I've been looking for example projects of people making flame detectors with microcontrollers. Most people seem to be using the IR type of sensors and it seems fairly easy to replicate. However I am interested in a more rugged flame detector and was thinking of flame rods like those used in furnaces or UV scanners like those from honeywell.
Basically I'd want an arduino or similar microcontroller to act as a flame relay, with a normally open and a normally closed contact. Flame relays seem to be common in process industry but expensive for hobbists.
The flame rod principle seems easy enough... just have to measure some microamps from the flame rod to ground. Perhaps an ADC with a programmable gain amplifier?
Now of course, safety equipment like this is better something not be DIYing. But I am curious if it can be done and maybe I could build a testing rig just because.
The idea is to build the detector using a flame rod. Any flame rod will do, but for example this one: flame rod
A flame rod itself is just a thick wire made out of a temperature resistant material like Kanthal. A flame acts like a diode so the complex part is the electronics to detect flame rectification. A transformer will be needed to produce the high voltage AC needed to produce the microamp current flowing from flame rod to ground. And then using some op-amps, ADCs and other electronic components the Arduino will somehow measure the presence of a flame.
OK I see 280VAC in your picture. Is there a large gap between rod and ground? What if you used a small gap like a sparkplug in the flame. I'm guessing that the voltages could be a lot lower and thus easier to manage with an arduino. Maybe?
If you get a small transformer , make an oscillator and connect to the low voltage side of it , you will generate a high voltage on the high voltage side .
Use that connected to your flame rod.
Check your circuit doesnt ‘see’ a flame is there is a short .
I built such a thing around 30yrs ago , but can’t recall the circuit . I think I charged a capacitor with the dc current .produced .
Careful with the high voltage and do not use to control any appliance - for curiosity only !!
Google will find a couple of circuits
“ low voltage rectification flame detector”
230VAC is German/European mains, so I guess that here in North America 110VAC could also work or a 1:2 isolation transformer could be used. The author remarks:
With the flame on, I measured a current of about 10 μA with a mains voltage of 230 VAC: this is more than enough to make the neon lamp glow. It's not very bright, but it glows. If a useful (digital) signal is needed, one can create a simple optocoupler by putting a photodiode or a photoresistor next to the glow lamp and shield them from ambient light in a black enclosure. This allows driving a microcontroller or a logic circuit.
This would work in theory but I think there should exist an easier way of detecting 10uA than creating a DIY optocoupler with the neon lamp. Perhaps a lower valued current limiting resistor to be able to drive 1mA and measure that current with a small resistor and a good ADC like the ADS1115?
These kind of UV scanners can be found for about $150 on eBay. Would definitely be cheaper to use a flame rod.
Fireye has a nice document showing how these UV sensors work, seems similar to some radiation meters where the ionization from radiation enables current to flow between electrodes and is detected as pulses.
Use Ohm's law, a shunt resistor and a suitable opamp. 330k gives 3.3V at 10uA. You'll have to protect the opamp input from the HV shenanigans, but that's not too difficult if you use a fairly manageable HV of 110-230V.
