Here is the controller.
After a half day chasing my tail over a bad op amp chip, I had to scrap the original board and find another way.
My second attempt involved trying to utilize the ZM chip in the infamous Ghurd controller (one that didn't quite make it to "dump controller" status). While tinkering with the controller, I realized that Ghurd thinks WAY outside the box, and what he has done with that IC is simply genius. It wasn't intended to do what he has very successfully made it do in the dump controller, but he managed to exploit very well in his design.
I wasn't as lucky with the ZM chip for my application, but the idea of using a chip for something completely different than it's original intended purpose inspired me to think a little deeper.
What I came up with works very well, and is about as unorthodox as it gets. I used an LM2575T adjustable switching power supply control chip to acquire my "trip" sequence. This is a 5 pin device in a TO-220 case, and has supply voltage, output, ground, feedback, and an "on/off" pin.
The exploit is in the on/off pin. It has a very sharply defined trip point along with just a touch of hysteresis. The thermister is tied in with this pin as part of a voltage divider. As the thermister crosses the internal threshold (about 1.4V), the chip comes to life/shuts down. The only other requirement to make it do what I needed it too do was to ground the feedback pin. This causes the chip to go full throttle to bring the output to the positive rail, as the original use for the chip uses the feedback pin to vary the duty cycle as necessary to regulate the output voltage. Keeping this pin low tricks it into 100% duty.
So, you have a voltage controlled switch essentially.
The output from this chip drives the gates of the MOSFETs, which in turn control the solenoids.
The tip/bump sensor is a normally open "switch" that closes momentarily (or longer) depending on if the unit is bumped or tipped over, respectively. This shorts the gate of the safety control MOSFET to ground, causing the solenoid to release.
The lockout switch is held closed by the safety solenoid, and if the solenoid lets go for any reason, these contacts will remain open, interrupting the gate control signals for the MOSFETs, keeping the unit locked in shutdown mode.
The minimum catalyst temperature is set by a pot on the board, and can be varied between about 180F and 260F. This setting corresponds to the lowest temperature the catalyst can go down to and still reliably go to full power when the valve is suddenly opened up full throttle. This temp was determined by trial and error, cycling the temperature several times and letting it sit and stabilize at progressively lower and lower temperatures until opening the valve completely resulted in the emission of odorant in the "exhaust", indicating incomplete combustion. When finally put into use, this threshold will be set about 30F higher than the temp where reignition failure was noticed, for an extra safety margin.
The 5V reference pilot just indicates that the stable reference chip is operating.
The catalyst thermistor is the sensor at the edge of the heaters guard which senses the temperature of the catalyst.
12V power in, pretty self explanatory. Just plug it in to any one of the many 12V sources in the van and go.
The solenoid connectors go to the safety and "full throttle" solenoids. When the full throttle solenoid is disengaged but the safety solenoid is engaged, the valve is held at the "idle" position by set screws adjusted to hold the catalyst temp just above the low temp cutoff threshold. Likewise, maximum catalyst temp is adjusted by a set screw that limits the degree that the valve can be opened. This isn't a safety function, it is purely an economic feature. Above a certain level, the temp doesn't rise much higher, but the hissing in the valve progressively increases. The result is propane that doesn't get turned into heat energy (and may pose a potentially dangerous situation if more gas comes through the valve than the catalyst can process). Not likely, but just in case, I'm limiting it for both reasons.
The power switch... It's... a... power switch
The control board has 3 LEDs and a momentary push button switch. The button is for resetting a fault condition caused by bumping/tipping the unit if the heater is "righted" and the button pushed before the catalyst temp drops below cutoff. During this time, a piezo beeper alarms with a steady tone until either reset is pressed, or low temp cutoff results in full shutdown. During/after any fault condition, the red LED is lit. The audible alarm is only on if the heater can simply be "reset".
From cold start or fault condition resulting in/from low catalyst temp shutdown, the button is pushed and held, then after several seconds an ignition attempt should be made (lighter, etc). There is a brief "flash" when the unit ignites, followed by immediate catalytic action. The button is held until the green LED illuminates, at which point the controller will keep the safety solenoid engaged.
Any time the button is pressed, the detection circuitry is bypassed and both solenoids engaged (to facilitate ignition).
The yellow LED on the control board indicates the "full power" solenoid is engaged.
I have partial schematics scattered throughout, I'll compile them into a single drawing and put that up soon.
Steve