Strange new heating problem

Discussion in 'Calibration, Help, and Troubleshooting' started by Dr Jeep, Oct 16, 2016.

  1. Dr Jeep

    Dr Jeep Well-Known Member

    Joined:
    Sep 12, 2015
    Messages:
    530
    Likes Received:
    169
    Had a successful print running last night, new print with the same profile but different part today yields this problem.

    https://s3-eu-west-1.amazonaws.com/ws-share/Printer_crash.tiff

    A few minutes into a print the printer just gives up heating the bed (note the target temp doesn't change, except for the manual change to 81C I made to try and kick it back into heating again). It's done it on 3 consecutive attempts (the 3rd after a power down reset).

    You might think it's a connection problem or fet driver overloading on the board etc...but when not printing the printer will happily keep the board heated.

    Eventually it crashes (similar error on the printer screen saying heating failed and reset required)

    Running RC7, no recent changes have been made for about the last 30 successful prints.
     
  2. Dr Jeep

    Dr Jeep Well-Known Member

    Joined:
    Sep 12, 2015
    Messages:
    530
    Likes Received:
    169
    Forgot to add, If I abort the print before it fails and then try manually heating the bed, it works.
     
  3. Dr Jeep

    Dr Jeep Well-Known Member

    Joined:
    Sep 12, 2015
    Messages:
    530
    Likes Received:
    169
    Right, I think this is marlin running out of ompfh on the rumba :(

    I noticed that I could only repeat this with the small part I was printing and the larger parts I was printing last night still run through ok.

    So it seems that something about short layers (this is a small part) makes life harder for marlin and it seems to run out of time when it can manage the PID loop for the Bed. I have LIN_ADVANCE enabled but even setting the K value for that to 0 doesn't fix it (I think LIN_ADVANCE still needs the cpu ticks to do the maths even when you set that to 0, looking at the code it doesn't truly disable it)

    If I print 6 of the same part then I can get through the print but you can see a dip in the temp when the layers are at their fastest (about when it recovers is when some supports start and the layer time goes up).

    https://s3-eu-west-1.amazonaws.com/ws-share/Printer_crash_2.tiff

    Probably I think I'll have to either give up on LIN_ADVANCE or think about using BangBang for the bed heaters to get some cpu ticks back. Though I am going to post an issue on the marlin tracker as well.

    Another thing of note is the shed is a bit colder than I usually have it when I am printing (though not as cold as I have had it when printing on my old K8200 or on older releases of Marlin on the BB)
     

Share This Page