Pellet Stove Sensor(s)

It is my hope that some here can suggest one or more *inexpensive* sensors that could be used to improve pellet stove design.


I have recently improved the controller code on an existing line of pellet stoves. I had to use the existing !@#$% microcontroller and the existing !@#$% PCB. Given the existing stove and PCB designs the only input the microcontroller has to work with is temperature. One sensor gives the temperature of the heat exchanger metal, the other the temperature of the flue where it leaves the stove. The microcontroller uses that along with about a dozen timers to control the electric igniter and the fans.

However, none of that gets to the heart of the combustion process. The stoves have several heat output settings with pellet feed rates and forced air combustion draft "in balance." That balance is achieved by tweaking fan speeds and pellet feed rates to achieve an appropriately lean fuel / air balance for the various heat output settings. ("Lean" refers to less fuel than air.)


My client company wants to have settings for wide range of heat production. Perhaps as wide as 8:1. The easy part of this is the fuel feed rate has to vary by 8:1. That easy part will use timers and variable motor speed settings. They also want to have a nearly continuous heat output adjustment. That is also easy given an updated PCB, control panel and microcontroller.

The problem in all this is getting the forced air combustion draft in balance with the pellet feed rate. To do this right my guess is that they need some kind of sensor in the outlet gasses. Perhaps they could sense O2, CO or CO2. Or would some kind of ionization sensor do the trick? It would also be nice if the sensor provided a better way to detect ignition than temperature rise. (The stoves have electric heaters to initiate combustion.)

It would be ideal if no specific altitude adjustment was needed. The current products are tweaked for the specific altitude when installed. (Thinner air requires more draft.)

These sensors have to work in a hostile environment. In addition too high temperatures and normal combustion products these sensors have to tolerate soot and other crap. (The flue gas temperature sensor avoids this problem by being connected to the outside of the flue.) Optical methods need to be immune to soot although a quartz window in the combustion chamber may be self-cleaning.

Price target? $100 is too high. $10 (each in 100 unit lots) is closer to their target price. This price should include any oddball microcontroller interface requirement.

Ideas, anyone?

Joe Green