Fire Alarm System Interaction with Energy Efficient Lighting

[tdunruh]

We have changed out incandescent lamps with two-lamp T-8 with electronic ballasts in several buildings. Now the fire alarm system keeps sounding false messages. We get nearly all messages. The fire system sensors are connected via a twisted pair and are directly addressable ("smart System&quot. The fire panels are fed 120 V power with three wires in metallic conduit. We have installed an isolation transformer for the fire system to obtain a clean Neutral-Ground voltage, with no improvement. Lighting ballast manufacturer indicates no history of problems, as does fire alarm company. We have installed millions of square feet of this retrofit without any issues. Need help 1) similar problems and associated solutions; 2) troubleshooting methods to find problems; 3) measurement methods. Thanks

 

[busbar]

This info may be a bit basic, but:

http://tis.eh.doe.gov/web/oeaf/lessons_learned/ons/sn9901.pdf

Hopefully this has not been a problem in your facilities:

http://www.hanford.gov/lessons/sitell/ll98/199805.htm

 

[Lewish]

Two things to note:
1. Your sensors are "smart" and therefore are susceptible to external effects.
2. Electronic ballasts put out significant high-frequency energy in their close proximity.

Thus, even though your wiring is twisted pair, if it runs too close to a fixture with an electronic ballast, you might be picking up enough energy to falsely trigger the sensor on that pair.

Try putting a 0.01uF capacitor across the signal pair right at the sensor. This will kill off the high frequency energy that the pair might pick up. You may also need a capacitor at the panel.

Let us know if this helps.

 

[mpparent]

Is your fire alarm wiring "free" and shielded or is it in conduit??

Mike

 

[dpc]

It sounds as though the fire detector wiring is acting as an antenna, picking up radiated noise from the electronic ballasts. As mpparent notes above, you may need to look at the noise isolation of the fire alarm signal wiring. Shielded twisted pair will help, and metal conduit would help more.

If the frequency of the fire alarm signals isn't too high, you might try using some ferrite beads or toroids on the fire alarm signal wiring where it enters the fire alarm panel.

I would lean on the fire alarm equipment supplier pretty hard. They probably have had the problem before and they are the ones who need to help you come up with a solution. I wouldn't count on much help from the ballast supplier.

This could be a tricky one to sort out. Good luck, and let us know what you figure out.

 

[kuwait]

HI TDUNRUH,
THIS SURE IS A SEPTEMBER 11 PROBLEM FOR YOU .DID YOU CHECK IF THE SHIELDING IS EARTHED AT BOTH ENDS - THEN THE COUPLING CAPACITORS TO EARTH MIGHT CREATE A PROBLEM OR IN SOME JOINTS.

 

[busbar]

This isn’t easy advice, but please tread carefully. Capacitor addition [or any mods not specifically sanctioned by the alarm producer] to a fire-alarm circuit or system represents a potential compromise in successful operation—even if you find it reduces or eliminates the false annunciations. This could affect the necessary risetime of the "good" signal and create more problems than it cures in non-reporting of sensor status. I would not make the decision alone—you know/appreciate the legal aspects in your particular setting.

Use of a fast, differential-mode/double-insulated oscilloscope to look at the supervised loops with lighting on/off might be revealing.

Equipment in repeated “crying wolf” situations breeds contempt for the critical system, but you also need to protect your long-term source of income. I would get any requests or approvals clearly understood, [in both directions] and signed off by (at least two) superiors.

 

[busbar]

As a last-ditch effort, you may want to contact the listing/labeling agency [UL, FM, NRTL…] and directly, formally and most seriously ask them if the particular alarm system has been evaluated for what you perceive to be a very common and prevalent source of RFI. I would do my best to find the technical contact within the listing organization that deals regularly with that specific standard and related components they are contracted to evaluate. Search ul.com or the applicable NRTL site for salient information prior to initial contact.

DO NOT let the NRTL contact excuse the matter by telling you that you and they "…are not bound by contract or purchase order with your firm, so call the alarm manufacturer." Be warned—they can feign secrecy and act extraordinarily compartmentalized about any dealings with their original client; in this case, the alarm-equipment manufacturer. At times, they also may try to shield their technical people from dealings with “the general public;” in this case, you.

It won’t be easy, but you may be able to get municipal building inspection or fire-prevention officials in the loop. If there was a contractor involved in the system, drag them into it, too. CC everything to everybody with detailed transmittal cover pages. Be bloodthirsty about a complete and accurate signup sheet at every meeting, and forward in print your best recollection of any meeting content and phone conversations.

In the case of nonresolution, be damn sure you make clear to your associates that the manufacturer (and possibly NRTL) are gold-plated, blue ribbon, scuzzball operations [but maybe not necessarily in writing.]

 

[tdunruh]

Thanks, all for the suggestions. I like the double-insulated scope thought, and that was my next plan as well. Right now, we are doing the obvious. Isolation transformer to the 120V power supply and moving sesors away from lighting. Some are as close as 1-2 inches from the ballast.

 

[plcsavvy]

Make sure that the shielded cable is grounded at one end only. grounding both ends can result in a ground loop which can cause you major problems. PLCSAVVY