0 N-s/m5 49 psi/cisFixed resistance, Rf 1.491x10 N-s/mr 36 psi/cis
FLUIDIC SERVOVALVE*
Amplifier
Stage LPA a No. of sections
1 x1505 1.000 5
2 x61505 0.667 4
3 x61505 0.333 12
Pressure Reducer
"" Stage Nozzle No. of sections
1 5221A-2( 6
2 5221A-20 8
3 direct supply
*
Supply pressure, P= 10.343 kPa (1500 psi)
58
- - -~. . . . . . . .. - . .A
TABLE 5 VALUES OF PARAMETERS OF FLUIDIC POSITIONSERVO
ACTUATOR
Area, A 432 mm2 0.67 in 2r
Volume (single side) V 8.19xi03 ,m 0.5 in 3
Oil Bulk Modulus, i 1.38x106 kPa 2xlO5 psi
LOAD
Mass, m 16.65 kg 0.095 lb-s 2 /in
Damping constant, d 29.6 N-s/m 0.169 lb-s/in
FLUIDIC SUMMER
Steady state gain, K 5Pure time delay, T 0.35 m
s
MECHANICAL-FLUIDIC DISPLACEMENT TRANSDUCER
Transducersensitivity, Kt 11 kPa/mm 40.5 psi/in
FLUIDIC SERVOVALVE
Steady state flow -9 5gain, G 1.207x0- m /N-s 0.5 cis/psiqs
Pure time delay, T 0.65 msv
1st order timeconstant, 5.3 msSq
59
P sd
lda
Flui-lic Suwmmer Schematic
%I.
P d Supply condition
P P - 448.2 kPa [65 psi]
as -6 3=s 4.64x10 m /sec [0.29 cis]
T = 29.4*C [85*F]
P 2d +0.0287P s
0 P 2d -0.0287
0.5 1.152.5 3
a I'd x10 2
P. ass
Figure 28. Fluidic summer schematic and static characteristics.
60
. ..~ - 0--. 4- L
0
0 10
0
000
*0 0 0
in 0
0
0
00
0O 0
09 0
up l~ub ans~ld BZT QUIO sagbou UT aieq
610
Temperature 29C [846F 0.5 dt
Supply pressure 345 kPa st 0-5
Leakage flow 1.632x10 -5 /sec[1.02 cis],'.- ::0.4
x 12.7 mn1-2 [ inch]
0.3
0.2
0.1
1.0 -0.75 -0.5 -0.25 0.25 0.5 0.75
X,•-, x
"-0.1 m
Sensitivity - 10.86 kPa/um
0[40 psi/inch]
.. , -0.2
h.°"." -0.3
-0.4
Figure 30. Displacement transducer blocked-load staticcharacteristics.
62
%',1
•°
4J4
V4
0 Li
-' 0'00
p-I
00
N.,
00
00
- 0 in Y9-nLLn4
RP Ue OX.4" tiU 84
:4'63
QLK0.4---Ps 10342 kPa s
[1500 psi]
Q s xi0 cis-m3/sec 0.3
T 29.4C[85F]
0.2
0
0.1
-3 -2 - 1 2 3
p id 3x 1o
-0.1 5Gqs 1.207x10 N-S.
(0.5 cis/psi]
S-0.2
-0.3
-0.4
Figure 32. Fluidic servovalve no-load staticcharacteristics.
64
00
00
*0
00
r- LUP 'TW5 %OI PBZTV=N 9Qfi~ UT gw1
7".
653
0.1 1.
20 insec
% 1.0
0.8 Fluidic position servo0.8 0. mm
I ~ IElectrohydraulic position servo,
8
'-4 .6 *(Lee and Wormley)0.6a
x'; 0.32 mm
IComercial fluidic-input servo'U (Lee and Wormley)
Oorýxm =0.304 mm0.4
V0.
0-20 40 60 - I0 1'20Tire in msec
Figure 34. coIuparison of step response between fluidicand commercial position servo.
66
"r.. I I
0 E
411
.941
i4 C
44 41
0) IH
,ftU
ft., H I'
bVA
* 0'4
C4tf OD% U
148 H
67
- - . f ft ft - - .-_-Af
pressure and temperature which are characterized in terms of the modified
Reynolds number. The point of transition-to-turbulence of three standard
LPA configurations namely, HDL 63020, HDL 72010 and HDL 61505 occurs between
modified Reynolds numbers of 100 and 120. The useful operating range of
LPA's has been determined through the experimental program to be
40 ' < 100.
The relationship between the supply conditions of the individual
stages and that of the final stage of the gain block has been derived and
verified experimentally. Compensated fluidic gain blocks and servovalves
are sensitive to temperature variation at constant supply pressure. The
temperature compensation technique, based on the supply pressure scheduling
to maintain an approximately constant final stage modified Reynolds number,
significantly suppresses the temperature sensitivity of the blocked-load
pressure gain of the gain block and servovalve.
The fluidic gain block, summing amplifier and feedback transducer
have been used with an actuator and load mass to construct a closed loop
position control systems. Static and dynamic tests of the servosystem have
shown its performance comparable to an electrohydraulic servoloop. This
development effort has demonstrated the capability to develop high per-
formance position servo components from standard integrated component
fluidic elements and to interconnect the components into a closed loop
servo with performance comparable to high performance electrohydraulic
commercial components.
In the current fluidic servo, the maximum load pressure differen-
tial and flow gain of the servovalve are primarily limited by the LPA
characteristics. Future effort is merited to optimize the LPA design to
achieve high pressure recovery, to reduce the ratio of quiescent to maximum
load flow and to further minimize the overall power-to-weight ratio.
68
- .j •. 2
APPENDIX A. -- GAIN BLOCK AND SERVOVALVE CONSTRUCTION DETAILSIn this appendix the gain block stacking order is sunmmarized as given in
tables A-i and A-2 with the components shown in figure A-i.
TABLE A-1 GAIN BLOCK STACKING ORDER DESCRIPTIONS
Stacking HDL HDL
Order Part No. Orientation Quantity Description
ValveBaseplate
1 5018 C (*)2 5047 G3 5221A-20 H Stage
5040 F 6ýpairs Frstte5 5221A-20 H Reducer6 5018 H 31pairs7 5200 C8 5200 H9 5021 F
10 5216 F211 5339A F 6 sections Third Stage12 61505 F Amplifier13 5339A F and14 5216 F 2 Vent
Assembly15 5040 E16 5221A-20 7pairs Second State6 5Pressure17 5018 F 21pairsR18 5222A-20 C Reducer19 5021 C20 5046 B21 5021 D22 5216 C 223 5216 H 724 5046 D25 5018 A"26 5046 B27 504628 5239 H Second State29 5137 H Amplifier
and3 sections Vent
Assembly30' 72010 H31 5137 C32 5237 C33 5216 H34 5021 A35 5200 H36 5200 C37 5047 D
1 unless otherwise stated
69
TABLEA-(Cn.
CAIN BLOCK STACKING ORDER AND DESCRIPTIONS
Stacking HDL HDLOrder Part No. Orientation Quantity Description
38 5239 D 239 5018 H40 5216 F 241 5237 F42 5236 F43 63020 F44 5236 F45 5237 F46 5216 A First Stage47 5237 A Amplifier48 5236 A and49 63020 F Vent50 5236 F Assembly51 5237 F___________
52 5216 F256 5046 E___ ____
* Input manifold
77
70
TABLE A-2
SERVOVALVE STACKING ORDER AND DESCRIPTIONS
StackingOrder Part No. Orientation Quantity Description
Supply/output
manifold baseplate
1. SP5
2. SP3 feedbackmainfold
3. HDL 5043
4. HDL 5026 [4 pairs feedback
5. HDL 5040 resistors
6. 3 stage ampli ier Refer to amplifier stacking order (1-56)
7. SP2 summingmanifold
8. HDL 5040
9. HDL 5027 feedbackresistors
10. HDL 5112 2
11. HDL 5196 inputresistance
12. HDL 5040
13. SP4
14. HDL 5040
15. HDL 5196 inputresistor
16. HDL 5040
17. HDL 5112 2 inputmanifold
71
AA
AAoo;
--
I Section A-A
2B
SP2 Suneaing Manifold (includinq cover)
o 0
00
,0 0
S !4 Se P
Figure A-1. Servovalve components descriptions.
72
I4-
NOMENCLATURE
A ram arear
B LPA control port minimum widthc
B LPA average control port channel width
B normalized nozzle throat width, bA of ih stageni n Siste
B LPA outlet port minimum widthB LPA average outlet port channel width0
Bsp LPA splitter width
B LPA downstream control edges spacet
B LPA jet-control edge space
B normalized nozzle average channel width, b /bC empirical constant in channel resistance model
Cd discharge coefficient
diC di discharge coefficient of i th stage LPA
Ce momentum flux coefficient
G servovalve blocked-load pressure gainP
G servovalve blocked-load pressure gain at steady statepsGpLPA LPA blocked-load pressure gain
G servovalve no-load flow gain|°b% qP qG servovalve no-load flow gain at steady state.. qsG servovalve output admittance:. qPH Hi normalized nozzle height, hn /bs
601 K fluidic servo open-loop steady state gain
K ith stage LPA blocked-load pressure gain, i-12,3
K gain block blocked-load pressure gainpK K gain block no-load flow gain:• q
K fluidic summer gain
K fluidic summer steady state gain"S
Kt displacement transducer sensitivityL normalized entry length, Re/b
L channel fluid inertance•g~i C
L transducer fixed channel fluid inertancefL transducer variable channel fluid inertance
v
NR Reynolds number
N channel Reynolds numberRc
Naf gain block final stage Reynolds number,:., Rf
73
.~:.....:. ..... . . ...:-: ,...- .•." " ........ -.._. .,_ . . - . . .-..-.---
t'-th
N gain block i stage Reynolds numberN'
Rit LPA modified Reynolds number
N, gain block final stage modified Reynolds numberRf th
NRi gain block i stage modified Reynolds number
P control pressure differentialcd
P saturation control pressure differential* cds
A P bias control pressureco
P output pressure differential of transducer- .- :dt
- id input pressure differential
Pd maximum input pressure differential
"" PL defined in figure 1
Pod output pressure differential
- ods saturation output pressure differentialP main supply pressure
Psd defined in figure 28S.' supply pressure of ih stage, I 1 1,2,3
PsiP! ass supply pressure of sumer
Pst supply prescure of transducer
""P ld input 1 of summr
P 2d input 2 of summer
:. Q volumetric flow rate
load flow
QLs saturation load flow
Qflow through ith stage nozzle
Qs supply flow
Qsi supply flow of ith stage LPA
R channel resistance
Snormalized resistance defined in equation (8)
""R gain block input deflection resistanceaRa LPA input resistance for summeras
Rf fixed resistance in transducerRfpR fp servovalve feedback resistance
Ri servovalve input resistance
Ril 1 st stage input resistanceR 2 nd stage input resistance
i2 rd3R 3 stage input resistance
7.4
"= ~74
•. •2
R nozzle resistance of ith stagenisR stat
Rol 1 state output resistance
R 2 ndstage output resistance02 r3rd
R 3 stage output resistance03 tR supply resistance of ith stagesi
R variable resistance in transducerV
,R1 summer input resistance
T temperature
T reference temperature
X channel length
X LPA control port channel lengthc
X normalized nozzle channel length of stage
X 0LPA outlet port channel length
X LPA supply nozzle-splitter distanje
X th LPA supply nozzle throat length
Z gain block input complex impedancea
Z servovalve input complex impedance
Zfp servovalve feedback complex impedance
Lower Case Letter
b channel width
b average channel width
b b LPA supply nozzle throat width
bsf final state LPA supply nozzle throat widthith
b h stage LPA supply nozzle throat width
d damping coefficient
f frequency
h channel height
meentry lengthe
1 mass
mi number of LPA in parallel in ith stage
number of nozzles in parallel in ith stage
S Laplace transform operator
x valve input in section 1
channel length in section 2
displacement in section 4
x half stroke, maximum travelm
75
x dimensional LPA supply nozzle-splitter distance
Greek Letter
OL aservovalve parameter, defined in equation (18)
bulk modulus, table 5Y y servovalve parameter, defined in equation (19)
- ~ A viscosity-temperature coefficient
x defined in equation (24)0
P. absolute viscosity of oilS•V kinematic viscosity of oil
.V• kinematic viscosity at reference temperature
damping ratio
P density
a aspect ratio
a• channel aspect ratio, h/b
a c aspect ratio of ith stage
ST total time delay of fluidic servo
T defined in equation (34)T 1t order time constant of G q()qqT time delay of sumera
T ' time delay of servovalve
w natural frequencyn
Subscripts
tr. transition
900 900 phase shift
L.
"76
.*2",- '-'.-, ---- S - - -t - - - - - --- .,.-
"DISTRIBUTION
ADMINISTRATOR ENGINEERING SOCIETIES LIBRARYDEFENSE TECHNICAL INFORMATION CENTER ATT1 ACQUISITIONS DEPARTMENTATTN DTIC-DDA (12 COPIES) 345 EAST 47TH STREETCAMERON STATION, BUILDING 5 NED YORK, NY 10017ALEXANDRIA, VA 22314
COMMANDER IDDR&ECOMMANDER PENTAGON, ROOM 3D 1089US ARMY RSCH a STD GP (EUR) ATTN G. KOPCSAKATTN CHIEF, PHYSICS & MATH BRANCH WASHINGTON, DC 20310FPO NEW YORK 09510COMMDER OFFICE OF THE DEPUTY CHIEF OF STAFF FORRESEARCH, DEVELOPMENT & ACQUISITIONUS AR1MY MATRIEL VOP T D TENT OF THE ARMY
READINESS COMMAND ATTN DANA-ARP-pATTN DRCLDC, JAMES BENDER ATTN DANA-CSS-N5001 EISENHOWER AVENUE WASHINGTON, DC 20310ALEXANDRIA, VA 22333
US ARMY R&D GROUP (EUROPE)COMMANDER BOX 15US ARMY ARMAMENT MATERIEL ATTN CHIEF, AERONAUTICS BRANCHREADINESS COMMAND ATTN CHIEF, ENGINEERING SCIENCESATTN DRSAR-ASF, FUZE & FPO NEW YORK 09510
MUNITIONS SUPPORT DIVATTN DRSAR-RDF, SYS DEV DIV-FUZES US ARM RESEARCH OFFICEATTN DRSAR-RDG-T, R. SPENCER PO BOX 12211ATTN DRSAR-ASF ATTH R. SINGLETONATTN DRSAR-LEP-L, TECH LIBRARY RESEARCH TRIANGLE PARK, NC 27709ROCK ISLAND, IL 61299
BM ADVANCED TECHNOLOGY CENTECOMMANDER PO BOX 1500US ARMY MISSILE & MUNITIONS ATTN J. PAPADOPOULWS
CENTER & SCHOOL HUNTSVILLE, AL 35807ATTN ATSK-CTD--FREDSTONE ARSENAL, AL 35809 COMMANDER
DIRECTORUS ,•%R IY FOREIGN SCIENCEDIRECTOR a TECHNOLOGY CENTERUS ARMY MATERIEL SYSTEM4S FEDERAL OFFICE BUILDINGANALYSIS ACTIVITY ATTN DRXST-SD1ATTN DRXSY-MP ATTN DRXST-IS3, C. R. MOOREABERDEEN PROVING GROUND, MD 21005 220 7TH STREET, NE
CHARLOTTESVILLE, VA 22901DIRECTORUS ARMY BALLISTIC RESEARCH LABORATORY DIRECTORATTN DRDAR-TSB-S (STINFO) APPLIED TECHNOLOGY LABORATORYABERDEEN PROVING GROUND, MD 21005 ATTN GEORGE W. FOSDICK, DAVDL-ATL-ASA
FT EUSTIS, VA 23604US ARMY ELECTRONICS TECHNOLOGY& DEVICES LABORATORY COMMANDER
ATTN DELET-DD US ARMY MATERIEL & MECHANICSFT MONMOUTH, NJ 07703 RESEARCH CENTERATTN R. KATZHQ, USAF/SAMI WATERTOWN, MA 02172
WASHINGTON, DC 20330
COMMANDERTELEDYNE BROWN ENGINEERING US ARMY MISSILE COMMANDCUMMINGS RESEARCH PARK ATTN REDSTONE SCIENTIFIC INFORMATIONATTN MELVIN L. PRICE, MS-44 CENTER, DRSMI-RBDHUNTSVILLE, AL 35807 ATTN DRSMI-RG, WILLIAM GRIFFITHATTN DRSMI-TGC, J. C. DUNAWAY
ATTN DRCPM-TOE, FRED J. CHEPLENREDSTONE ARSENAL, AL 35898
77
hyJ
DISTRIBUT10N (Cont 'd)
CONMANDER COKKRIDERUS ARMY MO0BILITY EQUIPMENT R&D CENTER NAVAL AIR DEVELOPMNIT CENTER (Cante4)ATTN TECHNICAL LIBRARY (VAULT) ATTN CODE 81 34, LOIS GUISEATTN DRD**-**44 Re N. WARE ATTN D. KEYSER, 60134
FT BELVOIR, VA 22060 WARNINSTER, PA 18974
COMMANDER C ING OFFICERUS ARMY ARRADMBCI NAVAL AIR ENGINEERING CUSTIEATTI SARPA-TS-S #59 ATTN ESSD, CODE 9314, HAROLD GUI TATTM DRDAR-LCN-C, A. E. SCHRIDLIN LAKEHURST, NY 08733ATTN DRDAR-LCC-E, J. CONNORSATTN DRDAR-SCF-IC, V. BAUNBARTH NAVAL AIR SYNSTERIS CMCNT
ATTN PBM-DPI4 (TAGLAIRINO) DEPARTMENT OF THE NAVYDOVER, NJ 07801 ATTN CODE AIR-5162C•, J. MNS
ATTN CODE AIR-5143J, D. HOUCKCOMMANDER WASHINGTON, DC 20361WATERVLIET ARSENALATTN SARWV-RDT-L COMNANDERATTN DRDAR-LCB-RA, R. RACICOT PACIFIC MISSILE TEST CUSTER
WATERVLIET ARSENAL, NY 12189 ATTH CODE 3123, AB 3. GARRUWATTN CODE 1243, A. ANDERSON
COMMANDER POINT MUGU, CA 93042US ARMY TANK AUTOMOTIVE COMMANDARMOR & COOP DIV, DRDTA-RKT COMMANDERBLDG 215 NAVAL SHIP ENGINEERING CESTERATTN K. WHITMORE PHILADELPHIA DIVISIONWARREN, MI 48090 ATTN CODE 6772
PHILADELPHIA, PA 19112COMMANDERATTN STEWS-AD-L, TECHNICAL LIBRARY COMMANDERWHITE SANDS MISSILE RANGE, NM 88002 NAVAL SURFACE WEAPONS CUSTER
ATTH CODE 413, CLAYTO! ICKISURACOMMANDER/DIRECTOR WHITE OAK, MD 20910ATMOSPHERIC SCIENCES LABORATORYUSA ERADC0M COMMANDERATTN DELAS-AS (HOLT) NAVAL ORDNANCE STATIONATTN DELAS-AS-T (R. RUBIO) ATTN CODE 51 23C, K. ENGLAMDERWHITE- SANDS MISSILE RANGE, NM 88002 INDIANHEAD, MO 20640
OFFICE OF NAVAL RESEARCH NAVAL SHIP RES & DEV CENTERDEPARTMENT OF THE NAVY CODE 1619, K. READERATTN STANLEY W. DOROFF, CODE 438 BETHESDA, MD 20084ATTI D. S. SIEGEL, CODE 211
4 ARLINGTON, VA 22217 NAVAL RESEARCH LABORATORYATTN S. SEARLES, 117 BG A68
DEPARTMENT OF THE NAVY WASHINGTON, DC 20375R&D PLANS DIVISIONROOM 5D760, PENTAGON NAVAL SEA SYSTEMS COMMASATTN BENJ R. PETRIE, JR. SA05R31
OP-987P4 ATTN J. H. HARRISONWASHINGTON, DC 20350 WASHINGTON, DC 20362
COMMANDANT CO0MMNDERUS NAVAL POSTGRADUATE SCHOOL NAVAL WEAPONS CUNTERDEPARTMENT OF MECHANICAL ENGINEERING ATTN CODE 533, LIBRARY DIVISIONATTN CODE 69 Nn(NUNN) ATTN CODER 3636, C. SURMEISTEtMONTEREY, CA 93940 CHINA LAKE, CA 93555
COMMANDERNAVAL AIR DEVELOPMENT CENTERATTN R. MCGIBONEY, 60134
78
4'
.. . . ... . . . . . .
DISTRIBUTION (Cont 'd)
COMMANDER DEPARTMENT OF COMMERCEAF AERO PROPULSION LABORATORY, AFSC NATIONAL BUREAU OF STANDARDSATTN LESTER SMALL, AFWAL/POTC ATTN JAMES SCHOOLEY, CHIEF,WRIGHT-PATTERSON AFB, OH 45433 TEMPERATURE SECTION
ATTN T. NEGAS, SOLID STATECOMMANDER CHEMISTRY DIVISIONAIR FORCE AVIONICS LABORATORY ATTN RAY DILS, RN B-254, BLDG 221ATTN AARA-2, RICHARD JACOBS ATTN GEORGE BURNS, R4 B-222, BLDG 221WRIGHT-PATTERSON AFB, OH 45433 WASHINGTON, DC 20230
DIRECTOR DEPARTMENT OF COMMERCEAF OFFICE OF SCIENTIFIC RESEARCH BUREAU OF EAST-WEST TRADEATTN NE OFFICE OF EXPORT ADMINISTRATIONBOLLING AFB, DC 20332 ATTN WALTER Jo RUSNACK
WASHINGTON, DC 20230CONNANDERAIR FORCE FLIGHT DYNAMICS LABORATORY DEPARTMENT OF ENERGYATTN AFWAL/FIGL, H. SNOWBALL C-156, GTN (OART)ATTN AFWAL/FIER, R. J. DOBBEK ATTN ROBERT ROBERTSWRIGHT-PATTERSON AFB, OH 45433 ATTN SANDY DAPKUNAS
WASHINGTON, DC 20545COMMANDERAF WEAPONS LABORATORY, AFSC DEPARTMENT OF ENERGYATTN SUL, TECHNICAL LIBRARY FE-22KIRTLAND AFB, NM 87117 ATTN T. K. LAU
WASHINGTON, DC 20545COMMANDERARMAMENT DEVELOPMENT & TEST CENTER DEPARTMENT OF ENERGYATTN ADTC (DLOSL), TECH LIBRARY F-317, GTN (COAL GASIFICATION)ATTN DLMA, DAVID T. WILLIAMS ATTN JIM CARREGLIN AIR FORCE BASE, FL 32542 WASHINGTON, DC 20545
AIR FORCE FLIGHT TEST CENTER FEDERAL BUREAU OF INVESTIGATION6510 ABG/SSD J. EDGAR HOOVER BLDGATTN TECHNICAL LIBRARY ATTN ROBERT WILLISEDWARDS AFS, CA 93523 WASHINGTON, DC 20535
AF INSTITUTE OF TECHNOLOGY, AU DEPARTMENT OF JUSTICEATTN LIBRARY AFIT (LD), IMMIGRATION & NATURALIZATION SERVICE
BLDG 640, AREA B 425 "I- STREET, NNATTN AFIT (ENN), MILTON E. FRANKE ATTN NEILL ICKAYWRIGHT-PATTERSON APB, OH 45433 WASHINGTON, DC 20536
HQ, AF SYSTEMS COMMAND SCIENTIFIC LIBRARYATTN SGB, CPT GEORGE JAMES US PATENT OFFICEANDREWS AFB, DC 20334 ATTN MRS. CURETON
WASHINGTON, DC 20231ARGONNE NATIONAL LABORATORYAPPLIED PHYSICS DIV, BLDG 316 NASA AMES RESEARCH CENTERATTN N. M. O'FALLAN ATTN HS 244-13, DEAW CHISEL9700 S. CASS AVE MOFFETT FIELD, CA 94035ARGONNE, IL 60439
NASA LANGLEY RESEARCH CENTEROAK RIDGE NATIONAL LABORATORY ATTN, MS 494, H. D. GARNERCENTRAL RES LIBRARY, BLDG 450ON, RM 175 ATTN MS 494, R. R. HELLBAUKPo BOX X ATTN MS 185, TECHNICAL LIBRARYATTN E. HOWARD HAMPTON, VA 23665ATTN C. A. MOSSMANATTN R. E. HARPER"OAK RIDGE, TN 37830
79
; r . ,¥. . .c°.. .. ru ".--.., -'-rr- r- *; ,i-•- m .r•r-,-u. -'r ... r ,- -- r-v r•. r r|--r i u-
DISTRIBUTION (Cont'd)
NASA SCIENTIFIC & TECH INFO FACILITY FRANKLIN INSTITUTE OF THE STATEPO BOX 8657 OF PENNSYLVANIAATTN ACQUISITIONS BRANCH ATTN KA-CHEUNG TSUI, ELEC OCR DIVBALTIMORE/WASHINGTON INTERNATIONAL ATTN C. A. BELSTERLINGAIRPORT, 1D 21240 20TH STREET & PARKWAY
PHILADELPHIA, PA 19103UNIVERSITY OF ALABAMACIVIL & MINERAL ENGINEERING DEPT HUGHES HELICOPTERS"P0 BOX 1468 DIVISION OF SUMMA CORPORATION
""t ATTN HAROLD Re HENRY CENTINELA & TEALS STREETSUNIVERSITY, AL 35486 ATTN LIBRARY 2/T2124
CULVER CITY, CA 90230
TECHNOLOGY CAMPUS lIT RESEARCH INSTITUTEPO BOX 3017 ATTN K. E. MCKEEATTN PAUL C. MCLEOD 10 WEST 35TH STREETLITTLE ROCK, AR 72203 CHICAGO, IL 60616
UNIVERSITY OF ARKANSAS JET PROPULSION LABORATORYMECHANICAL ENGINEERING ATTN JOHN V. WALSH, #S 125-138ATTN JACK H. COLE, ASSOC. PROF. 4800 OAK GROVE DRIVEFAYETTEVILLE, AR 72701 PASADENA, CA 91103
CARNEGIE-MELLON UNIVERSITY JOHNS HOPKINS UNIVERSITYSCHENLEY PARK APPLIED PHYSICS LABORATORIESATTN PROF. W. T. ROULEAU, ATTN MAYNARD HILL
NECH ENGR DEPT ATTN THOMAS RANKINPITTSBURGH, PA 15213 ATTN JOSEPH WALL
LAUREL, IC 20810
CASE WESTERN RESERVE UNIVERSITY
ATTN PROF. P. A. ORNER LEHIGH UNIVERSITYATTN PROF. B. HORTON DEPARTMENT OF MECHANICAL ENGINESRINGUNIVERSITY CIRCLE ATTN PROF. FORBES T. BROWNCLEVELAND, OH 44106 BETHLEHEM, PA 18015
THE CITY COLLEGE OF THE CITY LINDA HALL LIBRARYUNIVERSITY OF NY ATTN DOCUMENTS DIVISION
DEPT OF #ECH ENGR 5109 CHERRY STREETATTN PROF. L. JIJI KANSAS CITY, I4 64110ATTN PROF. G. LOWEN1 39TH ST. AT CONVENT AVE LOS ALAMOS SCIENTIFIC LABNEW YORK, NY 10031 PO BOX 1663
ATTN FRANK FINCH, MS 178
CLEVELAND STATE UNIVERSITY LOS ALAMOS, NM 87545PENN COLLEGE OF ENGINEERINGATTN PROF. R. COMPARIN MASSACHUSETTS INSTITUTE OF TECHNOLOGYCLEVELAND, OH 44115 ATTN ENGINEERING TECHNICAL REPORTS,
R14 10-408DUKE UNIVERSITY ATTN DAVID WORMELY, MECH ZNGR DEPT,COLLEGE OF ENGINEERING RM 3-146ATTN C. H. HARMAN 77 MASSACHUSETTS AVENUEDURHAM, NC 27706 CAMBRIDGE, MA 02139
ENGINEERING SOCIETIES LIBRARY MICHIGAN TECHNOLOGICAL UNIVERSITYATTN HOWARD GORDON LIBRARY, DOCUMENTS DIVISIONATTN ACQUISITIONS DEPARTMENT ATTN J. HAWTHORNE345 EAST 47TH STREET HOUGHTON, MI 49931NEW YORK, NY 10017
iii
80
N--• ''':...- '",.• • -'-",'._ - -• . _-% ." ... - _ L . • . . • . . - - _,- " ' ."- " " - " " - _ . - • _/ ,• ." _ .•, . .•. . : . • .-' .V-V'2'~--- --.. ....
- -- " .r'-' rr. •-W:-r- -. l r... r4 • .;, -CC rF- - -U -.-.-- . -_ ,. q--r r • .r~ - .-
* 4.
DISTRIBUTION (Cont' d)
UNIVERSITY OF MISSISSIPPI PENNSYLVANIA STATE UNIVERSITYATTM JOHN A. FOX ENGINEERING LIBRARY201 CARRIER HALL, DEPT OF UCKH ENGR ATTN N. BENNETT, ENGINEERING LIBRARIANUNIVERSITY, MS 38677 201 HAMMON BLDG
UNIVERSITY PARK, PA 16802MISSISSIPPI STATE UNIVERSITY
*-.,DRAWS ME PORTLAND STATI UNIVERSITYATTH C. J. BELL, UCH KING DEPT DEPT OF EDGINEERING &STATE COLLEGE, MS 39762 APPLIED SCDICE
PO BOX 751MISSISSIPPI STATE UNIVERSITY ATTN PROF. P. I. CHEIDEPT OF AEROSPACE ENGINEERING PORTLAND, OR 97207ATTN DAVID IURPHREEMISSISSIPPI STATE, MS 39762 PURDUE UNIVERSITY
4 SCHOOL OF MECHANICAL ENGINEERINGUNIVERSITY OF NEBRASKA LIBRARIES ATTH PROF. VICTOR W. GOLDSCH14IDTACQUISITIONS DEPT, SERIALS SECTIONS ATTH PROF. ALAN T. MCDONALDATTm ALAN GOULD LAFAYETTE, IN 47907LINCOLN, NE 68508
ROCK VALLEY COLLEGEUNIVERSITY OF NEW HAMPSHIRE ATTN KEN BARTONMUCH ENGR DEFT, KINGSBURY HALL 3301 N. MULFORD ROADATTH PROF. CHARLES TAFT ROCKFORD, IL 61101ATTN PROF. DAVID LIMBERTDURHAM, NH 03824 RUTGERS UNIVERSITY
LIBRARY OF SCIENCE & NICINRUNIVERSITY OF N * CAROLINA ATTN GOVERNMENT DOCUMENTS DEPTINSTITUTE OF MARINE BIOMEDICAL RESEARCH SANDRA R. LIVINGSTONATTM MICHAEL E. SHEEHAN NW BRUNSWICK, Wx 08903
• WILMINGTON, IC 28401SYRACUSE UNIVERSITY
NO JERSEY INSTITUTE OF TECHNOLOGY DEPT OF RMD & AEROSPACE ENGINEERIMGDEPARTUIDT OF MECHANICAL DIGINEERING ATTN PROF. Do S* DOSANJHAT"N R. Y. CHER 139 E. A. LINK HALL323 HIGH STREET SYRACUSE, NY 13210NEWARK, NJ 07102
UNIVERSITY OF TUINESSEEOHIO STATE UNIVERSITY LIBRARIES DEPT OF MECHANICAL ENGINEERINGSERIAL DIVISION, MAIN LIBRARY ATTN PROF. G. V. SMITH
46 1858 HEIL AVENUE KNOXVILLE, IN 37916COLUMBUS, OH 43210
UNIVERSITY OF TUNNESSEE SPACE INSTOKLAHOMA STATE UNIVERSITY ENERGY CONVERSION DIVISIONSCHOOL OF RMDC A AEROSPACE ENGR ATTN MARY AMl SCOTTATTH PROF. KARL N. REID TULLAHOMA, TN 37388STILLWATER, OK 74074
UNIVERSITY OF TEXAS AT AUSTIN
MIAMI UNIVERSITY DEPT OF MECHANICAL DIGINEERINGDEPT OF ENG TECH ATTN A. J. HEALEYSCHOOL OF APPLIE SCIDNCE AUSTIN, TX 78712
ATMN PROF. S. B. FRIEDMANOXFORD, ON 45056 THE UNIVLRSITY OF TEXAS AT ARLINGTON
MECHANICAL ENGINEERING DEPARTMENTPENNSYLVANIA STATE UNIVERSITY ATTN ROBERT L. WOODSATTN J. L, SHEARER ARLINGTON, TX 76019215 MECHANICAL DNGINEERING BUILDINGUNVERSITY PARK, PA 16802 TULANE UNIVERSITY
DEPT OF MECHANICAL ENGINEERING
ATTN H. F. HRUBECKYNW ORLEANS, LA 70118
81
--------------------. 4
DISTRIBUTION (Cont'd)
UNION COLLEGE BENDIX CORPORATIONMECHANICAL ENGINEERING ELECTRODYNAMICS DIVISIONATTN ASSOC. PROF. W. C. AUBREY ATTN D. COOPER
MECH MR DEPT. STEINMETZ HALL 11600 SHERMAN WAYSCHENECTADY, NY 12308 NM HOLLYWOOD, CA 90605
UNIVERSITY OF VIRGINIA BENDIX CORPORATIONDEPT OF 1ECH & AEROSPACE ENGR RESEARCH LABORATORIES DIVATTN DAVID LEWIS BENDIX CENTERCHARLOTTESVILLE, VA 22090 ATTN Co J. AIIERN
ATTN LARL TAPLINVIRGINIA POLYTECHNIC INSTITUTE SOUTHFIELD, NI 4875
OF STATE UNIV
MECHANICAL ENGINEERING DEPARTMENT BOEING COMPANY, THEATTN PROF. H. MOSES PO BOX 3707BLACKSBURG, VA 24061 ATTN KIEIRIK STRAUB
SEATTLE, WA 98124WASHINGTON UNIVERSITYSCHOOL OF ENGINEERING BOWLES FLUIDICS CORPORATIONPO A ox 1185 ATTN VICE PRUS/ZNGRATTN NW M. SWANSON 9347 FRASER AVENUEST LOUIS. NO 63130 SILVER SPRING, MD 20910
WEST VIRGINIA UNIVERSITY R.e . BOWLESMECHANICAL ENGINEERING DEPARTMENT 2105 SONDRA COURTATTN RICHARD A. DAJURA SILVER SPRING, MD 20904
NORGANTOWN. WV 26505CHAMBERLAIN MANUFACTURING CORP
WICHITA STATE UNIVERSITY EAST 4TH A ESTHER STSATTN DEPT AERO ENGR, E. J. RODGERS PO BOX 2545WICHITA, KS 67208 WATERLOO, IA 50705
UNIVERSITY OF WISCONSIN CONTINENTAL CAN COMPANYMECHANICAL ENGINEERING DEPARTMENT TECH CENTERATTN FEDERAL REPORTS CENTER ATTN P. A. BAUERATTN KORMAN H. DEACHLEY, DIR 1350 W. 76TH STRUT
DESIGN ENGINEERING LABORATORIES CHICAGO, IL 606201513 UNIVERSITY AVENUEMADISON, WI 53706 CONTROL SYSTEMS INNOVATION
ATTN N. F. MKCIAWORCESTER POLYTECHNIC INSTITUTE 517 EAST ORION STREETATTN GEORGE C. (GORDON LIBRARY (TR) TEMPE, AZ 85283ATTN TECHNICAL REPORTSWORCESTER, MA 01609 CORDIS CORPORATION
PO BOX 428AVCO SYSTEMS DIVISION ATTN STEPHEN F. VADAS, K-2ATTN W. K. CLARK MIAMI, FL 33137201 LOMLL STREET
WILMINGTON, MA 01887 CORNING GLASS WORKSFLUIDIC PRODUCTS
BARNES EIGINEERING CO ATTN R. H. BELLMANATTN FRED SWEIBAUN HOUGHTON PARK, B-230 COMMERCE ROAD CORNING, MY 14830STAMFORD, CT 06904
CHRYSLER CORPORATIONLHELICOPTER COMPANY PO BOX 118
PO BOX 482 CINS-418-33-22ATTN R. D. YEARY ATTN L. GAUFT WORTH, TX 76;01 DETROIT, NI 48231
"82
.• i, ,•; .' .W ,U ,-•. S,. W.,, • • -, , .• , .. . 7 V . , , ., - .. V-, . , 7, -. . - .% . - . . . ,. . ,'* - -.W
'47
DISTelr 'bUTION (Cont'd)
JOHI DEERE PRODUCT ENGINEERING CENTER GRUMMAN AEROSPACE CORPORATION (Cont'd)ATTN V. S. KUMAR ATTN JACK LEONARD, MS B1535WATERLOO, IA 50704 SOUTH OYSTER BAY ROAD
BETHPAGE, L. I., NY 11714ELECT'hIC POWER RESEAI,}H INSTITUTE:0 B)X 10412 HAMILTON STANDARD
ATTN MS. M. ANGWIN, DIVISION OF UNITED AIRCRAFT CORPORATIONP. M. GEOTHERMAL ENERGY ATTh PHILIP BARNES
3412 ;ILJVIEW AVE WINDSOR LOCKS, CT 06096PALO AL'1D, CA 94303
HONEYWELL, INCFLUIDICS ýUARTERLY ATTN J. HEDEENPO BOX 2989 ATTN W. POSINGIESATTN D. H8 TARU9OTO 1625 ZARTHAN AVESTANFORD, CA 94305 MINNEAPOLIS, MN 55413
FOXBORO COMPANY HONEYWELL, INCCORPORATE ATTN RICHARD STEWART, MS 200RESEARCH DIV 1100 VIRGINIA DRIVEATTN JAMES VIGNOS FT WASHING-lTON, PA 19034ATTN J. DECARLOATTN JOHN CHANG JOHNSON CONTROLS, INCATI74 TOM KEGEL ATTN WARREN A. LEDERMAN38 NE'ONSET AVE ATTN GEORGE JANUFOXBCRO, MA 02035 507 E. MICHIGAN
MILWAUKEE, WI 53201GARRETT PNEUMATIC SYSTEMS DIVISIONPO BOX 5217 LEEDS & NORTHRUP COATTN GLRY FREDERICK ATTN ERNEST VAN VALKENBURGATTN TREVOR SUTTON DICKERSON ROADATTN TOM TIPPETTS NORTH WALES, PA 19454ATTN C. APBOTT
111 SOUTH 34TH STREET MOORE PRODUCTS COMPANYPHOENIX, AZ 85010 ATTN R. ADAMS
SPRING HOUSE, PA 19477GENERAL ELECTRIC COMPANYSPACE/RES DIVISIONS MARTIN MARIETTA CORPORATIONPOBOX 8555 AEROSPACE DIVISIONATTN MGR LYIRARIES, LARRY CHASEN ATTN R. K. BRODERSON, MP 326PHILADELPHIA, PA 19101 PO BOX 5837
ORLANDO, FL 32805GENERAI ELECTRIC LOMPANYKNOLLS ATfOMIC POWER LAEORATORY MCDONNELL AIRCRAFT COMPANYATTN J. KROMMENIOEK GUIDANCE & CONTROL MECHANICS DIVISIONSCHENECTADI, 61. 12301 ATTN LOYAL GUENTHER
ST LOUIS, MO 63166'GENERAL .0TORS CORPORATIONDELCO ELECTRONICS DIV MODONNEL DOUGLAS ASTRONAUTICS COMANFRED C WRIGHT PROPULSION DEPARTMENTNEW COMMERCIAL PROnUCTS ATTN V. E, HALOULAXOS (t3-226)PO BOX 1104 ATTN J. D. SCHWEIKLE (A3-226)
ATTN R. E. SPARKS 530. BOLSA AVENUEKOKOM-,, IN 46901 HUNTINGTON BEACH, CA 9264?
CRUMMAN AEROSPACE CORPORATION NATIONAL FLUID POWER ASSOC.TECHNICAL INFORMATION CENTER ATTN JOHN R LUEKEATTN C. W. TURN'R, DOCUMENTS DIR 0.' TECH SERVICES
LIBRARIAN 3333 NORTH MAYFAIR ROADATTN TED SORENSEN, MS B1535 MILWtJKEE, WI 53222
83
DISTRIBUTION (Cont'd)
NEOS, INC SCIENCE & TECHNOLOGY ASSOCIATES, INC3711 AIR PARK RD ATTN DR. T. DRZEWIECKIATTN A. J. OSTDIEK 1700 N. MOORE ST., SUITE 1920LINCOLN, NE 68524 ARLINGTON, VA 22209
NORTHRUP CORP, ELECTRONICS DIV SIKORSKY AIRCRAFTATTN DESMOND NELSON, ATTN J. R. SOEHNLEIN
SENOIR ENGINEER NORTH MAIN STREETORGN C3133, W/C STRATFORD, CT 06602
2301 W. 120TH STHAWTHORNE, CA 90250 STEIN ENGINEERING SERVICES, INC
5602 E MONTEROSAPATSCENTER INTERNATIONAL PHOENIX, AX 85018ATflN MR. JOHN CLINE
707 ALEXANDER ROAD TRANS-TECH, INCPRINCETON, NJ 08540 ATTN L. DOMINGUES
12 MEEI AVEPLESSEY AEROSPACE LTD GAITHERSBURG, MD 20760ATTN A. ROSENBERG1700 OLD MEADOW ROAD TRITEC, INCMCLEAN, VA 22102 ATTN L. SIERACKI (2 COPIES)
PO BOX 56PROCON, INC COLUMBIA, MD 21045ATTN HERB MARCHOUP PLAZA UNITED TECHNOLOGIES RESEARCH CENTERDES PLAINES, IL 60016 ATTN R. E. OLSON, AGR FLUID
DYNAMICS LABORATORYPROPULSION DYNAMICS 400 MAIN STREETATTN T. HOULIHAN E. HARTFORD, CT 061082200 SOMERVILLE RANNAPOLIS, MD 21401 VOUGHT CORP
PO BOX 225907ROCKWELL INTERNATIONAL CORPORATION ATTN KELLEY FLINGCOLUMBUS AIRCRAFT DIVISION, PO BOX 1259 DALLAS, TX 75265ATTN MARVIN SCHWEIGERATTN LOUIS BIAFORE US ARMY ELECTRONICS RESEARCH4300 E. 5TH AVENUE & DEVELOPMENT COMMANDCOLUMBUS, OH 43216 ATTN TECHNICAL DIRECTOR, DRDEL-CT
ATTN PUBLIC AFFAIRS OFFICE, DRDEL-INSANDIA LABORATORIESATTN WILLIAM R. LEUENBERGER, DIV 2323 HARRY DIAMOND LABORATORIESATTN JERRY HOOD ATTN CO/TD/TSO/DIVISION DIRECTORSATTN NED KELTNER ATTN RECORD COPY, 81200ATTN ANTHONY VENERUSO, DIV 4742 ATTN HDL LIBRARY , 81100 (3 COPIES)ALBUQUERQUE, NM 87185 ATTN HDL LIBRARY (WOODBRIDGE), 81100
ATTN TECHNICAL REPORTS BRANCH, 81300SCIENCE APPLICATIONS, INC. ATTN LEGAL OFFICE, 97000ATTN J. ISEMAN ATTN CHAIRMAN, EDITORIAL COMMITTEE8400 WESTPARK DR ATTN CORRIGAN, J., 20240MCLEAN, VA 22102 ATTN CHIEF, 13000
ATTN CHIEF, 13400 (20 COPIES)
S84a8
9 : . - ' ' - ..- 5- I - -'-. - - - - - - - -
Recommended