I can't imagine PID melting anything, obviously but I was just curious as to whether we want the curve to match the target so literally in ALL cases. For instance, O2 is a lot more dynamic and changing (rapidly) than idle for instance. So I was just wondering if a perfect target response is ideal in all cases or not - I bet the answer is "it depends"
i can. with a mahoooooooooooooooosive P and nothing else. be aware.
no - as i said regarding the 1 6 lump... based on characteristics... you can get sorta kinda close-ish with kinda generic numbers as some of you are finding out now (see - once you see it - it's simple) but if you started burning say welch's pancake stuff or whatever you'd probably want different flow control
VR6Turbo wrote:I wish Allan would sort that out because it's confusing.
coz the rest was a doddle, right? hehe
Totally
What I meant by that, was if you have say, P-80, I-20, D-0 in a normal PID loop. Does that then become P-800, I-200, D-0 in Allan's '0-1000' world? In other words, has he rescaled the PIDs, or just increased the numbers allowed in the cells?
VR6Turbo wrote:I wish Allan would sort that out because it's confusing.
coz the rest was a doddle, right? hehe
Totally
What I meant by that, was if you have say, P-80, I-20, D-0 in a normal PID loop. Does that then become P-800, I-200, D-0 in Allan's '0-1000' world? In other words, has he rescaled the PIDs, or just increased the numbers allowed in the cells?
It's not just a ratio. If I input the same PID numbers as on my posted trace above but devide them by 10 (P=10.8, I=1.1, D=0) the results are noticeably different. They seem to work OK but much slower...
