Hi
Early on in this sub thread I did point out that if you have a heat source, none of these methods are optimum. Having the sensor in a bottle of water or on the wall of the chamber *both* will give you an error if you have active fermentation. The only way to really control the temperature of an actively fermenting fluid is to sense in that fluid.
Bob
Hi
Actually you have completely missed the point altogether by not
reading (or thinking) it through.
For ferming, I said to use the wall of the vessel; and for serving, I said how and why it was advantageous (and optimum for this type of controller) to use a bottle of liquid in the same way for maintaining serving temps. This is due mainly to it being a primary indicator, as opposed to your method using a secondary indicator (which causes more issues with control system design that the lag or overshoot in a keezer). Besides not having to experiment with where to place the sensor, and what differential to set, you don't have to worry about any bottles freezing (many people store bottles in their serving keezers). As long as the sensor vessel is as large as the smallest vessel, you get optimum (minimum) cycling as well.
I already stated that your method of using a secondary indicator (wall of freezer) can give the same results as using a primary indicator (wall of vessel). The difference is all of the experimentation (time) involved in tuning the system, which is not something most people would want to do. In fact, it is possible for a secondary indicator to provide superior performance when its temps lead the controlled vessels temps (due to external ambient temps in this case). The rub is that it would require even more tuning and different regimes for any seasonal (or daily) swings in the ambient temps, and no freeze protection either.
That was all based on a passive wall (like a fridge). Placing the probe directly on a evaporator coil, there is no way in real practice you would ever get a system to perform anywhere close to the bottle of beer method.
Any more control theory talk needs to focus on primary vs secondary indicators, because that is biggest difference between your method and everyone else's. And quit mixing ferming and serving strategies.
The only way to get any better performance with a standard freezer is with some kind of predictive control system, which is what placing the sensor on the passive wall of the freezer mimics. The main issue with these type of systems is the difficulty of tuning them, especially with varying ambient temps (garage), and is made even more difficult when the only tuning parameters available are moving the probe to a new spot on the wall or adjusting the temp differential.
A great idea to really get to the bottom of this would be to start a new thread and invite Cat22 to have his 'probe in air' (even for ferming) do battle with your 'probe on wall'. He was equally as passionate that his way was right as well.