Strange Bed PID tune

boelle

So i wanted to pid tune my rebuild printer before i start the task of collecting Z distortion data

when i pid tune at 90 degree it works ok, but when i try to tune for 100 it never stops and i get a timeout. it starts by heating and stops heating 2 times but after that it just heats and the temp slowly goes up. i use the C8 parameter in both cases.

Q... how much above does it need to go for being successful? what i think is that it just cant get high enough above 100 for a successful tune

Command send: M303 P1 S100 C8

output:

13:18:10.272 : Info:PID Autotune start

13:19:43.100 : bias: 165 d: 89 min: 99.18 max: 100.48

13:20:10.939 : bias: 234 d: 20 min: 99.22 max: 100.55

after that the bed LED just blinks and the temp hovers arround 102... if i dont reset the printer it will timeout

what could be wrong here?

settings in firmware:

pid drive min: 80
pid drive max: 255

bed pid max value: 255

i tried to let the pid tune run, if going for 100 degrees it will at least reach 110 before timeout

13:55:34.902 : Info:PID Autotune start

14:03:59.665 : bias: 175 d: 79 min: 99.15 max: 100.48

14:04:30.651 : bias: 235 d: 19 min: 99.15 max: 100.58

14:14:31.986 : Error:PID Autotune failed! timeout

Repetier

It is not a question of how much but more a question of how long. But a slong a syou can exceed the temperature you should be able to autotune it. From your log I see that you even got the first 2 iterations, but then it seems to not swing any more so you get the timeout after 10 minutes (dev 20 minutes).

boelle

yep... after the 2 iterations is stays pretty much at 101-102

is there anything i can do to "fix" it ?

Repetier

I think you are hanging here:

        if(heating == true && currentTemp > temp)   // switch heating -> off

        {

            if(time - t2 > (controllerId < NUM_EXTRUDER ? 2500 : 1500))

            {

                heating=false;

                pwm_pos[pwmIndex] = (bias - d);

                t1 = time;

                t_high = t1 - t2;

                maxTemp=temp;

            }

        }

You have set it to cooling with pwm pwm_pos[pwmIndex] = (bias - d); whcih is enough to keep > 100 so it never goes down. So I guess the algorithm does not work for that timing and you can do nothing.

YOu might try with 95°C and use these values. SHould be pretty close.

boelle

oki... will try that (95 degrees)

just curious is this a software bug or is it my hardware that somehow plays a trick on me?

boelle

so i send M303 P1 S95 C8

same thing happend. and just before timeout temp was 108-109

will try to load dev version and see if its the same thing

boelle

hmmm... just going through firmware settings

could it have anything to do with HEATER_PWM_SPEED ?

boelle

same thing on dev version, but the timeout is not 20 mins. its the same 10 mins

but then i  tried M502 and M500 to make sure eeprom is not screwed up

sucess at 95 degrees

boelle

no success at 100 so think this is as far it goes

Repetier

It is no software bug. It is how this pid tuning works. In the code part I shows you see it uses results from previous iterations to adjust. So heater off is not off just lower power - and in your case enough to not go down to < 100 which is why it times out. So just a not working combination between speed/max temperature/algorithm. Does not mean pid doe snot work with 100°C just autotune does not work at that high level.