NASA Technical Reports Server (NTRS) 20110011412: Crew Quarters (CQ) and Electromagnetic Interference (EMI) Measurement Facility Combined Impedance Study

0.66 MB·English

by  NASA Technical Reports Server (NTRS)

Abstract for Crew Quarters (CQ) and Electromagnetic Interference (EMI) Measurement Facility
Combined Impedance Study
This report documents an investigation into observed failures associated with conducted
susceptibility testing of Crew Quarters (CQ) hardware in the Johnson Space Center (JSC)
Electromagnetic Interference (EMI) Measurement Facility, and the work accomplished to
identify the source of the observed behavior. Investigation led to the conclusion that the
hardware power input impedance was interacting with the facility power impedance leading to
instability at the observed frequencies of susceptibility. Testing performed in other facilities did
not show this same behavior, pointing back to the EMI Measurement Facility power as the
potential root cause. A LISN emulating the Station power bus impedance was inserted into the
power circuit, and the susceptibility was eliminated from the measurements.
Crew Quarters (CQ) and Electromagnetic
Interference (EMI) Measurement Facility
Combined Impedance Study
Robert C. Scully
NASA Johnson Space Center
Engineering Directorate
2101 NASA Parkway
Houston, Texas 77058-3696

Abstract— This report documents an investigation into observed designed to emulate an output impedance very similar to that of
failures associated with conducted susceptibility testing of Crew the Station power bus impedance, eliminated the observed
Quarters (CQ) hardware in the Johnson Space Center (JSC) failure conditions. Following this testing, it was postulated that
Electromagnetic Interference (EMI) Measurement Facility, and the cause of the original failure was related to the combination
the work accomplished to identify the source of the observed of the EMI Facility power supply output impedance and the
behavior. Investigation led to the conclusion that the hardware CQ power input circuitry input impedance. An analysis was
power input impedance was interacting with the facility power
performed on this basis, and the results of that analysis are
impedance leading to instability at the observed frequencies of
contained in this report.
susceptibility. Testing performed in other facilities did not show
this same behavior, pointing back to the EMI Measurement II. PREPARATION FOR TROUBLESHOOTING
Facility power as the potential root cause. A LISN emulating the
Station power bus impedance was inserted into the power circuit, A. Preliminary Considerations
and the susceptibility was eliminated from the measurements. Pursuant to the decision to engage in troubleshooting of the
initially observed anomaly, an Engineering Evaluation test
Keywords-conducted susceptibility; LISN; instability
procedure was discussed and written by engineering
representatives from the cognizant design organization, the test
I. INTRODUCTION
support contractor, and engineering representatives from
This report documents an investigation into observed
NASA JSC Avionic Systems Division. The test was designed
failures associated with conducted susceptibility testing of
first to replicate the successful CS01 test conducted in Building
Crew Quarters (CQ) hardware in the Johnson Space Center
49, followed by a repeat of the original CS01 testing performed
(JSC) Electromagnetic Interference (EMI) Measurement
in Building 14. The primary difference in configuration
Facility. In brief, during conducted susceptibility testing [1] in
between these two setups was the use of standalone power
August 2008, over the frequency range from 30 Hz to 50 kHz
supplies in Building 49, and the use of the laboratory power
as described in test method CS01 of International Space Station
supply in Building 14. Following the repeat of the two initial
requirements documents SSP 30237 and SSP 30238, hardware
setups, troubleshooting would proceed with investigation into
susceptible behavior was observed between 880 Hz and 1 kHz.
faulty grounding, excessive supply voltage drop, excessive
Subsequently, the same testing was repeated in the JSC
applied ripple voltage, and reduced current limit setting for the
Acoustic Test Facility [2], with no observed failures. Testing
laboratory power supply. In the event that all of the foregoing
was repeated in the Acoustics Facility because the hardware
was unsuccessful at replicating the observed anomaly, the
was already removed from the EMI Facility, and logistics
engineering representatives involved would then determine the
considerations precluded relocation of the hardware. Moreover,
next steps, if any, either to continue with additional unplanned
it was desirable to repeat the testing in a different facility, in
testing, or terminate the testing and declare an unexplained
order to determine if the problem was localized to the EMI
anomaly. In all cases, the intent of the testing was to replicate
Facility. The testing was repeated a second time for credit in
the originally observed anomaly.
the ATF, and the hardware certification was enabled to move
forward on the recorded positive results. Discussions among a B. Test Configurations and Sequencing
team comprising membership from the cognizant design The engineering evaluation test was planned to move in
contractor, NASA civil servant staff, and EMI Measurement sequential order from one configuration to the next, based on
Facility support contractor, determined that additional whether the original anomaly was reproduced or not. For each
engineering evaluation testing [3] in the EMI Facility was configuration, a CS01 test would be performed from 0.1 kHz to
warranted. This testing was able to replicate the originally 1.5 kHz at a standard ripple voltage level of 5 VRMS,
observed failure conditions. Substitution of a Line Impedance maximum. In the event the original anomaly was captured, the
Stabilization Network (LISN) as specified in SSP 30238, test would move to a more freeform investigative phase in
whiich the team wwould be at liberty to troubleeshoot the systeem IV. DISCUSSION OF POWERR SYSTEM STAABILITY
to ffocus in on thee possible anommaly root causse. Table 4.1 liists
Powwer system stabbility has beenn a concern for International
the planned configgurations.
Space Station (ISS) ppower systemss since the inception of the
The requiremeent for this testting was taken from SSP 302337, Programm. A typical ccause of systeem instability iis a negative
andd is stated as follows: “5 VVRMS or 10%% of the suppply load immpedance. As long as a neegative load iimpedance is
volltage, whichevver is less, ffrrom 30 Hz tto 2 kHz, thhen powereed by a voltage source that exhibits a veryy low output
deccreasing log-linearly to 1 VVRMS or 1%% of the suppply impedaance across freequency, the system will rremain stable
volltage, whicheveer is less. Thee requirement iis also met whhen throughhout its operaational enveloppe. In cases iin which the
the audio source specified in SSP 30238 adjuusted to dissipaate negativve load impedaance is poweredd by a source wwhose output
50 watts in a 0.5 oohm load, cannnot develop thee required voltaage impedaance amplitudde is larger than the neegative load
at tthe equipment under test (EUUT) power input terminals, aand impedaance at one or mmore frequenciies, the system may become
the EUT is not susceptible to thee output of the signal source.””. unstablle at those freqquencies. Givenn this situation, and the fact
that ISSS systems wouuld be provideed by multiple vendors and
TABLLE I supplieers, it was deterrmined early onn to adopt the cconcepts first
PLANNED CONFFIGURATIONS FFOR CQ CS01 ENGINEERINGG popularrized by Middlebrook [4], annd this adaptattion was well

EVALUATIOON TESTING describbed for large dcc power systemms by Gholdstoon, et al. [5].
As disscussed in [4]], situations mmay arise wherein one is
Coonfiguration Description
confronnted by a “blacck box” powerr converter sysstem, and the
1 RReplicate Bldg 49 Setup Using Two Stand Alone Powwer
SSupplies need too analyze the bbehavior of the combination oof that “black
box” annd any possiblee filtering that might be introoduced on the
2 RReplicate Bldg 14 Setup Using One Stand Alone Poweer
SSupply and the EMMI Facility Power Supply power input. This describes very nnicely the situaation with the
combinnation of the CCQ hardware ppower supply iinput and the
3 PPower Supplies Saame as #2; Remove Electrical Bond
CConnection Betweeen the CQ Electriccal Panel and the CCQ EMI FFacility power supply outpuut combinationn. Given this
CChassis similariity, techniquess as described iin [4] and [5] were used to
4 PPower Supplies Saame as #2; Replacee Electrical Bond examinne the charactteristics of thhe CQ and EEMI Facility
CConnection Betweeen the CQ Electriccal Panel and the CCQ combinned power bus iimpedance.
CChassis, Insert a 6 Ohm Series Resisstance to Simulate
IInput Voltage Drop Thee techniques uused were first derived in [[4] for small
5 PPower Supplies Saame as #2; Remove Series Resistor, signal considerationss of combinattions of source and load
RRun CS01 With Ripple Voltage Settting Increased to 66 impedaances, in partiicular the interaction between a dc-dc
VVRMS convertter and its inpput EMI filter. From fundammental control
6 PPower Supplies Saame as #2; Reducee the Current Limitt theory, the transfer fuunction for the system shown in Fig. 1 can
SSetting on the EMI Facility Power SSupply, Run CS01 be writtten as
wwith Standard Rippple Voltage Settinng of 5VRMS Leveel
F FF

F = S L (1)
SL ⎛Z ⎞
1+⎜ SS ⎟
IIII. HARDWARRE INVESTIGATIION ⎝ Z ⎠
L
The anomaly was observedd during the teesting employiing
where F = the source transfer fuunction
connfiguration 5. TThe test entereed the troublesshooting phasee at S
F = the loaad transfer funcction
thiss point. The firrst step was to insert a 10 μFF capacitor at tthe L
F = the syystem transfer ffuunction
outtput of the EMMI Facility powwer supply. Hoowever, it provved SL
nott possible to opperate the signnal source at 6 VRMS with tthe ZS = the source output immpedance
cappacitor in placee, and in the aattempt, the siignal source wwas ZL = the loaad input impeddance
dammaged. The cappacitor was remmoved, and a nnew signal sourrce
wass substituted innto the circuit. The test wass repeated at 8800 Thee ratio of source to load impeddance can be coonsidered as
Hz,, and the anommaly was againn observed. A sstandalone powwer the loopp gain for the integrated systeem. The systemm loop gain is
suppply was substituted in placce of the EMI Facility powwer used to determine the stability of thee system.
suppply at this poinnt, the test repeeated, and the anomaly was nnot
obsserved at 700 HHz, 900 Hz, or 1000 Hz. The sstandalone powwer
suppply was remooved from the circuit, and replaced with tthe
EMMI Facility poower supply. A Line Impeedance Stabilizzer
Nettwork (LISN) was inserted in series in botth the power aand
retuurn leads betwween the CQ and the EMII Facility powwer
suppply. The test wwas repeated at 700 Hz, 800 Hz, 900 Hz, aand
10000 Hz, and thhe anomaly waas not observeed. The test wwas
repeated a secondd time with thee signal source set to maximuum
outtput voltage caapability. At thhis point, a fulll standard CSS01 Figure 1. Impeddance and Transferr Function Relationnship
testt was run on both the powwer and returnn leads, with no Between A Source and Looad Subsystem.
anoomaly observedd.
If |ZZ | < |Z | for alll frequencies, thhen the systemm is stable. If
S L
|Z | > |ZZ |, further anallysis is needed to determine ssystem
S L
stability. The Nyquist criterion can then be applied to the loop crossover frequencies, as discussed previously, and a
gain to determine system stability. corresponding gain and phase margin exist at each point.
Stability requirements may be expressed in terms of gain
and phase margin. To establish gain and phase margin of the
transfer function, TM = ZS/ZL, the first step is to establish for Load Impedance Magnitude |ZL|
which frequencies, if any, |Z (s)/Z (s)| = 1. This condition is
S L
represented on a Bode plot by an intersection of the magnitude
with the 0 dB line, or on a Nyquist plot by an intersection of
the plot with the unit circle. The frequencies at which these
crossings occur are known as crossover frequencies, fc. Once
Source Impedance Magnitude |Z|
the fc are determined, phase margins at the fc are determined S
by identifying the phase angle of the transfer function at each
Figure 4. Source and Load Impedance Crossover.
fc and adding 180 degrees. Gain margin at each fc is then

defined as the difference between the magnitude of T and the
M
0 dB line on a Bode plot. On a Nyquist plot, gain margin is Bus impedance peaking may occur at such points of
identified as the distance between the intersection of TM with overlap. The bus impedance, ZBUS, is defined as the parallel
the negative real axis. Fig. 2 illustrates gain and phase margins combination of Z and Z .
S L
as shown on a Nyquist plot.
⎛ ⎞
⎜ ⎟
Z 1 (2)
Z =Z ||Z = S =Z ⎜ ⎟
BUS S L ⎛Z ⎞ S⎜ ⎛Z ⎞⎟
1+⎜⎝ SZL⎟⎠ ⎜⎝1+⎜⎝ SZL⎟⎠⎟⎠
Peaking, a manifestation of bus resonance, occurs at the fc,
and is a function of the phase margin. When the phase margin
is small, the factor on the right hand of equation (2) is less than
1, and peaking will occur. Conversely, if the phase margin is
large, the factor is greater than 1, and no peaking occurs. This
effect is illustrated in Figs. 5 and 6.

Figure 2. Nyquist Plot Gain and Phase Margins.
According to the Nyquist stability criterion, small-signal
system stability is simply determined by inspection of the
Nyquist plot for the system under study. If the Z /Z circles the
S L
point (-1,0) in the s-plane, shown as the “X” in the left-hand
Figure 5. Small Phase Margin Figure 6. Large Phase Margin
side of Fig. 3, the system is prone to be unstable. Otherwise, Bus Z Peaking. Bus Z Peaking.
the system is unconditionally stable. The right-hand side of Fig.
3 shows the equivalent Bode plot representation of the Nyquist
plot on the left-hand side. V. ANALYSIS RESULTS
Data collected from all CS01 testing strongly suggested the
EMI Facility power supply was entering into a current limit
state. In particular, measurements and observations made
during the engineering evaluation testing illustrated this effect,
even though the current limit of the supply was set above the
expected current maximum, the maximum being controlled by
means of an inline fuse. This behavior in turn suggested the
combination of the Crew Quarters and the EMI Facility power
supply bus impedances was such that bus peaking was
occurring between 880 Hz and 1 kHz, such as described in the
Figure 3. Impedance Criteria.
preceding theory. In order to pursue this train of thought,

references [6], Summary of the EMI Lab Power Supply Output
An important aspect of this theory is the realization that Impedance Test, and [7], Quick Look Data Package Revised
when the load impedance magnitude exceeds that of the source Crew Quarters (CQ) Electrical Power Quality Test, were
impedance, a negative load impedance condition exists. This is obtained, as well as the impedance – frequency relationship of
illustrated in Fig. 4, wherein the load and source impedances the Space Station LISN [8], and the individual and combined
can be seen to overlap at two different frequencies. These are bus impedances were determined as a function of frequency.
The resulting plots of these curves are shown in Figs. 7 and 8.
Whhat the first curvve illustrates iss the comparisoon of the Crew VI. CONCLUSION
Quarters input immppedance and thhe EMI Facilityy power supplyy
Bassed on the foreggoing analysis,, coupled with tthe
outtput impedancee, and the combbined bus impedance of the twwo.
informaation containedd in the precediing theoretical discussion, it
It ccan be seen thatt between approximately 1 kHHz and 2 kHz,
is conclluded the obserrved anomalouus behavior of tthe Crew
the EMI Facility ppower supply ooutput impedannce is well abovve
Quarterrs hardware in the EMI Faciliity was in fact tthe result of
the Crew Quarterss input impedance, a region oof negative loadd
the inteeraction of the CCrew Quarters input impedannce and the
imppedance for this combination.. Indeed, as theeory would
EMI Faacility output immpedance. Subbstitution of thee Space
predict, just beloww 1 kHz, the coombined bus immpedance showws
Station LISN, designeed to emulate thhe impedance oof the Space
a mmarked peakingg, or resonant, rresponse. This rresonance is
Station bus in orbit, elliminated the aanomalous behaavior
almmost certainly thhe cause for thee EMI Facility power supply
entirelyy, and together with the resultts from the Buiilding 49
currrent limit respoonse during thee Crew Quarterrs CS01 testingg.
testing,, further demonnstrated the Creew Quarters haardware is in
Thee second curve illustrates the Crew Quarterss input
compliaance with its CCS01 EMI requuirement.
imppedance and thhe output impeddance of the Sppace Station
LISSN that was insserted in-line wwith the EMI Faacility power REFERENCES
suppply output durring the engineeering evaluatioon testing. It is

cleaar from this currve that for all frequencies thee LISN
[1] D. M. Tran, “EVV5-07-EMC-020R - EMI Test RReport for the
imppedance is welll below the Creew Quarters inpput impedance,
Cerrtification of the International Spaace Station (ISS) Crew Quarters
andd the resulting ccombined bus iimpedance doees not exhibit anny (CQQ)”, Johnson Spacce Center, Houstonn, Texas, 10 Octobber 2008.
peaaking behavior..
[2] H. Ngo, “EV5-08-EMC-021R - Ameendment to EV5-007-EMC-020R -
EMMI Test Report forr the Certification of the Internationnal Space Station

(ISS) Crew Quarters (CQ)”, Johnson SSpace Center, Houuston, Texas, 29
Octtober 2008.
[3] Tassk Performance Shheet (TPS) EF082XXXX, “Engineerinng Evaluation of
the CQ Rack for CS01 EMI Testing”,, Johnson Space CCenter, Houston,
Texxas, 3 November 22008.
[4] R.DD. Middlebrook, “Input Filter Considerations iin Design and
Appplication of Swwitching Regulatoors”, in Proc. IEEE Industry
Appplications Society Annual Meeting, Chicago, 1976.
[5] E.WW. Gholdston, K. Karimi, F.C. Leee, J. Rajagopalann, Y. Panov, B.
Maanners, “NASA TTechnical Memoraandum 107281-Staability of Large
DCC Power Systems Using Switching Converters, Withh Application to
the International Sppace Station”, preesented at the 331st Intersociety
Eneergy Conversion EEngineering Confeerence, Washington, DC, Aug. 11-
16, 1996, Paper IECEEC 96-96079.
[6] J. CC. Neeley, Lockhheed Martin Memorandum, “Summmary of the EMI
Labb Power Supply Output Impedancce Test”, Johnsonn Space Center,
Figure 7. Crew Quaarters Input Impedaance, EMI Facility Output Impedancee, Houuston, Texas, 7 Seeptember 1999.
and CCombined Bus Imppedance, without LLISN [7] Eneergy Systems Test Branch Report EESTA-DP-7X019QQ, “Quick Look
Datta Package Revissed Crew Quarterrs (CQ) Electricall power Quality

Tesst”, Johnson Spacee Center, Houston,, Texas, 21 Octobeer 2008.
[8] Figgure 3.2.1.3.2-1 LISN for CE07 MMeasurements, “SSSP 30238, Space
Staation Electromagnnetic Techniques, Revision E”, Johnson Space
Cennter, Houston, Texxas, 31 July 2002.

Figgure 8. Crew Quarters Input Impedaance, EMI Facilityy Output Impedancce,
and Combined Bus Immpedance, with LISN