The following text is taken from an email dated 10/19/04 from Glenn Birket, P.E. to part of the membership of the NFPA Technical Committee on Special Effects. In his email, Mr. Birket relates a story supporting a change to NFPA 1126 allowing non-isolated power supplies, and requiring that both sides of the match circuit be broken. It was written after he accepted that the committee would not be willing to add the requirement for breaking both sides of the match circuit for this revision cycle.
An Unintentional Pyro Ignition Experience
Mr. Birket writes: Why do I pursue this issue despite the skepticism of so many? One respondent who called me rather than emailing, John Noonan, asked “Why you are passionate about this? You must feel like you are pushing on a string.” He’s right. This is why I am passionate about this:
Eleven years ago, at Treasure Island in Las Vegas, Birket Engineering, Inc. installed a pyro system, the first that we had designed and installed entirely on our own. It was a simple system using relays controlled by a PLC , but it was a widely distributed system over a large show with a lot of matches. We designed and installed the system just as we had seen it done on other shows designed by the best in the business. We also read the code, and followed it, including the requirement for an isolated power supply.
Sometime after the show opened two actors were injured as they stood near a charge that wasn’t supposed to go off for another couple of minutes. As I understood it, one was burned on his arm; the other lost his vision temporarily. Both made a full recovery. They were very lucky. We were very lucky. Our industry was very lucky.
I got on an airplane and spent most of two weeks considering every possible reason why this could have happened. Of course I started looking in the wrong places: software and controller problems, then EMI, then galvanic effects. Gradually I concluded that it did not happen because of a legitimate command from the system. Then I gradually concluded that there existed no other sufficient source of energy such as EMI or a naturally occurring voltaic cell. That only left our own firing supply – but how, since I had concluded that the match was a shunted at the time that it ignited. So, I started taking our firing circuit apart one ohm at a time. First, by redrawing the circuit showing the resistance of the wire I realized that it is quite possible to ignite a shunted match if there is enough current and voltage available, depending on how close the shunt is to the match, etc. Then, finally, the “two ground fault” thought hit me between the eyes. I realized that the requirement for an isolated supply allows ground faults to exist without penalty. So they do, at times. Given enough time, the right two ground faults will be present, allowing a deliberately ignited match to take another match with it, unexpectedly. Careful study of the circuit shows that there are other possible faults such as “fire on arm” and “spontaneous fire”.
Having reached this awareness, all I had to do was to walk the site with an ohm meter while shaking and pulling on wires to find the intermittent ground faults. Both ground faults were clearly a result of changes the local staff had innocently made using available wire. One was wire with low temperature insulation run through a conduit next to a very hot gas flame effect. The other was zip cord across an expanded metal surface that could be walked on. (I’m told that the blocking of the show had also changed, placing the actors closer to the errant pyro device than originally intended.) We already knew that the errant pyro ignited at the time as another legitimate ignition of a device in an other area. I found intermittent ground faults in the temporary wiring on the positive side of both of these match circuits. I was then able to convincingly explain and conclusively demonstrate to Thaine Morris, Jules Lauve, Bob Bauer, and June Fields exactly what had happened. As a solution, we implemented double pole contacts in the areas of concern. We would have liked to have grounded the supply but the code did not permit that. Everyone was satisfied that the problem had been identified, and a solution implemented.
In the weeks that followed, being unable to ground the supply, I attempted to implement a ground fault detection system for that show. I spent several days monitoring ground faults. I learned that high resistance ground faults are always present throughout such a system, and they combine in parallel into the net effect of a low resistance ground fault. I also found that from show to show it was not unusual to find one or more points in the system temporarily grounded, probably related to how to pyro was loaded for that show. It simply is not possible to control what happens beyond the match terminals after the designer/installer leaves the site. It should be noted that Treasure Island may be a particularly problematic show in this regard because of the water and temperature cycling, but in other ways it is probably typical of other large permanently installed distributed firing systems.
With regard to ground fault monitoring I concluded that while the occasional low resistance ground fault is often distinguishable, often it is not due to the presence of all of the other high resistance ground faults that exist because one side of the system is wired in common all the way out to the match. A GFCI outlet don’t help of course because it is on the wrong side of an isolation transformer. Thus the only near-certain solution is double pole relays (or another approach, including solid state, that breaks both lines to the match). Being unable to ground the supply, we went back and installed more double pole relays. On future systems we implemented individual isolated one-per-match firing sources, and continuous testing for grounds and cross-wire conditions at each match.
I have wondered since why the industry has been slow to learn from this incident, and why I encounter so much resistance when I explain it. I have to remind myself that it took me two weeks to find the problem on this show, and that understanding it requires a solid grasp of several electrical concepts for which most people have no good use. It is also too easy to attribute an unintended ignition to EMI or some other mysterious cause. Only I have had the benefit of a first hand experience, a successful and exhaustive field investigation, and then hashing this over and over in my head for eleven years.
On the other side of the fence is that prevailing (but erroneous in this case) feeling that “isolated” means “safe” because it is an NEC solution to a shock hazard. Finally, there is one other factor that I believe to be involved. The industry has developed very responsible attitude that says “it’s pyro, it can do bad things, so you have to stay away from it or somebody might get hurt.” Unfortunately I’ve heard that idea submitted as the “theory ” for an incident when science does not quickly present an easily digestible theory about an unintentional ignition. There is no NTSB of pyro. Had it not been for the support of the four aforementioned individuals, and my willingness at the time to do nothing else until we understood exactly what happened, the incident at Treasure Island would have been attributed to the “well, pyro can hurt people” theory and the solution would have been to “change something until we think it won’t happen again” and to change the show blocking so that if it did happen again there would be no one nearby – until complacency set in again.
Later John Noonan wrote, “The skeptic would say that you are trying to change the rules to match your unique product. Personally, knowing you, I would guess that you are just trying to make the world a little safer.”
A completely fair question is, “Since Birket is the only one pushing this, does Birket have something to gain?” Possibly, but realistically, not much, and it is nothing that is not available to everyone else in the business. Is it a reason not to call attention to an important safety issue? No. In fact, I’ll explain exactly what we have so that others can have it too.
In the months and years after the Treasure Island incident we took it upon ourselves to develop a firing system that tests every match terminal for continuity to ground and to other match terminals using the same current-limited source used to test match continuity. We were counseled that it was patentable, but we did not feel it appropriate to pursue exclusivity on an important safety concept. So, go for it – it is a good idea. Our system marks each and every fault detected with an LED. We also use a isolated and pre-charged capacitor as the firing source for each match circuit. No possibility of cross-talk between circuits. And, we keep every match shunted until the very moment of intended ignition.
In other words, you bet, we took what we learned and designed a system to make sure that an unintentional ignition would never happen again on a system of our design. We feel that anything else would have been an abdication of responsibility after that event at Treasure Island . There are now several installations of our system around the world but they probably represent less than two percent of the pyro firing systems out there. They represent some larger, but still small, percentage of the fixed installations. Pyro is a relatively small part of what we do, and we are not in the mainstream of the pyro community. We have never done a temporary installation, and do not expect to have that opportunity. The likely market for our system will be for use in close proximity to personnel. The existing, trusted and popular systems have a lot of market momentum in this industry, and they have an excellent safety record. We would be naïve to think that we could push a change to the standard that would suddenly make our system the one to buy. By pressing this issue however, we have raised the industry’s awareness on an important safety issue, and that is the right thing to do. It is also an important function of this committee.