Find out why BICSI is today's fastest growing Cabling Association. Our industry expertise, technological resources and Certification programs keep you on the competitive edge. Let us serve you to our mutual success!


Home | HOTS | Environmental Info | News & Issues | The Router

Calendar of Events | About Us | E-mail Us



Article 06 Page 1b


Comparison of Communications LAN Cable Smoke Corrosivity

J. Thomas Chapin, Ph.D.
Lucent Technologies, Bell Laboratories
2000 Northeast Expressway
Norcross, GA 30071 USA

Pravin Gandhi, Ph.D., P.E.
Underwriters Laboratories
333 Pfingsten Rd.
Northbrook, IL 6062 USA

L. M. Caudill
Chestnut Run Plaza
Wilmington, DE 19880 USA

Submitted to National Fire Protection Research Foundation

Fire Risk & Hazard Research Assessment Research Application Symposium
San Francisco, CA, June 25-27, 1997


This paper reviews various US and International smoke corrosivity methods to assess their ability to predict the reliability of digital electronic equipment exposed to smoke effluent. The mechanisms of digital electronic equipment failure are presented and the ability of the smoke corrosivity methods to assess electronic reliability are judged. To validate the assessment, seven commercially available local area network (LAN) communications cables having a wide range of fire behavior were tested with four representative US and IEC test methods (ASTM D5485 (metal loss), ISO DIS 11907-3 (metal loss), IEC 754-2 (acid gas) and a newly developed leakage current test). The test data for each cable and test method is presented. Existing methods were found to be ineffective at predicting the reliability of smoke-exposed digital equipment. The leakage current method which has been derived from dust exposure studies (employing an interdigitated test pattern) was found to be most suitable for predicting reliability. Cables with the highest fire performance characteristics (low heat release and smoke release properties) had the lowest leakage currents in the study. It is concluded that these cables should provide the most protection to digital equipment in fire situations.


The evaluation of the corrosivity of combustion products from polymer insulated copper wires, polymer coated glass fibers, and cable jackets is a topic of vigorous discussion in standards organizations world wide. At present, the smoke corrosivity methods found throughout the world assess smoke corrosivity by measuring changes in pH or conductivity of water extracts, or by measuring the metal loss of copper targets. These methods fail to accurately predict the electronic reliability of digital equipment, since it has been found that failure usually arises from the degradation of insulation resistance rather than direct metal loss. These conclusions are based on the experience from dust contamination studies and analysis of contamination arising from telecommunications equipment fires. Further, pH and conductivity methods assess only one characteristic of fire effluent - acid gas - and not other smoke components such as particulates.

Another limitation of these small scale corrosivity methods is that they are material, rather than product-based. Materials tests do not allow the assessment of the smoke effluent arising from the co-combustion of all materials found in the manufactured product. At present, no standardized smoke corrosivity test assesses the effects of smoke by measuring surface insulation resistance changes on circuit board substrates.

Several test methods have been standardized to assess the corrosion hazards arising from the evolution of gases generated in a fire involving these products. However, many of these methods are not suitable for assessing the hazards to electronic equipment in particular, since the relevant property related to electronic equipment reliability is not measured directly. A summary of current test methods is provided in Appendix A.


There are damage mechanisms to electronic equipment associated with fires that cannot be adequately assessed by determining the pH or conductivity of water extracts or by metal loss targets. This is because in a fire, ionic contaminants associated with fillers, flame retardants, colorants, processing aids and/or impurities in the polymers, or by-products of the polymerization reactions may be released and deposited on circuit boards. The impact of these deposits are anticipated to be similar to those from similar airborne contaminants.1 The most common cause of equipment malfunction following exposure to smoke from cable combustion is not loss of thickness of structural metals or metal circuitry from direct deposition of corrosive gases, but rather electrical shorts and arcing that cause cross-talk and malfunctioning components.2 The low insulation resistance associated with smoke-related contamination leads to metal migration (dendrites), electrolytically corroded conductor lines (quite distinct from the direct corrosion caused by deposition of corrosive smoke on conductors), and other electrochemical degradation processes.3-5 The full array of contaminants from the smoke includes the halide gases, of course, but also includes other ionic contaminants, organic gases, and, in some cases, graphitic carbon. High humidity exacerbates the effects of ionic contaminants.3,4 However, if graphitic carbon is formed in a fire, it has been found to be conductive at all humidity levels. Data showing the effects of these kinds of contaminants on electrical leakage are shown in a published paper2 and also in Figure 1. The figure shows that ionic contamination and the resultant leakage currents measured on a standard interdigitated test pattern.

Figure 1
Influence of Relative Humidity on
Leakage Current for Dust Particulate
from Various Sources

These data reveal several important characteristics of the dependency of leakage current dependency on relative humidity. For example, there appear to be a threshold of relative humidity where the leakage current is very low (less than 10-9 amperes). Beyond the threshold, the leakage current is exponentially dependent upon the relative humidity. However, the threshold relative humidity may be influenced by the nature of particulate smoke as may be seen of the smoke deposits from the Kuwait fires (see Figure 1). In this case, the leakage current is relatively higher even at low humidity levels. The smoke particulate from the Kuwait fires were found to be graphitic in nature and thus were conductive irrespective of the relative humidity.

The threshold behavior of leakage current is due to moisture absorption characteristics of the ionic constituents of the particulate matter. On the other hand, the exponential nature of leakage current with respect to the relative humidity and temperature has been postulated and developed theoretically by Comizzoli1 as shown in the following equation:

I = Ioexp(b.RH) exp (-E/kT)

where I is the surface current; Io is pre-exponential constant; b is a constant; RH is relative humidity; E is activation energy; k is Boltzman constant; and T is absolute temperature.

Smoke Corrosivity Measurements

Based on these data shown in Figure 1, and the experience with smoke damaged electronic equipment, a test program was developed to investigate the leakage current characteristics of smoke generated from a number of commercially available copper communication cables and deposited on a standard interdigitated circuit. The cables were selected to include typical material arrangements (insulation and jacket materials) currently in usage.


Prior to performing the corrosion tests, the cable jacket and insulation materials were characterized by various analytical methods. Also, the samples were tested by cone calorimetry (ASTM E1354) to characterize overall fire performance of the cables. Primarily, heat and smoke release data was used to compare the cables. Corrosivity data for each cable was compared.

Metal loss corrosivity data were obtained from ASTM D5485 and ISO DIS 11907-3. The IEC 754-2 combustion tube furnace for the measurement of pH of evolved gases. In addition, the IEC 754-2 apparatus was modified for the measurement of leakage currents using the newly developed leakage current test.

Test Samples

The samples included seven typical commercial local area network (LAN) communication cables were used as test samples. The cables consisted of four insulated, twisted pair 24 awg conductors surrounded by a thermoplastic jacket material. The samples are identified in Table 1.

Table 1 - Cable Test Samples
Test Sample
UL 910-CMP
UL 910-CMP
UL 910-CMP
UL 1666-CMR
UL 1581 VW-1-CMX
IEC 332-3C
IEC 332-1

The hierarchy of fire performance of LAN cables in the US as compared to the IEC fire test methods is shown in Table 2

Table 2 - Comparison of US and IEC Cable Fire Tests
IEC Testb
UL 910 Plenum - CMP, OFNP
UL 1666 Riser - CMR, OFNR
UL 1581 Vertical Tray - CM, OFN
IEC 332-3C
UL 1581 VW-1 - CMX
IEC 332-1

a Copper and fiber optic non-metallic cable descriptions found in Appendix B
b No IEC cable markings required, no equivalent IEC test

Quantitative and semi-quantitative elemental analysis were performed on the jacket and insulation materials to understand the chemistry of the materials being tested and their combustion products. The results of the analysis are presented in Table 3 and Table 4.

Table 3 - Composition of Plenum (CMP) Cables
C, F
C, F
C, H, O, Cl, Br, Al, B, Ca, Fe, Pb, Mg, Mo, P, Si, Ti, Zn
C, F
C, H, O, Cl, Br, Al, B, Ca, Pb, Mg, Mo, P, Zn
C, H, O, Cl, Br, Al, B, Ca, Fe, Pb, Mg, Mo, P, Si, Ti, Zn

Table 4 - Composition of Riser, CMX, IEC Cables
C, H, O, Cl, Al, Sb, B, Ca, Pb, Mg, Mo
C, H, O, Cl, Al, Sb, Ca, Pb, Mg, Zn
C, H, O, Cl, Al, Sb, B, Ca, Pb, Mg, Mo
C, H
C, H, O, Ca, Mg
C, H
C, H, O, Al, Ca, Mg, Si, Ti
C, H

The same type of fluoropolymer (FEP fluorinated ethylene propylene) is used as the insulation in cables A and F. A slightly modified fluoropolymer was used as the jacket in cable A. Several different smoke suppressed (LS) PVC materials were used as the jackets in cables F and I. Different plasticized PVC materials were found in cables I, G and B. The variations in PVC materials are reflected in the elemental analysis differences found in Table 4.



This test apparatus consists of the following components: a conical-shaped radiant electric heater, a specimen holder, an exhaust gas system with oxygen monitoring and flow measuring instrumentation, ignition source, data collection and analysis system, and a load cell for measuring mass loss. A photograph of the test apparatus is provided in Figure 3.

Figure 3
Cone Calorimeter Apparatus

The Cone Calorimeter tests were conducted with a horizontal sample holder, edge frame and grid. The heat flux was 50 kW/m2. The typical burn duration is 40 minutes.


The apparatus was identical to that described in IEC 754-2 test standard, described herein and shown in Figure 4. It consisted of a furnace, silica tube, combustion boat, air supply system, and two glass bottles for collection of combustion products.

Figure 4
pH Test Apparatus (IEC 754-2)

The tube furnace had an inside diameter of 60.3 mm and an effective heating zone of 600 mm. The test temperature was controlled by an electronic temperature controller. The silica tube was 900 mm long, 36 mm inside diameter and had a wall thickness of 2.5 mm. The silica tube was placed in the tube furnace such that it extended 80 mm from the rear connected to a dry compressed air supply. The rear end of the tube was ground, fitted to an adapter and connected to wash bottles. At the exit of the glass tube, the gases passed through the two wash bottles filled with 450 ml of distilled water. A magnetic stirrer was positioned at the bottom of the first wash bottle to produce a swirling motion which induced better absorbance of the combustion gases. Porcelain combustion boats with inner dimensions of 80 ± 5 mm long by 12 ± 1 mm wide and 9 ± 1 mm deep were used to hold the test sample during the test.

The solution pH was measured using a Cole-Palmer digital solution analyzer Model 5800-05 with a pH scale with a range from 0 to 14.0 pH. A dip cell type conductivity electrode Model 1481-62 and a glass pH electrode Model 5997-10 (reference cell type Ag/AgCl) were used in the measurements.


The test apparatus consists of the following main components: a conical-shaped radiant electric heater, load cell, a specimen holder, an exhaust gas system with oxygen monitoring and flow measuring instrumentation, ignition source, gas sampling system for corrosion, data collection and analysis system. A schematic of the test apparatus is provided in Figure 5.

Figure 5
Schematic of the Cone Corrosivity Apparatus

The gas sampling system for corrosion testing, as shown, consisted of a combustion product collection device, an 11.2 liter polycarbonate exposure chamber, a particulate filter, flow meter and a pump. A corrosion target Model No. 030788-SO.35-8061 (span 45,000 Angstrom) from Rohrback Cosasco was employed to measure the corrosion on the targets. The corrosimeter has a resolution equal to target span/1000. The test method for the cables was in accordance with the ASTM D5485 standard.


The apparatus consisted of a moving furnace, a quartz tube, a sample holder, air supply system, and corrosion detector. This apparatus is illustrated in Figure 6.

Figure 6
ISO DIS 11907-3 Test Set-Up

The annular furnace consisted of a 100 mm heated section, with an annular diameter of 51 mm. The furnace had an electrical heating system capable of attaining temperatures of 900°C. The annular furnace was mounted on a track capable of moving it coaxially at a speed of 10 ± 0.5 mm/min. The combustion tube was manufactured from quartz glass which is resistant to the action of corrosive gases. The quartz tube had a length of 1000 mm with an outside diameter of 40 ± 1 mm, and a wall thickness of 2 ± 0.5 mm.

The sample holder was a curette. The curette was a half quartz tube 400 ± 2 mm long, 15 ± 1 mm high, and with a wall thickness of 1.7 ± 0.2 mm. There was a 2 mm high lip at each end of the curette. The air supply system consisted of compressed air cylinders connected to one end of the quartz tube. The cylinders were provided with pressure regulators and air flow meters to attain a flow of 100 liters/hour.

The corrosion detector consisted of a printed wiring board (PWB) circuit with 36 continuous copper tracks, each 52 mm long, 0.3 mm wide on a laminated epoxy base-plate. The thickness of the copper was 17 mm. Figure 7 shows a schematic of the corrosion detector.

Figure 7
Corrosion Detector

The corrosion detector was mounted on a non-conducting holder in a vertical position as shown in Figure 8.

Figure 8
Corrosion Detector Mounting

The test consists of setting a tube furnace to a specified temperature with a controlled flow of air running through the tube. A test specimen is downstream from the furnace in the tube. The furnace is moved at a controlled rate over the test specimen causing an even burning rate. A corrosion detector is positioned downstream from the burning sample where it was exposed to the products of combustion. Each test cable sample was 1.0 ± 0.05 inches in length. The test method for the cables was in accordance with ISO DIS 11907-3.


The leakage current apparatus consisted of a tube furnace and exposure chamber as described herein.

The combustion tube furnace is identical to one described in IEC 754-2, and consisted of furnace, silica tube, combustion boat, air supply system, and an exposure chamber for the combustion products. The tube furnace had an inside diameter of 60.3 mm and a heating zone of 300 mm. The test temperature was controlled by an electronic temperature controller. The silica tube was 1600 mm long, 47.5 mm inside diameter and had a wall thickness of 2.75 mm. The silica tube was placed in the tube furnace such that it extended 400 mm from the rear end of the furnace. The rear end of the tube was ground and was fitted with a glass adapter connected to an air supply from a dry, compressed air cylinder. A porcelain combustion boat, 97 mm in length was used to hold the test sample during the test. Each test cable sample was 1.0 ± 0.05 inches in length.

The exposure chamber was made from polymethyl methacrylate (PMMA), with dimensions of 310 x 310 x 340 mm. A stainless steel plate was attached to the inner side of part of the chamber connected to the silica tube. The purpose of the plate was to protect the PMMA surface from flames or hot gases emanating from the silica tube. The top of the exposure chamber had a blowout panel to release excessive pressure. The chamber had a 50 mm opening at the bottom of one of the sides to permit exhaust of combustion products to a smoke abatement system. The exposure chamber was placed 350 mm away from the end of the tube furnace, such that 55 mm of the silica tube protruded inside the chamber. A photograph of the test assembly is shown in Figure 9.

Figure 9
Test Apparatus for Leakage Current Experiments


Commercially available leakage current targets, Model PCBS P/N REMOD were supplied by Precision Prototypes, Inc. The comb pattern spacing was 12.5 mils with all copper surfaces solder tinned and mounted on a printed circuit board laminate 0.062 inches thick. Prior to testing, the target was cleaned in an ultrasonic bath with 75% isopropyl alcohol, followed by rinsing in de-ionized water for 30 seconds, and drying with bottled nitrogen. The interdigitated test pattern used in these tests is shown in Figure 10.

Figure 10
Corrosion Detector for Leakage Current Measurements

After one hour of exposure to combustion products, the targets were removed from the exposure chamber and placed in a controlled humidity chamber. This chamber can accommodate three targets exposed to combustion products, and one target not exposed to combustion products and used as a reference target. The leakage current on the target was measured at a potential of 50 volts D.C. using a Keithley picoammeter/voltage source Model 487. Switching between targets was accomplished with a Keithley Model 7001 switch system. The 30% to 90% controlled %RH environment was produced using a General Eastern humidity generator Model C-1.


The results from the tests conducted on the LAN cables are presented in the following sections.


The ASTM E1354 test results included heat and smoke release data. The peak heat release rate and total heat released presented in Figure 11 and Figure 12. The effective heat of combustion data are presented in Figure 13.

Figure 11
Peak Heat Release Rate

Figure 12
Total Heat Release Results

Figure 13
Effective Heat of Combustion

The peak smoke release rate and total smoke release data are presented in Figure 14 and Figure 15 respectively. For these cables, it may be observed that the peak heat release rate, total heat released, and the effective heat of combustion decrease with improved fire performance as determined by the chemistry of the jacket and insulation materials. Smoke generated from cables A, F, and I (plenum-rated cables) were found to be generally lower in the ASTM E1354 tests as compared to the CMR,CMX and IEC rated cables (G, B, H and K cables). Further, only the CMP rated cables (cables A, F, and I) have combined low heat and smoke release properties.

Figure 14
Peak Smoke Release Rate Results

Figure 15
Total Smoke Release Results

IEC 754-2 (pH) RESULTS

Results obtained from the IEC 754-2 tests are presented in Figure 16.

Figure 16
pH Results from IEC 754-2

Each cable component (J=Jacket and I=Insulation) was tested separately. From Figure 16, it may be observed that the solution pH values are consistent with the presence or absence of halogens in the cable materials (see Tables 3 and 4). It may also be noticed that the method cannot distinguish between high and low performance halogenated materials.


The results included resistance change of the target at the end of 24 hours after smoke exposure. The results are shown in Figure 17.

Figure 17
ISO DIS 11907-3 Results

The lowest percentage corrosion values (R. Corr.) were obtained from one plenum (cable A) and two non-halogen cables (cables H, and K). Higher percentage corrosion values were observed for the four PVC jacketed CMP, CMR, and CMX cables (cables F, I, G, and B).


The ASTM D5485 corrosion results included metal loss of the targets exposed to combustion products after a 24 hour post test exposure at 85%RH. The results are depicted in Figure 18.

Figure 18
ASTM D5485 Results

The lowest metal losses were obtained from two plenum cables (A and F) and two non-halogen cables (H and K). High metal loss values were obtained from a plenum cable (cable I), CMR (cable G) and CMX (cable B) cables. Presumably, high metal losses arise from the attack of chlorine on the copper surface.


The leakage current characteristics measured with an applied voltage of 50V are presented in Figure 19. The leakage current results for non plenum rated cables are shown in Figure 20.

Figure 19
Leakage Current Measurements on Plenum Rated Cable

Figure 20
Leakage Current Measurements from Non-Plenum Rated Cable

The plenum cables (cables A, F, and I) exhibited the lowest RH-dependent leakage currents. Low leakage currents were measured for materials with higher fire performance (FEP, and LSPVC). Non-halogen riser, CMX and IEC (non-halogen) cables exhibited high leakage currents (cables G, B, H, and K). The higher leakage currents measured the non-halogenated cables may have been due to either carbonaceous soot, ionic or acidic species in the smoke.


In this investigation, four smoke corrosivity methods were used to test seven commercial communications cables. The relevance of the methods to predict electronic reliability of digital electronic equipment was discussed, based on the basic mechanisms of digital electronic failure. The results may be summarized as follows:

The IEC 754-2 pH test indirectly measures only one characteristic of fire smoke (pH) and cannot differentiate between high and low performance cable materials (such as FEP that has a heat of combustion of 4,856 kJ/kg versus plasticized PVC with a heat of combustion of 21,500 kJ/kg).

The corrosion data from the ISO DIS 11907-3 test (using the serpentine copper resistance target) are difficult to interpret due to the combined effects of metal loss and circuit bridging. This is due to the formation of conducting regions between the metal lines. These regions are conducting independent of %RH and may result from the deposition of bridging structures of presumably conducting smoke or from formation of conducting corrosion product bridges. Metal loss results in a resistance increase while bridging results in reduction of insulation resistance.

The ASTM D5485 copper metal loss test appears to be more sensitive to the reactions of chloride-forming compounds as compared to fluoride and non-halogen (e.g., metal hydrate) compounds.

The leakage current test employing the interdigitated printed circuit test pattern was able to differentiate high and lower fire performance cables. The smoke effluent from the cables was deposited on the test patterns and the dependence of leakage current on RH was measured. Plenum-rated cables exhibited the lowest leakage currents of the cables tested.


Various US and International test methods are currently available to measure the effect of smoke effluent on copper test specimens. Other methods exist which measure the change in solution pH and/or conductivity with smoke effluent. Both types of methods either claim or are intended to provide an assessment of the "corrosivity" of smoke effluent. This report concludes, through comparative measurements, that none of the existing standard smoke corrosivity methods are adequate in assessing the effects of smoke effluent on digital electronic equipment. This is based on the following reasons:

  1. Based on years of research at Bell Laboratories, Murray Hill, the failure of digital circuits is due to elevated leakage currents (low surface resistivity) and not due to metal loss or contact resistance degradation.

  2. Those methods that measure insulation resistance changes do so with test substrates that are primarily assessing metal loss (ASTM D5485 and ISO DIS 1907-3).

  3. Most methods are material rather than product-based, which prevents the evaluation of smoke arising from the co-combustion of all materials in the product.

  4. Acid gas-based methods are not true corrosivity methods at all (measuring pH and conductivity changes) and fail to measure the effects of particulates in the smoke effluent.


The authors wish to thank Mr. Gabriel Miller and Ms. Tracey Cash of Lucent Technologies for assistance in manuscript preparation and Mr. Bob Backstrom of Underwriter's Laboratories for data acquisition and analysis. The ongoing technical guidance of Dr. Doug Sinclair and Mr. Bob Comizzoli of Lucent Technologies, Bell Laboratories, Murray Hill is also greatly appreciated.


Discussion of Existing/Proposed Methods for Corrosivity Testing

Throughout the world there are a number of test methods that assess the corrosivity of smoke arising from the combustion of plastics and cables. These methods are often cited in cabling specification documents. A list of smoke corrosivity methods is found in Table A-1.

Table A-1 - Smoke Corrosivity Test Methods
Test Standard
Halogen Gas
IEC 754-1
BS 602 Part 1
SAA AS 1660.5.3
Acid Gas
JCS C No. 53/No. 397
CSA 22-2 No. 0.3M
IEC 754-2
BS 602 Part 2
VDE 9472 Part 813
pH, conductivity
pH, conductivity
pH, conductivity
ISO DIS 11907-2
metal loss
ISO DIS 11907-3
metal loss
ASTM D5485
metal loss
ASTM E5.21.7
metal loss


Acronym List

ASTM- American Society of Testing and Materials
BS- British Standard
CM- "Communications Metallic" (copper communications cable for use in
vertical tray spaces as defined by the US National Electric Code)
CMP- "Communications Metallic Plenum" (copper communications cable for
use in plenum spaces as defined by the US National Electric Code)
CMR- "Communications Metallic Riser" (copper communications cable for
use in riser spaces as defined by the US National Electric Code)
CMX- "Communications Metallic" (copper communications cable for use in
residential applications as defined by the US National Electric Code)
CSA- Canadian Standards Association
DIN- Deutsche Industrie Normen
DIS- Draft international standard
FR PO- various nonhalogen, flame retarded polyolefin materials
IEC- International Electrotechnical Commission
ISO- International Standards Organization
JCS- Japanese Cable Standard
LSPVC- PVC materials that have been sufficiently modified with additives to
yield low peak and average smoke characteristics when measured by
the UL 910 cable fire test method
NIBS- National Institute of Building Science
OFN- "Optical Fiber Nonmetallic" (dielectric optical fiber cables for use in
vertical tray spaces as defined by the US National Electric Code)
OFNP- "Optical Fiber Nonmetallic Plenum" (dielectric optical fiber cables for
use in plenum spaces as defined by the US National Electric Code)
OFNR- "Optical Fiber Nonmetallic Riser" (dielectric optical fiber cables for use
in riser spaces as defined by the US National Electric Code)
PVC- various standard, plasticized PVC materials


  1. J. D. Sinclair, L. A. Posta-Kelley, C. J. Weshler, and H. C. Shields, "Deposition of Airborne Sulfate, Nitrate, and Chloride Salts as it relates to Corrosion of Electronics", The Electrochemical Society, Vol. 137, No. 4, 1990.

  2. Frankenthal, R.P., Siconolfi, D.J., and Sinclair, J.D., "Accelerated Life Testing of Electronic Devices by Atmospheric Particles: Why and How", Journal of The Electrochemical Society, Vol. 140, No. 11, November 1993, pp. 3129-3134.

  3. Sinclair, J.D., "Corrosion of Electronics", Journal of The Electrochemical Society, March 1988, pp. 89C-95C.

  4. Comizzoli, R.B., Frankenthal, R.P., Milner, P.C. and Sinclair, J.D., "Corrosion of Electronic Materials and Devices", Science Vol. 234, 1986, pp. 340-345

  5. R. B. Comizzoli, "Surface Conductance on Insulators in presence of Water Vapor", Materials Developments in Microelectronic Packaging: Performance and Reliability, Proceedings of the Fourth Electronic Materials and Processing Congress, 1991, pp. 316-331

  6. J. T. Chapin, et. al., 44th International Wire and Cable Symposium, 1995, pp. 432-437

  7. J. T. Chapin, et. al., 45th International Wire and Cable Symposium, 1996, pp. 184-193
  Friday, July 18, 2003
Copyright ©