I have a project in mind, and part of it is sensing whether a grill is turned on. The grill is of the propane-fueled variety. Originally I thought I could use a temperature sensor or thermocouple to sense the grill's temperature, but the grill would take too long to cool down. (Though I could use some calculus shown here to make a predictive filter) Then I though of putting a CdS cell to look at the light level inside the grill and figure out if there was a flame, though this could have interference if light leaked in. Another idea I had was to put pots or encoders on the valves to figure if they were on or not... Not sure.
Anybody have any ideas? Has anyone tried something like this before?
P.S. The system can't get too hot b/c I need to have an ATtiny and Nordic nearby to send the results inside.
Some commercial devices like stoves and hot water heaters use a thermocouple in the flame to sense when to shut down the gas in the case of a blowout. Some other systems, for example, those associated with piezoelectric pilot-less systems use an IR/UV sensing arrangement. I can't remember off hand whether it's IR or UV. Pretty sure it's UV because IR would not react fast enough. That's my US$0.02.
Fire can conduct electricity, being that it is a plasma; so, if you have a couple of conductors spaced apart that can withstand the heat of the flame, when the flame bridges the gap, you should see a complete circuit (not sure on what resistance there would be - I am not sure how good of a conductor a flame is, nor whether different flames from different burning material conduct differently)...
The conductive fire looks interesting... Only problem is that the flames on my grill are very small (but very hot). I'm not sure I could get an accurate reading out of that.
For the bimetallic strip/thermocouple, only problem is that it would take a while for the material to cool down. Does anyone know if the link I posted above about the predictive meat thermometer would work? The demo video seemed to work...
No idea how feasible it is, but I wonder whether there's some sort of gas detector that'd work. If your flame has any carbon monoxide output, then maybe cannibalize a First Alert. If it works, there might some lag on the off side, depending upon how quickly the grille evacuates fumes.
The thermocouple is placed so that it is in the flame whenever the gas is burning. It works fast enough to prevent explosions on probably millions of gas water heaters in the US. The especially nice thing about them is that they are available in practically every hardware store as replacement parts for gas water heaters. They put out about 750 mV when hot. They have little mass so they cool quickly when the flame goes away. .75 V is high enough to use without conditioning on the Arduino. Just use the internal 1.1 V reference.
Which route would you recommend? The LPG seems nice b/c that's what the grill runs off of and it will be no-contact with the flames, but, as someone said, thermocouples are used in water heaters, and they work. I'm not sure.
P.S. Also, which one uses less power? The thing will be controlled by an ATtiny(44) + a Nordic to send the results inside. The nordic uses 15mA max while transmitting, and the Attiny will not use much current, so I was thinking of powering everything of a CR2032, so the sensor can't draw too much current.