First Print

Discussion in 'Getting Started' started by Spoon Unit, Aug 22, 2019.

  1. Spoon Unit

    Spoon Unit Well-Known Member

    Joined:
    Sep 6, 2015
    Messages:
    1,546
    Likes Received:
    481
    [​IMG]

    Essentially moved to get a single tool online and tested to make sure everything was working from that perspective.
     
    Greg Holloway likes this.
  2. mhe

    mhe Well-Known Member

    Joined:
    Sep 5, 2015
    Messages:
    279
    Likes Received:
    76
    That surface finish is damn gorgeous!
     
  3. orcinus

    orcinus Well-Known Member

    Joined:
    Oct 20, 2015
    Messages:
    336
    Likes Received:
    113
    Printed the pre-sliced benchy from the github repo, mine’s pretty bad…
    Going to slice some test cubes with KISSlicer and try to figure what’s up.

    Layers are super uneven, top surfaces are way worse than my BigBox, and there’s insane amounts of stringing.

    Might be insufficient cooling, we’ll see.
     
  4. orcinus

    orcinus Well-Known Member

    Joined:
    Oct 20, 2015
    Messages:
    336
    Likes Received:
    113
    As an aside - is it normal for the A, B and extruder steppers to get so hot?
    I never drove them that hot on the BigBox.
    Z stepper is completely cold, though.

    Are the currents set a bit too high?
     
  5. Spoon Unit

    Spoon Unit Well-Known Member

    Joined:
    Sep 6, 2015
    Messages:
    1,546
    Likes Received:
    481
    I did wonder about the currents. I haven't run a super long print yet, so haven't really tested the heat of the motors. First question I guess is going to be; what is a reasonable temperature for those motors. I bet we can look that up. I seem to recall some recommendation that you run as little current as is required to have dependable function and no more.

    I tried a benchy only briefly and didn't have good enough adhesion at the time. I've switched to 'quality' filaments while I try to tune this thing. Just had my first successful multi-material print that's close to perfectly aligned:

    Front
    [​IMG]

    Right
    [​IMG]

    The slight bulge on the layer line is, I think, probably now down to bowden tuning and not having a perfectly primed nozzle. The bulge's are equivalent for each material, so I don't think this is down to Z alignment between the nozzles.
     
    orcinus likes this.
  6. orcinus

    orcinus Well-Known Member

    Joined:
    Oct 20, 2015
    Messages:
    336
    Likes Received:
    113
    I’d recommend staying away from glitter filaments while tuning the machine for multiple reasons.
    It obscures layer unevenenness and can sometimes cause weird issues with top surfaces that have you chasing ghosts.

    Also, upping the temps will make the prints shinier which helps in nailing down layer consistency, flow rates, ghosting (when tuning acceleration and jerk) etc.
     
  7. orcinus

    orcinus Well-Known Member

    Joined:
    Oct 20, 2015
    Messages:
    336
    Likes Received:
    113
    (Very nice print, by the way.)
     
  8. orcinus

    orcinus Well-Known Member

    Joined:
    Oct 20, 2015
    Messages:
    336
    Likes Received:
    113
    No go, my cube’s also suffering from pretty bad layer inconsistency.
    I’m going to try playing with the flow rate / Esteps/mm, although i doubt that’s the issue, as top surfaces are fine.
    Plus, how would i end up with a different Esteps/mm anyways, i’m using the same Titans the default config files were built with.

    Although…
    I’m using the hardened hobs.
    Any chance they have a slightly different diameter than the regular ones?
     
  9. orcinus

    orcinus Well-Known Member

    Joined:
    Oct 20, 2015
    Messages:
    336
    Likes Received:
    113
    Noticing a lot of little (and not so little) glitches in RRF/Duet3D FW.

    For example, i’ve been printing with the bed set through 56C from the slicer.
    I’ve just sliced another file with bed temp set to 0C to eliminate bed bang bang as the source of banding/layer irregularities, uploaded it, hit print from the PanelDue and - it set the bed to Active and started heating it up to 56C again. I’ve double checked, and the Gcode file very clearly says M140 S0.

    Edit: scratch that, it looks like the upload of the updated gcode file failed somehow.

    Then there’s the fact i’ve reenabled “Run free macro”, “Run pre macro” and “Run post macro” in the web settings, but T0 issued from Gcode file run from web still completely ignores those macros, while T0 issued from PanelDue works fine. I guess the settings are cached - i’ll see if it goes back to normal after a restart.

    There’s also weirdness that happens when pausing a print, running a tool change, and resuming print - it seems to lose the coordinate system (or gains some kind of an offset) during the tool change, so when you hit resume, it goes god knows where and crashes.

    Edit: here’s another one - sometimes, fast G1 diagonal travel moves get split up into very audible segments. Usually noticing it when cancelling a print and the tool head zooms back in front of the dock.

    I must admit i’m a bit disappointed. I’ve expected a more mature, stable platform, not another Marlin…
     
    #9 orcinus, Aug 25, 2019
    Last edited: Aug 25, 2019
  10. Spoon Unit

    Spoon Unit Well-Known Member

    Joined:
    Sep 6, 2015
    Messages:
    1,546
    Likes Received:
    481
    I think with the bowden setup, pressure advance is going to be key to getting good results. A couple of times I've enabled it and the whole machine starts acting very weird. I haven't tested yet, but i think I've found the culprit.

    https://github.com/e3donline/RepRapFirmware-SD/blob/master/sys/config.g#L46

    Code:
    M566 X400 Y400 Z8 C2 E2:2:2:2                ; Set maximum instantaneous speed changes (mm/min)
    These settings for E seem to be resulting in the whole slow down. This, for comparison is what I have for that on the BigBox:

    Code:
    M566 X600     Y600    Z30  E300:300      ; Maximum instant speed changes mm/minute (sometime miscalled 'jerk')
                                             ; Alex9779 uses M566 X480 Y480 Z24 E300:300 Elmoret had E20
                                             ; DC42 once said in a discussion of pressure advance (M572) that
                                             ;    "200mm/min extruder jerk is unusually low, so the associated acceleration
                                             ;     limit for printing moves that it causes might have an effect."
    The commentary is thanks to @Ephemeris. In any case, I've made a change and will try this next:

    Code:
    M566 X800 Y800 Z30 C2 E300:300:300:300            ; Set maximum instantaneous speed changes (mm/min)
    Will report results.

    @Greg Holloway did say that @dc42 had picked up a fully-built TC from them. It's possible they've decided to invest in the Duet expert in order to get the setup a little more on rails. I ended up going back to the original homex, homey, and homec because they seem to work 100%.
     
  11. Greg Holloway

    Greg Holloway Administrator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    984
    Likes Received:
    622
    This is most likely due to mesh leveling. As the bed moves at a slower rate to the X/Y axes it has to slow down to allow the bed to 'catch up'. This is why mesh leveling is only enabled during a print. The pause script does not turn off mesh levelling as this is only done so at the end of a print.

    I have updated the cancel.g code with

    ; Disable Mesh Compensation.
    G29 S2


    While I understand your frustrations please try to remember the ToolChanger is a very new system and there has been no platform for software to be developed on. We, E3D, took the decision to build the hardware knowing that the firmware and slicers were most likely not able to fully use the machine as we intended. The ToolChanger and Motion System are a reference platform to develop ToolChanging. As the number of machines being used goes up, more and more features and functions will become available and made reliable. Consider yourselves pioneers of what will, in a few years time, be the next generation of integrated additive manufacturing.
     
    Ntesla66, mhe, orcinus and 1 other person like this.
  12. Spoon Unit

    Spoon Unit Well-Known Member

    Joined:
    Sep 6, 2015
    Messages:
    1,546
    Likes Received:
    481
    Actually, I reckon the problem with the chunked motion might also be resolved with some change to M566. We used Z10 on the BB, but on the TC, it was set to 8 initially. Given the speed that think can move, I suspect we could move up from 10 to say 20. Could be some other setting though. There are so many settings.

    Anyway ... results with bowden were a bit poor

    Front:
    [​IMG]

    Side
    [​IMG]

    On the bright side, an extra 0.05mm nudge seems to have fully aligned the print, although the random stop points (to accentuate bowden settings while tuning), mask that a little. It's really close now at any rate.

    Enabling bowden did work after changing M566 to allow faster speed changes on E. It didn't sound brilliant, with the motors sounding like they were grinding a little. The sight of the 'zitty' print did make me think that this sort of print could also be good input for the idea of @Amr to setup some machine learning to give early warnings of this sort of result.

    I think the wiper I put together is not functioning as well as it could. Perhaps it's too floppy. Also, with any remnant held there, it can be picked up by the next tool. Some serious wiping tests are needed to develop a good result. Also the debris fall onto the spools below, which is not ideal, though I don't think there's much chance of a serious problem resulting from that.

    By the way, on that print, K was set to 0.1 lower down, 0.2 in the first blue, then down to 0.05 and lower. I'll probably revert to the slicer-directed retraction/wiping for now as that was so much easier on the eye.
     
    orcinus likes this.
  13. orcinus

    orcinus Well-Known Member

    Joined:
    Oct 20, 2015
    Messages:
    336
    Likes Received:
    113
    That's still way better than what i'm getting.
    I'm not getting zits and blobs, i'm getting whole layers that stick out 0.5mm further than others.
    And whole layers that are inset by 0.5mm or spottily extruded.

    And no, they're not misaligned, they're *that much misextruded*.

    It's not the bed (disabled heating to eliminate that), it's not the fans (disabled to eliminate "freezing"), it's not the extrusion width being too small or too big, or the extrusion multiplier. I've pretty much eliminated everything except the firmware itself, the drivers, and the extruder+hotend. It was too late to try different tools, hoping to do that tonight.

    @Greg Holloway thanks for the mesh bed leveling fix.
    And yeah, i understand things are very experimental, but there are things that have nothing to do with toolchanging that are either just odd and clunky, or just don't work as expected. For example, if you hit print on the PanelDue, you get the print progress on the PanelDue, but the web sometimes UI perpetually reads just "processing 0%" for the g-code file. I know i sound whiney, i'm just overly sensitive to quirks and bugs after all the issues with Marlin and IDEX-es (IDices?), and the switch to 2.0.x which was supposed to fix things, but instead just made a greater mess :)
     
    #13 orcinus, Aug 25, 2019
    Last edited: Aug 25, 2019
    Greg Holloway likes this.
  14. Greg Holloway

    Greg Holloway Administrator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    984
    Likes Received:
    622
    If you have constructive criticism and bug reports to give, the duet forums are the best place for it. I know both David, Tony and Chris are on there, plus all the other Duet users too. They will probably be able to better advise as they'll have heaps more experience :)

    https://forum.duet3d.com/
     
    orcinus likes this.
  15. orcinus

    orcinus Well-Known Member

    Joined:
    Oct 20, 2015
    Messages:
    336
    Likes Received:
    113
    Will do.

    Switched to the second tool, results seem much better, although still not perfect.
    I still get the occasional overextruded or underextruded layer, but it's at the edge of what i'd call acceptable.

    I wonder what's wrong with tool 0...
    I've rechecked the PTFE tubing, adapters, everything. It's all tip-top.
    A this rate, it's got to be either the Titan, its stepper, or the hotend/nozzle.

    I'm running 0.3mm Nozzle X with a nickle plated block on Tool 0.
    Tool 1 is a 0.6mm Nozzle X also with a nickle plated block.
    I'll give Tools 2 and 3 a try too later, those are regular blocks and 0.4 and 0.25mm Nozzle Xs respectively.

    Edit: bizarrely, the top surfaces on both tools turn out smooth as silk... in fact, i think this is the first time in my 7ish years of 3D printing that i'm getting top surfaces that are smoother than side surfaces.
     
  16. orcinus

    orcinus Well-Known Member

    Joined:
    Oct 20, 2015
    Messages:
    336
    Likes Received:
    113
    Interesting.

    The faster i print, the more consistent the extrusion.
    I vaguely remember having that happen once years ago on some machine, but i'm too senile to remember what the cause was...

    Extruder jerk or acceleration?
     
  17. Spoon Unit

    Spoon Unit Well-Known Member

    Joined:
    Sep 6, 2015
    Messages:
    1,546
    Likes Received:
    481
    Grip setting on the Titan? Truly horrible old filament? Retraction settings on the edge of sanity? Thinking about it, I did have some really weird results, which as far as I can tell boiled down to half-jams. So you'd have no extrusion, and than all that pressure in the bowden tube would really and you get a horrible gush of material. Funnily enough, that also happened for me at points of slow printing and the second time I printed, I sped it up to avoid the slow speed and the problem was resolved.

    I think we'll ultimately find that retraction of 4mm might be right on the edge of sanity. A transparent nozzle and hot end would really help to be certain.
     
    #17 Spoon Unit, Aug 25, 2019
    Last edited: Aug 25, 2019
  18. orcinus

    orcinus Well-Known Member

    Joined:
    Oct 20, 2015
    Messages:
    336
    Likes Received:
    113
    Tried everything from no retraction to 2.5mm retraction.
    Absolutely no effect (apart from the blobs and strings, of course).

    Tried loose grip and tight grip, no effect.

    Truly horrible old filament is a possibility... Good point.
    I've been using some old filament from the bottom of the drawer to print all the tests.
     
  19. Greg Holloway

    Greg Holloway Administrator
    Staff Member

    Joined:
    Sep 4, 2015
    Messages:
    984
    Likes Received:
    622
    I was going to suggest the idler tension too..... Also check the hob teeth.
     
  20. Amr

    Amr Well-Known Member

    Joined:
    Jun 2, 2019
    Messages:
    130
    Likes Received:
    30
    I have not started printing Multi Materials yet, I am trying to get the kinks out of a single material first, bur I did notice a huge improvement in print quality when I speed things up, I am currently printing at 60 mm/s flawlessly. I have been away from Simplify3D for a while and I am kinda still getting back to remember where and how to do things but the profile that I downloaded from the git hub was a good starting point. PS: I never had any blobing issues I only had stringiness on my prints which disappeared by speeding up things and bumping up temps a notch, don't know why should increase stringiness o_O.


    It defently would help. but I will need a bunch and while printing not after the fact, the idea is to create an early warning system of sorts so it needs to be able to identify issues while printing I will set up a share folder some where, if any one is interested in helping they are more than welcome to upload pictures(Alot of them maybe even time lapses) for good and bad prints, I have worked out a good part of the solution (The hardware part of it and some of the software) so the more images I have the faster the AI Model is to be ready.
     

Share This Page