›

DEADTIME control

Tinchus2009Tinchus2009
in Repetier-Firmware
Hi. I have upgraded from 092 to altest firmware.  But now my hotend temperature is not as stable as used to be. I use deadtime control, and with 0.92 I use to have a variation of 0.5 degrees in temperature or even less
Now, still useing deadtime control, useing same values, same extruder, all the same, temperatures goes around like +-2 degrees. I have oovershoots going on, and also temperature then drops at least 1 degree, at least. I have tried to recalibrate, but with no success
So, has something changed in the way of calculateing the deadtime value or anything that can help me understand what is going on?
Also, I tried going back to PID, but after the autotunning, I try to introduce the values manually executeing : 
M301 Px Ix Dx and I get in the console "unknown command" I havent used PID control for a Looooong time, may be repetier is not useing the standar sintax? if not please point me in the correct direction
Thanks in advance

Comments

  • RepetierRepetier
    Yes, we made a change making pid work betterm but on dead time control it had the opposite effect. Use eeprom editor to set new pid values or tell m303 to store them directly. See in repetier.ino file for syntax and supported gcodes.
  • Tinchus2009Tinchus2009
    ok, great. another question: can the manual tuning of PID values be entered WHILE running a test print so I cant tune the parameter with the hotend doing real" print (tune parameters while hotend is moveing and extrudeing)?
  • RepetierRepetier
    Yes, you can do it during a print. There might a tiny pause when eeprom is read, so do it not during a quality print.
  • DeJeDeJe
    I am facing same issue. PID control isn't really working at all. At least if changing temperatur during print I have over/undershoots of about 5-6 degrees +-!
    Deadtime I was using a long time now reacts better than PID but nowhere near it was before. Variations about +-2 Degrees.

    I myself ask why fix something which is not broken. ;) Deadtime was working nearly perfectly.
  • RepetierRepetier
    Yes, pid is not perfevt and has always over or undershoot on changes, that is beacuse of how it works. Dead time was working good for some and not so good for other cases. In v2 firmware i want to improve on the problem and come back to old plus a cooling delay which might differ. Hopefully that will be even better then the old dead time.
  • DeJeDeJe
    OK, is it possible to test v2 already? My printer is not used for production so I could test. ;)
     

  • RepetierRepetier
    V2 does not have dead time so far and other things are also not ready, but you can see on github in the dev2 branch.
  • Tinchus2009Tinchus2009
    Just a uestion: why dead time has been abandoned? so far worked really really well for hight temperatures, at least that was what I observed: for high temperatures worked much better than PID. PID worked much better for lower temps like 210 and lower
  • RepetierRepetier
    It is not abadonned, but now works on shorter time lookback - we not have the last 4 temperatures in 0.1s interval any more only the last with 1s resolution. And that seems to make it not as stable. Thinking of it i might have missed changing a factor for the time used now 1s instead of 0.3s before. The increased time made pid more stable but deadtime worse. In V2 each method has it‘s own class so each can have the best solution regarding time table.
  • DeJeDeJe
    @Repetier
    Is it possible to just change it back in source? Or there are other changes involved which makes it hard revert?
    If yes for easy change, where do I change it?
  • RepetierRepetier
    It is hard to revert because we had before a temp array with last 4 entries, now we only have the last and update only every second to new temperature and not every 0.1s. So parameters for the algorithms have changed.
  • DeJeDeJe
    Ah I see.
    Thanks for answer.
  • DeJeDeJe
    edited August 2018
    @Repetier, Sorry, I can't edit anymore...
    In general I don't care about PID control...so would it help to shorten the intervall from 1s to something 0.2s? I think this could help deadtime control, isn't it? Even if we don't have 4 tempstamps control loop would be much faster. ;)
  • RepetierRepetier
    Yes might help. I added some extra code to cope with the new problem, but that also did not really help. So you can also remove that part.
  • DeJeDeJe
    Can you direct me at which source I need to change something? Especially where do I find the 1s interval?
    I will let you know here if it works better for dead time control. :D

    Thanks in advance.
Sign In or Register to comment.