Time averaged temperatures

[bergerrings]

Afternoon!

Due to the way I have set up for fermentation control system I need my temperatures to be very precise. A feature I would LOVE to see would be TIME AVERAGED TEMPERATURES. Something along the lines of a check box or two next to each temperature probe input where you can 'check' if you would like to use real time temps (nothing checked) or say a 30s or 120s time averaged. This would only really by useful for fermentation due to the long time frames but it would surely fix the minimal 'bouncing temps' that we all see when we stare at the temperature readings for more than a few seconds - which I know surely pisses off others too!
This would then allow me to really dial in my fermentation controls as currently I have a rather large temperature dead band (+/- 0.6 C) so the system doesn't fire off after a couple of temp bounces. Having just typed out my dead band I will have to admit that it isn't really that big but it could be smaller! haha
I currently use a couple of timers to cancel out some bouncing but there isn't enough states per process to cancel out all of the bouncing and plus this doesn't address the displayed values.

With the temps time averaged over say 30 seconds I think I could get my dead band down to +/- 0.2 or possibly even 0.1. This is pretty anal but it would be sweet to be able to control my fermentation/largering within 0.1 C for four months!

[JonW]

With the temp bouncing you've described, I'm going to guess that you haven't grounded the BCS. Correct? You should run a jumper from a BCS ground connection to the AC ground. Any bounces at that point should be 0.1.

[bergerrings]

I have grounded everything and isolated it all to the best of my abilities but even the minor jumps are annoying. This would easily be fixed with time averaged temps...super pedantic but hey, all you can do is ask?!

[brahn]

All of the temperatures on the BCS are averaged over time. It's actually done at two levels. In the hardware it's an average of the last 64 samples and then in the software we average the last 8 readings of the hardware values, dropping the highest and lowest.