[3.4.0 beta 3] Heater 0 fault

dc42

@siam there was a bug in previous firmware versions whereby for slow heaters, the heater fault detection would never kick in. That is fixed in 3.4beta3.

Your bed heater has an abnormally low dead time (M307 D parameter). That is probably contributing to the problem. When you turn the heater on, is there a noticeable delay before the temperature display starts rising? If so then I suggest you increase the D parameter, to be greater than that delay time.

SIam

@dc42 The first 50C° will heat the bed without problems the error occurs first on higher than 50C°

dc42

@siam what is the approximate air temperature? A low ambient temperature could affect the behaviour.

SIam

@dc42 the room temp. Is at 20C°

SIam

Here is a video from the heating process

https://siam-net.de/temperror.mp4

small correction the problem begins at 40C° not at 50 C°

Phaedrux

Can you provide some details about the heater?

Can you try a D value of 2?

SIam

@phaedrux it is a silicon heater (400x400 clued on an ALU plate 420x420x8) with integrated thermistor 220V / 500W connected with an SSR. I have yesterday played around with the D value but this without success.
One thing wonders me why i sometimes get this error:

Error: Heater 0 fault: temperature rising much more slowly than the expected 0.3°C/sec

and sometimes this one:

Error: Heater 0 fault: temperature rising much more slowly than the expected 0.4°C/sec

with the same settings

deckingman

I'm wondering if this is related to the integrated thermistor and the thick(ish) plate. I found that having the temperature sensor close to the junction between the heater and the plate is a bad idea. What tends to happen is that the heating process starts, the temperature at the junction between the heater rises very rapidly (hence the very low dead time) but the plate itself, especially the upper surface is still relatively cold. Despite this, the heater turns off because the junction where the thermistor is fitted is hot. After a period of time, the junction temperature falls, the heater turns on but due to the rapid increase in temperature at the junction where the sensor is located, it quickly turns off again. The process repeats until eventually the top of the plate reaches the desired temperature but it takes an awfully long time due to the on-off-on cycling of the heater during what should be the fully on phase of the heating process.

I cured mine by drilling a 3mm dia hole, as deep as possible into the edge of the plate and as close as possible to the top surface and putting the thermistor in there. For info, I also have a 400 x 400 plate but it's 12mm thick and I also have a 240 V heater via an SSR but mine is 800Watt (IIRC). My D value (the dead time) is 28.16 seconds (tuned in January 2021 using RRF3.2).

SIam

@deckingman Thank you for your details, it sounds so far logic and i will give it a try when i get no other solution. What me wonder is that the printer never had a problem in the last 5 years and now with the latest beta occurs this problem

deckingman

@siam said in [3.4.0 beta 3] Heater 0 fault:

@deckingman Thank you for your details, it sounds so far logic and i will give it a try when i get no other solution. What me wonder is that the printer never had a problem in the last 5 years and now with the latest beta occurs this problem

Yes it looks like something has changed in firmware, born out by what DC42 has said. Moving the thermistor will be beneficial but it doesn't explain why you get the error now but you didn't before. I mentioned the position of the thermistor because that might help to understand how that interacts with whatever has changed in the firmware. Looking at your video, it seems that the temperature rises rapidly (as I would expect with the sensor being fitted where it is) but then the temperature rise seems to pause before rising again. Almost as if the heater turned off because of the rapid rise in temperature at the sensor, then of course the temperature rise became slow which triggered the error. If the firmware looks at the rate of change of temperature more frequently that it did before, that would explain why you see the error now and you didn't before with earlier firmware versions.

hestiahuang

Duet2wifi, 350W power supply.

I hitted the same issue when heating the BED. I have tried to run auto PID for it, also tried BANG-BANG mode for it, and I increased the dead time to around 10 too. It doesn't work. However I have used my printer for two years, I never saw that before. I think it may be the firmware issue, V3.4 b3. Then I have fallbacked the related change and rebuild, it comes normal now. Hope this can help you, if you doesn't print with PLA only, it will hapen when you need high temperature BED, like ABS, PC... Purge high performance PC, need around 140degree BED.

Phaedrux

@deckingman said in [3.4.0 beta 3] Heater 0 fault:

it doesn't explain why you get the error now but you didn't before.

I believe it's due to this bug fix in the beta

Detection of initial heating failure did not happen for heaters with slow heating rates

https://github.com/Duet3D/RepRapFirmware/wiki/Changelog-RRF-3.x-Beta-&-RC#reprapfirmware-340beta3

It may now be catching faults that were missed before and may be too sensitive.

hestiahuang

@phaedrux
I guess so, but it must be not make sense that it doesn't work even in BANG-BANG mode.

hestiahuang

Can adjust or tuning the two conditions below temporary, in LocalHeater.cpp
if (derivative + AllowedTemperatureDerivativeNoise < expectedRate * 0.75
&& (float)(millis() - timeSetHeating) > GetModel().GetDeadTime() * SecondsToMillis * 2)

Phaedrux

@hestiahuang said in [3.4.0 beta 3] Heater 0 fault:

@phaedrux
I guess so, but it must be not make sense that it doesn't work even in BANG-BANG mode.

Bang bang mode hasn't been very practical for quite a while. the firmware needs some idea of the heaters performance to allow the heater fault watch dogs to function. So PID tuning is highly recommended and basically required.

I assume that @dc42 will be able to tune the heater fault detection to be less sensitive to false positives.