ML20217M659

Engineering Analysis Hrrm Coaxial Cable Testing

ML20217M659
Person / Time
Site:	San Onofre
Issue date:	06/04/1996
From:	Stilwagen S, Trotta K SOUTHERN CALIFORNIA EDISON CO.
To:
Shared Package
ML20217M648	List: ML20217M642 ML20217M651 ML20217M659 ML20217M664
References
NUDOCS 9708250130
Download: ML20217M659 (35)

Text

Page Iorl7 ENGINEERING ANALYSIS Illgh Range Radiation Monitoring (IIRRM)

Coaxial Cable Testing SAN ONOFRE NUCLEAR GENERATING STATION UNITS 2 & 3 t

t 4

4 6!f Tb Prepared By:

Date:

/

Ken Trotta l

M Date:

dM %

Reviewed By:

Steve StilwageV e

d 4

h825013097032o ADOCK 05000361 p

Page 2 of17 TAllLE OF CONTENTS SECTION TITl4E

.PAGC 1.0 P URP O S E.....................,..................

............ 3 2.0 IGSULTS/ CONCLUSIONS and IGCOhihfENDAT10NS................

3 3.0 ASSUhfPTIONS

............. 5 4.0 hiETHODOLOGY..

l 5.0 RE FE REN C E S................................................. 6 1

6.0 N0hiENCL ATURE............................................... 6 7.0 INVESTIGATIONS AND RESULTS.....................

7 TABLES 1

Signal Current and Equivalent Dose Rate................................... 5 FIGURES 1

Coaxial Cable Test Current Response to LOCA/hiSLB Thermal Shock........ 13 2

SONGS 2&3 Installed Configuration LOCA/htSLB Indicated Dose Rate........ 14 3

Second Test Sequence - LOCA/htSLB Temperature Curve.......

............15 4

Second Test Sequence - Chamber Penetration Design...................... 16 APPENDICES A

Preliminary Analysis of Radiation Detector Anomalies (4 pages)

B Kaman Sciences, "hfI Cable Qualification" (7 pages)

C

" Notes on Cable Testing, 11/09/95"(3 pages)

D

" Synopsis of Cable Testing, November 16/171995 at CGI Laboratories"(4 pages)

E Coaxial Cable LOCA Simulation Test Procedure and Results for hionitoring Electrical Parameters, Second Test, hfarch 25 through 30,1996 (28 pages)

F "Wyle Test Results hiarch 29,1996" (8 pages)

G

" Notes on Thermally Stimulated Depolarization Currents"(2 pages)

}{

"Second Test Sequence - Trace ofinduced Current verses Time"(117 pages)

(Not included in distribution, original sent to CDht.)

Page 3 of17 1.0 PUIU'OSE The purpose of this engineering analysis is to document the results of work performed to determine the transient effects of high temperature and pressure conditions (such as hiain Steam Line Dreak (htSLD) or Loss OfCoolant Accident (LOCA)) on the coaxial cable used with the IUgh Range Radiation hionitors (11RRhi), Equipment ID's 2(3)RE78201 and 2(3)RE7820-2.

The scope of this analysis is limited to the inside containment HRRhi coaxial signal cable installed at SONOS Units 2 and 3.

2.0 RESULTS/ CONCLUSIONS AND RECOhlhlENDATIONS Several conclusions were reached as a result of this analysis. The conclusions involve: 1) spurious signals generated during LOCA or hiSLB; 2) the effects of containment heat up and cool down on the installed cable; 3) comparisons between cable types and cable routing methods (electrical conduit verses cable tray); 4) operator training and 5) licensing and industry considerations.

2.1 Spurious Signals The Rockbestos RSS.6104/LE coaxial cable used with the SONGS 2&3 HRRhi will undergo significant induced positive and negative currents as a result of exposure to transient temperature conditions. The induced current is the result of temperature stress, specifically the rate of temperature change. In the presently installed configuration, operating at the detector " keep alive" signal of one R/hr, the HRRhi will provide a false high radiation reading, for a duration of approximately 15 minutes, when exposed to extreme temperature transient conditions inside containment such a: LOCA or MSLB. The magnitude of the false reading may be in the range of the low thousands of R/hr. The low thousands ofR/hr indication spike will last less than one minute, then drop down to hundreds of R/hr, then return to below the alarm set point (<10 R/hr) in approximately 15 minutes. Following the initial temperature transient, the containment atmosphere will begin to cool, and spurious " fall" signals may occur as described in the next -

section.

Note that if significant radiation dose rates do exist, the resultant detector signal strength will be additive to the spurious signal. Example: If the coaxial cable thermally induced current is equivalent to 200 R/hr, a detector signal of 200 R/hr will be additive, with the RP2C reading 400 R/hr. However, the design basis LOCA dose rates are expected to be on the order of IE6 R/hr (Reference 5.6), which will overcome the spurious signal..

2.2

- Containment Heat Up and Cool Down The sign of the current generated by the cable is a function of the temperature gradient across the cable insulation. When the cable is being heated, the induced current signalis positive, when the cable is cooling, the current signal is negative. A current signal with a negative charge will cause the HRRM to alarm " fail." Consequently, when the cable begins to " cool" following the initial temperature transient, the effect on the HRRht system will be to indicate a false " fail" signal on the RP 2C readout. This is true only if the cooling transient induces a current of sufficient magnitude to overcome the " keep alive" and/or detector dose rate signal. At high dose rates, the

Page 4 of17 cooling effect may not cause a fail alarm, it was determined the pressure associated with steam line break conditions does not directly cause or influence the magnitude or sign (positive or negative) of the induced current.

Note that if significant radiation dose rates do exist, the resultant detector signal strength will be additive to the spurious signal. Example: If the coaxial cable thermally induced current is equivalent to -200 R/hr, a detector signal of +200 R/hr will be additive, with the RP2C reading 0 R/hr. llowever, the design basis LOCA dose rates are expected to be on the order of IE6 R/hr (Reference $.6), which will overcome the spurious signal.

2.3 Comparisons of Coaxial Cable Type and Conduit verses Tray Various induced current performance tests, including high temperature steam testing. were performed on three cable types:

1.

Rockbestos RSS 6104/LE (currently installed) 2.

Rockbestos RSS 6105/LE " low noise"(proposed installation) 3.

Brand Rex CS 75146 (previously installed)

The results of the various testing described in Section 7.0 indicate the presently installed Rockbestos RSS 6104/LE provides the best overall performance in terms of recovery time from spurious signals and maintaining a positively charged signal.

- A comparison of steam test results for 250' cable lengths indicate that cable routed completely in conduit does not provide relief from thermally induced charges. The test specimens routed completely in conduit did experience a delay time of approximately 60 seconds, but the resultant -

thermally induced charge was of equal magnitude (thousands of R/hr range) to the one half conduited specimens.

It is concluded that the installed configuration of Rockbestos RSS 6-104/LE coaxial cable, routed approximately one half length in conduit, the other halfin tray, is an acceptable configuration.

2,4 Operator Training The results of this testing program demonstrate false high and " fail" indications will be provided to the Operators during LOCA or MSLB environmental conditions inside containment. The operator training initiated as a result of Non Conformance Report (NCR) 95110073, Disposition Step 2, should include the following two highlights: 1) The duration of the spurious signal will be approximately 15 minutes and may range into the thousands of R/hr and; 2) The spurious high R/hr signal will return to normal, and as the containment environment cools, a spurious " fall" signal may also occur as the thermally induced signal changes sign from positive to negative. -

masking any authentic lowlevel dose rate. When accident temperature conditions stabilize, and/or significant dose rates exist, indicated radiation dose rates will be more accurate;

Page 5 of17 2.5 Licensing Considerations The llRRhi is a Regulatory Guide (RG) 1.97 post accident monitoring instrument. The RO 1.97 accuracy requireme... for this equipment are that it be accurate within a " factor of two" over the entire range. It is concluded that the HRRhl does not meet this require ent during the traasient portion of high temperature environmental conditions such as LOCA or htSLD. Appropriate licensing actions should be taken to address the operation of the liRRhi outside it's required accuracy range. It should be noted that the HRRhi already does not meet the accuracy requirements, and licensing actions are ongoing to address this issue.

2.6 Industry Considerations All commercial nuclear power plants have a RG 1.97 required HRRhi detector manufactured by General Atomic (Sorrento Electronics), Victoreen, Kaman Sciences, or other detector manufacturer. Since these devices operate using extremely low signal strengths, and were tested separately from their associated signal cables, it is probable the issue of thenaally induced coaxial cable signals is not addressed by existing qualification testing or documentation. A synopsis of the SONGS investigations of this issue will be presented to the Nuclear Utility Group on Equipment Qualification (NUGEQ) during their next meeting in June 1996.

3.0 ASSUntPTIONS None.

4.0 AfETIIODOLOGY This section is intended to be a summary of the detailed information prosided in the Appendices and References cited in this report.

4,1 Introduction The IIigh Range Radiation hionitor (HRRhi) is a R.G.1.97 post accident monitoring device used to monitor high radiation conditions inside containment during post LOCA environmental conditions. The primary function of this device is to estimate core damage following a LOCA.

The HRRht operates using detector current signals ranging from lE ll (0 R/hr) to IE-3 (lE8 R/hr) amps. See Table I for the signal current verses equivalent dose rate.

During a periodic test of the Unit 3 containment emergency coolers, the HRRh1(3RE7820 2) was noticed to " fail" low for periods of up to 15 minutes. This faillow period coincided with the initiation of the emergency cooler test. After completion of the cooler test, an upscale reading of approximately 1 R/hr occurred, decaying back to zero in about 15 minutes. The " keep alive" source within the detector maintains an indicated reading of 0.8 R/hr. Non Conformance Report (NCR) 950800177 (Reference 5.1) was prepared to address the HRRh1 signal problem and provides more detailed background information. NCR 950800177, Cause and Corrective Action Step 2, called for an evaluation of the transient response of all channels of the HRRhi under postulated accident conditic,n:. This document presents the summary, conclusions and recommendations based on the required transient response evaluation.

Page 6 of17 Table 1 Signal Current and Equivalent Dose Rate Signal Cunent (amps) I Dose Rate (R/ht) 1 lE 3 IE8 l E.4 IE7 lE 5 IE6 l E.6 lES lE.7 lE4 l E.8 IE3 1E.9 1E2 lE.10 lEl lE.11 1

4.2 Investigations and Testing Performed to Date Several terting methods have been used to investigate the thennally induced charge phenomena.

Testing and investigation methods include in service testing, simple bench tests such as mechanical motion and heat gun, heat oven and steam chamber testing._ In addition, existing qualification test reports and published literature were reviewed. Discussion regarding these investigations and testing are provided in Section 7.0 of this report.

5.0 REFERENCES

5.1 NCR 950800177 5.2 Equipment Qualification Data Package (EQDP) M37609, General Atomic High Range Radiation hionitor (HRRhi) 5.3 SCE hiemorandum To File,

Subject:

High Range Radiation hionitoring (HRRhi)

Coaxial Cable Test Results and Recommendations, L. Conklin, Dated January 24, 1996.

5.4 WYLE Test Report 45145, LOCA/htSLB Simulation Test Program on Coaxial Cables, Revision 0 5.5 NCR 951100073 5.6 -

Bechtel Calculation N 114013, Post LOCA Equipment Doses Inside Containment, Revision 11

- 5.7 SCE hiemorandum to File,

Subject:

Brand Rex Coaxial Cable used in the TEC Pressurizer Relief Valve hionitoring System, L. Conklin, Dated February 5,1996

~

Page 7 of17 6.0 NOMENCLATURE 6.1 Trioboelectric Effects Triboelectric effects are defined as an internally generated voltage which is created when a coaxial cable shield moves slightly over the wire insulation when the cable i in motion. Changing temperature conditions cause the cables organic insulation to expand and flex. When the shield moves over the insulation numerous minute voids momentarily form, disappear, and re form. The opening and closing of these volds has the same effect as many capacitors charging and discharging.

6.2 Thermally Stimulated Depolarization (TSD)

The manner in which polarization can occur within a dielectric is by charges creating dipoles at impurity moleculo sites within the dielectric. These dipoles require an electric field to initiate the dipole. It is theorized that the electric field is applied during dielectric withstand (5000 VAC) and insulation resistance (500 VDC) production testing. TSD is defined as ^e mechanism where current is released frorn a dielectric as a result of an increase in temperature.

7.0 LNVESTIGATIONS AND RESULTS The following sections summarize the investigations and testing perfonned to date, with further details provided in the associated Appendix as identified below.

7,1 In service Testing (November 1995)

Initially the HRRM signal problem was found to be linked to the testing of containment emergency cooler 3ME400. It was determined that the emergency cooler power cable was routed in the same cable tray as the HRRM signal cable. In rush currents of approximately 800 amps were measured in the power cable, lasting only milliseconds. The in rush current was ruled out as the source of the 15 minute HRRM signal current transient. A Cs-137 test calibration source of 10 R/hr was placed in front of the detector, then the emergency cooler was started. The indicated reading dropped from 10 R/hr down to 9 R/hr, recovering back to 10 R/hr in.ibout 15 minutes. When the emergency cooler was shut off, the indicated reading went up to 11 R/hr, recovering back to 10 R/hr in about 15 minutes.

It is speculated that when the cooler operates, cool air is drawn over the HRRM coaxial cable located in nearby tray, creating a temperature gradient across the cab!c insulation, causing the cable to generate a negative charge resulting in a " fail" indication. As the negative charge de:ays, indication returns to normal. Further information on in service testing is presented in Section 3 of Appendix A.

7.2 Review of Existing Qualification Test Reports (November 1995)

Environmental qualification testing of the SONGS 2&3 HRRM detector was perfarmed by General Atomic (now Sorrento Electronics)in 1981 and also Sandia Laboratories ?n 1988 (Reference 5.2).

Page 8 of17 During the General Atomic (GA) testhg, only 18 inches of coaxial signal cable was included in the steam chamber. Although not discussed in the GA test report, an examination of the test recordings by SCE personnelindicate at the beginning of the transient a four minute up scale reading of about 25 R/hr occurred. The HRRht detector was tested using only the " keep alive" source and the acceptance criteria was based on steady state operation. Further information on the GA qualification testing is presented in Section 4 of Appendix A and Reference $.2.

The Sandia Laboratories testing included simultaneous exposure to radiation and pressure.

temperature conditions. The test report did not identify the length of coaxial cable included in the test chamber. Test laboratory personal were contacted and could not recall the approximate cable l

length included in the test chamber. The Sandia testing resulted in severallarge radiation readings (spikes) when only temperature and pressure were being applied. TF-negnitude of at least one of these spikes was in the ten thousands of R/hr range. The duration of these spikes was for approximately three minutes Subsequent radiation exposure produced detector currents large enough to mask any temperature induced currents ( > 10,000 R/hr). After 50 plus hours several negative excursions were noted. The report stated that the cable behavior during the test was not completely explainable. Further information on existing qualification testing is presented in Section 4 of Appendix A.

A review of Rockbestos, Brand Rex and Raychem coaxial cable environmental qualification testing indicates testing was performed using voltage and current levels eignificantly greater than the HRRht application. Coaxial cables are typically LOCA tested while canying signals at the milliamp level (1E.3), while the HRRhi application operates near the picoamp level (IE.11).

No conclusions can be reached regarding cable performance near the picoamp range.

hiineralinsulated cable performance was also investigated. Previous testing of mineralinsulated (hil) cable by Kaman Sciences for thermal transient behavior indicates that hil cable also experiences significant transient current effects. There would be no benefit to replacing the existing organically insulated cable with hil cable since hil cable exhibits only marginally better performance. Further information on hil cable transient behavior is provided in Appendix B.

7.3 Mechanical Movement (November 1995)

By mechanically n.oving samples of the Rockbestos coaxial cabies, the induced current corresponded to only a few R/hr. This indicates that significant induced current is not the result of cable motion effects. Further information on mechanical movement (motion) testing is prosided in Section 6.1 of Appendix A and Section 2 of Appendix C.

7,4 Heat Gun Testing (November 1995)

Using an electric heat gun, Rockbestos coaxial cables were subjected to heated and unheated air flow over various lengths to monitor the induced currents. It was determined that peak currents were not a function of cable length subjected to heating. Applying heat produced positive charges (equivalent to 150 R/hr), and cooling (removal of the heat gun or blowing cold air) produced negative induced charges. - At steady temperature, induced current decay times were approximately 15 minutes. Further information on heat gun testing is provided in Section 1 of Appendix C, a

i

Page 9 of17 7.5 Heat Oven Tec.ng (November 1995)

Samples of Rockbestos RSS.6104/LE (100'), RSS 6-105/LE (50') and Drand Rex CS 75146 (100') were subjected to a series of heat oven tests to determine the thermally induced current.

The temperatures ranged from ambient to approximately 300F and were held for five to 10 minutes. Testing was limited by the rise rate of the heat oven (4F/ minute). This heat rise rate is signif:cantly less than the postulated LOCA or MSLB conditions of >300F/ minute, it was determined that the cables would generatt >nly one or two decades of current (equivalent to tens or hundreds of R/hr) when exposed to the slow temperature rise rate. Further information on heat oven testing is provided in Appendix D.

7.6 Steam Chamber Testing First Sequence (December 1995)

The primary objectives of the fmt steam chamber test sequence were to:

Quantify the maximum values ofinduced current in the cables from rapid temperature changes.

Determine if the induced charge is a function of cable length.

Steam chamber testing of various lengths of Rockbestos RSS 6104/LE, RSS 6105/LE and Brand Rex CS 75146 coaxial cables was performed during the week of December 11,1995 at Wyle Laboratories in Huntsville, Alabama. The results of this testing are documented in an SCE Memorandum to File (Reference 5.3) and Wyle test report (Reference 5.4). The conclusions and recommendations may be summarized as follows.

Due to troubles with the test chamber penetrations (moisture intrusion and low insulation resistance), not all test objectives were met. Based on extrapolation of the test results, it was concluded that erroneous indication of dose rates as high as 2,000 R/hr may occur as a result of LOCA or MSLB temperature transients. NCR 951100073 (Reference 5.5) was revised to include -

a disposition step for the Operations Disision to include training for all operators regarding accident temperature induced spurious signals on the HRRM. It could not be concluded that the induced charge is a function of cable length.

The Brand Rex cables failed at their test vessel penetration, shorting from shield to ground, and were not considered suitable for further te; sting. In light of this failure, the safety related EQ applications of Brand Rex CS 75146 coaxial cables were reviewed and found to be acceptable-(Reference 5.7). More detailed information on the first sequence of steam chamber testing is provided in References 5.3 (test monitoring) and 5.4 (environmental conditions). Additional LOCA testing was recommended to empirically simulate the installed HRRM inside containment coaxial cable configuration (Rockbestos RSS 6104/LE), and also a proposed future installation oflow noise cable (Rockbestos RSS 6105/LE) contained completely in conduit. The results of the supplementary LOCA testing are documented in the following Section (7.7).

7.7 Steam Chamber Testing Second Sequence (March 1996)

The second sequence of LOCA testing had three objectives:

Page 10 of17 Identify the magnitude and duration of the spurious signal that would be produced for a typical SONGS 2&3 inside containment HRRhi coaxial cable configuration during LOCA or hiSLB environmental conditions. Specifically, that configuration is represented by a 235' Rockbestos RSS 6104/LE cable, located approximately one halflength in conduit, one halfin cable tray.

Determine whether a " low noise" cable such as Rockbestos RSS 6105/LE provides superior performance over the installed Rockbestos RSS 6104/LE.

Determine whether there is some improvement in coaxial cable performance when the cable is routed entirely in electrical conduit.

As described below, all three objectives were met.

Steam testing was performed during the week ofhfarch 25,1996 at Wyle laboratories in Huntsville Alabama. The intent of this testing was to:

li Empirically model a typical SONGS 2&3 inside containment HRRhi coaxial cable run. A typical coaxial cable run was represented by a 235' of Rockbestos RSS 6104/LE routed one halflength in conduit, the other halfin cable tray. A review of the Nuclear Consolidated Da:a Base (NCDB) indicates that most HRRhi signal cables routed inside containment are approximately 50% in conduit, 2.

Empirically model possible future installations of:

a) 250' of Rockbestos RSS 6-104/LE 100% in conduit b) 250' of Rockbestos RSS 6105/LE " low noise" 50% in conduit,50% in tray c) 250' of Rockbestos RSS 6-105/LE " low noise" 100% in conduit This would provide practical information on cable performance during postulated LOCA and -

hfSLB environmental conditions. Rockbestos RSS 6104/LE and RSS 6105/LE coaxial cables as identifica below were included in a LOCA/hiSLB simulation as described in the Wyle Test Report (Reference 5,4),

250' Rockbestos RSS 6104/LE (100% in condult) 235' Rockbestos RSS 6-104/LE (50% in conduit) 250' Rockbestos RSS 6-105/LE (100% in conduit) 250' Rockbestos RSS 6105/LE (50% in conduit)

Control Penettation (18" Loop of RSS 6104/LE)

The use of a control penetration provided a baseline measurement to differentiate between what would be considered a cable failure, verses a test vessel penetration failure, The second steam test used an improved test vessel penettstion design, since the previous steam test results were considered suspect due to test vessel penetration failures limiting the useful information relating to cable performance, It should be noted that several test vessel penetration designs were tested and abandoned prior to the second sequence steam test. The final design does not include features that were discarded

Page 11 of17 such as the use of PVC and Tefbn insulated test leads. The PVC dielectric soflened at high temperature, causing the center conductor to migrate through the dielectric, and short to the shield. Teflon twisted shielded pair test leads had a strong sensitivity to triboelectric etTects, producing large spurious signals that could erroneously be attributed to the test specimens, The improved test vessel penettation design in presented is Figure 1 of the Wyle Test Report (Reference 5.4). The final design consisted of the coaxial cable specimen, covered with a Raychem sleeve, encased in an epoxy filled pipe nipple, with a BNC connector one foot outside the test chamber (see Figure 4).

7.7.1 Second Sequente Steam Test Results The test :pecimens were arranged in the simulated conduit and cable tray fixture as described above and detailed in the SCE test monitoring procedure (Appendi:: E) and Wyle Test Report (Reference 5.4). Test monitoring is described in detailin Appendix E of this analysis. To summarize, the test specimens were energized with a current signal of IE.ll amps, simulating the keep alive source of the HRR.M detector. Thermally induced current signals were monitored and recorded over time. The thermelly induced signals are presented graphically in Figure lof this report. The complete recordings ofinduced current verses time during the two and one half hour warm up, and four hour steam test, for all test specimens is provided in Appendix H.

The te:t chamber was taken from ambient temperature to approximately 120F and held for two and one half hours. This allowed the specimens to sti,ilize since even the small rise in temperature from ambient (80F) to 120F produced dgnificant thermally induced currents.

Following test specimen signal stabilization, steam was supplied to the test chamber raising the temperature to approximately 420F in less than three minutes. The temperature was held at approximately 400F, sometimes reaching 425F for a total 915 minutes. The steam test temperature profile is presented in Figure 2 of this report. As shown in Figure 1, all four cable test specimens (not the control channel) experienced thermally induced currents equivalent to the low thousands of R/hr within 100 seconds of steam initiation, Test specimens routed completely in conduit experienced a delay time of approximately one minute before reaching their peak induced signal. This indicates that the conduit provides only a thermallag effect, not complete protection. The cable samples in conduit experienced a secondary induced current spike of approximately 50% of the main spike occurring at approximately 250 seconds. The timing of this secondary spike corresponds to a sudden increase in test chamber temperature f:om 360F to 400F. The induced current quickly drifted down to the tens of R/hr and as low as the " keep alive" 1 R/hr signal within 15 minutes. All test specimens provided reasonable performance from approximately 15 to 30 minutes into the transient.

At approximately 30 minutes into the transient, all test specimens signals became erratic. It was discov,ned that moisture had migrated into the BNC connectors used to connect the test messarement leads to the test specimens located at the test vessel penetration. An examination of the test vessel penetration revealed the Raychem sleeve was being extruded through the penetrations epoxy potting. Due to the high test chamber pressure (70 psig) and temperature, moisture migrated through the cable jacket, along the cable braid, through the penetration, and into the BNC connectors, An attempt was made to dry out the connectors by disconnecting and shaking out the moisture. Only test specimen five responded to the drying efforts, and the signal

page 12 of17 immediately returned to normal (keep alive signal 4 i R/hr).

Since all test specimens signal became erratic at approximately the same time, and all connectors contained moisture (a few drops), it is reasonable to conclude that moisture within the connectors caused the erratic readings 7.7.2 Performance of the SONGS 2&3 Simulated installation Since the test objective was to predict IWGi sienal caw performance, interfaces such as terminations (connectors) with the RD.23 detector and containment electrical penetration assembly were not included in the test set up. The SONGS 2&3 inside containment HRRM coaxial cable installation was simulated using a 235' piece of Rockbestos RSS 6104/LE coaxial cable routed one halfits length in conduit, the other halfin cable tray. In the test monitoring procedure (Appendix E), it is identified as test specimen number 4. The signal behavior of this specimen is considered representative of the installed condition, and may be used to predict the performance of the SONGS HRRM during LOCA or MSLB environmental conditions.

The following is a narration of the information of the signal behavior during the first 900 seconds (15 minutes) and is graphically presented in Figure 3. The indicted radiation dose in the first 60 seconds spiked to 3,800 R/hr, dropping down to approximately 900 R/hr in 90 seconds. A small secondary spike back up to 1150 R/hr occurs at 120 seconds. The indicated dose rate drified down into the tens of R/hr in approximately nine minutes. At 13 minutes and until approximately 17 minutes the indicated dose went into " fail" due to a negative signal resulting from " cooling" of the test chamber. At 17 minutes the indicted dose rate returns to the tens of R/hr and less, approaching the keep alive source value (one R/hr) until approximately one half hour. At this point a negative signal was generated, and became erratic until the test was terminated at approximately four hours. As discussed above, it is concluded that moisture migrating into the test lead BNC connectors caused the erratic readings.

At SONGS 2&3, the inside containment HRRM coaxial cables are terminated at their Westinghouse containment electrical penetration assemblies using bulkhead HN connectors.

Because the connector are terminated inside containment, no pressure differential or driving force operates to force moisture through the cablejacket, into the HN connector. It should be noted the bulkheed HN connectors are covered w;th an environmentally qualified Raychem heat shrink kit dt. si to prevent moisture intmsion. However, if moisture enters the cable through the jacket, the,e is a significant elevation difference between the detectors and the penetration, which may cause a gravity feed into the electrical penetration HN connecters. This is considered a separate issue from the coaxial cable performance and is being tracked by Action Request MR) 951100073.

7.8 Theory of Thermally Induced Cnarges (May 1996)

Recent investigations reveal that a phenomena exists where currents are created from dielectric relaxation processes. This phenomena is called thermally stimulated depolarization (TSD) currents. TSD is defined as the mechanism where current is released from a dielectric as a result of an increase in temperature. The manner in which polarization can occur within a dielectric is by charges creating dipoles at impurity molecule sites within the dielectric. The current released is

i Page 13 of17 dependent on cable length, polarizing E field, number ofimpurity sites, dipole activation energy and rate of temperature change. There are complex equations that can be solved to predict the current resulting from rapid temperature change, however, the many variables related to cable dielectric impurities make these calculations impractical in this application.

The TSD phenomena is a realistic explanation of the thermally induced currents that occur in the coaxial cables associated with the IDUW. Funher information on the TSD phenomena can be found in Apr 7. dix 0.

I

i 4E-08 0

C="

{

l n

3E-08

=

b o

-4 I

E 2

n E 2E-08 y

O S

E l

m

=,.

$ 1E-08

\\g

=

n o

ft j gi2

//--~,Q~,.

i:

/

-..'.% D*

.{ --.. ~....~~ m 0- ^ -

~

o

--N,/

zt::e -

FO O

-1 E-08 9

~

tn 0

200 400 600 800 1000 Q

C Time in Seconds a

=_

250' RSS4-105/LE 100% Conduit in 250' RSS-6-105/LE 50% Conduit 5

2 T

250' RSS-6-104/LE 100% Conduit -

235' RSS-6-104/LE 50% Conduit I

o f

Page 15 of17 Figure 2 SONGS 2&3 Installed Configuration LOCA/SISLH Indicated Dose Rate a

8 i

l l

1

- 8m a E

8 s

n>

EF 8v 8

/

N

/

<L

~

O n

n n

JLl/M ul esoO lueleAlnb3 i

i

1 Page 16 of17 Figure 3 Second Test Sequence - LOCA/51SL11 Ternperature Curve N

i e

a:

l V

'f 4

5 4

g s

R i

~

mw F-Dz 1

H E

w g w o 1 o

H o

J 1

a O

l v

M i,n m to I

wa i

>- N l

m i

<a g

wUN n

UOM (o.J O

+

d i

i i

I I

i e

i i

8 8

8

.o d

p 8,

R 8

8 0

o 8

m n

(._ d ' 1 3_0 L _

aen.Lv.e_adH.JA_

Page 17 of17 Figure 4 Second Test Sequence Chamber Penetration Design 1~

'lf u

l V

ll 1

pi p;';

&l t

I l'; ' :1;'.'

E

[

r. l g$

9 It I' il ic I t

1

_g v

w

.1 3-

.l1..' I E

n Et

., J -ll

,)

':!f s.

^

i,,'

4 I

}/*

I I

ii I

iil'

/l l

l 7

!j % x 3

3 g]/'

13

~

W ',

E15,h -

E e

tb.x n n i f

ll Preliminary Analysis of Radiation Detection Anomalies 1

Introduction During a periodic test of the emergency coolers located within the containment, High Range Area Containment Radiation Monitor, RE7820-2, was noticed to go out of operation for periods up to 15 minutes. This period of non operation coincided with the initiation of the cooler test. After completion of the cooler test, the cooler was switched off. This action resulted in an upscale reading of monitor RE7820-2. The upscale reading produced a radiation equivalent wse rate of approximately 1R/hr and decaying back to zero in about 15 minutes.

2.

Pre Start-Up Modifications During the refueling outage, new coaxial cables were pulled for use with RE7820-1 & -2 monitors. These cables are partly located within conduit runs and partly in cable trays. The sections of cable run within the cable trays were laid on the top of the cable tray with no novering on the tray. This section of cable tray is in close proximity to the intake of the emergency cooler.

3.

Initial in-situ tests in order to better understand the nature of these anomalies, several tests were performed in the containment. The cooler power cabling is run n the same ccble tray es the radiation monitor coaxial cables. In-rush currents of approximately 800 amps were measured in this power cable at the start of the coolers. The timing of the current, which lasted milliseconds in duration, was eliminated as a cause of the radiation reading anomalies.

A further test was performed using a Cs-137 test calibration source. This source was used to bias up the lon Chamber detector with a field of about 10 R/hr. The emergency cooler was operated and the same magnitude negative and positive changes in radiation readings occurred. However, instead of going off scale, low when the cooler was turned on, the reading dropped from the 10 R/hr level to about 9 R/hr. This value recovered in about 15 minutes as before. When the cooler was switched off, the reading read about 11 R/hr, again recovering in about 15 minutes to 10 R/hr.

4.

Review of Other Qualification Testing Results The equipment was manufactured by the then, Electronics Systems Division of the General Atomic Company, now known as Sorrento Electronics. A visit was made to their facilities in San Diego. An examination of their qualification test data for this equipment showed that during seismic testing, the variations in

b

.:. h 2h

/

2 recorded field values varied in the range of12 R/hr. Temperature and Pressure tests were redormed during qualification testing. An examination of the recording showed only one anomaly at the begining of the testing. This value was an up scale reading at about 25 R/hr. The duration of this reading was about 4 minutes. This occurred at the begining of the testing. General Atomic qualification acceptance criteria was for values obtained during steady state.

This value was not used in the acceptance / failure decision.

A review of a series of tests performed by the Sandia Labs as described in report NUREGICR-4728 dated February 1988 was conducted. Several large radiation readings were noted during initial periods where only temperature and pressure testing was being conducted. The magnitude of one of these 5

anomalies was in the 1 x 10 R/hr range. The duration of the spikes was approximately 3 minutes. Subsequent values of radiation remained high for the duration of the temperature and pressure testing. After 50+ hours of test, several negative excursions were noted.

In the summary of the Sandia report it stated that the nature of the cable behavior during the test was not completely explainable and suggested further testing be performed. The appendix described that the cabling had split during the test at about 53 hours6.134259e-4 days <br />0.0147 hours <br />8.763227e-5 weeks <br />2.01665e-5 months <br /> with moisture probably seeping into the dielectric region.

5. Review of Available Test Notes / Literature A review of the previous test data notes from General Atomic Co., the Sandia Lab report and conversations with people involved in previous testing led to a postulated explanation for these anomalies. To explain the long time constants that have been observed in the containment measurements and accounts of short duration spikes ( seconds) while at high temperature and pressure, required long time constants that changed value at high temperature, If charge was being produced by static forces termed TriboeleWie and charging up the distributed capacitance of the cabling, the discharge path would be through the insulation resistance of the cabling. The insulation resistance changes with high temperature. The IR value reduces about 3 orders of magnitude over the temperature range. This would then explain the long and very short time constants that had been noted. (Refer to secion 6.2)

6. Preliminary in-house Tests in an attempt to support the distributed capacitance and insulation resistance theory, a simple set of tests were conducted. These consisted of several motion tests to determine magnitudes of induced currents. Several tests were performed by heating various lengths of cable to about 350 'F and adding

4::o. A

c. 3/

J external resistors to alter the effective insulation resistance. These tests were to determine the magnitude of the induced current and to see if the time decay constant would vary with the change in IR value.

The results of these tests were that:

6.1. The magnitude of current produced by motion is limited to approximately a reading of 5-6 R/hr. The frequency response of the current is proportional to the frequency of the motion input. Therefore it was concluded that the anomalies seen where not as a result of motion effects.

6.2. The distributed capacitance and Insulation resistance discharge path theory did not hold up. The addition of 100 Meg ohms and then 10 Meg ohms in shunt across the cable did not substantially change the discharge times. As a result of these tests, the internal resistance had some effect but was not the dominant factor in determining the decay path of the charge in the cable.

6.3. The positive going, induced current decayed in about 14 minutes while still l

subjected to a heat source of about 350'F. This decay time remained fairly constant for each length of cable and for each change in insulation resistance.

6.4. When the heat source was removed, the current rapidly decayed but then reverted into a decay pattern similar to that shown during the application of heat to the cable.

6.5. The maximum value obtained during these heat tests was 75 R/hr. The value was reached after about 1 minute of heat and decayed off af'.er about 15 minutes while heat was continuing to be applied.

7. Interpretation of in-House Testing A possible explanation for these long time constants is that charge is generated in the cable as a result of a temperature differential across the dielectric As the external heat source initially produces a high delta temperature, a large displacement current is produced in the dielectric. As the heat flow occurs in the cable toward the center conductor, the temperature differential across the cable reduces in magnitude. The current is either bled off by external sources or by recombination within the dielectric. Some small delta temperature will always occur across the cable so that a small offset current remains.

When the heat source is removed a delta temperature exists from the center to the cutside. An initial charge change appears to be produced in the opposite direction fairly rapidly but then the decay follows the same decay time constant as the heat flow occurs from the center to the outside.

i

Ag. A p 4 4

/

The heating of the cable was performed on a portion of the cable as it was colled in about a 12 inch diameter. The heat was therefore applied at about 3 foot intervals along the cable. Heat transfer from the point of heat application may have been moving out to the unheated sections of the cable. This maybe a better representation of the actual application of heat to the uncovered portion of the cabling within a cable tray than a total cable immersion into a heated test oven.

The heating and cooling explanation supports the observed responses in the containment monitor. The emergency cooler could be removing heat from the outer surface of the cable causing a negative charge to drive the picoammeter off scale-low. The thermal time constant of the cable is about 15 minutes. After reaching an equilibrium value, the cooler is turned off causing heat flow from the outside toward the center of the cable creating a positive charge. As thermal equilibrium is reached current ceases to flow from the induced charge.

The bench tests produced a maximum current value equivalent to 75 R/hr at

- 350*F. An oven test on a portion of the cable could support whether this value is consistant from both types of testing.

Prepared by: A.T. Hyde 11/13/95 filename: 7820 data \\jcodata

// g d n 8 p.

V ;e NI CABLE QUALIFICATION Thermal Transient Investigation During LOCA testing the signal current being recorded displayed a transient increase coinciding with the temperature ramp from 150*F to 385'F in thirteen seconds.

See Figure K-1.

The tra ns ient effect decayed in an exponential fashion after reaching a peak value of approximately 150 pA. The nonnal signal was approximately 10 pA plus or minus a factor of two.

Subsequent tests run on the ion chamber element did not show a similar behavior.

However, subsequent tests run on the MI cable confirmed that the transient behavior could be repeated.

An experiment was designed whereby three different lengths of cable were exposed to a rapid thermal transient to three different temperatures.

Each cable-length / temperature combination was tested twice to improve the statistics. A total of 18 runs was made, with an addition-al 3 runs made to confirm some threshold effects.

l The characteristic transient behavior appeared to break into two effects.

which will here-in be called an initial and a secondary response.

The initial response is characterized by a higher amplitude and faster decay effect, while the secondary response is characterized by a lower amplitude but longer decay effect.

This evaluation will begin by postulating the cause of the effects

observed, then empirically deriving an algorithm to predict the transient behavior, and conclude with suggested methods to minimize the effect of the transient.

postulated Cause of the Transient Current The initial and secondary responses are postulated to be due to thermally induced stresses on the dielectric.

The construction of the subject cable is triaxial with a copper center conductor and metallic coaxial'and triaxial sheaths insulated from each other by dielectric material.

At thermal equilibrium, no transient stresses are present. However, due to impurities or defects in the dielectric material, charge traps exist within the band gap.

Stresses that are applied (such as a thermal transient) which are sufficient to increase the energy of the trapped charge carriers may release those charges, giving rise to the thermal transient behavior.

As a thermal wave front hits the triax sheath, either the expansion of the triax sheath occurs or the thermal transient itself places the dielectric under a transient stress, releasing trapped charge into the return circuit (coax sheath). A picoanneter across the center conductor / coax sheath circuit measures this current as an offset from reference.

Since the triax sheath is uninsulated from the thermal transient, and the thermal conductivity of the stainless steel sheath material is relatively high, this stress and resultant current are l

essentially concurrent with the step deltc-T applied.

Therefore, the fast rising edge of the transient current is observed.

to a re-trapping of free charge as the transient stress equilibrates.The subsequent decay is i

X-1 U

>, =.

,9,,

s n

T'

.. i..

_r.. 3 ;...

1_,__-

a, -

7.

,..3

....r-----

4 ; ; i

,i,,.:.

l,

,'.....---+-._ _ _.. + -..

_i-

-. _. = -

i,,-r.,

...1 c

,. i h.-

..r.-

4,_

l_ ; l. a t

l

.F, 4.

c..., e,. +. r 1-r.

e..

t

7. '_..

4 i-

.{

t

..Qor

[ 3. ;..

? M..t 7 r.. P.. P.

T.,

.? w

.. h..."

+H.c r

_. g., :F. M.

.. i., r ;.,

i

- r.&..r ii i

..r d..

..f'..,. : t',

.,.3

. ~..

A L.

p-

.. ' L.,. 6 4.

i. 3

. L

4..'. ', '...... -

1.

...W

--I

- M, ;,i-d-i-$

'i.

l'

-.g. !

.-- i-e 2,- "

, w t-

, -h t-

-*t-

,tr w,,,.

r.-

er "

1--

.-r.

r ~.- -

. r r--+;*~ -. *--

-- > 1 ri P* ' ri.

, i,

.x:

,, n s

e iei N:

e W.

,i

,,i 1

.. t

, i,,

_ -- _ _. _ _...i..,

, i

.t J p..L p.L.:...__ :;..u :-

a.:..J...!

v---. e. _

i

,, ~, w.

.-r

,,--r.

r-r-

.z -.-, r. r v~

t

, y c. n,, m,,

,...i

--.- T r *- -

,i., a m.~~.,

n win, i i,

-e

.--- ri,,

s,

..:~.

t

....w 6 1

1,i 1,

6

(

l l

'l l '@

1 6

'l

! e i

eQL j,.

• L g,

• e

(

jf.,f.,t. 9 -f-_. +iD %-!

l'l' 1

.4 2

,l

. 4_..... 3.L.

g.

__r._ - ;,

.g. 4.*.: g..p ;..

,; A...

<..d-.---

'- ',' i.-. L..<r- -..g.

. '., _. -.;. g.:., b,., ;

.r r

i

-- n*

': ;l l

l.

. l m,,,;

,,' t + - +.", t.H _r -

c

.. i i,

r.

l 4

4

.-i _.

, i.

i

+.

+

_t.

u i..'..i i

. i.

. - r. ;.,..u..

'.R.L...: p2...I'L:.

.i

.i a..

.L v,

u, ty y.ia

.4.. n.

rt ' - -, - '

,r.

,.- r. - - - - -r- :.

4.,

- i

.... z. 1,,.

- : r-r 10 20 30 40 50 60 70 80 90 picoamperes Chart trace recording of ion chamber S/N 22710 during LOCA temperature transient.

16 Nov.1984.

Chart speed = 8 in/hr. Data recorded Figure K-1 1

i i

K-2

_ _ _. ~. ___ _

N P,> '3 fJ'/

r i

A similar event occurs when the thermal front reaches the coax sheath, but 4

4. '

is delayed by the thermal rMistance of the dielectric.

l

_The Initial-Response Characteristic Data was obtained during the thermal transient experiment that suggested 4

the peak current generated is a function _of the step temperature change applied, its rate of application, and the length of cable exposed to the transient.

The transient current then decays in a classical exponential fashion.

The initial peak current per cable-foot generated as a function of the temperature change (delta-T) as the parameter time rate of application (t) increases, is plotted in j

Figure K-2, M

co -

g' O

.= M t,6'

/

t\\

0 i'

6

/

y j

/

.f

.. /

/

/,

o so in in 2:o no soo CC.TA-T @t irtEE5 FArftDttCT)

Figure K-2 Initial i /ft as a function of delta-T and the rate parameter t.

pk

_The Secondary Response Characteristic Similarly, the-secondary response transient peak is dependent on the cable length, delta-T, and ' rate, fiowever, since the secondary response is an effect of the coaxial sheath, the insulating dielectric delays and dampens the peak The delay of the peak is consistent with the cable diameter, and is current.

essentially constant for all L delta-T, and t.

The peak current value is likewise reduced because of the reduction in the rate of temperature increase of the coaxial sheath.

Further reduction of the peak c:agnitude will be realized as the rate of application of the temperature transient is reduced.

Figure K-3 plots the secondary i parameter.

Pk per cable-foot as a function of delta-T using t as a K-3

-e w

m m

e w +

App.13 pA/7 I?

tA 13 14 13

/

12 f 11

/

l 10 E

9 -

', ( $,*

,s

? -

2 -

(*

3

• , so.

f e*

/

s-

/

l 2-

~

o j

/'

a s

c so 100 ISO 200 230 300 DCLTA-T @C2t25 ;'ArtettCT)

Figure K-3 Secondary i /f t as a function of delta-T and the rate parameter t.

_ j pk isolated nature of the coax sheath (i.e. insulated on both sides material) tends to equilibrate in a longer time than the initial peak.

to reduce the transient current to less that 10 pA is a parameter of interest The time as a detector in a radiation field of of two specified accuracy at that level of noise current.IR/hr would then read within the required for secondary i = 10 pA yields results as shown in Figure K-4. Plotting the time e

t 4

K-4

A p p. i? p. sj'

\\

. 1.

e 3i,.

3,si -

/

!!!2 l

7 ises -

/

f' im -

.w-i m-

n -

h ire -

. ix.

s3. -

/

i_

23 -

I

/

se -

/

i.>

.c 4

ico

. ::c.

.as c.

m e.:

w, TMC 1,5E:DCS)

Figure K-4 Time to attain secondary i f 10 PA for various secondary i pk' Transient Pmp, tion.

The cc oined response for the initial and secondary transient events was analyzed and an algorithm was developed to predict the t'cansient-response for any cable length, delta-T, and rate conibination.

A typical transient response curve for a 150' foot cable exposed'to a 200'F delta-T in '10 sec. is presented in Figure K-5.

e

-K-5

i A p-l3

p. 6/g p

j 2.

~

1.s -

l.O 1.7 j

1.6

(

is -

f3 i$)

I s.f 12 t.:

y, w3 1

it c.s -

f, o.s -

y c.7 -

8 g

c.s -

s os -

f*.4

f

.,'N L

o.:

o

.c o

m m

a xo soo a

m m

SECONCS SIPCC T-BMX Figare K-5 Noisecurrent(1)vstimesinceLOCA(typical)

Discussion I '

The effects. described here-in have been observed by others, not only in mineral-insulated cables,1,2 but also in polymer dielectric materials 3 7

used.in traditional cable.

The noise current produced by the rapid thermal transient-l will be additive to any signal current being generated by the detector when the cable is utilized as a signal cable.

The effect is insignificant when the cable is utilized as a high-voltage cable, as the small current generated is well

(

within the regulation range of the high voltage power supply.

The ramp to LOCA temperature may or may not-be accompanied' by a ramp increase in the dose rate at the detector.

If there are accompanying and comparable chan R/hr,or more, ges in the-dose rate, say of several tens to several thousand then the noise current due to the thermal transient may be insignificant (i.e. less than factor of two error).

If there is little or no accompanying dose rate increase with the LOCA temperature ramp, then a dose rate error condition.of greater than a operator.'

factor of two may be presented to the The magnitude of the noise current is dependent on the rate at which maximum 1.0CA temperature is attained, that is more accurately, the rate at which tne triax sheath temperature is increased.

It has been shown in Figures K-2 and K-3 that for rates of sheath temperature increase of greater thar, approximately 30 seconds to the maximu : LOCA temperatureq typical of. most plants, that the

/!

noise current generation is sufficiently reduced to becore insignificant, even f

at low-(-1R/hr) dose rates.

The plant may want to explore methods by which the i

rate of heating of the H.I. _ cable can be reduced (e.g. thennal insulation-i vis-a-vis steam or hot water pipe lagging or other thermal bar,riers such as a

{

s:

K-6

+

a

\\

,mpp.Bp.v/7 conduit or wire ' tray).

Studies have shown that for LOCA t,aperature ramps of 300* F increases in less than 1 'second that a quarter-inch of insulating material with a thennal conductivity of 0.05 BTU /hr-ft *F will limit the. heating rate of the M.I. cable sufficiently to prevent interfering noise currents.

Since each plant must be considered on-an individual basis due to unique cable lengths, unique LOCA temperatures and ramp rates, and unique accompanying dose rates, it is suggested that tne plant utilize the information provided in this report to determine the adequacy of the supplied equipment to meet individual requireents.

References, 1.

G.F. Lynch and R.B. Shields, "On the Use of Ng0 as Insulation for Coaxial Signal cables in a Reactor Environment", Atomic Energy of Canada Limited document AECL-4827, June 1974.

2.

C.P. Cannon, " Comparative Ganna Radiation and Temperature Effects on 510 and Ng0 Insulated Nuclear Cable". IEEE Transactions on Nuclear Science, 2 l

1982.

j Jason Wilkenfield and Vesa Junkkarinen, " Thermal and Radiation l

Depolarization of Persistent Charge Stored in Polymer Dielectrics", IRT l

Corporation, Intel-RT 8124-005, August 13, 1976.

1

( ~

277-stoo e

1 K-7

'b

A :e m < L- [. ' -

3 Notes on Cabla Testing Prior to formal testing of the Rockbestos cables, several informal test were conducted to get some first hand experience of the kind of noise that is generated by this cable and some level of understanding of how the noise is initiated. They fell into the following:

1. Heat Related Effects 1.1 Movement of unheated air over a coil of cable (-35'), at a flow of ~2-3 scfm, caused displacement currents. These currents were in the negative direction. The time for the maximum swing in current value was ~1 minute, guestimate. By maintaining a constant air flow on the cable, the current value recovered to its start value. This value was in the order of 10s of rr.inutes. This is the phenomona that was previously analyzed as charge creation within the cable sheath with the discharge mechanism l

governed by the internal capacitance and internal insulation resistance.

Further investigation should proove this or not.

1.2 Removal of the air flow during the recovery process createc: a chrge in the opposite direction that tended to counteract the previous charge. The l

time response was comparable with the initial charge change - 1 minute.

l The sum of the charges at this point was a net positive charge that caused overshoot with an accompanyir;g slower discharge time as before.

1.3 The magnitude of the response was about a factor of 2-3 larger than the steady state value of ie-11 amps.

2. Mechanical Movernent 2.1 Observation was that the frequency of the current generated in the cabling was proportional to the frequency of the cable movement.

2.2 The magnitude of the current generated in the cable was proportional to the magnitude of the displacement of the cable.

2.3 The ringing or naturally occuring oscillation generated within the cable were determined by the length of the free, unrestricted cable that was struck. By reducing the length of the restrained cable and stricking it, higher frequency oscillations were crested. As the free length of the cable was extended. The oscillation created by stricking went down in frequency.

2.4 The cable generated current acted in all ways as the mechanical oscillations would occur in a naturally hanging cable. The steady state

14 pp C

[< Ifg current in all these test was at 1e-11 amps. The estimated magnitude of these oscillations / impacts were about i to 1.5 decades up from steady state, Need some more refined tests to nail down exact values t

Notes 11/10/95 ath & dgw d

Further observations from test data taken in maintenance lab.

a 1.

Mechanical Motion Effects A correction to yesterdays results was that output current resulting from motion is proportional to the rate of change of motion not the change in motion.

T~.a magnitude of current generated from motion is a function of cable length.

With a 210'+ cable length, maximum displacement current is 1e-10 amps equivalent to 10 R/hr. This was generated with excessive force over '.he entire coli.

2.

Heat Generated Current Effects Data gathered showed the following:

2.1 Added values of shunt resistance did not effect the time response of the decay curves of various cable lengths after a heating sequence.

2.2 Maximum values of current generated with the cable at temperatures of

~350F were ~1,5e-9 amps. This equivalent to 150 R/hr. This test was a reasonable approximation of time response at the source, 2.3

. Time decays were in the order of tens of minutes for both lengths of cable tested (30' & 210') with or without shunt resistors added.

2.4 The generated current peaked after -1 minute in all cases.

2.5 The generated current decayed down from this peak with a constant heat being maintained on the cable.

2.6 Decay from peak current did not reduce dback down to the previous zero or offset value.

2.7 Removal of the heat source after decay, further reduced the current in the circuit at a more rapid rate.-

bj;>f b

?. 3 3-Speculation as to what this all means 3.1 The capacitance charging theory does not hold up.

3.2 Suspect that explanation of previous data results is that the heat source produces charge in the cable within a minute.

3.3 _

Decay mechanism is governed by thermal decay constant rather than electrical, 3.4 -

Explanation of inferred data by GA in one of their notebooks where they talked about noise pulse in the order of seconds after heat test can be explained by moisture on the surface of the cable drastically effecting the heat transfer function. Our test did not have any moisture therefore not able to test this theory.

Prepared by: A.T. Hyde 11/09/95 Filename: 7820 data \\ note 1109

\\

h:p-a.x h

f. I/s.

Synopsis of cable testing Nov 16/17/95 at CGI i_ abs 1.

Tests performed on the following cables:

100' lengths of Rockbestos RSS-6-104/LE and Brand Rex Cable No. ????

50' length of Rockbestos RSS-6-105/LE 2.

Tests Series of tests in oven cycling temperature from ambient to ~300F holding for 5-10 minutes and cycling back to ambient temperature. Temperature rates of rise and cooling determined by oven capabilities. Oven was manufactured by Tenney.

Approximate inside dimensions were 3' x 3' x 3'.

3.

Results of tests See tables for synopsis of test observations, voltage and current conversions and estimates of maximum deviations from 0 R/hr and non operate state times.

Nature of GA picoammeter module is that it is designed for a positive polarity input.

When subjected to a negative input it takes a long time to recover back to zero. The Keithley doesn't suffer from such recovery process since it is bipolar in design.

Prepared by: A.T. Hyde 11/22/95 file: 7820\\testdata I

J

Table of Cable Test Results Date of tests:

Heat rise rate:

Cool Down rate:

Instrument 1: KetNey 6000 Electrometer Instrument 2: GA RP-2C Picoammeter Nov 16 & 17 4F/ min 14 F/ min 1995 l CableType Test No.

Instrument Nc.

Hesit up cycle comments Cool Down cyc8e comments tm RSS-6-104/LE 1

GA RP-2C Negative offset to start, at 133F reverted +0 9 v rising to Dropped from 2.4 volts to O voits in 6 6 min at 260F. Held at max 2.55 vous at 182F. falling i>ack to 1.5 vots at 240F. 33 just negative to 195F went fus -ve. Stayed fuB -ve through mins later at 253F reading verd -ve. 6 min later at -280F co.npietson of test to ambient temp.

+ve reading Output increased to 2.4 volts at 300F RSS-6-104/LE 2

GA RP-2C Started with 2e-12 offset (causes slight -ve value). At 80F Started cool cycle at +ve offset. Held constar A -ve value value of 0.75 volts reaching + value of 1.2 volts at 140F &

unta 190F werd fun -ve. Remained fun -ve und ambient max value of 1.65 voRs at 230F. Recovered to cross 0 at 280 F. Crossed back +ve at 300F going to 1.5 voRs RSS-6-104/LE 3

Keithley O current at start +ve value of-7.5e-10 at 140F. recovering Held steady at 0 unti 190F. High frequency norse large +/-

to 3 Oe-10 at 190F. High frquency norse present. Increased we sp&es. Went fus -ve at 154F sts with targe noise to +ve vaEue of ? 38e-9 at -240F. Noise reduced at 260F spikes. Recovered to near 0 at ambient Ste noese Recovered to 0 at 290F Held at 0 present RSS-6-10NLE 4

GA RP-2C trrattic at begining pnur to heat. After heating began.

Reverted to 0 voPs at 300F At 195 raodfy went -ve.

almost immediatety went to +15 vots at 80F. Went to 2.25 Required +ve offset currents rangmg from 5e 10 to Se-9 to vots at 120F & stayed at this vis!ue up to 220F. increased to bring reading back on scate.

2.65 volts at 250 F. Reduced in value urda constard at 0.15 vots at 315F. Began to increase to 0.9 volts at 325F l

RSS-6-105/LE 4

Kethley No change in offset current urdd ~300F with I offset at -ve Reverted to 0 amps at 300F. At 190F noisy -ve max value 1.5e-12 amps going to 2.6e-11 amps at ~325F of Se-12 amps unta 180F reverted to 0 amps at 140F RSS-6-105/LE 5

GA RP-2C Offset of 2e-12 amps ( offset slighDy -ve & po! arty).

Offset currert at 2e-12 but reading sts fun scale -ve. Came Reading constant -ve(sBghtfA throughout beat cycle to back to singht we read ng at 230F. At 190 F increase in -ve 240F. Slow -ve increase unts at 315F fu3 -ve reading.

direction but recovered at 160 F. Wes behaved through rest of test et st' iht -ve vabe.

v Brand-Rex 1

Keithley Smag -ve offset -1.5e-11 at start 92F. Crossed 0 at 190F.

On cooldown max +ve vt.tue of 4 Se-11 amp at -300F

+ve man value of 4 65e-10 amps at 220F. Reached +ve 3e-value reached O at -15CF Maintained 0 throughotd rest of

)

11 at 300F test Brand-Rex 2

KeRNey Started at O amps At 230F slight +ve value reaching max Startang at 165e-10 d= creased to J at -150F Held at 0

_U 1.65e-10 amps at 310F.165e-10 held to 321F through rest of cycle

$I Brand-Rex 3

GA RP-2C Started slightly -ve (offset of 2e-12). Went ius -ve at -150F.

Sta ted at 1.2 vots +ve crossed 0 at 140F Remained (f

Recovered to slight -ve at 165F. Crossed 0 at ~260F going silgh7y-ve through rest of test.

to max +ve value of 1.2 vous at 325F.

Brand-Rex 5

Keithley No observed changes up to 280 F. Went +ve to value of

+ve value of g Oe-11328F gradually revertrng to 0 at 150F 3 De-11 at 300 F. Max +ve vstue of 6e-11 at 320F. Some

& remaining at 0 to ambiert g

sma9 noise at these temp.

'.S

th fb D C

?^-

i

/

li e

~

9 i 6,

6, o,-

6 o

6 h

6 o

y e

e e

e e

e 3

M M

M M

M M

M M

M l

.g e

e e

e e

e e

e b

o b

o b

o b

'o b

e w

-e e

e e

e e

e g

M M

M M

M M

M M

M e

e e

e e

e e

e e

s Cf k

a m

m o-N m

m o

m o

N M

M o

~

N e

m m

N e

m e

h ]~ v h

3 Nt e

lc n

y n

in n

e C

in im im m

im m

m n

n n

0 in n

l 5

5 im 5

im im M

im o

it l

o 5

1 1

2 6

ia C

8 8

8 D

0 0

?

0 F

e ta rep O.

xaM e

lc e

y t

n a

C in i

u a

m im i

im im im im i

m in t

n n

n n

n n

n m

m 6

m s

e 3

E H

1 5

0 0

0 2

2 0

4 0

ecna m

ro fre P

e l

in cy n

C r

r r

r r

r r

r r

r o

h h

h h

h h

h h

h h

l it o

/

/

/

/

/

/

/

/

/

/

a o

R R

R R

R R

R R

R R

iv C

0 0

0 0

0 0

0 0

0 0

eD

'i m

rh u

u m

q x

E i

a r

M h

/

f R

o e

l s

m c

r r

r r

r e

u y

h r

h h

r h

h r

m im C

h

/

/

h

/

/

h t

/

/

r R

R a

R

/

r R

R

/R R

h R

h t

x a

5 8

2 5

5

/

/

6 a

e 0

1 3

3 6

R 6

6 1

R s

M H

1 2

1 1

2 0

4 1

9 6

E tne y

y y

y y

m C

C e

C e

C le u

2 2

h 2

h 2

h hle C le l

l 2

h r

t it P-it i

ie P

i t

t t

s P

P e

e P

e e

n R

R K

R K

R K

K R

K I

.oN tseT 1

2 3

4 4

5 1

2 3

5 E

E L

L

/

/

e 4

5 x

p 0

0 e

y 1

1 T

R e

6 6

-d l

b S

S n

a S

S a

r C

R R

B j

L