Spurious heater faults again

trobison

Thank you very much for taking the time to explain that. This is very useful. I have set my initial temperatures in the config.g file. Once again, thank you for your time.

Phaedrux

@dc42 Related thread: https://forum.duet3d.com/topic/29113/pid-d-bed-keeps-failing-on-heat-up?_=1660537110945

Tinchus

Im @dc42 . My problem is with the heated chamber. after tunning and new values added, thes values eems to not work and I always get the "temp rising too slow"... even repeating the tunning, the values are the same but alter they fail.
Tunning doesnt fail, temperature is reched and tunning gives me the values. I add them to config.g.
Then these values fails

Notepad

I also appear to have started having this bug with the new firmware 3.4.0 and 3.4.1.
Bed type is a silicone heater slapped onto a 3mm plate, 400w heater.
Most common temperature for it to fail is 75°C when trying to heat to 100°C, and the bed is PID tuned.

It makes me think that the sensor is getting a momentary bad read, and the fault sensitivity is too high. Ill get a message saying something like expected temperature rate 0.9°C/s measured rate 0.2°C/s. However after looking at the temperature graph it shows the expected rate - this is why I believe the fault algorithm is slightly too sensitive.

Tinchus

Here I have the latest data from a tunning try today: it is my heated chamber

I ran M303 H0 P0.88 S120

The P0.88 is because I need to lower the power input to the resistor in order to not burn it.

The tunning finished OK:

8/17/2022, 2:39:15 PM Auto tuning heater 0 completed after 3 idle and 14 tuning cycles in 8036 seconds. This heater needs the following M307 command:
M307 H0 R0.125 K0.058:0.000 D44.76 E1.35 S0.88 B0
8/17/2022, 1:10:13 PM Auto tune starting phase 3, measuring
8/17/2022, 12:47:41 PM Auto tune starting phase 2, settling
8/17/2022, 12:25:23 PM Auto tune starting phase 1, heating up

I saved the values, I restarted the board, I checked new values are being used.

Now I try to heat the chamber and I get:

Error: Heater 0 fault: temperature rising too slowly: expected -0.02°C/sec measured -0.03°C/sec

This has happened with every try to tune PID.
The PID was working on this same chamber , same resistors, same PWM in version 3.3

Actual version is 3.4.1, board duet3 in SBC mode

ctilley79

@dc42 I'm still having this problem with 2.4.2 RC 2. When cold, the bed warms up fine. However, If I turn the heater off before the bed temp is reached and then turn the heater back on less then a minute later, the temp will rise about 10C before throwing the error.

One thing to note is I'm using a heating pad with a built in thermistor, and this type of setup results in the pad heating up well before the bed. So my guess is it has something to do with the bed being warm when the heater is turned back on and that's why the PID tune isn't seeing the values it expects. I'm not sure what changed in 3.4.2 that is causing this.

deckingman

@ctilley79 If you have a reasonably thick bed (which you should have) then having the thermistor at the junction between the heater and the bottom of the bed is the worse place you can put it. When you turn the heater on, the temperature at that junction rises rapidly, so it turns off. Then when the temperature at the junction drops, the heater turns on but soon turns off again. Meanwhile, the temperature at the top of the plate is still cool. This constant one-off cycling plays havoc with the PID tuning and results in excessively long warm up time (for the top of the bed). What I did was drill a small hole in the edge of the bed, as deep as possible and as close to the top surface as possible, and put a thermistor in there. No more probs.....

deckingman

@tinchus This doesn't answer your question but using 88% pwm will not limit the power to your heater. It just means that it'll have full power for 88% of the time.

ctilley79

@deckingman said in Spurious heater faults again:

cycling

I had zero problem for 2 years. The pid tune seems to compensate well enough, until 3.4.2 RC2. Duet provided instructions for using these types of heaters with built in thermistors so I know the less than ideal placement of the thermistor isn't the issue. I do have to heat soak my chamber for a good 30 minutes to allow the heat to transfer to the entire bed. But it worked. Not ideal, but it worked until upgrading to RC2.

Regarding PWM. If I recall, depending on the PWM frequency, it will result in slowing down the heating process via pulsing from high/low 60% of the time.

deckingman

@ctilley79 said in Spurious heater faults again:

Regarding PWM. If I recall, depending on the PWM frequency, it will result in slowing down the heating process via pulsing from high/low 60% of the time.

The PWM value is the percentage of time spent "On" vs "Off". So if you used a value of 0.6, then it would indeed spend 60% of the time on (at full power) and 40% of the time off. The PWM frequency is how fast it cycles. So for example if you used a frequency of 100 Hz, that would be one cycle every 10 milliseconds. So within each 10ms period it would be on for 6ms and off for 4 ms.