Slow hotend heating near target temp

[pkm]

Hello,
Since I switched to the new heating controller in 1.15, I often notice the hotend heating very slow when the temperature is 5-10C below the target.
Sometimes it takes 2-3 minutes to go from 230C to 238C!
I use autotuning and save the results. Did that a few times.
What could be the problem?

[Jackal]

Have you tried 1.17e or 1.18RC2?
There has been some changes to hotend tuning.

[pkm]

Yep, I just installed 1.18RC2, and before that I had 1.18beta and 1.17 something.

[burtoogle]

Here's something I discovered recently… e3d v6 hotend was heating up quickly all the way to the target temp and then I had to change the hot end fan. The new fan seemed to shift a lot more air than the old one and, subsequently, the hotend struggled to reach the target temp. I swapped the fan again to one that wasn't quite so aggressive and everything's back to normal again. Maybe your hotend fan is cooling the hotend too much?

[pkm]

If it matters, that's what the autotuning put to config-override.g recently
M307 H0 A290.3 C1404.5 D1.0 S1.00 B0
M307 H1 A722.7 C248.3 D10.0 S1.00 B0

[pkm]

I have e3D Volcano.
Not sure about the fan, it's been the same for a long time. And that tiny 30mm fan just can't blow too much…
Also, the hotend heats up really fast (24V 40W heater), but only up to 5-10C below the target temp, then the curve just gets flat.

[dc42]

@pkm:

Also, the hotend heats up really fast (24V 40W heater), but only up to 5-10C below the target temp, then the curve just gets flat.

That indicates that auto tuning has over-estimated the hot end gain. Try reducing the A parameter in the M307 H1 command by 10 or 20%.

[pkm]

Thanks, I'll try that!
IIRC autotuning said that the hotend would reach 750C at full PWM. Hard to believe, actually.
Also interesting that the dead time is 10s for hotend but 1s for bed… Does it make sense to decrease D to 2s?
I'll also try to measure average PWM at work temp.

[dc42]

10s for the hot end is higher than normal, it's usually about 5s.

1s for the bed is extremely low. Is it a silicone heater with the thermistor embedded in the element?

If it's a Duet WiFi that you have, try auto tuning with firmware 1.18RC2 and see whether it gives different results. Leave out the P parameter in the M303 command (it's no longer needed), but choose the T parameter to allow for up to 10C overshoot.

[pkm]

Those were settings from 1.18RC2 autotuning…
I just tried once again:

[[language]]
11:39:33M307 H1
Heater 1 model: gain 745.5, time constant 250.1, dead time 10.0, max PWM 1.00, mode: PID
Computed PID parameters for setpoint change: P6.0, I0.024, D41.9
Computed PID parameters for load change: P6.0, I0.235, D41.9
11:10:55Warning: Heater 1 appears to be over-powered and a fire risk! If left on at full power, its temperature is predicted to reach 765C.
Auto tune heater 1 completed in 328 sec
Use M307 H1 to see the result, or M500 to save the result in config-override.g
11:07:05Auto tune phase 3, peak temperature was 211.1
11:06:52Auto tune phase 2, heater off
11:05:32Auto tune phase 1, heater on
11:05:26M303 H1
Auto tuning heater 1 using target temperature 200.0C and PWM 1.00 - do not leave printer unattended

I think this might explain something:
Computed PID parameters for setpoint change: P6.0, I0.024, D41.9
This is exactly when the problem occurs, for example changing the temp from 230C to 238C is very slow.
So why is I 10 times less for a setpoint change?

[pkm]

The average PWM is 0.333 at 235C which corresponds to A650…

[dc42]

So use A650 in your M307 command. I suggest you also try reducing D in M307 to about 5, because I don't believe that 10 is the correct value. If this causes the temperature to oscillate, then increase D towards 10 again. Reducing D will reduce the response time of the PID control.

The I parameter is lower for a setpoint change because using a large I parameter usually results in a large overshoot. If the gain is set correctly then the set point will still be reached quickly.

[pkm]

Thanks David.
I did a few tests and M307 H1 A650 C200 D5.0 B0 proved to work well enough.
Not sure why autotuning was a bit off.

[pkm]

I think you can adjust autotuning to more agressive temperature rise, even if it leads to a small overshoot.