UNSOLVED Hiccups on 3HC expansion board

  • I'm getting a number of hiccups on a 3HC expansion board.

    Moves scheduled 495904, completed 495904, hiccups 43894

    Is this normal?
    I'm running extruders off this board.
    My microstepping is set to 16 with interpolation

    Microstepping - X:16(on), Y:16(on), Z:16(on), U:16(on), E:16(on):16(on)

    And I've only been using the second extruder recently.
    Anyone got any thoughts?

  • @jay_s_uk No, but if I get chance, I'll try it on my machine. All six of my extruders are on expansion boards.

    Edit. Are you on latest firmware? And what are your extruder steps per mm?

  • I'm on the latest firmware. 3.1.1 on the 6HC and 3.1.0 on the 3HC.
    Also running 3.1.1 of DSF in SBC mode.
    My steps per mm are 409 (I'm using a hemera which at the moment is printing like sh**).
    So I don't think I'm taxing the expansion board at all.

  • @jay_s_uk I'll try later. I can't run any prints but I'll try exercising the extruders by doing some G10/G11 retraction cycles (once I have disconnected filament from blocked heat break). How are you getting the report I'm guessing M122 B1?

  • Yes. M122 B1.
    I checked as I'd left a print running to see what sort of print quality I was getting from the hemera and when I came back, the filament had been fully retracted. I'm inclined to say that my poor quality of prints so far has been down to the number of hiccups I'm seeing on the expansion board.

  • administrators

    I think that's an issue with the expansion board firmware. I have an update planned. I'll try to get a beta available later today.

  • @jay_s_uk I'll try later and see if I have any similar issues. The filament being fully retracted is usually caused by having an extruder move like G1 E-5 but that move being with the firmware set to expect absolute extrusion whereas the slicer is set to relative. So instead of retracting by 5mm, it retract to the position of 5mm from when the extruders was set to zero which is usually at the start of the print.

    That might explain your other problems too. If your slicer is set to produce relative extruder moves, make sure that you have M83 in your configuration file to "tell" the firmware this is the case. Conversely, if your slicer is set to produce absolute extruder moves, make sure your configuration has M82 rather than M83.

  • administrators

    @dc42 said in Hiccups on 3HC expansion board:

    I think that's an issue with the expansion board firmware. I have an update planned. I'll try to get a beta available later today.

    I meant the hiccup issue, not the filament retraction.

  • @deckingman
    I've check the gcode file and I'm definitely in relative mode.
    If extrude steps are being missed but retraction steps aren't, that would manifest itself as eventually the filament being fully retracted.
    I will report back as soon as dc42 releases a beta for testing

  • @jay_s_uk Yes, but is your firmware configured to expect relative moves? Do you have an M83 in your configuration file. If so, is there an M82 anywhere else that could override it? i.e. in your start gcode or config-override.g.

  • I have M83 in my config, an M82 at the start of my gcode and then a switch straight back to M83.
    Its setup the same as my other machines.

  • OK, so I just tried a few G10/G11 cycles to exercise my extruders. To be clear, the reason why I did that is that I can't run a print right now so have no other way of exercising the extruders. Also firmware retraction happens at a much faster speed than "normal" extruder moves and exercises all 6 extruders simultaneously.

    After about 20 cycles, I ran M122 B1 and the result shows zero hiccups. My steps per mm are around 410, micro-stepping is 16X. Firmware retraction is set to use a feedrate of 3600.

    So I don't think this is a firmware issue. Also, I did manage to get a print out of my machine yesterday using 3 extruders in mixing mode which wasn't half bad.

  • @dc42 any update on a beta build to sort the hiccups out? I'm happy to carry out testing etc

  • administrators

    It's built, but I've not had a chance to test it yet.

  • @jay_s_uk Have you managed to sort out why the filament gets fully retracted at the end of the print? Are you absolutely sure that there are no problems with your configuration or the gcode file that you are trying to print?

  • @deckingman
    It's not at the end of the print, its mid print.
    And I use the same sliced files on all my other machines which run duets

  • @jay_s_uk Perhaps you have a blocked, or partially blocked nozzle, causing the extruder to skip steps on extrusion moves but not on retraction moves? I've only done limited printing with my experimental hot end, but I can't replicate your problem.

    Edit. Try disconnecting the filament from the hot end and run that print again to see if you still get hiccups.

  • @deckingman
    I'm running a print now without filament and I'm still getting hiccups

  • @jay_s_uk What does extruder movement look like? Normal or erratic? Can you upload the gcode file somewhere and post a link to it. I should have my machine back together soon so I can try running your file and see if it gives me the hiccups too 🙂

  • Hmm, I've swapped from driver 1 to driver 2 and the problem has gone away.

  • For info, I've just finished a print of about half an hour, ran an M122 part way through and another at the end - no sign of any hiccups

    Annotation 2020-07-01 142259.png

    Annotation 2020-07-01 140614.png

  • Ok, the issue is still there after swapping drivers so ignore that

  • administrators

    I'm sorry for the delay. The new firmware was affected by a chip feature that didn't cause a problem in a debug build but did in a release build. It's taken me more than half a day to track it down and fix it, and now I need to review and remove most of my earlier attempted fixes. I should have a beta tomorrow.

  • @dc42 no rush

  • administrators

    I've put a binary of the new EXP3HC firmware at https://www.dropbox.com/s/7rqp6pul9ip3yam/Duet3Firmware_EXP3HC.bin?dl=0. Please monitor it carefully when using it, it's had only a little testing.

Log in to reply