What am I doing wrong?

Discussion in 'Getting Started' started by wHack, Sep 13, 2019.

  1. Spoon Unit

    Spoon Unit Well-Known Member

    Joined:
    Sep 6, 2015
    Messages:
    1,546
    Likes Received:
    481
    I do have the same waves in the same direction. Possibly not quite as strong, but definitely present.
     
  2. wHack

    wHack Well-Known Member

    Joined:
    Aug 20, 2019
    Messages:
    96
    Likes Received:
    16
    Ok. Just checking. I am guessing you will find they are repeatable as well. I'm going to start a hotter print at like 230 which seems too hot for PLA but I will do it anyway.

    Also just an interesting side note for the gcode I uploaded. The actual gcode for layers 5-66 is identical, which is basically all but the first few layers of the square part. So the waves in that part are clearly independent of the gcode itself.
     
  3. Krayn

    Krayn Well-Known Member

    Joined:
    Aug 11, 2019
    Messages:
    49
    Likes Received:
    13
    One of the things I'll try is to see if I can mount a dial indicator to measure the positional accuracy of the machine while printing. I'm curious if the waves are coming from belt backlash or something like that. Specifically I'll point it at the X axis and see if it returns to the exact same position every time it tries to print the same part of the wall. I just want to rule it out before I start trying to investigate the hotend.
     
  4. wHack

    wHack Well-Known Member

    Joined:
    Aug 20, 2019
    Messages:
    96
    Likes Received:
    16
    The layers aren't shifted. They don't stick out on one side and in on the opposite. If they stick out they stick out on all sides. If they stick in they stick in on all sides.

    That's why I think the problem is in Z if it is in the motion system.
     
  5. Krayn

    Krayn Well-Known Member

    Joined:
    Aug 11, 2019
    Messages:
    49
    Likes Received:
    13
    Well backlash isn't always the same going both ways, I have to imagine that running in core X/Y also makes that harder to pin down as both motors do contribute.

    I will mount it on the build surface and probe a couple different ways. I can try raising and lowering the surface and aiming the indicator at a perpendicular block to see if I see anything wavy as it moves.
     
  6. wHack

    wHack Well-Known Member

    Joined:
    Aug 20, 2019
    Messages:
    96
    Likes Received:
    16
    So I printed it at 230 and it's shinier but still has the same waves.

    It's hard for me to imagine how backlash in X or Y or Z could cause this but by all means check it out.
     
  7. Spoon Unit

    Spoon Unit Well-Known Member

    Joined:
    Sep 6, 2015
    Messages:
    1,546
    Likes Received:
    481
    Good to confirm. So then it doesn't seem like hotend-constrained flow either.

    What about inconsistent extruder behaviour. I don't think we're running the extruder with any interpolation:

    M350 E8:8:8:8 C8 I0 ; Configure microstepping without interpolation for extruders and coupler

    Maybe that's worth setting to interpolated?
     
  8. Krayn

    Krayn Well-Known Member

    Joined:
    Aug 11, 2019
    Messages:
    49
    Likes Received:
    13
    Doesn't necessarily have to be backlash but any other wonkyness that we haven't accounted for. I was throwing ideas out while I was eating lunch. I'm still going to do some measurements and see if I can ram into the indicator at different speeds to see if the movement is consistent.
     
  9. wHack

    wHack Well-Known Member

    Joined:
    Aug 20, 2019
    Messages:
    96
    Likes Received:
    16
    So is this just a flaw of this printer? Can either one of you guys print one of these without the ripples with any settings? @Krayn @Spoon Unit
     
  10. Spoon Unit

    Spoon Unit Well-Known Member

    Joined:
    Sep 6, 2015
    Messages:
    1,546
    Likes Received:
    481
    It might be a flaw with that print, with those settings, with the current tool setup, rather than with this 'printer' which is actually a platform, not a printer (but I'm splitting hairs). The ripples could be the result of the material, electronics, bowden, mesh levelling (still electronics), granularity of mesh (certainly a thought), or the bed not providing perfect stickiness (we have different beds).

    Have you got another printer which doesn't produce any ripples with this print when viewed this way?
     
  11. wHack

    wHack Well-Known Member

    Joined:
    Aug 20, 2019
    Messages:
    96
    Likes Received:
    16
    My Form 2. But it's an SLA printer so not exactly apples to apples. All that aside, are you able to print that object without ripples on this printer? There is nothing special about the object to explain the problem. I would love some settings that actually work. As for material three different people using different materials all get the ripples on three different machines using different settings. And it is a printer that came in a kit, it is based on a platform, but it is a fully implemented printer based on that platform. Anyway I don't mean to be argumentative, I just would like to know if this printer is capable of printing this and other similar objects without those ripples.
     
  12. Spoon Unit

    Spoon Unit Well-Known Member

    Joined:
    Sep 6, 2015
    Messages:
    1,546
    Likes Received:
    481
    Honestly. I'm stumped and without several clear days to sit and play purely on that problem, I don't honestly know if they can be eradicated altogether, mainly due to the fact that the filament itself is not perfect, which is to say that filament diameter changes over the course of 100mm, perhaps insignificantly, but that insignificant amount is being weaponised by casting shadows in this way. The SLA is completely different in this regard, because there's just no filament feeding; it's all there already and is crystalised by the process used onto the existing model.

    Perhaps the thing to do would be to get some filament from Prusa as an attempt. They claim to have filament width to within 0.02mm difference along the entire spool along with a QR code to show you the exact results for your spool when it was made.

    All this said, you might be able to determine whether that has any chance of success already. If you print this five times, with the stipulation that bed adhesion must be perfect for the print so as not to be the cause of any rippling, are those ripples in the same place on every print. If they are different each time, perhaps it's the filament. If they're the same each time, it more likely to be something mechanical.
     
  13. wHack

    wHack Well-Known Member

    Joined:
    Aug 20, 2019
    Messages:
    96
    Likes Received:
    16
    It’s not inconsistent filament. If it were then the ripples would be different every time. But they are not. Also I have some Hatchbox which is +-0.03. I have spent days on the problem. Only to find no solution through settings. This is why I tried to enlist your help but you have also come up short. So this is why I ask if it is just a problem with the printer. Which could be hardware or electronics or any number of other things.
     
  14. wHack

    wHack Well-Known Member

    Joined:
    Aug 20, 2019
    Messages:
    96
    Likes Received:
    16
    It's also not bed adhesion. For the same reason. The ripples are the same each print with or without perfect bed adhesion.
     
  15. Spoon Unit

    Spoon Unit Well-Known Member

    Joined:
    Sep 6, 2015
    Messages:
    1,546
    Likes Received:
    481
    I'm not an expert. Just another owner. One who's just as interested in finding any solution, if one exists. Good that you've ruled out the filament too.
     
  16. Spoon Unit

    Spoon Unit Well-Known Member

    Joined:
    Sep 6, 2015
    Messages:
    1,546
    Likes Received:
    481
    So. I printed a smallish square in vase mode, 0.3 layers, 0.4 extrusion width on a 0.4 nozzle using E3D Edge. I can see no waves at all with this print. Repeating that using 0.3 layers, 0.8 extrusion width on a 0.4 nozzle and then I do see waves.

    My take from this is that the printer seems to exhibit no challenges with positional accuracy and seems to have no movement in X and Y that might result from any issues with the lead screw. Extrusions wider than the actual nozzle do show that attempts to extrude beyond the nozzle width seem to naturally result in some ebb and flow in the deposited material that is likely to differ over time.

    Given that a vase mode print shows the printer positional accuracy, at least on mine, I would then begin to suspect that, if the model you're printing shows the same waves on different machines, that the waves are the result of the combination of model, slicer settings, and the inaccuracies inherent in FFD.
     
  17. orcinus

    orcinus Well-Known Member

    Joined:
    Oct 20, 2015
    Messages:
    336
    Likes Received:
    113
    Sorry for disappearing, i was away on a trip (and leaving again on monday, and leaving again next sunday, ugh...)

    I had 0 luck with PA tuning.
    I was getting wildly different and completely nonsensical results and finally gave up completely.

    I've got it turned off at the moment, because even with relatively conservative settings i'm getting worse print quality than with it off.

    Regarding the waves - these are usually either due to positional / microstepping error in the extruder stepper, some kind of interpolation in the firmware, or some kind of moire effect between the resolutions of *some* two parts of the system. Try:
    1) killing PA
    2) changing the steps/mm to see if the pattern changes (if it doesn't, it's mechanical)
    3) changing the Z resolution slightly
    4) changing microstepping/interpolation for the extruders and/or X and Y axes

    Obviously, one by one, or you won't be able to narrow it down.
     
  18. wHack

    wHack Well-Known Member

    Joined:
    Aug 20, 2019
    Messages:
    96
    Likes Received:
    16
    I have PA turned off as well. I did the test script thing and got a value of 1.8 but it was worse than with it off. I then tried a whole range of values and could not find anything that came out as well as just having it turned off. I will have to try some of those suggestions when I get a chance. Thanks for adding to the conversation as this problem is frustrating.
     
  19. Spoon Unit

    Spoon Unit Well-Known Member

    Joined:
    Sep 6, 2015
    Messages:
    1,546
    Likes Received:
    481
    Maybe this is another thread to pursue. The Z axis motor and leadscrew is based on this: https://e3d-online.com/leadscrew-motor-with-pom-nut. This means that if you change your Z motion to full step only, 100 steps are required per mm, and 1 step moves 0.01. In this case, you have moved your Z motion to an exact number of full motor steps if you have a layer height of 0.15, for example and furthermore. The fly in the ointment is mesh bed levelling, which is mostly required to get a nice flat first layer. So if you inject into you GCODE a G29 S2 at the layer 2 interval, you're then in a position where Z is only chunking on precise layer intervals. This makes no sense for vase-mode printing, but some sense for layer-based printing. I find the waves are much reduced by this and I tested with pretty crappy filament.

    If you want to try this, you'd need to change the following lines:

    Code:
    M350 E8:8:8:8 C8 I0                     ; Configure microstepping without interpolation for extruders and coupler
    M350 X16 Y16 Z16 I1                        ; Configure microstepping with interpolation for primary axes
    M92 X100 Y100 Z1600 C100 E420:420:420:420    
    to

    Code:
    M350 E8:8:8:8 C8 Z1 I0                     ; Configure microstepping without interpolation for extruders and coupler
    M350 X16 Y16 I1                        ; Configure microstepping with interpolation for primary axes
    M92 X100 Y100 Z100 C100 E420:420:420:420    
    Be prepared for a much noisier machine if you try this. Given the noise, I'm guess this completely negates silent step stick on the TMC drivers.
     
  20. wHack

    wHack Well-Known Member

    Joined:
    Aug 20, 2019
    Messages:
    96
    Likes Received:
    16
    @Spoon Unit it's hard to tell if that gives better results or not. If so it's only slightly better. It's definitely not a silver bullet.
     

Share This Page