Hi
If you tape the probe to a keg, and then swap around kegs, your controll point is going to change every time you do anything. Regardless of where the controll point is, you *will* need to calibrate things. That calibration will only be valid for a single control location. Why do it the hard way? There is 100+ years of engineering experiance out there on how to do this...
Bob
If you tape the probe to the freezer wall, you will need to adjust the controller temperature every time the ambient temp outside the keezer changes.
There is no re-calibration needed if switching between 1 keg and another, because it is a direct indicator. This is exactly why it is the preferred method, along with numerous other benefits. In this case, the 'control location' never changes, and is irrelevant anyway, because we are using a direct indicator regardless of where the beer is positioned.
Your approach isn't even using a secondary indicator for feedback, it is using the primary input (cold freon) as the indicator.
I don't know what control theory you are using, but I have never heard of one that favors staying as far removed as possible from the primary indicator (the beer). Even worse, one that uses the primary input (cold freon in this case) as a secondary indicator.
And I understand your position that the beer acts as an integrator, which makes it possible to control the temp more tightly using a secondary indicator. I also understand that would involve much more tuning and dependence on ambient conditions, something which you repeatedly/conveniently leave out, and neither of which the average person on these forums has either the desire to do, knowledge of, or control over. They want the simplest and most suitable solution.
Using your logic and theory, using a small bottle will give the same results as your approach, without having to worry about cooling coil location and without needing any temp offsets- just adjusting the differential for acceptable cycling freq.
RE:control loop reacting fast- We don't want the control loop acting fast for compressor based systems. This is exactly why everyone (else) recommends putting the probe on something massive, setting a small differential (~1F) and the desired temp (~38F), then calling it good; while you, confusingly, recommend setting an abnormally high differential (8F) with the probe on some object whose temp is far removed from the beer temp, then go about searching for what temp to set on the controller to get the beer temp you want.
Does that about sum it up?
Maybe you and Cat22 should battle it out over 'probe in air' (even for fermenting according to him) and 'probe on cooling coil'. The common element to both of your arguments is the insertion of another controller (yourself) into the loop- Cat22 with himself adjusting the control temp offest multiple times a day/week during ferment, and you fiddling about trying to determine the proper temp offset and differential because you are using the primary input as the indicator.