Thanks Rowland. The heater has been revised. I had plans to run it via a spare channel of my dew heater, but it started to snowball as usual (my famous motto of keeping things simple, lasts for about 30 seconds.) So it is now an always on series (5 x 47 ohm) giving a poultry 0.85W @ 13.8v whenever the TEC is energized. Seems to be effective enough at this stage.
So far very noticeable results when comparing a few dark batches of 3 x 300sec subs @800 at room temp and then again after a 15 min cool down period, actually exceeding my expectations. My last remaining issue before putting it into service is a very random and getting to be somewhat annoying 99 error. No luck chasing it down yet, but seems to generate at the end of an exposure as it initiates the download to PC. It will not be a nice problem to have if I can't find the cause.
George the Error 99 is a general catch all error which is not easily debugged. Sometimes it can be cleared by powering down and removing the battery or power source, then unplugging the memory card. However if not cleared it can mean a connection failure between the sensor and processor. The ribbon connector locking bars are notorious for causing open circuits if they btrak or fail to lock properly. I'd suggest checking all ribbon cable connections.
Thanks Glen. Yeah I have a good handle on how unhelpful the error 99 message is. I have been going about a process of elimination to narrow it down, at first it seemed to be the micro switch on the CF card door, so I hot wired both that one, and the one on the battery door and it seemed to improve but it kicked off again. The only real pattern to it now is it is more likely to occur when the sensor is cooling, it very rarely throws it with the cooling off. This probably points to your suggestion of the ribbons and connectors. I will target this area in the next time I am sitting in front of it.
Did you earth the cold finger? The finger has to be isolated electrically. If you read through Rowland's build thread it shows how we used plastic shims to hold the finger away from the sensor pins. Its probably something simple. Good luck.
The camera chassis, sensor frame and cold finger should all be bonded to ground. Sometimes the sensor to chassis bonding strap screw loosens after the first cool. If there is no other path to ground the sensor won't work. Tighten up the small screw as tight as you can.