Re: [EE]: Flame rectification

At 01:37 AM 2025-12-19, you wrote:
>I was doing a cleaning of my gas-fired forced air furnace, and it 
>got me thinking about how wonderfully designed it must be. It has a 
>Honeywell S8610U "Intermittent Pilot" module that combines both 
>high(15KV-ish) voltage ignition and low(80VAC-ish) voltage, 
>low(5uA-ish) current flame detection on the same wire and electrode. 
>Fantastic. Especially since this very module has been running since 
>1997, with very little maintenance, and has not missed a beat.
>
>Has anyone reverse engineered the in and outs of this device and can 
>describe the circuitry, and how flame detection doesn't die the 
>first time the high voltage comes on? And how it can toggle between 
>the two modes so quickly and reliably?
>
>For those who haven't encountered it, the module runs on 24VAC, 
>generates a recurring high voltage spark at twelve Hz, energizes a 
>pilot gas valve, senses the pilot ignition on the same wire using 
>flame rectification (in between the high voltage sparks) and 
>energizes the main gas valve if there is a pilot flame present, then 
>stops the HV and continues to monitor the flame. It sounds like 
>bzz-click-bzz-thunk-whoosh when starting. Quite obsolete by today's 
>standards, of course. It was built to last almost forever :)
>
>Thanks! Bob
>--

I've analyzed one that used a flame rod / flame rectification to 
detect the flame and an ignition transformer,
but I don't recall whether it  used the same rod for ignition, 
probably not. This was ~20 years ago so the memory
is a bit dim. It worked from the mains directly  and used a little 
TO-92 SCS (silicon controlled switch- a
diac-like device with 3 pins) to generate the spikey waveforms for 
the HV transformer primary. Possibly for
something like a water heater or clothes dryer rather than central air.

The flame rectification circuit was simple as I recall it- some 
discretes and passives. No MCU at all in the unit.
The only unusual feature (besides the conformal coating) was that 
there was built-in redundancy with extra components
so no single point of failure could lead to the gas valve being 
energized without a flame being present (that being
the sort of thing that could turn one's home into an unruly heap of 
bricks and kindling). The beauty of the flame
rectification method is that disconnection, shorting, contamination 
etc. does not mimic a flame.

Gas furnaces, even those of similar vintage to yours, may use an MCU. 
This introduces another level of complexity when it
comes to design compliance, of course, because the firmware is 
probably safety-critical.

For central air gas furnace, the controller starts up the small draft 
inducer fan for some time to purge, verifies the air
pressure and pressure differentials are high enough to be safe 
(probably just a permissive chain in the
simplest/most reliable way possible) as a proxy for flow, then turns 
the gas valve on and turns on the ignitor (a spark or
perhaps an HSI hot surface ignitor). If the flame is not detected in 
a certain time, it shuts down, purges again and
retries a few times before quitting with an error.

I don't think it would be hard to tickle the primary of a HV 
transformer to get ~100V for the flame detection after
giving it beefy pulses to (hopefully) ignite the flame. You'd need 
some HV-capable high value resistors and series
capacitors to prevent the kV ignition pulses from causing issues with 
sensing. You could probably troll relevant patent
files for more information.

Best regards,
Spehro Pefhany