Kitedemon-
Accuracy is important but I think many also get too hung up on it. In the end the thermostat wants to be as or more stable than the element being used and any more than that is really not getting us anywhere. The element being used is generally heat tape and while neat stuff it is not NASA ready if you know what I mean. Fortunately our animals like to regulate themselves if given the opportunity. In my seemingly endless hours of taking notes I found the moving to a higher resolution than the .25 worked well with less aggressive heaters in larger cages - i.e. things were very slow to change and the algorithm could respond smoothly. That set up was not really a good representation of real world. When I used it on racks like most of us are playing with it was actually less stable than the lower .25 resolution. It was in constant fluctuation- a bit of a fight with itself. I had actually thought about an "incubator" model with the very tight resolution as well as a real world version but after wresting with the response time of the algorithm and the resolution I got where it is now. So far I have not found a scenario that it does not do very well in.
Was not being sensitive on the price comparison, just did not understand it? The way I went at this was to design regardless of price then see where it could be beaten down to then see where it landed. If it was not competitive then I was going to abandon project. Like I said, no shortage of nice stats in the world so unless this was going to add something there was not much point. You would likely pass out if you knew how much money I had to tie up in this.
As for your multi probe tracking gadget. The short answer is yes. Already got all of that built in to the box. I won't get in to too much detail about it yet but I think it will meet the requirements you are putting forth. On that we can dial up the resolution quite a bit more as we are recording, not acting on the data. As for the accuracy- we are still subject to the 1% or so on the digital sensor itself. A component called a thermistor could be used and "calibrated" which would get us better than that but the problem then is that exact probe must be used in that exact port and that calibration can drift over time and need to be redone. Not really realistic for this market.
The sine wave issue is a tough one on triac devices. I have tested these on the two generators I have here. One a household 18KW and the other my 5kw Honda. Worked fine on both of them but unfortunately there can be a lot, and I mean alot, of variation from generator to generator. We selected a triac and driving components that would tolerate the most sloppiness and would deal with the most different types of loads (resistive elements, pumps, etc)- we did not go with a single triac driver device but rather built our own driver. This was one of the considerations. The "out of the box" driver would have been a little easier and about the same money but I have also seen complications with them. Tough to say on a case by case basis so try it out first before you need that generator. The power supply for the processor is very stable and will tolerate about anything so the brains of the operation should be ok- just that pesky triac.
Reply With Quote