1. Got a question or need help troubleshooting? Post to the troubleshooting forum or Search the forums!

Layer Alignment

Discussion in 'Troubleshooting' started by ammulder, Feb 6, 2017.

  1. ammulder

    ammulder Member

    Joined:
    Feb 3, 2017
    Messages:
    37
    Likes Received:
    2
    The layers don't seem to be especially well aligned on my C2, leading to rough vertical surfaces. I thought it might be just an issue with the settings used for the test print on the device, but from the look of it, my real job is suffering the same issue (which I prepared with the Robo Cura download for Windows and printed via USB stick).

    Here's the vertical straight edges from the test print. If you bring up the full images you can see one looks OK, and the other, not so straight:

    IMG_9609.JPG IMG_9611.JPG

    (I'll print a side by side with the real print and the same done on another printer in an hour or two when the job finishes. But I assume the issue is the same.)
     
  2. ammulder

    ammulder Member

    Joined:
    Feb 3, 2017
    Messages:
    37
    Likes Received:
    2
    Here is a sample print from a couple angles (same filament, same temp, .15 mm, no infill). In each case the C2 is on the left (from Robo Cura Windows download and printed via USB stick), and a print from a Replicator 2 is on the right (from the MakerBot software, which doesn't seem any better than Cura in general).

    IMG_9623.JPG
    Above: Look at the right-hand side of the left print (C2) compared to the right-hand side of the right print.

    IMG_9621.JPG
    Above: the back side isn't as bad, but you can clearly see layer lines on the left print (C2), and to a much greater extent "mid-brick" than the right print.

    I did four different models in this print job, and they all came out similarly. If I pick up matching tiles and just look at them in my hand, it's clear from the layer lines which one came from the C2.

    Any ideas on this?

    P.S. The top surfaces are equivalent, perhaps slightly better on the C2. It's the sides that are the problem.
     
  3. mark tomlinson

    mark tomlinson ༼ つ ◕_ ◕ ༽つ
    Staff Member

    Joined:
    Feb 21, 2013
    Messages:
    23,671
    Likes Received:
    7,316
    I can't specifically address the C2 hardware (do not have one to examine) but on the older models this would be something causing that axis (with the ripples) to have a slack drive. Belt or motor drive cog loose. If you can check those that might give you some insight. Additionally it can be compounded by over extrusion (if it is happening on all sides that would be more likely) so calibrating the extruder would be a great start.

    I have pinged RoboHQ to have someone with C2 knowledge to comment on this as well (I reached out to @Harry ).

    You can try slowing down acceleration and print speeds and if that has a positive effect you know it is something loose. Slowing down accel on the R1 was always a good idea because the numbers were insanely high for that style of drive. If that applies to the C2 or not I am not certain, but it is worth experimenting.
     
  4. WheresWaldo

    WheresWaldo Volunteer ( ͠° ͟ʖ ͡°)
    Staff Member

    Joined:
    Feb 18, 2015
    Messages:
    5,905
    Likes Received:
    3,593
    Well from the pictures it appears the C2 is over-extruding a bit. I don't know how it is set up as only a few people here have messed with the C2 and none of the old time users own one. It also could be that the firmware in the C2 may need a few tweaks to get it perfect. As an experiment you could try printing it a bit slower on the C2 after you have calibrated the extruder.
     
  5. ammulder

    ammulder Member

    Joined:
    Feb 3, 2017
    Messages:
    37
    Likes Received:
    2
    Thanks for the help. Can you elaborate on "calibrating the extruder"? I'm not sure what that means or how to do it.

    Likewise, I'm not sure how to check belt/motor cog tightness. All 4 belts give a little twang if I tweak them, and if I try to give them a little back-and-forth jiggle, I don't see any slosh in the gears. So they doesn't seem obviously loose to me but I'm not sure that's a scientific evaluation.

    I'm reluctant to slow down the print speed, because the ones I looked at (the defaults for the Robo Cura for Windows) were already set equal to to what I used on a Printrbot Play, and that wasn't the fastest machine. And, with those settings, the C2 is what I'd characterize as really slow. The 4-piece job I did would take 11-12 hours on the Replicator 2, Cura estimated it at 13:47, and it actually took over 16 hours. So I'm not very happy with the "class-leading speed" but I figured I'd start with the print quality issue. :)

    Finally, I've sent a couple e-mails to Robo support and gotten zilch except for the automated response. Hopefully you'll have better luck getting somebody's attention.
     
  6. mark tomlinson

    mark tomlinson ༼ つ ◕_ ◕ ༽つ
    Staff Member

    Joined:
    Feb 21, 2013
    Messages:
    23,671
    Likes Received:
    7,316
    Tom handles that:

     
  7. mark tomlinson

    mark tomlinson ༼ つ ◕_ ◕ ༽つ
    Staff Member

    Joined:
    Feb 21, 2013
    Messages:
    23,671
    Likes Received:
    7,316
    No luck so far, but it can take a day or two. Sorry.

    You might want to consider alternative slicers. Some are better at generating faster toolpaths than others.
     
  8. ammulder

    ammulder Member

    Joined:
    Feb 3, 2017
    Messages:
    37
    Likes Received:
    2
    I will try the calibration video, thanks.
     
  9. ammulder

    ammulder Member

    Joined:
    Feb 3, 2017
    Messages:
    37
    Likes Received:
    2
    I did the calibration routine. He recommended extruding 100mm and measuring how much filament it used. I measured 99mm. I don't think that difference is enough to account for the poor sidewall print quality (and certainly not *over* extrusion).

    I put on the latest Robo Marlin firmware, and I still get the jaggy edge and uneven sides on the test print.

    I'm going to try the Cura setup recommended by another user here, and see if that makes any difference for me.

    Otherwise, I guess I'll try to block out some time to spend on the phone with Robo.
     
  10. ammulder

    ammulder Member

    Joined:
    Feb 3, 2017
    Messages:
    37
    Likes Received:
    2
    Interestingly, I see the settings provided by @Cory Harpell have 90% flow rate... Maybe I should try that too.
     
  11. Geof

    Geof Volunteer Moderator
    Staff Member

    Joined:
    Nov 9, 2015
    Messages:
    6,729
    Likes Received:
    2,330
    The flow rate usually only helps if your not calibrated, if the process was done correctly it shouldn't help.
     
  12. ammulder

    ammulder Member

    Joined:
    Feb 3, 2017
    Messages:
    37
    Likes Received:
    2
    So I talked to Robo on the phone, and the person there said he'd send the issue to Level 2 and they'd e-mail me back within a few business days.
     
  13. ammulder

    ammulder Member

    Joined:
    Feb 3, 2017
    Messages:
    37
    Likes Received:
    2
    Here are my pics to try to clarify this issue.

    I printed some with the Robo Cura 2.3.1 distribution for Windows, at .15 layer height (same as I used on the Replicator 2). I printed some with the Robo "Normal Quality" profile for Cura 2.3.1, but using the regular Cura distribution for the Mac. And I printed the others with a Replicator 2 using the "Standard" quality preset in the MakerBot software but overriding the layer height to .15 All three were done with Inland Gray PLA filament.

    The worst of the bunch, a corner tile:
    corner-comparison.jpg
    The wall tile isn't as bad, but I still see pretty obvious layer lines from both print jobs:
    side-comparison.jpg
    Just looking at the top surface, it did a better job at .15 than at .1 The .1 result looks like it's made up of strings, whereas the .15 one seems to be made of plates. I was actually pretty happy with the top surface at .15 on the C2.
    top-comparison.jpg
     
  14. WheresWaldo

    WheresWaldo Volunteer ( ͠° ͟ʖ ͡°)
    Staff Member

    Joined:
    Feb 18, 2015
    Messages:
    5,905
    Likes Received:
    3,593
    @ammulder it is obvious the Replicator is just more precise and there may be more backlash in the axes for the C2. Obviously without the old timers having one we cannot provide definitive answers, but some of that can be mitigated by changing the acceleration and Jerk settings in the firmware, I just don't know which way to tell you to go with it, I think they are pretty low already.
     
    mark tomlinson likes this.
  15. mark tomlinson

    mark tomlinson ༼ つ ◕_ ◕ ༽つ
    Staff Member

    Joined:
    Feb 21, 2013
    Messages:
    23,671
    Likes Received:
    7,316
    Bear in mind that as you play with layer heights you may eventually have to play with nozzle sizes.
    Also, speed is a factor (the smaller the layer height the slower you need to print).
    Lastly it is possible with the right combination of nozzle size, layer height, etc. for you to hand the slicer a model that can have "impossible" sections of detail. Something that is too small (for example) for the nozzle to deliver. In that case almost all slicers will simply omit that section (the old "I can't print it so I will ignore it" approach).

    Not saying that is the case in this specific instance, but keep it in mind.

    In your case I would go with what @WheresWaldo suggested for a start and slow things down to see if they improve.
     
  16. ammulder

    ammulder Member

    Joined:
    Feb 3, 2017
    Messages:
    37
    Likes Received:
    2
    Well, it is true that the Replicator was a costlier printer. I just figure that today's printer should be at least as good. I've heard such stellar things about the print quality on the Mk2, but looking at e.g. the Make Magazine reviews it's not that far above the rest of the current crop in print quality. (Though again, many are costlier.)

    I have done some of these tiles on a Printrbot Play as well (which was a $300 printer recently), and I feel like they were comparable in quality, but unfortunately I don't have a way to know which ones were done on the Play, and it's out of commission at the moment so I can't do a side-by-side test.

    I believe the C2 is already printing slower than either of those others, though I will have another look at any acceleration/jerk settings I can get at from Cura and maybe try another print tonight.

    When I sent the photo to Robo, the person confirmed that he has seen and would expect better print quality from the C2. And the other person here who said he had two C2s didn't sound unhappy with the performance.

    So I feel like something is wrong -- either loose parts or just unlucky with the motors that went into mine or something.

    Probably I've titled this thread wrong -- though it manifests as layer lines, it may be the x/y positioning accuracy that's the problem -- when it goes to the "same" spot again it ends up off a little. Anyway, rambling now...
     
    mark tomlinson likes this.
  17. mark tomlinson

    mark tomlinson ༼ つ ◕_ ◕ ༽つ
    Staff Member

    Joined:
    Feb 21, 2013
    Messages:
    23,671
    Likes Received:
    7,316
    Well shoot the usual suspects first -- acceleration and speed. If that does NOT make it better at all (try dropping them 50%) then we can rule that out.
     
  18. ammulder

    ammulder Member

    Joined:
    Feb 3, 2017
    Messages:
    37
    Likes Received:
    2
    OK, I see loads of speed, acceleration, and jerk settings in Cura.

    My plan is to halve print speed, print acceleration, and print jerk (and check the boxes for enable acceleration control and enable jerk control).

    Is that sufficient or do I need to adjust the other settings individually as well (such as Wall and Top/Bottom Speed/Acceleration/Jerk)? I'm not sure whether it will default to the "Print" setting or take the lower of the specific and the "Print" setting, or whether all the settings are really independent and therefore each require adjustment.
     
  19. mark tomlinson

    mark tomlinson ༼ つ ◕_ ◕ ༽つ
    Staff Member

    Joined:
    Feb 21, 2013
    Messages:
    23,671
    Likes Received:
    7,316
    Been a while since I looked at other slicers, S3D gets the raw movement numbers in terms of speed and then everything else is a % or that (100%^, 60%, etc). What you are suggesting sounds right. Take teh printer speed, accel and jerk numbers down by 50% and see if anything changes (beyond the print time getting longer).
     
  20. ammulder

    ammulder Member

    Joined:
    Feb 3, 2017
    Messages:
    37
    Likes Received:
    2
    OK, I dug up a small test model on thingiverse, and printed one (left) with the Robo Normal profile in Cura 2.3.1 on the Mac (forgetting that it would add a raft), and then dialed down the print speed, acceleration, and jerk by 50% and did it again (right) without the raft. This print is about 1" long:

    IMG_9681.JPG

    I see some prominent layer lines on both. I'm second guessing myself as I'm writing this, because I know you don't expect to get NO layer lines. But at 100 micron, I really was hoping for smooth walls, and there are obvious layers that are either sucked in a bit or bumped out a bit. The box part on the opposite side shows this more clearly, with the one on the left having layers sucked in and the one on the right having layers bumped out:

    IMG_9682.JPG

    Meanwhile, Robo tier 2 support never got back to me.

    I'm currently planning to return the C2 (my 14 day window is up this week). I would be willing to have another look in 6 months and see if any manufacturing and software and support issues have been sorted... assuming MicroCenter is willing to let me. :)
     

Share This Page