According to Section 12.2.4.4, NFPA 72, 2010, "All fire alarm systems shall test free of grounds. Exception: Parts of circuits or equipment that are intentionally and permanently grounded to provide ground-fault detection, noise suppression, emergency ground signaling, and circuit protection grounding shall be permitted."
Real-Life Troubleshooting Tale
A few weeks ago I was asked by our electrical division to troubleshoot an apparent ground-fault problem with a fairly new fire alarm system installed in a dormitory. The vendor, Siemens, had already dispatched a fire alarm technician to investigate the problem. He concluded that the problem was, indeed, a legitimate ground-fault. Because the conduit and cables were installed by our electrical division, it was our company's responsibility to find the problem and fix it, not Siemens.
The dormitory is made up of four floors equipped with analog addressable automatic smoke detection, manual fire alarm boxes, sprinkler flows and tampers, and multiple air handlers - four per floor (16 altogether) equipped with duct-type smoke detectors. The cable our electricians installed was power-limited fire alarm riser (FPLR) in conduit. A local Siemens representative was there to assist me.
The first thing I did was verify a ground-fault truly existed. I examined the panel and its LCD readout. The panel not only identified the problem as a ground-fault, it also identified the cable as a signaling line circuit (SLC) - as well as the exact conductor within the cable. In this case, it was the negative side of the return wires.
The SLC was a Class A circuit with a feed and return. Even with all of this going for me, the prospect of locating a single ground-fault with all of this conduit, metal boxes, and numerous addressable devices was not pleasant. I quickly formulated a plan of attack and, with the help of the Siemens' rep, within two hours I was able to localize the problem and fix it.
Finding, Solving the Problem
A ground-fault can come from a piece of metal that has penetrated the insulation of a conductor within a cable. It can occur when a bare wire comes in contact with metal that’s bonded to an earth ground.
As stated earlier, the first thing I did was verify that the problem existed. In this case, I disconnected the return side of the SLC from the FACP. Using a standard, digital volt-ohm meter, I placed one of the leads to the return circuit's negative wire and the other to chassis ground. I set the meter to ohms and measured the resistance. It fluctuated between 80M ohms and 135M ohms.
Because of the cyclic fluctuations in resistance, I began to contemplate the possibility that the problem might be physically near one of the HVAC units, probably in a duct-type smoke detector. Because of the intense heat we had during the day, which began about the time the ground-fault appeared, this was well within the realm of possibility. I also contemplated the possibility of condensation within some of the conduits that feed the sprinkler system as well as the duct-smoke detectors.
I physically traced the return wire through the various conduits on the lower level. The first general area involved was the sprinkler room where there are flow and tamper switches and multiple metal junction boxes.
Working my way along the conduit, I used the handle of a large screwdriver to strike the conduit, inducing vibration into all nearby junctions. When I reached the vicinity of the sprinkler room, the Siemens' rep indicated that the fluctuating readings on the meter had changed. This is when I knew we were close.
The fluctuating resistance had disappeared, and now the meter indicated a voltage existed between the negative conductor and the panel's chassis ground, which meant the ground-fault had cleared. I removed all the box, flow and tamper covers, and I repositioned wires inside of each. I also checked to assure that all wire nuts were on tight and that individual conductors were not stripped beyond the lower edge of the wire nuts.
After some additional effort, it was decided that the ground-problem was gone, at least for now, and it has not yet returned. But if it should, I am prepared to execute an even more aggressive method of circuit tracing.
Al Colombo is an award-winning writer who has covered electronic security and life safety since 1986. Visit his Web site at www.firenetonline.com.