DMX pyro

[Dazedm]

Hi all,

 

I'm just about to start my final year at university and I've started thinking about my independent project. I'm hoping to make a pyro ignition unit controlled by DMX.

 

Many people have rejected the idea outright due to safety concerns but the system would have several safety features such as:

• A dead mans handle and key switch
• Multiple channels (currently 3) at specific values for each effect
• Specialist universe for pyrotechnics

The unit could be placed close to the pyro (10cm or so) to reduce the length of the wire from the control unit reducing the risk of EM interference, the wire resistance and simplifying wiring.

It would only accept a specific value out of the possible 511 for each channel which means it would be accurate to 0.2% and the likelihood of EM interference of hitting that specific value on three separate channels is very small.

 

As with any pyrotechnic system it must be controlled and checked by a human but this system would simplify the firing process. They would already be cued into a lighting desk so that the technician responsible only has to worry about if the area is safe to fire or not and reducing the risk of firing the wrong effect.

 

If anyone has any guidance, suggestions or criticisms I would be most grateful.

 

Cheers

 

Dan

[Andrew C]

If anyone has any guidance, suggestions or criticisms I would be most grateful.

 

Advice; do a VERY thorough search of the archives here. We have gone into a lot of detailed discussion in the past. And for the rest of you, please keep this on track, and try not to go over old, contentious, ground again.

[timsabre]

It's not clear how you propose controlling this and how your "dead mans handle" fits in with "they would be already cued into a lighting desk".

 

Are you building a dedicated firing unit? (in which case why use DMX rather than something a bit more robust)

Or are you having some type of "go/inhibit" unit for someone who is standing near the pyro, which will allow the lighting desk cue to fire it only if safe?

 

The basic problem with DMX is that it has no error detection, so even with your three channels there is a small possibility of firing the device if enough random noise gets onto the line... by definition noise contains all possible digital codes, so given enough noise, you can generate a valid signal.

 

A small illustration from a previous life...

I used to work on the railway at the time they were introducing solid state signalling (SSI), this had all the safety checking back at the signal box and used two identical digital data links to send simple instructions to the trackside to move points, change signals etc. Because of the danger involved in accidentally moving points or giving the wrong signal, the data links had very robust Hamming error checking/detection on them. However, despite this there was an incident near Berwick-on-Tweed where an electrical fault caused a high level of noise on the data links which managed to fool the error detection into moving a set of points under a freight train, derailing it. And that was WITH error detection.

[mark_s]

I also wouldn't feel at all comfortable about pyro being controlled being the lighting op from the same desk as the lighting. Ignoring all of the concerns of the suitability of DMX, how many times have you accidentally programmed the houselights into a show? Or messed up and ended up with live moves where there shouldn't be any? If the pyro is fired from the desk, at some point the desk will need to contain the programming which allows the pyro to be fired, and if this is recalled at an inopportune moment the result would not be fun. Add in an operator clearly focused on something else and the possibility for a serious accident is high. I'd much rather see it controlled by an entirely separate desk, and a different op. But if that's the case, why bother with DMX at all?

 

You'd also need to design in some form of inhibit hardware which can be located on stage (let's face it, firing from FOH really isn't ideal either), which won't allow anything to be fired unless it's safe to do so. You'd probably also need one of these per circuit of pyro - having one 'global' dead man's handle won't stop the lighting op recalling the wrong cue and firing something entirely different which isn't safe to fire. I see you've already thought about that on some level, but it's more complicated than just having a dead man's handle.

[ImagineerTom]

I'll stear clear of the safety issues for a moment and ask a much more important question.... "Why?"

 

What you're proposing is nothing groundbreaking or revolutionary and in control terms is rediculously simple to do so you need to ask why none of the (many, many) pyro companies have ever produced something similar. Designing and building something that doesn't innovate any new technology or process and doesn't actually solve any real world problems or demands isn't going to get you as good a mark as designing some new technology or solving an actual problem.

 

 

[timsabre]

Having said all that, if this is just a project and will not be used in a real-life situation, the engineering challenge of coming up with suitable error detection systems and design of safety features would probably make it a good project with plenty of scope to demonstrate your ability. You can get into all sorts of areas like suitable transmission protocols and fail safe design.

 

Just don't use it on stage.

[Ynot]

The trouble with that approach is that if it's more than just a theoretical project and the box o' tricks does get built, then somewhere down the line it may very well get used in anger.

 

And with that, Andrew's caution hasn't been heeded and we are indeed hashing over ground already covered ad nauseam, so be wary that the dancing padlocks may appear very soon.

 

 

[PJG]

I have a serious comment that I do not think has been madein another thread.

 

Lets say you have one dead man’s handle and you have pyro all overthe place for different cues.

 

Your look out is happy that the cue that is about to happen canbe fired safely so has their finger on the button but the board op for whateverreason jumps to the wrong cue which also contains pyro in a different part of thestage that the dead man is not watching then what happens.

 

To make it 100% safe you would need a separate dead man’s handlefor every cue or area of the stage.

 

 

[Ynot]

 

To make it 100% safe you would need a separate dead man's handlefor every cue or area of the stage.

That has indeed been made elsewhere, yes - references to the likes of theme park shows where cast members HAVE to be standing on a platform-operated or a certain footswitch they MUST be on before their programmed effects will fire - BUT I'd be pretty confident in saying that these wouldn't be DMX operated...

[Ken Coker]

the engineering challenge of coming up with suitable error detection systems and design of safety features would probably make it a good project with plenty of scope to demonstrate your ability. You can get into all sorts of areas like suitable transmission protocols and fail safe design.

 

I would concur with this. Given that part of your project will be to "investigate", you'll also end up with pretty in-depth knowledge of DMX.

 

 

I would also run this past Simon & Kit before you go much further and I would also go and look at the dissertation on this very subject that a bloke called Michael submitted in the very first third-year of SLLET; it's in the KR library and I'm sorry I've forgotten his surname.

 

(As it happens, writing a risk assessment that is approved by a staff member is part of the dissertation process - or it certainly was.)

 

KC

[gyro_gearloose]

It would only accept a specific value out of the possible 511 for each channel which means it would be accurate to 0.2%

DMX Values are 8 bit and therefore range between 0 and 255, not 0 and 511 as you seem to think. Also, even if DMX values were 9 bit, there would be 512 possible values as you start counting at 0 not 1. I would venture to suggest that since you don't even appear to understand the fundamentals of the DMX protocol, you shouldn't really be designing a safety-critical system which uses it, even if it is just for a University course.

 

I also don't think you understand EM interference. You say that your unit would be closer to the pyro to reduce EM noise. How would this work? Since your unit is closer to the pyro, it must by definition be further from the control desk necessitating a longer DMX run. It is the DMX cable which is most susceptible to EM noise as this carries high-speed digital data, not the cable to the pyro as this only carries voltage to trigger the pyro.

[Jivemaster]

Without real cutting edge RDM DMX the main problem with a DMX pyro circuit is that there is NO continuity indication, you will never know whether a cue is good to go.

 

Now IF you can impliment good DMX and RDM you could probably get the proper handshaking between controller and firing unit

 

However the DMX still says not to use it for pyro and other safety critical events.

[rossmck]

The general consensus on this is covered in the wiki;

 

http://www.blue-room.org.uk/wiki/Effects_FAQ#Can_I_use_DMX_as_a_control_protocol_for_my_electronic_pyro_controller.3F

 

The actual DMX Specification specifically states "DMX512 is not an appropriate control protocol for hazardous applications." Their website also states "DMX512 and DMX512/1990, with their lack of error detection, is NOT appropriate for some applications - such as pyrotechnic control and scenery automation".

So any such device would not comply with the DMX standard, whatever safety measures you put in place.

[smalljoshua]

IIRC, Kit mentioned that a DMX Pyro firing unit has been a final year project before.

 

There are also DMX CO2 canons and DMX Flame Jets commercially available which are arguably just as important to do safely.

 

A standard system requires an operator to turn a key to arm and then push a separate button to fire each pyro.

 

A DMX system, requiring a dedicated operator to turn a key to arm, push a button (per pyro) and then receive a command specific to that pyro would surely be safer as there'd actually be a 3rd level of security. I realise that there is the possibility of getting a DMX error on the line but when you set the protocol on the system to only fire on press of the firing button, even the most wide band noise would only fire on the press of that button! A system giving a visual indication of good communication with the lighting desk (using RDM perhaps) or perhaps integrating with ACN or ArtNet to keep an eye on packet loss too, would allow for a very clean signal. Combined with the requirement for an operator of some description would give a safe system employing a good amount of complexity.

 

Josh

[dbuckley]

As long as DMX is only commanding when to fire, not if it is safe to fire, then there is no safety issue with using DMX. The enabling of firing must be under manual control, with "safe" (preferably) electrics or electronics.

 

The worst that can happen is that due to DMX error that either the pyro fails to fire at all, or the pyro fires at the wrong time, neither of which look good, but neither of which are safety issues.