PeeBee wrote: ↑Wed Mar 21, 2018 7:02 pm
... I give up. ...
I can't really give up, that's not helpful to anyone. But I have had to blank "EI*****'s" posts because my head couldn't cope with the logic of the explanations (there are lots of things my head won't cope with!).
I remember earlier posts had a bad "analogy" between rooms full of very fluid air and a cauldron of ... well porridge isn't too far off a mash consistency. From this "analogy" it can be assumed "EI*****" is either a naive youth or forgetful older person - before cheap multiple remote controllers for central heating there was just one controller, and that controlled the boiler temperature! The boiler temperature is still controlled - the room thermostat controls the flow much like my "tongue-in-cheek" suggestion for HERMS in the earlier post.
But there is a point to be made from "EI*****'s" ramblings - HERMS systems are very poor at one aspect: Convergence of set temperature with measured temperature. This convergence can be speeded up by allowing a small potential or actual overshoot, either by tuning the PID controller, or temporarily and manually adjusting the set temperature to be a degree or two higher than the desired temperature, or else come up with some madcap scheme to control the temperature to give the illusion it works (the illusion is as good as the reality if the target is actually met).
When tuning PID controllers the easy mistake to make life harder is have the impression our systems are fast - they're not! Not at all, they are s-l-o-w when compared to other PID controlled system. Far better to ignore the "D" bit and go back to the earlier "PI" type controller (set "D" to zero, or the lowest value allowed - it's 2 on my Auber PIDs).
(24/3/18 18:00 - Minor edit to delete conflictory language).