PID Issues 3.4 RC1

Reefwarrior

Hi all,

So I have issued now M303 T0 S220....

Everything goes ok-ish; and after 24 minutes of 'tuning', I got:

Warning: heater behaviour was not consistent during tuning
Auto tune starting phase 4, measuring with fan on

This happens to both my hotend and my bed. Bed is AC powered by a relay.

Any help/advice?

Tinus

Reefwarrior

@reefwarrior Ok so the following:

Warning: heater behaviour was not consistent during tuning
Warning: Auto tune of heater 1 failed due to bad curve fit (R=3.805 K=9.435:0.000 D=8.61)
31/01/2022, 06:45:31 Auto tune starting phase 3, measuring
31/01/2022, 06:42:56 Auto tune starting phase 2, settling
31/01/2022, 06:42:06 Auto tune starting phase 1, heating up
31/01/2022, 06:42:01 M303 H1 P0.5 S220 Y3
Auto tuning heater 1 using target temperature 220.0°C and PWM 0.50 - do not leave printer unattended

I've tried various different P values - all result in the same. 'behaviour was not consistent...'

Reefwarrior

@reefwarrior also running: Firmware: RepRapFirmware for Duet 2 WiFi/Ethernet 3.4.0beta7+7 (2022-01-14)

Phaedrux

Can you describe your hotend and bed setup?
What's the fan setup like? How big is the bed? Where is the thermistor mounted?
It's basically saying the tuned results don't meet expectations.

Reefwarrior

@phaedrux Morning. My setup is an Ender3 converted to a switchwire setup.

For reference the link is:

https://github.com/VoronDesign/VoronUsers/tree/master/printer_mods/Gizzle/ender-3_(pro)_switchwire

Whole setup is 24V and heater cartridge is 40W. Thermistor for both bed and hotend is 100k (NTC 3950) type. Bed has an AC silicon heater mat (with the thermistor being integrated in the middle of the bed, and controlled by AC relay from the duet board's 'BED' connectors); accompanied by the original Creality 3D glass bed (approx 235 x 235 x I think about 4mm). The hotend is a E3D V6-like type; with thermistor in the heat block (it looks like a round cylinder of prob 15 to 20mm in length).

Current parameters for the 'hot section' in the configuration file is:

; Heaters
M308 S0 P"bedtemp" Y"thermistor" T100000 B3950 ; configure sensor 0 as thermistor on pin bedtemp
M950 H0 C"bedheat" T0 Q10                          ; create bed heater output on bedheat and map it to sensor 0
M307 H0 R1.476 K0.523:0.000 D5.70 E1.35 S0.90 B0 P0.8
M140 H0                                        ; map heated bed to heater 0
M143 H0 S120                                   ; set temperature limit for heater 0 to 120C
M308 S1 P"e0temp" Y"thermistor" T100000 B3950  ; configure sensor 1 as thermistor on pin e0temp
M950 H1 C"e0heat" T1 Q10                          ; create nozzle heater output on e0heat and map it to sensor 1
M307 H1 R3.135 K0.382:0.302 D8.78 E1.35 S0.80 B0 V24.1
M143 H1 S280                                   ; set temperature limit for heater 1 to 280C

The parameters used in both M307 commands were the ones that was given after the PID tuning; and are stable enough to provide a constant temperature for both bed and hotend, but PID tuning still resulted in the odd behaviour.

Hope my description provides a bit more clarity?

Rgds
Tinus

Reefwarrior

So for what it's worth; I upgraded firmware to RC1; still experiencing the same issue when trying to do a PID tuning.

dc42

@reefwarrior the message about inconsistent behaviour is just a warning. It's not uncommon when tuning a bed header in which the thermistor is in better contact with the heating element than it is with the bed plate. It is much less common when tuning a hot end. Can you provide a photo of the hot end?

Reefwarrior

@dc42 Hi; I will provide a photo shortly. I have in the meantime exchanged everything except for the control board - new heater cartridge, new thermistor - and same result. I have changed output of the hotend cooling fan to only 50% of the PWM; and even the part cooler fan I have defined to have a maximum of 70% PWM - still the same issue. Other thing is it takes like 30 to 40 minutes for a single PID session (only to have the warning) --- is that normal?

For what it is worth, here is the printer in action: https://www.reddit.com/r/voroncorexy/comments/snin2s/ender3_converted_to_voron_switchwire_serial

It doesn't have an issue with a short print, however, I have tried a 3 hour print yesterday, just to have it failed basically because the 'rise in temperature' was slower than that which is expected - I forgot the exact wording of the error.

I will provide a photo soon of the hotend itself.

Rgds
Tinus

Phaedrux

@reefwarrior said in PID Issues 3.4 RC1:

D8.78

The dead time for the hotend seems high to me.

Reefwarrior

@phaedrux
@dc42

Sorry I have not uploaded a photo yet - been a bit busy- only got the printer fully assembled again this evening - weird thing now is the temperature stays about 10 to 15 degrees HOTTER than requested.

Did a tuning with 200C and PWM of 0.8 and this was the result:

M307 H1 R2.723 K0.545:0.322 D10.96 E1.35 S0.80 B0 V24.1

Rgds
Tinus

Phaedrux

That dead time of 10 seconds seems unusually high for an E3D V6 style hotend.

Reefwarrior

@phaedrux That's the 'D' parameter in the line? To what do you suggest should I decrease it?

Phaedrux

It means it's taking 10 seconds for the sensor to register a change in temp between the time heat was requested and when it's detected. For my V6 it's closer to 2 seconds. So I wonder if there is something off with assembly causing that.

Hence the request for photos to see what is happening or if it's an unusual design etc.

Is the hot block in direct contact with the cold end heatsink or something? Or the heater cartridge or sensor not firmly secured?

Reefwarrior

@phaedrux Everything is firmly secured - did double check it

The only thing that may be a little weird is the gap between the nozzle and the heat block, may be a tad big.... but that should not be an issue, should it?

Phaedrux

Maybe? We're looking for anything non-standard, so...

OwenD

@phaedrux
Have you done a PID autotune with ~~2.4rc1~~ 3.4rc1?
I'm getting D values in the 10 second range with my E3D V6 as well.
As you probably know I've been experiencing issues with wild temperature fluctuations with later releases.

Reefwarrior

@owend you mean 3.4rc1?

OwenD

@reefwarrior Yes. My fingers are dyslexic

OwenD

Not wanting to hijack this thread......
A search reveals that in November 2021 in V3.4.0b6 , I was getting PID results of

M307 H1 B0 R1.594 C358.1:217.5 D6.05 S1.00 V24.3

With same hardware in 3.4rc1

M307 H1 R2.848 K0.325:0.197 D10.51 E1.35 S1.00 B0 V24.3

Different algorithm of course, but I wouldn't have thought the dead time would change that much as a result and it may explain the over shoots I now get.

Reefwarrior

@owend I will try a different firmware version tomorrow, however 3.4 is needed to keep my mini12864 up and running. My overshoot is typically 10 to 15 degrees at the moment...