Smoke machine setting off smoke detector heads

[Tom Sutton]

On the issue of the smoke machine setting off the smoke detectors in Night Clubs. Does anybody have a solutions, other than throw out the smoke machine, I would imagine it comes up quite often?

[John Webb]

These smoke machines can be a problem - a hirer in my local church hall some years ago used one and I had to go down twice to reset the system. There were no SDs in the hall but there was one in the storeroom off the hall and enough got in there to trigger it. The Conditions of Hire were hurriedly rewritten to ban their use with the threat of a £25 'fine' if the alarm was set off. (No reset until the money was paid!) But this has never had to be enforced, I am pleased to say.

There is however no such easy solution for a night-club where such effects are expected by the users. I've never had to deal with one myself but there are parallels with some large churches when numerous candles may be lit, incense burners in use etc. I understand the technique is to deploy the SDs on a circuit that can be isolated for a timed period (for older systems) or with addressable systems program in a lower sensitivity on the SDs (again the programming of the system can include time limits for this reduction) or use dual sensing comparisons - ie cool smoke from a machine does not cause a significant rise in temperature whereas smoke from a fire will raise the temperature as well.

[Tom Sutton]

Thanks John

Has anybody had experience of Advanced Multi-Criteria Fire Detector's and could they deal with this problem.

Check out http://www.esmagazine.com/articles/95284-fire-detectors-system-sensor and www.systemsensor.com/pdf/COPTIR-performance-report.pdf

Honeywell have one check out http://www.notifier.com/tech/adv-sensing.htm

[Phoenix]

This is a common issue, as you stated.

The generally accepted protocol is to turn off all detectors that might be affected for the period during which smoke is present.  They usually aren't required as the smoke generators are only used when there are hundreds of people around, at least one of whom should notice a fire!  But the means of escape should be re-assessed taking into acount the lack of detection and it must be demonstrated to be satisfactory, particularly from side rooms, alcoves and upper levels. 

There is also the issue of re-instating the system at the end of the night when everyone goes home, this is something that must not get overlooked and I have seen timers to automatically turn the detectors back on at the end of the night - not sure how resilient the timers might prove to be - but probably better than a forgetful nightclub manager.

Stu

[kurnal]

Other systems include the use of pagers for the house manager with hourly messages to rremind them that the detectors are disabled.

[Mike Buckley]

I have come across this problem in some hotels. The solution I suggest is a Permit to Work system covering the use of smoke machines, part of the sign off procedure is to reinstate the fire alarm system.

[Wiz]

Other methods I have seen are a 'device disabled' warning indicator adjacent to the exit used when locking up the premises, and even an interface that prevents the premises intruder alarm from being set until the fire alarm device disablement is re-enabled.

[Tom Sutton]

Thanks guys but I have concerns with isolating the fire alarm and relying on human detection. The reinstating the fire alarm is not my principle objection because as Kurnal Mike and Wiz have shown there are ways to address that problem, but what  if the detectors are located in concealed places. I have had personal experience, as you approach a floor with no signs of fire then in an instance the ceiling collapses and the whole floor is involved, and then imagine this happening in a club with all those party goers, doesn’t bear to think of.
 
I have had other suggestions from the disco kids, capping the offending detector heads or use plastic bags. A more acceptable solution was to put all the offending detector heads on one loop and design it to reduce the sensitivity from the control panel, having two modes, normal and smoke machine mode assuming that is technically feasable.

I think the best solution would be replacing the offending heads with advanced multi-criteria fire detector heads if it was technically feasible and the cost/availability was closer to the standard equipment used today. As they always say what happens in America today will happen here ten years later.

[Wiz]

Tom, we often build-in a simple keyswitch to the system that automatically isolates the required detection and, at the same time, provide various visual and/or audible warning that the system is in a 'partially disabled' state. We do this for the detectors over the stage in many theatres.

There are many detectors (I prefer the Apollo Discovery) that has adjustable sensitivity controllable from the control panel of an analogue addressable system. You can make the detector more or less sensitive (in five steps) and you could use this with a timer and/or the sort of keyswitch system I mention above. This is available now in the UK (and has been for more than a decade) and not just America!

The problem with generating smoke-like stimulus is that it is hard to control it at a level that will definitely not affect your 'adjusted less sensitive' detector. You end up still getting some, if as not as many, unwanted alarms. It is far more practical to temporarily totally isolate automatic detection in the smoke machine affected areas during those times the areas have people in them.

We must also always remember that automatic fire detection is primarily for early detection of a fire to give the occupants enough time to effect escape. If there are people in the area of a fire, they will normally 'detect' it far sooner than any any automatic detector. It doesn't matter that we have isolated the automatic detection in areas containing many people, just as long as we remember to re-enable it when there isn't going to be people about to notice the fire.

In your disaster scenario it would appear that we have a ceiling void where you suggest a fire could eventually cause the ceiling to collapse on to party-goers below. I would suggest that this type of ceiling must be of the type where the people below have no idea of what is happening above it i.e. it is solid. Obviously this type of void wouldn't be affected by smoke machine smoke anyway (no openings in the ceiling) and therefore any automatic detection in the void wouldn't ever need to be isolated.

From what I have read, deaths resulting from fires in nightclubs etc. are rarely to do with the lack of early detection of a fire but more to do with the occupants not being able to evacuate the premises quickly because of inadequate or blocked fire exits.

[Tom Sutton]

Wiz does all the detector heads likely to be affected need to be on one loop or can individual detectors be adjusted from the panel, so by operating a switch, all the offending heads can be switched off and have all those other safe devices you mentioned activated?

[Wiz]

Tom, that the devices in question may be on different loops is not a problem on all the addressable system that I have used.

With the power of the configuration software for modern addressable systems this sort of control is potentially very easy to achieve.

I can't answer for every manufacturer with all their various different configuration software capabilities but, for example, if I were to use an Ampac FireFinder fire panel, one of the many ways I could achieve the same end result would be to put all the devices I wish to control into a GROUP (this is a notional grouping within the software and can include devices on any loop and/or allocated to any zone etc of the system being controlled). I would then write the function as to how that notional GROUP would be affected i.e. DISABLE or MODE CHANGE (different sensitivity) and finally I would decide the CONTROL condition i.e. which causes the function to operate. This could be a key switch being operated or a signal from a timeclock. This is a basic summary of what I would write as a 'cause and Effect' programme in the configuration software.

Tom, I can expand on the above, or write it with more explanation, if you need me to.

[Tom Sutton]

That's fine Wiz and thank you very much. The reference to America was about the System Sensor Advanced Multi-Criteria Fire Detector and the Honeywell Notifier IntelliQuad I could not find many details about them in the UK especially about the cost.

[Wiz]

I Googled them Tom, and I can see why you are interested, they seem very high-tech and potentially useful.
The manufacturers say that both versions sense four different potential elements (smoke, heat, IR & CO) of a real fire, and need valid 'fire' information from at least two of these elements to give an alarm condition. I have not heard of anything similar available in the UK yet. Both companies have a presence in the UK so surely it is just a matter of time. If they work well, I'm sure all manufacturers will have their own version, in time.
p.s. - I wonder if both 'versions' you highlight are actually the same manufacture being marketed under different names

[kurnal]

Doesnt the Gent Vigilon / S Quad offer similar functionality Dr Wiz?

As i understand it each of the S quad sensors (smoke heat CO) have a range of states that can be programmed and the C&E can be configured to give a prealarm or fire signal on a single or combination of the sensors.

Or maybe I have it wrong?

[Tom Sutton]

Notifier intelli-quad is one of Honeywell's which have bases in the UK and for the Notifier I think it is in West Sussex. The do have systems installed in UK Liverpool Hard Days Night Hotel and I think there is one in Halifax and the Festival Hall in London. Check out http://www.notifierfiresystems.co.uk/case_studies.asp?id=465 for Liverpool.

The System Sensor Advanced Multi-Criteria Fire/CO (AMCF/CO) detector appear to be based only in America.

Check out http://www.systemsensor.com/multi/ and watch the demo.