New heater tuning algorithm

[sozkan]

@dc42 said in New heater tuning algorithm:

@LB, I suggest you try the new algorithm on your bed heater. The new algorithm doesn't need the long cooldown period at the end, so it won't necessarily take longer.

Hi,
In my case, it took more than 2 and a half hours as seen I have to give up and postpone for later.

How many times it cycles I don`t know.
IT might be from our bed size;

However, the previous PID was very short. The reason I am doing it again because it fail to start heating after the firmware upgrade.

By the way on PID, Dyze dual heater and dual sensor kind of struggling me.
https://docs.dyzedesign.com/typhoon.html#liquid-cooling
It keeps failed. All was working on Duet2and3Firmware-3.1.1

Kind Regards

[sozkan]

@sozkan Hi,
It is my second attempt at the trial of Autotuning heater 0.
It is started at 7:55 and it is still continuing now at 12:55. Isn`t something wrong?

The latest deal of; 12/8/2020, 9:13:27 AM Autotune starting phase 3, fan off. It keeps cycling heat on and off!

it is Duet 2 WiFi/Ethernet 3.2-beta4 (2020-11-26) installed on Duet2Wifi board.

[jay_s_uk]

@sozkan it will go to a maximum of 30 i think.
because your bed is so large i would just leave it to keep going

[dc42]

Beta 4.1 has an improved algorithm, that is sometimes much faster on large bed heaters.

[sozkan]

@jay_s_uk
Basically, it keeps cycling at stage 3. It was started at 7:56 and come to stage 3 at 9:13 the time now 15:44. I don't see any reason to continue for now.

Is there any way to set manually? I appreciate it if you suggest the changes because for me to figure out will take more time.

You can see on the table I define 5 important stages as I thought.
The recorded variables;

[dc42]

It will cycle for at most 30 cycles in stage 3. But the algorithm used in RRF 3.2beta4.1 is often quicker.

If it is now cooling down again, it has finished and you should see a result in the console. Send M303 without parameters to check.

[sozkan]

@dc42 said in New heater tuning algorithm:

It will cycle for at most 30 cycles in stage 3. But the algorithm used in RRF 3.2beta4.1 is often quicker.

If it is now cooling down again, it has finished and you should see a result in the console. Send M303 without parameters to check.

My heating table heats for about 3 minutes. from 35 to 40celsius. But cooling is 43 minutes on average due to the insulation and volume. In total each cycle taking 55 minutes on average. If 30 cycles, it is mean it should take more than 28 hours to complete.
@sozkan
By the way, this is the preview of the incomplete test of 4.0;

So I should go for 4.1 then immediately and share the results.
Thanks for the help.

[dc42]

If it takes 43 minutes to cool by 5C but only 3 minutes to heat up, perhaps you have too much insulation? With most print surfaces, you need the bed to cool down before you can release the print.

Why are you using such a low target temperature (40C)? All the materials I print with need 60C or more indicated temperature.

[sozkan]

@dc42 you right but, It is just fresh to build a handcrafted heated bed with high voltage. I am afraid to run a higher temperature for safety concerns. therefore, I am aiming to test it first at low temperatures. But I cannot test before it runs smoothly, later I can PID for normal operating temperatures again.

[sozkan]

@dc42 I was testing Autotuning with firmware latest 4.1 version, It took 7 hours, but still did not complete. The thing I have faced again, the motor cracking sound with this version too.
So I revert back to firmware 3.2.3b that works smoothly on the motor axis. But the temperature tuning things that I have seen, not for my machine type. I am trying to figure out the M301, M303, M307 M950 parameters to set up manually with the values I find. If you could help me with it I appreciated.
By the way, these parameters not valid in this firmware, wich was Firmware version 3.1.1 AutoTuning found values

M307 H0 A233.3 C508.3 D1.3 S1.00 V24.3 B0

Thanks.

[jay_s_uk]

I am beta4.1 on a Duet 3 running in standalone mode.
I have retuned my bed using the new algorithm.
Whenever I heat up the bed to print, it faults with heating slower than expected.
Here is my M307

M307 H0 R1.468 C146.948:146.948 D2.43 S1.00 V27.0 B0

I have a 500mm bed (10mm aluminium) running a 240v 2kw heater.
Anything else I can provide?
It takes the full 30 cycles to tune and fails.

[jay_s_uk]

I've rolled back to beta 2 to be able to tune the bed with the old algorithm and this was the result.

 M307 H0 A63.2 C50.0 D1.9 S1.00 V27.0

[A Former User]

This post is deleted!

[sozkan]

Extra Long Hot ends ( Like Dyze Typhoon or pulsar ) blocks with multiple sensors heater on the same block need to be calibrated together. Is there any possibility to add in the new algorithm?

[A Former User]

Good morning. I've jut finished tuning the Heater 1 at 230°C with M303 H1 S230. This is my first print after using the new calibration algorithm....I'm experiencing a lot of oscillation around the set point as you can see:

Even arouns 235°C (first layer) and around 230°C (second layer) as well.....I made the autotuning with tha fan on for the first 2 stages and turned it off for the third stage. Oscillationa are sometimes more than 1°C above and belows

Never seen oscillations like those with the older algorithm

As you can see the issue goes on for all the print:

I've been able to tune only the Heater 1 because the tuning for the bed looked to go in loop...After 40 minutes I've turned it off...as you can see the graph for the bed is very stable

What's happening?

Uh, I was forgetting....Beta 4.1

[wilriker]

@the_dragonlord said in New heater tuning algorithm:

I made the autotuning with tha fan on for the first 2 stages and turned it off for the third stage.

The tuning process will automatically control the fan. It seems you confused it by manually changing fan speed. Please tune again and don't touch the fan.

I've been able to tune only the Heater 1 because the tuning for the bed looked to go in loop...After 40 minutes I've turned it off...as you can see the graph for the bed is very stable

The new tuning algorithm takes up to 30 heat-cool-cycles and can take very long on large heated beds. So what you describe sounds pretty normal in the first place.

[A Former User]

@wilriker said in New heater tuning algorithm:

@the_dragonlord said in New heater tuning algorithm:

I made the autotuning with tha fan on for the first 2 stages and turned it off for the third stage.

The tuning process will automatically control the fan. It seems you confused it by manually changing fan speed. Please tune again and don't touch the fan.

I've been able to tune only the Heater 1 because the tuning for the bed looked to go in loop...After 40 minutes I've turned it off...as you can see the graph for the bed is very stable

The new tuning algorithm takes up to 30 heat-cool-cycles and can take very long on large heated beds. So what you describe sounds pretty normal in the first place.

the first attempt I've made the tuning didn't control the fan at all neither in the beginning neither in the end... maybe there's an option to control it?

[jay_s_uk]

@the_dragonlord you need to tune as a tool rather than a heater.
So M303 T0 S200

[A Former User]

@jay_s_uk said in New heater tuning algorithm:

@the_dragonlord you need to tune as a tool rather than a heater.
So M303 T0 S200

uh ok ok... thanks! What's the tool number for the bed?

[jay_s_uk]

@the_dragonlord for the bed you tune as normal as a heater