.... to make thing as complicated as possible.
No matter which (HERMS)-system you choose, you will always have problems with the mash temperature control, if only one probe is used.
- if the probe is in HLT only, there's no way of telling what the mash temp really is and the stratification comes a real problem. Any steps in mash temp are slow to do and overshoots are really difficult to control. If you overshoot the HLT temp it takes for ages it to cool down. It works best when the HLT has a really small amount of water in it + a powerful stirring.
- probe in the wort output only. This limits ramp times as you can't heat wort over the set mash temp without complicated control system. I also noticed that the mash temp starts to creep up over time. I think this is the best solution, if only one probe is used. With a separate small HERMS-kettle this is a good way to control the mash.
- probe in the mash only, with a careful control (PID?) of the HLT temp (or a tiny HERMS-kettle) it might be doable. Any steps in mash temps are really difficult to do without overshooting and the lag time in probe feedback makes things pretty much impossible?
My solution is to use 2 probes for the control. One in the mash and one in the wort output (I used to have 3th one in the HERMS-kettle before moving into CFC-HLT-HERMS). The control sketch in the Arduino is made so that if mash temp is way off what the set mash temp is wort can be hotter than the set mash temperature. I reduce the amount of wort "overshoot" as the set mash temp is closing in and I finally set the wort temperature same as mash. This enables fast steps (with the CFC-herms) and there's no problems with the overshoot once the settings are right.
2 point temperature control can be done also with 2 "normal controllers". I used 2 Sestos's in series before the Arduino stepped into my life. 1 was 12V output and the other one with a relay output....
One thing that comes to mind ... it's really important to check what's your free flow out from the mash tun is. It can be quite slow and if too powerful pump is used it can create suction in the mash tun and stuck mash is more that likely outcome of that (been there) especially if false bottom is in use, braided SS seems to work better. If the hoses on the input side of the pump start to jiggle, the pump sucks
I've about 3L/min flow out from mashtun with the grains in (if I remember it right). The flow also limits the temperature step times.
I just made a new version of the control sketch. I added pump on time limiter when the mash is <0.5C from the set point. The pump can run only 3000ms and can be turned on again after 10000ms break. This seems to be taking care of the overshoots when HLT temperature is much higher than the mash. I got +-0.15C accuracy with 20L water in the mash tun.
One more thing that is great with the CFC-HERMS. There's no need for powerful heating element to get the mash step time short. Just a big kettle full of thermal mass.
Edit: A short plot of the test with the new settings. It's easy with the water only to keep temperatures in check, but this looks promising:
Mash drops 0.4C from the set point and wort goes down 0.7C max (because of heat losses in HERMS line).