NASA Technical Reports Server (NTRS) 19910020953: Test and evaluation of load converter topologies used in the Space Station Freedom Power Management and distribution DC test bed

10 Pages·0.59 MB·English

by  NASA Technical Reports Server (NTRS)

_. vlr- --
_-7 2 -I
NASA Technical Memorandum 105217 _
.--_
Test _d Evaluation of Load Converter
Topologies Used in the Space Station
Freedom Power M_agemem and
DistributionDC Test Bed
_ _ Ramon C_ Lebron and Angela C. Oliver
Lewis Research Center
Cleveland, Ohio
- - and
RobertE-Bodi ......
Analytical Engineering Corporation
NorthOlmsted,Ohio _ ___.__.........._....:...._...._._..:......._ .= -:_:--;-.-.-
Prepared for the
26th Intersociety EnergyC0nvergion Engineering_Conference _ '- .........
cosponsored by the ANS, SAE, ACS/_, ASME, IEEE, and AIChE
Boston, Massachusetts, August 4'9, 1991 -
_ (NASA-TM-I05217) T_ST ANO EVALUATI°N OF N91-302c_7
LnAO CO_!VmRTER TOPOLOGIPS USED IN TNE SPACE
=
STATION FREFCqM POWER MANAGEMENT AND
.......... 0 1STR Ii_U T T3,N _C TEST _D (NASA) 9 p Unclds
CSCL _IH G3120 0037961
NASA
:;41.

TEST AND EVALUATION OF LOAD CONVERTER TOPOLOGIES USED INTHE SPACE STATION FREEDOM POWER
MANAGEMENT AND DISTRIBUTION DC TEST BED
Ramon C.Lebron Angela C. Oliver Robert F. Bodi
National Aeronautics and SpaceAdministration Analytical Engineering Corp
Lewis Research Center Space Station Engineering and Integration Contractor
Cleveland, Ohio 44135 NorthOlmsted, Ohio 44070
ABSTRACT Unit and Battery Charge/Discharge Unit) to convert and
regulate solar array or battery voltage to 160Vdc. The test
The Power Management and Distribution (PMAD) bed secondary subsystem utilizes two dc to dc converters
DC Test Bed at the National Aeronautics and Slface which provide power to secondary and tertiary power
Administration Lewis Research Center inCleveland, Ohio, distribution units, and three load convener units. The dcto
is aunique facility fortesting power components hardware dc converter unit provides regulation by converting the
in support of the Space Station Freedom dc Electrical primary voltage to 120V dc for secondary distribution.
Power System (EPS). One type of breadboard hardware Finally, dc load converter units provide the last stage of
tested at this facility is the dc Load Convener Unit, which regulation by convening the secondary voltage to 28 Vdc.
constitutes the power interface between theelectric power The final regulating unit is required on the space station
system and the actual load. These units are dc to dc because a majority of user loads require voltage levels
conveners that provide the final system regulation before different from the 120 Vdc secondary bus voltage.
power is delivered to the load.
General specifications were given forthedevelopment
Three load convener topologies have been tested inthe
of load converters, however, some specifications were
test bed and inthe Power Electronics Laboratory These are
listed as goals so that the developer had the flexibility to
a series resonant convener and a series inductor
include features deemed appropriate for this type of
switchmode converter, developed by TRW,and aswitching
application. The following is a partial list of the
full-bridge forward convener developed byV_stignhouse.
specifications for the load converters: Input _ltage 120
This paper describes, in general, the topologg operation
Vdc nominal (+/- 18 Vdc); Output 'Voltage 28 Vdc(+/-
principles, and tests results. A comparative analysis of the
1%); Output Power 1kV_, Input Current Ripple +/- 1%
three units is given with respect to etficiency, regulation,
peak topeak; Output Voltage Ripple +/- 1%peak topeak;
short circuit behavior (protection), and transient
90% Efficiency at full load (design goal);Output Short
characteristics. A fourth topology, the TRW zero voltage
Circuit current limiting; and Input Under/Over voltage turn
switching converter, will be tested in the near future and
off.
test results compared with information presented in this
paper.
TOPOLOGIES DESCRHrrION AND TEST
INTRODUCTION PERFORMANCE
The NASA Lewis Research Center is responsible for
Series Resonant Topology
the development of the Space Station Freedom (SSF)
Electrical Power System. In order to identify system level
Each converter topology is made up of three stages:
issues during the SSF Program design and development
input filte_ power stage, and output filter The series
phases, asystems test bed was assembled. Some of the
resonant design [3]is shown in Figure 1.It consists of four
objectives of this test bed facility are the evaluation of
power MOSFET switches. QI and Q2 are driven at
system efficiency, power quality, system protection and
100kHz, 50% duty cycle and are 180 degrees out of phase.
reconfiguration, and system stability.
Transistors Q3 and Q4 arepulse width modulated, with Q4
pulsed while Q1 is on, and Q3 pulsed while Q2 is on.
The test bed power system consists of a 160V dc Output voltage control is achieved by varying the pulse
primary distribution system which is converted to a 120V width on Q3 and Q4 to control the current through the
dc for secondary distribution to the user loads [1]. In this transformer. The LC tank circuit results in a resonant
system, regulation is provided at three different locations sinusoidal current through the transformer which reduces
within the distribution system. Primary distribution switching losses. The unit control power is obtained from
regulation is provided at the sources [2] (Sequential Shunt the 120 Vdc input power source.
inputoftheloadconverter toallow theinput filtertochafe
bVoq
up while the unit was still off. The unit was then
commanded oninto a resistive load (36 A). Atthe instant
of turnon a current spike of approximately 10 A was
observed decreasing to 0 A and increasing again to a
O
constant value of approximately 9 A in 16msec. Output
voltage built up to a value of 28 volts in approximately
imsec. Output current, in a similar way, increased
smoothly toitsfull loadvalue of 36Ain 16msec. Theunit
had asimilar behavior forturn onintoa50% or100% load.
When the 120 Vdc supply contactor was closed into the
discharged input filter with the unit inthe on state (100%
V'm
load connected), approximately 488 msec elapsed before
the output voltage and currentstarted tobuild up.
3O
25
V20
O O 15
L I0
T
,I I I I
S
Figure I. Series Resonant Topology Power Stage
(2Milliseco/ndDsiv.)
This unit's operating features include output current
40 OLrI'PUCTURRENT : =
limitprotection, and input under-voltage andover-voltage
tripcapability. Itscontrol commands are ON/OFF, output __t "
30 _J
voltage setpoint, and output current limit setpoint. These A /
commands can be applied either manually or through a M "
DataInterface Unit (DIU) for operation with a controller
p 20 / . ,
and a Mil. Std. 1553B databus. Theunit, when operated S 1o
with the D1U, allows monitoring of the inputvoltage and
current, output voltage and current, input/output poweg o
current limitsetpoint, status, and temperature. ( 2Millisecon/dDsiv.)
Figure 2. Series Resonant Converter Load Step Response
Efficiency.. Ripple. and Voltage Re_lation Tests
The performance test setup consisted of a 120 Vdc
Forload dropouttests the outputloadof the converter
power supply input to the load converter , and a
was reduced from 100% to 10%(36 A to 3.6A) using a
programmable loadbankinresistive mode connected toits
switch to disconnect partof the load. The result of these
output. Steady state efficiency tests wereperformed onthe
tests demonstrates that this load converter is able to
series resonant converter, increasing the load level from
withstand alargedrop inloadwithout significant effecton
10% (100W) to fullload (1kW). The unit§ efficiency at
theoutputvoltage. Whenloadwas restored toitsfull load
full load was 86.4% and varied between 85% and 87%
value(see Figure 2), inputandoutput currentsincreased to
throughout the load range. The measured output voltage
theirnominal values, settling in 16msec.
regulation was0.3%whichmetthe 1% regulation specified
forth-t-m-let.ihputcu-_ n_e-at fullloads/asl_-_ak
Short Circuit Tests
topeak,andthefullloadoutputvoltagreipplweas 1.6%
peaktopeak. Shortcircuit testswereconducted ata50% outputload
condition (18 A) toevaluate the protective features of the
unit. After the outputcurrent had reached steady state, a
23"dnsientTests
short circuitwas applied attheload converter outputusing
Turnon, turnoff,short circuit,and loadstepwansient almife switch. As soon astheshort circuitwas applied (see
tests werealso performed ontheseries resonantconverter. Figure3), outputcurrent increased to 120 A, butdropped
Forturnontests, the 120Vdc supply was connected tothe toitscurrentlimit setpoint of41.6 Ainless than1.6msec.
The unit continued to operate indefinitely at an output
voltage sufficient to sustain an output current of 41.6 A.
Input currem increased from its initial value of
approximately 4A to avalue of 7.5 A and oscillated down
to a constant value of 1.25 A with asettling time of 1.6
msec. When the fault was removed, output voltage
increased to 28 volts in approximately 0.1 msec. Output
current decreased from 41.6A to 0 A in 81 microsec and
started to ramp back up reaching avalue of 10 A in 2.86
msec.
0
30
25
V 20
V'm
O 15
L 10 Q3 m,_ Q4
T
S 5
0 O
¥-
-5
(800 microseconds/div )
Figure 4.Series Inductor Topology Power Stage
INPUT and OUTPUT CURRENT
130
120
1lO The operating features are equivalent to the Series
100 .
Resonant Converter, allowing the user to set the output
A 90
M 80 voltage, output current limit, and turnthe unit on oroffboth
manually and via aD1U. The protection and monitoring
s 50 \ wu_pu_ capabilities are identical to the series resonant converter.
40
30
20 Efficiency. Ripple, and Voltage Re_lation Tests
10 I
,_ ,---_ __ [Inpu.t.....
0 The same tests and test set ups were mn as onthe series
-10
resonant unit, with similar results. The efficiency of this
(800microseconds/d)iv
unit at full load (lkVO was 90%, dropping to alow of 75%
at one-tenth of rated load (100W).
Figure 3.Series Resonant Converter Short Circuit Test
The output voltage regulation, 0.17%, fell within
design goals. The output currem ripple was within 1%for
Series Inductor Switchmode Topology all loads tested, except the 10% of full load case forwhich
the ripple was just under 1.1%. Input ripple stayed around
The Series Inductor switchmode device [3] comists of
one-half of one percent, well within the required range.
the same three stages common toall of these load converter
topologies. The input and output filters are similar in
Transient Tests
design to the Series Resonant Converter The power stage,
shown in Figure 4, uses asimilar switching scheme. Inthis To test the start up characteristics of this device, the
case, Ql and Q2 are driven at 100 kHz with a 50% duty input filter was charged by applying 120Vdc to the input
cycle. Moreover Q3 and Q4 arepulse width modulated, but while the unit is off. The device was then switched on and
Q3 is pulsed during the Q1 on state while Q4 is pulsed the response observed. In this case, the input current spike
during the Q2 onstate. Because this topology has adiscrete was reduced (about 5 Amps) and the output voltage
inductor instead of aresonant circuit, the current through smoothly rose to 28V in less than 70us. Under 75% load,
the isolation transformer will approach asquare wave. The the output current rose exponentially to the required value
currem through this inductor will always flow in the same (27A) in approximately 12ms. Similar results were
direction because it will act as a current source when both observed using other load values. This test was also mn
Q3 and Q4 are off. with the input filter discharged, The unit inthe onstate, and
the input breaker switched on applying 120Vdc to the
30
device. This resulted in an input current spike of about I -I
100Amps. The output voltage suddenly switched on to 25
28Vdc after about one-half second. V 20 --
O 15
L
10
T
Load drop-out tests were performed with an identical S 5
test set-up as for the series resonant unit. These tests 0
showed some oscillation in the output current and the input -5
current when the load was switched from full load to 10% (33.3 Microseconds/Div. )
load. The output voltage, howeveK stayed steady.
Likewise, the restoration of the load showed effective
INPUT and oLrrPuT CURRENT
voltage regulation and much smoother current transitions 11o
10o
(see Figure 5). In both cases, the output currents reached
90
steady state in 8ms.
A 80 I
M 70
p 6O
S .SO
30 4O
25 3O
V 20 2O
0 15 l0
L 0
10
T -10
S ' 7-| ( 33.3 Microseconds /Div.)
o
( IMilfiseconds/Div. )
Figure 6.Series Inductor Converter Short Circuit Test
4o OUTPD'r.CURRE_
35
A
25 J _
M 20 _ Switching Full Bridge Topology
P
15 /
S to / --_ The design of this unit is similar to the Series Inductor
/
5 Switchmode converter. The magnetizing inductance of the
o -transfdrmer takes the place of the discrete inductor and the
(IMilfiseconds/Div). current isswitched directly through thetransformer with no
other components in series with the FETs. This device has
Figure 5.Seres Inductor Converter Load Step Response the simplest power stage design of the three topologies [4]
(see Figure 7). The switching frequency of the unit is
10kHz, onetenth the value of the other units. This unit also
contains a FET operated in the linear mode in series with
the input filter to provide input current limiting.
Applying a short circuit across the output of the unit
(see Figure 6) produced some oscillation in the input and The switching scheme is as follows: Q1 and Q2 are
output currents and the output voltage of about 3or4 KHZ. switched with square waves 180 degrees out of phase, and
The output current spiked to about 105Amps before Q3 and 04 are pulse width moduIated. 04 ispulsed-during
reaching the current limiting value in 200us. The unit QI ON state, while Q3 is pulsed during Q2 ON state. The
continued to supply the current limit setpoint current current through the transformer approaches a square wave
indefinitely (this setpoint was manually varied from 40 to as the PWM approaches 50%. The output voltage is
controlled by varying the pulse widths.
52Amps in multiple runs). The inputcurrent spiked toless
than 10Amps.
The unit has the capability to "soft start" by ramping
the output voltage to the desired setting in an adjustable
Recovery by removing the fault showed the same 30-800us turn-on period. The unit_ output voltage is
oscillating frequency, taking 610us to settle to the normal adjustable via a Mil. Std.1553B data bus as is the output
output voltage. The output current oscillated before current limit setpoint. The undervoltage input trip can also
reaching steady state in 610ns. he disabled, if desired.
4
By varying the slew rate of the slow start-up function,
Vo
) O the unit was set to reach asteady-state output voltage with
avariety of settling times.
The load drop--out tests, performed by dropping the
load from full rated to about 10% of rated, showed only a
O
slight dip in the output voltage for both the drop out and
recovery. For drop-out the output current dropped in two
steps with aslight rise between them, settling out in about
350us. The input current smoothly dropped down in about
550us. For load restoration step (see Figure 8), output
current increased to its nominal value (36 A) in 16msec,
after abrief oscillation with peak value of 20 Amps.
V'm
OUTPUT VOLTAGE
30
J
25
V2o
O 15
O L IO
T 5
S o
-5 i,,
(2.0 Milliseconds /Div. )
Figure 7. Switching Full Bridge Pow_ Stage
Efficiency.. Ripple. and Voltage Regulation "lF.slt OUTPUT CURRENT
4O
With atesting set-up similar to those used forthe other
units, this load converter unit was found to have an
efficiency in the mid 90% range (93%-97%). The output A 30 1
M ,,,,I
voltage regulation was accurate to 0.7% over afull range
p 20 f
of loads. The output current ripple was well within the 1% /-
S /
design goal, and the input current ripple was just atthe 1% 10
threshold.
(2.0Milliseconds /Div. )
2taageat..Tu_
Turning theunit on with afully chaged input filter into
avariety of resistive loads showed smooth input and output Figure 8.Switching Full Bridge Conv. Load Step Response
current and output voltage response, with no oscillation.
This is due to the ramping capability that this unit provides.
The output voltage rose to rated value in approximately
700ms. The short circuit test data for the switching full bridge
unit isshown inFigure 9. Atthe time theshort was applied,
The same test without the input filter pre--energized an output current spike of magnitude over 170Amps was
also showed alack of any major oscillation in the output (to observed, quickly dropping down to 50Amps within 4ms.
a resolution of about 50us), and in this case the output During current limiting state, output current oscillated
voltage rose to the rated value in less than 10 ms, with a between 52 and 58 amps. At shut-down, the output current
maximum input current spike of 150 Amps. and voltage quickly dropped to zero.
converter, where part power efficiency is critical and the
3O
unitremains on orbit. The other two units also have high
25 efficiency at full load and approach the efficiency of the
V20 switching full bridge unit. This type of efficiency
O 15 characteristicalong withthere lowerrelative weight, males
L
T 10 ....... these units useful for a power supply load converter
S 5 application. Forthisapplication the power supply andload
o are carried toorbit and changed asneeded The maximum
--5 efficiency point can be chosen tocorrespond tothe applied
(2.0_seconds / Div.) load.
andOUTPUT
A
m"'m .....•.... m-
M \
\ -11
P \ 0.9
S O.ip., E
• /-'x F
F 0.8
I
V l_m_t
I ! I C
I I 1 I 0.7
E • -- 3_.rtes lntl]ctor
( 2.0Milliseconds/Div.) N
• - S-_riesRetonant
Figure 9. Switching Pull Bridge Cnvrtr. Short Circuit Test Oy 0.6 • - s Mtchingj FullBrilge
0.5
0 0.20 0.40 0.60 0.80 1.20
FRACTION of FULL LOAD
COMPARATIVE ANALYSIS
Figure 10. Load Converter '_pologtes Efflclendes
Itis difficult todirectly compare the unitsbecause of
differences in design capabilities and features. The
switching full bridgeunitcontains anumber ofmonitoring
Figure 11 shows steady state ripple and regulation
and protection features not available onthe otherunits. For
results at full load condition. Input current ripple for the
example, the unitcan current limit the input and rampthe
threeunitsmet thedesign goal of 1%. Theunitwithlowest
output voltage. This ramping capability allows much
input current ripple was the series inductor converter.
smoother start-up characteristics and is a definite
Outputvoltage ripple specifications (1%)weremetonlyby
advantage inapplications where load control isrequired.
the switching full bridge unit. The output filters on the
series resonant and series inductor units are appreciably
Despite the differences infeaturesand capability ofthe smaller than the switching full bridge unit. Further
units a comparison is made in this section using testdata. additions to the output filters of the series resonant and
Thiscomparison ismade forthose parameters considered series inductorunitscan be made with minorimpactonthe
critical to the user(i.e. regulation, short circuit protection) weight and efficiency ofthe units andisexpected tobedone
andpowersystem (i.e. efficiency, transiem behavior, power tomeet the design specification.All three units fell within
quality). the required limits of output voltage regulation, the series
inductorproved tohave the best outputvoltage regulation
characteristics.
Figure10shows themeasuredefficiencies forthe three
loadconverters atdifferent loadlevels. Theefficiency data
along withweight information canbeused toselect anideal Forload steptransients (load dropoutandrestoration)
application forthese load converters ina space station like the three units behaved well and all units had an
environment. Theswitching fullbridgeunit has the highest overdamped response. The switching full bridge unit
efficiency and maintains this efficiency backto less than responded theslowest taking approximately 16millisecs to
10% of full load. This converter also has the highest reach steady state, while theothertwo unitstookabouthalf
relative weight and therefore wouldmake anideal bur load thattime.
6
In addition, there maybeotherload converterdesigns
InpuCturrentOutputVoltageOutputVoltage
being developed in the Space Station Freedom Program
Ripple(p-p) Ripple(l_-p) Regulation
which will be evaluated as they become available. A
Series
Resonant 1%r 1.6 % 0.3 % variety of bulk and individual load conveners from the
! multi-kilowatt rangedowntothe tens ofwattsareexpected
Series
Inductor 0.42 % 3.57 % 0.17 % to be available in the near future.
Switching
FullBridge 1% 0.03 % 0.7 % All of these units need to be tested in a secondary
system environment, where their interactions and
compatibility with other system components will be
Figure11.FullLoadRippleandRegulation Parameters addressed. It will be the system tests that will provide the
data necessary fora detailed comparison of load converter
All oftbe units provide shortcircuit capability toallow types. The PMAD DC Test Bed isexpected to be usedto
for the incorporation of user protective devices in provide early evaluation ofload conveners insupport ofthe
downstream loads. The unitsalso incorporated fold back SSF Program.
characteristics toconvert from constant power toconstant
REFERENCF_
resistance, providing the necessary protection to the unit
from the power system.
[1] R.Beach, L.Trash,B.Bolerjack, "Description of
the PMAD DC Test Bed Architecture and Integration
Sequence", IECEC-91, August 4--9,1991, Boston, MA.
SUMMARY
[2] R. Button and A. Brush, "Development and
All of the conveners described inthis paper are first Testing of a Source Subsystem for the PMAD DC Test
generation unitsdeveloped foruseon the PMAD DC Test Bed", IECEC-91, August 4--9, 1991, Boston, MA.
Bed at the Lewis Research Center: The component level
tests demonstrated that these units metmost of the design
[3] D. Decker and L.Inouye, "Mulfikilowatt Power
goals and requirements specified. The requirements and
Electronics Developments for Spacecraft", IECEC-91,
specifications for these units were established in early
August 4---9,1991,Boston, MA.
1989,just afterthe SSFProgramchangetoa desecondary
Furtherwork intheprogram hasestablished morecomplete
specifications for load converter units. Modifications of [4] R. Thornton and D. Fox, "Power Management
these units will be made to comply with the latest and Distribution Equipment Development for Space
specifications asthey become available. Applications", IECEC-91, August 4--9,1991, Boston, MA.
FormApproved
REPORT DOCUMENTATION PAGE OMB No. 0704-0188
Public reporting burden for this collection of information is estimated to average 1hour per respor_se, Including the _me for reviewing Instructions, searching existing data sources,
gathering and maintaining the data nee<led, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this
collection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for information Operations end Reports, 1215 Jefferson
Davis Highway, Suite 1204, Arlington, VA 222024302, and to the Office of Management and Budget, Paperwork Reductlo_ Project (0704-0188), Washington, DC 20503.
1. AGENCY USE ONLY (Leaveblank) 2. REPORT DATE 3. REPORT TYPE ANDDATES COVERED
1991 Technical Memorandum
4. TITLE ANDSUBTITLE $. FUNDING NUMBERS
Test and Evaluation of Load Converter Topologies Used in the Space Station
Freedom Power Management and Distribution DC Test Bed
W'U- 474 - 42 - 10
6. AUTHOR(S)
Ramon C. Lebron, Angela C. Oliver, and Robert F. Bodi
7. PERFORMING ORGANIZATION NAME(S) ANDADDRESS(ES) 8. PERFORMING ORGANIZATION
REPORT NUMBER
National Aeronautics and Space Administration
Lewis Research Center
E-6532
Cleveland, Ohio 44135 - 3191
9. SPONSORING/MONITORING AGENCY NAMES(S) AND ADDRESS(ES) 10. SPONSORING/MON_ORING
AGENCY REPORT NUMBER
National Aeronautics and Space Administration
Washington, D.C. 20546- 0001 NASA TM -105217
11. SUPPLEMENTARY NOTES
Prepared for the 26th lntersociety Energy Conversion Engineering Conference cosponsored by ANS, SAE, ACS, AIAA, ASME, IEEE,
and AIChE, Boston, Massachusetts, August 4-9, 1991. Ramon C. Lebron and Angela C. Oliver, NASA Lewis Research Center; Robert
F. Bodi, Analytical Engineering Corporation, North Olmsted, Ohio 44070. Responsible person, Ramon C. Lebron, (216) 433- 8325.
12a. DISTRIBUTION/AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE
Unclassified -Unlimited
Subject Categories 20 and 33
13. ABSTRACT (Maximum 200words)
The Power Management and Distribution (PMAD) DC Test Bed at the National Aeronautics and Space Administration
Lewis Research Center in Cleveland, Ohio, is aunique facility for testing power components hardware in support of
the Space Station Freedom dc Electrical Power System (EPS). One type of breadboard hardware tested at this facility
is the dc Load Converter Unit, which constitutes the power interface between the electric power system and the actual
load. These units are dc to dc converters that provide the lrmal system regulation before power is delivered to the load.
Three load converter topologies have been tested in the test bed and in the Power Electronics Laboratory. These are a
series resonant converter and a series inductor switchmode converter, developed by TRW, and a switching full-bridge
forward converter developed by Westinghouse. This paper describes, in general, the topology, operation principles,
and tests results. A comparative analysis of the three units is given with respect to efficiency, regulation, short circuit
behavior (protection), and transient characteristics. A fourth topology, the TRW zero voltage switching converter, will
be tested in the near future and test results compared with information presented in this paper.
14. SUBJECT TERMS lS. NUMBER OF PAGES
Space station power supplies; Power converters; Power electronics 8
16. PRICE CODE
A02
17. SECURITY CLASSIFICATION 18. SECURITY CLASSIFICATION 19. SECURITYCLASSIFICATION 20. LIMITATION OFABSTRACT
OF REPORT OF THIS PAGE OFABSTRACT
Unclassified Unclassified Unclassified
NSN 7540-01-280-5500 Standard Form 298 (Rev. 2-89)
Pra_.-.tlbed by ANSI Std. Z30-18
298-102