Full-Scale UK Fire Tests of LAN Data Communications Cables
Used in Concealed-Space Applications, Rev. 1
This paper evaluates the fire performance of commercial data communications cables (and materials) in:
- Full-scale tests simulating current UK installation practices.
- Intermediate-scale Steiner Tunnels (NFPA 262-1990 or UL-910) used in Canada, Japan, Mexico, and the USA.2
- Small-scale Tube-Furnace tests in UK apparatus designed to evaluate fire effluent.
In buildings worldwide, the installation of combustibles in concealed spaces is an important concern because of the potential for undetected spread of fire and smoke throughout the building.1
A series of large devastating fires have recently occurred in buildings involving combustibles in concealed spaces (see Reference 11 for listing of fires).
In the UK, the proliferation of local area networks (LANs) in buildings can result in heavy concentrations of communications cables in concealed spaces.
Common LAN cable constructions in the UK include compounded PVC or compounded polyolefin (non-halogen) sheathing over polyethylene or polypropylene based insulations on the copper conductors. Non-halogen cable constructions are often designated LSZH (for Low Smoke Zero Halogen).7
Typically, LSZH polyolefin base polymers have high fuel loads and are highly combustible. Therefore, they are compounded with metal-hydrate fillers to delay ignitability until the water of hydration is exhausted. Vigorous combustion can then result.
In contrast, construction products used in concealed spaces are usually required to be non-combustible or limited combustible.
Full-scale fire tests simulating UK installation practices were conducted at BRE/FRS (Building Research Establishment/Fire Research Station) Cardington. The test program was designed to support the development of new performance data for hazard assessments, international fire test protocols and fire safety engineering.
The earlier tests were carried out in a burn-room/concealed-space re-burnable structure with a nominal 1-megawatt wood crib source-fire. The later tests used a nominal 1-megawatt gas burner. Fire scenarios, ventilation conditions, and LAN cable designs and configurations were varied.
Fire performance measurements included mass loss, pressure differentials, lateral flame spread, heat flux, vertical temperature profiles, smoke opacity, heat release, CO and CO2 generation and O2 depletion. Tests were documented with still and video photography in both IR and white light.
Most data were logged electronically (about every 10 seconds) for real-time on-line graphical monitoring, and then stored in spreadsheet formats to facilitate statistical analysis and computer modeling.
LSZH cables that pass IEC 332-1 and IEC 332-3 ignited readily and burned the full length of the concealed space configuration. A large fire-ball developed on the horizontal cable ladder and a pool of fire formed on the suspended ceiling beneath the cable ladder in the concealed space. Ceiling tiles often fell out during tests.
Under the same full-scale test conditions, LAN cables that pass NFPA 262 (Steiner Tunnel) test criteria showed no sign of flame spread and generated little smoke.2 These cables are designated CMP in North America, Canada, Mexico and Asia (see Experimental section below).
Other related tests were conducted in the intermediate-scale Steiner Tunnel and in the small-scale tube furnace/smoke box apparatus developed for the British Cable Makers Confederation (BCMC).
The cable fire performance (flame, spread, smoke and heat release) data from the Steiner Tunnel was relatable to the BRE/FRS full-scale simulations. The data from the IEC 332-1 and IEC 332-3 tests was not relatable.
Steiner Tunnel facilities have now been installed in Europe by BRE/FRS and by the LPC (Loss Prevention Council) to help companies develop higher fire performance cables intended for use in horizontal concealed spaces.
Surprisingly, dense dark smoke and forceful explosions occurred with just one gram of polyolefin cable materials from the LSZH cables in the BCMC small-scale apparatus tests. The CMP cable materials produced very little smoke and no explosions in the same small-scale BCMC tests. These explosions may relate to flashover phenomena.
Worldwide, ceiling and floor concealed spaces (voids) in commercial buildings are increasingly being used for utilities and ventilation. This design approach helps maximize flexibility in meeting changing tenant churn requirements.
Many new and refurbished buildings use concealed spaces to contain communication and power cables, plumbing, fire detection and suppression systems and similar mechanical and electrical services. Sometimes the same space is used for environmental air handling. Installing services in concealed spaces provides convenient access, easy alterations, lower construction costs and energy conservation for heating, ventilation and air conditioning (HVAC).
If these concealed spaces contain combustibles, they are potential sites for the undetected generation and movement of fire and smoke.11
Historically, this has meant that construction products exposed in the concealed spaces have been required to be (a) fire partitioned, or be (b) very low in fuel-load and combustibility, or be (c) protected by either fire resistant coverings or fire extinguishment systems. In the past, these options have been found to be acceptable.1
However, with the growing use of concealed spaces for cabling, new fire-path and fire-load problems are emerging (see Figure 26). LANs are growing at 25%/year in many regions and LAN cabling systems are being replaced every 3 to 5 years as personal computers (PC´s) become faster and more powerful. As a result, many concealed spaces are becoming filled with multiple generations of data communications cables with low or unknown aggregate fire performance.
Commercial cables were selected that met a range of available fire performance.
The tests were conducted with commercially obtained 4-pair sheathed unshielded twisted pair (UTP) communications data cables. These cables are typically installed in horizontal concealed spaces to connect PCs to LANs in Structured Wiring Systems. According to Lucent/AT&T surveys, the 200 seven-meter lengths of cable used in each test represent about one generation of cable in a typical one-floor open-plan office layout. More than one generation of installed cable is often present in actual buildings.
CMX is a low fire performance communications cable used in the USA that is required to be installed in protective metallic conduit in plenums (concealed spaces).3 CMX/T is CMX cable installed in capped steel trunking, sometimes used for communication cables in the UK.
CMP is a high fire performance communications cable (per NFPA 262 or UL-910) used in the USA in plenum cavity voids without requirements for protective metallic conduits or trunking. Low fuel-load fluoropolymers like Teflon® FEP are commonly used for insulations in CMP rated cables ( Figure 21).
LSZH communications cable per IEC 332-1 and IEC 332-3 is used in the UK and some European countries, often without trunking. High fuel-load polyolefins and flame retardant polyolefins (FRPE) are often used in LSZH cables (see Figure 21).
Full-Scale Tests (BRE/FRS Cardington, UK)
The full-scale test rig was a 7.4m x 5.7m x 4m high concrete block burn-room/ceiling-void re-burnable structure with a 2-hour fire rated suspended ceiling. The above-ceiling concealed space is 1m deep ( Figure 1).
The source fire was created with a nominal 150kg. crib of kiln-dried pinus sylvestris with 20% moisture. This fuel was stored in conditioning rooms until used. One hundred sticks, 60mm x 60mm x 1m, were stacked 10 per row in 10 rows to generate a nominal 1-megawatt fire over 30 minutes.
The 1-megawatt intensity was chosen to simulate a fire in an office workstation. Recent National Institute of Standards and Technology (NIST) research indicates such workstation fire energies can typically range from 2 to 6 megawatts.
The wood crib calibrations in the BRE/FRS full-scale calorimeter showed very linear mass loss after about 8 minutes from ignition. However, the energy output ramped up sharply until about 10 minutes into the burn and then plateaued until about 20 minutes. The energy output then ramped up again from 20 minutes until about 30 minutes.
The air extraction system was capable of 4.5m3/sec. The hot gas and smoke from the crib fire entered the cavity void through a breach (hole) in the suspended ceiling directly over the crib and were then extracted through vents at the far end of the overhead concealed space.
Cables were supported on a steel ladder 7.2m long by 0.38m wide. The ladder was located midway between the suspended ceiling and the structural ceiling of the test rig.
Thermocouples (TCs) arrayed vertically in the burn room provided data for mapping temperature profiles. Other thermocouples arrayed horizontally and vertically in the void space provided data for mapping both temperature profiles and flame spread.
The baseline performance of the source fire and test apparatus without cable combustibles was determined by operating the system with an insulated board on the ladder in place of cable.
Additional test apparatus details are presented in an Interflam ´96 paper entitled "Cables Fires In Concealed Spaces . . . A Full-Scale Test Facility For Standards Development" by P. Fardell, S. Rogers, D. Smit, R. Colwell, BRE/Fire Research Station, UK.
Full-Scale Results (BRE/FRS Cardington, UK)
The CMX/T and CMP cables exhibited very low smoke optical densities and heat release. No burning was visible beyond the suspended ceiling breach area.
The CMX and LSZH cables released comparatively large amounts of heat. These cables burned the full length of the plenum space and developed large fire-balls on the ladder and large pool fires on the suspended ceiling. The LSZH smoke density was higher than that of either CMX/T or CMP, but notably lower than CMX. In addition, LSZH cables also produced more smoke and CO than expected. CMP cables had the lowest smoke and CO.
Downstream peak temperatures along the entire length of the ladder exceeded 800°C when exposed LSZH cables burned (see Figure 18). CMX downstream peak temperatures averaged slightly under 775°C. Recent British Steel Technical fire tests at BRE/FRS indicate that temperatures exceeding 800°C can allow loaded structural steel beams to severely deform.8
CMP downstream peak temperatures averaged less than 335°C. Baseline downstream peak temperatures with no combustibles in the plenum void averaged slightly under 325°C in calibration tests.
Steiner Tunnel Tests (Underwriters Laboratories)
The intermediate-scale horizontal tray tests were conducted per NFPA 262-1990, Test for Fire and Smoke Characteristics of Wires and Cables2 (UL-910) which is a modification of the 7.6m Steiner Tunnel3 used for cables and construction products.
In this method, an array of cables 0.3m wide and 7.6m long is strung beneath the ceiling on a ladder extending for the length of the tunnel. A diffusion gas burner flame of 90kW (nominal) engulfs up to 1.5m of the cable at the far upstream end. Ventilation is supplied from the burner end at a linear flow rate of 73m per minute.
Cables for plenum use are required to spread flame no more than 1.5m past the burner flame tip and produce a peak smoke optical density no greater than 0.5 (logIo/I) and an average smoke optical density no greater than 0.15 (logIo/I) in the exhaust duct.1,2
These performance requirements were selected so that approved exposed cable would compare favorably with cable protected in metal conduit when exposed to the same test conditions.4,5,6 Additional measures of O2 consumption made to calculate heat release are not mandatory today.
All results were comparable to the full-scale results. Further thermal and statistical analyses are in progress.
The fire performance of the IEC 332-1 and IEC 332-3 rated LSZH cables were similar to each other in both the BRE/FRS full-scale tests and in the Steiner Tunnel (with relatable fuel-limited combustion conditions).
Tube Furnace Tests (Underwriters Laboratories)
Dense dark smoke and forceful double explosions occurred with just one gram of polyolefin insulation material from LSZH cables in small-scale tests utilizing the tube furnace/smoke box apparatus developed for the British Cable Makers Confederation (see Figure 23). These tests were being run to assess fire effluent damage on computer micro-circuits when the explosions unexpectedly occurred with LSZH cable materials.
Fluoropolymer insulation materials from the high fire performance CMP cables, tested under the same conditions, did not explode and did not generate dense dark smoke.
Further study is underway to see if these explosions with LSZH cable materials relate to similar explosive flashback, backdraft and flashover phenomena encountered in large building fires.
Combustion Toxicity Considerations
The comparative carbon monoxide (CO) generation shown in Figures 5 and 6 is one of many evaluations needed to assess toxic hazards over a range of fire scenarios.
Figure 24 shows a comparison of the bio-confirmed toxicity of the fluoropolymer category of electrical materials typically used in CMP cables versus the polyolefin category of electrical materials typically used in LSZH cables.
Similar data for PVC is shown in Figure 24a. This data was developed by the National Electrical Manufacturers Association (NEMA) using the NYS/NEMA protocol that subjects products to a continuous range of ramped fire scenarios.10
There was no significantly discernible difference in toxicity (LC50) between the fluoropolymer and the polyolefin categories. This is likely due to the tendency for polyolefins to generate large amounts of CO, especially in situations where they burn rapidly and quickly reduce available oxygen (i.e., concealed spaces). Toxicity information should be used as part of a relevant total fire hazard assessment where parameters such as ignitability, fuel load, heat release, flame spread and smoke opacity are also considered.
- The data obtained in the Steiner Tunnel was relatable to the full-scale BRE/FRS simulations.
- The fire performance of the exposed CMP cable was comparable to CMX cable in metal trunking (CMX/T) in the BRE/FRS full-scale and Steiner Tunnel tests.6
- The flame spread, heat release and smoke opacity results for exposed CMP cable were significantly lower than results for exposed LSZH and CMX cable.
- For LSZH cables, the high temperatures, high heat release rates, flame spread, fire-balls, pool fires and tube furnace explosions were unexpected considering their extensive use in concealed spaces in commercial buildings.
- There was no discernible difference in toxicity between the fluoropolymer category of materials used in CMP cables versus polyolefin and PVC categories of materials used in the LSZH cables per NEMA data.10
These and other cables are being evaluated with variations in source fires, ventilation, installation configuration, fuel loads, fire loads and fire scenarios. Data for concealed space fire modeling is also being developed.
AUTHORS AND PRESENTERS
Dr. J. Thomas Chapin and T. C. Tan, Lucent Technologies
Arthur Willis and Dr. Keith Pye, BICC Brand-Rex
James R. Hoover and Loren M. Caudill, DuPont
- F. Clark, J. Hoover, L. Caudill, A. Fine, A. Parnell & G. Butcher, "Characterizing Fire Hazard of Unprotected Cables in Over-Ceiling Voids Used for Ventilation", Interflam '93, page 259, 1993.
- "Test for Fire and Smoke Characteristics of Wires Cables", NFPA 262-1990, National Fire Protection Association, Quincy, Mass., USA, 1990.
- cf. "Standard Test for Surface Burning Characteristics of Materials", ASTM, Philadelphia, Pa, USA, 1987.
- L. Przybyla, E. J. Coffey, S. Kaufman, M. Yocum, J. Reed and D. Allen, "Low Smoke and Flame Spread Cables", Journal of Fire and Flammability 12, 177 (1987).
- S. Kaufman and M. Yocum, "Behavior of Fire-Resistant Communications Cables in Large-Scale Fires", Plastics and Rubber: Materials and Applications, November, 1979, 149.
- L. Przybyla, E. Coffey, S. Kaufman, M. Yocum, J. Reed, D. Allen, "Low-Smoke and Flame Spread Cables", The 28th International Wire and Cable Symposium Proceedings, Cherry Hill, N.J., USA, 1979.
- See Figure 20 for a relative ranking of fire performance among various fire tests, ref NFPA 70 and UL reports.
- Kirby, B.R., "British Steel Technical European Fire Test Programme", Fire, Static and Dynamic Tests of Building Structures; Armer, G. & O'Dell, T., (1997), pp. 111-126, Conference Proceedings.
- Hoover, J., "Full-Scale Fire Research on LAN Cables in Concealed Spaces", BICSI Presentation Summaries, January 1997, pp. 3-16, Winter Conference, Orlando, FL.
- National Electrical Manufacturers Association, 1987, "Registration Categories of the National Electrical Manufacturers Association for Compliance with the New York State Uniform Fire Prevention and Building Code", R. Anderson, P. Kopf, pub., Arthur D. Little, Inc.
- These fires included the Dusseldorf Airport in April '96, the Paris Credit Lyonnais Bank in May '96, the New York Rockefeller Center in October '96, the Hong Kong Golden Mile, Garly Building in November '96, the Bangkok Presidential Tower (36-story office complex) in February '97, among many others.
A special thanks is due to the following people who have provided valuable professioanl advice over the course of the work:
A. Fine, J. Hardiman and S. Kaufman, AT&T*, Norcross, Ga., USA
L. Przybyla, P. Gandhi, T. Ebert, W. Metes, R. Backstrom and J. Resing, UL, Northbrook, Ill., USA
A. Parnell and G. Butcher, Firecheck Consultants, Tonbridge, Kent, UK
D. Woolley, P. Fardell, J. Rowley, S. Rogers, R. Colwell, R. Mallows, D. Smit, and S. Vollam, BRE/FRS, UK
M. K. James, M. Cardona, J. Walnock and D. Benson, DuPont, Wilmington, Del., USA
E. Champney, EC Associates, Wilmington, Del., USA
R. Gottwald, SPI, Washington, DC
* Following this work certain AT&T personnel have become associated with Lucent Technologies.
DuPont (U.K.) Limited
Herts. HP2 7DP
Tel. (01442) 21 85 00
Telex. 825 713 DUPONT G
Telefax (01442) 24 94 63
Lucent Technologies (U.K.)
Global Commercial Markets
101 Wigmore St.
London WIH 9AB
BICC Brand-Rex Limited
Viewfield Industrial Estate
Fife KY6 2RS
The information set forth herein is furnished free of charge and is based on technical data that is believed to be reliable. It is intended for use by persons having technical skill, at their own discretion and risk. The handling precaution information contained herein is given with the understanding that those using it will satisfy themselves that their particular conditions of use present no health or safety hazards. Because conditions of product use are outside our control, we make no warranties, express or implied, and assume no liability in connection with any use of this information. As with any material, evaluation of any compound under end-use conditions prior to specification is essential. Nothing herein is to be taken as a license to operate under recommendation to infringe any patents.