Error: Heater 0 fault: temperature rising much more slowly ...

A Former User

@chuser said in Error: Heater 0 fault: temperature rising much more slowly ...:

@LB the temp diagram is shorter

Just to be sure: Do you have a config-override.g and if you have one does it get loaded e.g. at the end of your config.g?

chuser

@LB

yes i have a config-override.g the content is as follows:

; config-override.g file generated in response to M500 at 2021-03-15 13:34
; This is a system-generated file - do not edit
; Heater model parameters
M307 H0 R1.911 C66.100:66.100 D1.19 S1.00 V24.1 B0
M307 H1 R4.902 C144.300:144.300 D7.31 S1.00 V24.2 B0

and in the config.g at the end M501
is that correct?

Phaedrux

@chuser Photo of how you have the heater mounted?

chuser

@Phaedrux on OUT0 PIN 1&2

chuser

I now know that it is a semitec104gt thermistor. But I have not yet solved the problem.

Phaedrux

And you've updated your config.g to use the semitec thermistor values?

chuser

@Phaedrux yes i did. but did not help

Phaedrux

And did you re-tune after changing the values?

Phaedrux

It may help to increase the dead time. Try going from D1.19 to D10.

May also try decreasing the gain from R1.911 to R1

chuser

All these test were with the new config.

M308 S0 P"temp0" Y"thermistor" T100000 B4725 C7.06e-8

I have made a few tests with different PWM values.

M307 H0 R1.726 C184.233:184.233 D1.53 S1.00 V24.1 B0 ; 1.00 PWM
M307 H0 R1.712 C169.414:169.414 D1.31 S0.80 V24.1 B0 ; 0.80 PWM
M307 H0 R1.703 C173.799:173.799 D1.23 S0.75 V24.0 B0 ; 0.75 PWM
M307 H0 R1.651 C161.829:161.829 D0.99 S0.60 V24.2 B0 ; 0.60 PWM
M307 H0 R1.499 C169.336:169.336 D0.79 S0.50 V24.2 B0 ; 0.50 PWM

With S0.5 it works. But the heating up goes very very long.

@Phaedrux
Then I tested with increased dead time, but it did not work.

M307 H0 R1.726 C184.233:184.233 D10 S1.00 V24.1 B0

Then I tried to decreasing the gain, and this work.

M307 H0 R1.0 C184.233:184.233 D1.53 S1.00 V24.1 B0

I can't believe it.

Can I leave it set like this? Or does this still need to be optimized?

dc42

@chuser said in Error: Heater 0 fault: temperature rising much more slowly ...:

@Phaedrux the Heater is this one

24V ~60W

https://well-engineered.net/index.php/en/90-new-kapton-heater

Is that the thermistor in the centre of the heater?

If so then I think the problem is that the new algorithm doesn't work well with heaters having a large reservoir effect. Your temperature plot of the tuning supports this. By "reservoir effect", I mean that the thermistor is coupled fairly tightly to the heater, and less tightly to a large heat reservoir, in this case the bed plate. This means that the FOPDT model breaks down and there are in effect two values for the heating rate or gain: a short term value (which is what the new algorithm measures) which is the heating rate when the reservoir is close to the target temperature, and a long term value which is the heating rate when the reservoir is very different from the target temperature.

The ideal solution is to use a thermistor that is coupled to the bed plate (the thing you are trying to heat) instead of the heater. This will also give you better temperature control. However, as a workaround you can reduce the R value in the M307 command. Try reducing it by 30% or if necessary 50%.

I have it in mind to modify the algorithm to measure the long-term heating rate as well as the short term one, and use the long term rate when monitoring for a heater fault.

Phaedrux

@chuser said in Error: Heater 0 fault: temperature rising much more slowly ...:

Then I tried to decreasing the gain, and this work.
M307 H0 R1.0 C184.233:184.233 D1.53 S1.00 V24.1 B0

I can't believe it.
Can I leave it set like this? Or does this still need to be optimized?

What does the heat graph look like with those values? If the temps are stable, you can continue to use it.

dc42

Decreasing the gain or heating rate parameter will fix the "temperature rising too slowly" error. It might have the side-effect of causing the temperature to oscillate slightly. If so, then increase the dead time parameter enough to stop the oscillations.