Multiple thermocouples for flame detection

[Buriedcode]

Instead of trying to parallel the sensors, what about paralleling multiple valves?

Paralleling valves would indeed separate each thermocouple, but then of course it would mean all the thermo's would have to be up to temperature - rather than 'at least one'.  I'm fairly confident I can knock up a power-OR circuit with comparators and MOSFET's (at 25mV... gonna need very low resistance..).  The off-set input won't be much of a problem because I'm sure these valves 'hold' at >18mV.  That is of course, absolute worst case as far as I can see.

If this is not practical, then would it be possible to modify the solenoid to have three independent windings? You would need to maintain the ampere-turns and wire cross-sectional area for each winding, so the magnetic circuit might need to be modified.

Probably not a very practical idea overall, but it should still be fairly fail-safe.

I didn't think of that..separate electric circuits, paralleled magnetic! I may be able to get the coil assembly out and given the extremely low voltages, I can't see it being more than 20 turns.  I'll see how much room these is on the valve connector, as it'll have to have at least 3 wires (outer jacket is common, 3 thermos, solenoid coil commoned).


 I *do* like the idea of different sensors   I would have to drive the valves solenoid separately with say, a flame ionization device, and of course that would use up a fair bit of power driving the sensor itself.

I have yet to put it in a flame, but hope to do it tomorrow as the weather has been rather nice.  Blow-torch to see the peak output, then a dirty oil burner flame to get a ball-park figure.  Also need to dig out my crude DIY CC power supply to measure the resistance of the solenoid.

[johnwa]

Instead of trying to parallel the sensors, what about paralleling multiple valves?

Paralleling valves would indeed separate each thermocouple, but then of course it would mean all the thermo's would have to be up to temperature - rather than 'at least one'.  I'm fairly confident I can knock up a power-OR circuit with comparators and MOSFET's (at 25mV... gonna need very low resistance..).  The off-set input won't be much of a problem because I'm sure these valves 'hold' at >18mV.  That is of course, absolute worst case as far as I can see.

Sorry, perhaps I didn't explain this well enough. I meant having multiple thermocouples, each of which would be wired to its own valve. The valves would be plumbed side-by-side, in such a way that gas can flow to the burner even if only one of them is open. (i.e. gas is shut off only if all valves are closed). Of course, this solution means having multiple (probably expensive) valves, and also more pipe fittings that can leak.

[Buriedcode]

Sorry, perhaps I didn't explain this well enough. I meant having multiple thermocouples, each of which would be wired to its own valve. The valves would be plumbed side-by-side, in such a way that gas can flow to the burner even if only one of them is open. (i.e. gas is shut off only if all valves are closed). Of course, this solution means having multiple (probably expensive) valves, and also more pipe fittings that can leak.

Ahh I see, paralleled valves.  That's a fair bit of pipe-work, but I do see what you mean.  The regulator is of course in series so it wouldn't matter how many valves in parallel, the flow will be limited.  I got a bit carried away with the whole '3 thermocouples, 1 valve' thing (sounds like a very dodgy internet video..).

So far the solutions, in order of preference (and ease):

- Give a single thermocouple adequate thermal contact to the shield.  One thermo, one valve.  Hope it can maintain temperature to hold valve open.
- Three thermocouples, power-OR circuit, to a single valve.  Relatively easy circuit, but requires power (+9v), and annoying insulating connectors to prevent shorts.
- Three thermocouples, one valve, solenoid has 3 commoned windings.  Annoying forcing 3 thermocouples to one valve connector, requires modding the solenoid (not too tough, #22 enameled wire).
- Three thermocouples, each with their own valve, plumbed in parallel.  Expensive, requires pipe work, but *possibly* the most reliable solution.

The only other advantage of an active circuit is that is makes it much easier to add other forms of cut-off from other sensors.  A tilt switch could easily be used with the power-OR, and if its a circuit that drives the solenoid directly, one could add a delay between a sag in thermo voltage, and tripping the valve.  Its an overkill solution though.