Is Marlins PID-autotune reliable?

3D Modelling, Slicing, Hosts and Management
Post Reply
aliasmeum
Posts: 9
Joined: Thu Jun 15, 2017 11:33 am

Is Marlins PID-autotune reliable?

Post by aliasmeum » Fri Jul 07, 2017 6:51 pm

Hello,

I am changing the printbed and silicon heater for my Taz5, but first I want to learn something about PID settings.
After watching Toms video "3D printing guides - Using Marlin's PID autotune", I have run through the routine several times and for two different temperatures. The values I get are so inconsistent that I am wondering if the PID-autotune can be trusted?

These are the values i get after running the PID-autotune for the bed 4 times with: M303 E-1 C10 S60
Run 1: Kp: 342.25/Ki: 52.92/Kd: 553.37
Run 2: Kp: 968.09/Ki: 172.49/Kd: 1358.36
Run 3: Kp: 161.85/Ki: 17.39/Kd: 376.55
Run 4: Kp: 383.62/Ki: 68.87/Kd: 534.25

The values in run 3 is close to the default values in Configuration.h (162/17/378).

This is to random for me to trust, and I feel the right ting is to stick to the default values. Can anyone tell me if you use PID-autotune and how you choose your PID values?

Regards Robert

Stefan
Posts: 31
Joined: Thu Jun 15, 2017 12:46 am

Re: Is Marlins PID-autotune reliable?

Post by Stefan » Sun Jul 09, 2017 8:46 pm

Hi Robert. I did run the PID autotune for the bed (only once) and was very pleased with the results. Keep in mind that if you run the test multiple times, the conditions should be nearly the same each time you start (IE: make sure the bed cools to room temp). The values I got were vastly different than what were set as the defaults, but the end results were positive.

-S

aliasmeum
Posts: 9
Joined: Thu Jun 15, 2017 11:33 am

Re: Is Marlins PID-autotune reliable?

Post by aliasmeum » Sun Jul 09, 2017 9:50 pm

Thank you for your answer Stefan,

Can I ask you how you determine that the end results were positive?
What factors are you looking at, and how do you measure them?

Regards Robert

ronka
Posts: 41
Joined: Thu Jun 15, 2017 4:00 am
Location: Holland

Re: Is Marlins PID-autotune reliable?

Post by ronka » Mon Jul 10, 2017 12:22 am

after i beefed up my powersupply (was horrible underpowerd) the standard marlin pid value made the bed temp vary between +/- 5 degrees.
after autotune it stays rocksolid to the desired temp.. even with nozzlecooler on it. also the nozzle pid is now ever so steady.. so yeah..
:ugeek: trust the autotune :ugeek:

bedpid changed from 10.00/.032/305.4 to 340.19/24.20/1195.55

ps why did you measure 4 times when you ordered marlin to measure 10 times??
just try M303 E-1 C4 S60 and see the result..

Stefan
Posts: 31
Joined: Thu Jun 15, 2017 12:46 am

Re: Is Marlins PID-autotune reliable?

Post by Stefan » Mon Jul 10, 2017 8:13 am

aliasmeum wrote:
Sun Jul 09, 2017 9:50 pm
Can I ask you how you determine that the end results were positive?
What factors are you looking at, and how do you measure them?
The time to 90C was markedly shorter (under 5 min), the final temp was always within 2 degrees of the target, and I used the graphing tool in Pronterface to measure the results. Previously, I was getting many automatic shutdowns due to "thermal runaway" errors because the firmware was not able to respond to changes in the room's environment. Adjusting the PIDs fixed that right away.

aliasmeum
Posts: 9
Joined: Thu Jun 15, 2017 11:33 am

Re: Is Marlins PID-autotune reliable?

Post by aliasmeum » Mon Jul 10, 2017 11:50 am

Thank you guys,

I think I am starting to understand PID-autotune a little bit better now. I guess I do not have to think about the values not being close to the default ones, this is about the relation between the values Kp, Ki and Kd.
ronka: I did run the whole routine 4 times, with 10 cycles each time, expecting (wrongly) to get similar values each run. That is why I questioned the reliability.

I tried the autotune function to find out how it works before installing my new bed, not because I have a temperature control problem. With the default PID-values, my old bed heats quickly, right on to the set temperature and stays there. On the other hand, I notice that my hotend has a temperature overshoot of about 10 deg when heating up. I guess this is not that much and the temperature seems to be stable once it stabilizes, but I will definitely try running the autotune on the hotend to see if I can get a quicker stabilization of the temperature.

Regards Robert

BoriSpider
Posts: 5
Joined: Wed Jun 21, 2017 12:55 am

Re: Is Marlins PID-autotune reliable?

Post by BoriSpider » Tue Aug 08, 2017 9:56 pm

So my delta sat for a LONG while before I started messing with it recently. So I noticed the nozzle temp going crazy during printing(no heated bed here). Good thing I'm printing small stuff dialing in KISSlicer 1.6 beta3.1 settings. So yesterday I re-tried the PID auto tune since before(maybe a wk later) it didn't work, kept failing. So I did the M303 S195 C10 and got my p i d #'s and then a M500. Today while printing temp was all over the place again so I had to do another PID autotune. I hope M500 works this time. I don't want to have to the autotune every time I turn on the printer.

Post Reply

Who is online

Users browsing this forum: No registered users and 3 guests