Jump to content
Ultimaker Community of 3D Printing Experts


  • Content Count

  • Joined

  • Last visited

Posts posted by picaschaf

  1. Today I've found an interesting issue (possibly) with auto leveling. After some time I enabled it again because of wobble issues, but now it seems as the Auto Leveling at the begin of a print overrides my manual settings of the print bed.

    The back part of the bed was a little bit too high, so after Auto Leveling, during the print of the brim I adjusted the back bed screw so the layer was fine again. But the next time I started a print, after Auto Leveling the brim looked the same way it did before. After 100 prints the back screw is possibly falling off the bed :p

    Why does the auto leveling compensate my compensation? Does it have some always applying correction matrix or something else? If yes, why is then leveling necessary at all if it always "does its own thing"?

  2. Hi,

    Today I was setting up our new office printer (that classic 2D one ;)) and when I was in the setup dialog I found our Ultimaker 3 Extended offering itself as PostScript printer via Bonjour (ZeroConf). As there is currently a long running print I was too afraid to just add and test it. Does anyone know something about this? Is it a feature, and what does it do?

    Best regards,




  3. I’ve a little update on this issue.

    When I tried to run the gcode file to an UTF-8 to ASCII converter it stopped because of a unconvertable character (‚ä‘ as in März, March in german) in the file.

    It looks like S3D is using a system function to get a localized Date in short RFC format and if this contains a non ASCII character, it will automatically use UTF-8 for the export.

  4. Hi,

    I have a different kind of a problem. Could it be, that the new firmware silently changed the parsing of the G-Code file from UTF-8 to ASCII?

    When I try to print a S3D sliced model (exported G-Code is UTF-8) it immediately stops with an error in line 1 and the log says:


    Mar 29 15:17:22 ultimakersystem-ccbdd3000c63 python3.4[443]: UnicodeDecodeError: 'ascii' codec can't decode byte 0xc3 in position 418: ordinal not in range(128)Mar 29 15:17:22 ultimakersystem-ccbdd3000c63 python3.4[443]: 2017-03-29 15:17:22,877 INFO     faultHandler    warning(, Error while reading file 'Spray_Can_Rack_Improved.gcode', 28, {'file_line_nr': 1})Mar 29 15:17:22 ultimakersystem-ccbdd3000c63 python3.4[443]: 2017-03-29 15:17:22,880 WARNING  faultHandler    added Fault: L:WARNING C:FILE_ERROR M:Error while reading file 'Spray_Can_Rack_Improved.gcode'


    When I slice a model with Cura Beta, it outputs ASCII and works fine.

  5. Are you sure the nozzle scratched the glass?  Brass is MUCH softer than glass.  Even steel is softer than glass and shouldn't scratch it normally.  Could it instead be a very very thin layer of pla?


    No, the nozzle didn't let scratchmarks on the glass, but it touched it so heavily that it made bad noises when the print head moved ;)

  6. Picaschaf,

    i have 3 UM glass plates now and all of then are warped, some more, some less. One that came with the printer, one from conrad.de and one from 3dmensionals..

    My solution was: checking the yellow pages for a local business selling glass. It ended up a tiny big thicker (4mm), doesn't have the spiffy rounded edges and set me back 80,00 Euros.. but now I got a mostly flat one. I gave up trying to get one from Ultimaker.


    Thank you for sharing your experiences! Maybe I'll try this as well. But how thick is your glas? My original glas is already exactly 4mm thick.

    At the moment I'm in heavy contact with Tom. It's crazy how seriously the guys @ Ultimaker take this issue. At the moment we are discussing different possible causes. The print problem itself with my original glas disappeared after moving the printer back in room where it was before and after a series of auto and manual calibrations.

    It seems that the firmware is capable of working around small warps (like my original glas has) but not big warps >2mm (like my new glas has).

    I've got some tests to do from Tom and after we figured out the cause of the sudden nozzle scratch I will update you all in this thread.

    • Like 1
  7. Indeed it should not. And maybe my problem was the metal and not the glass. But it did work and on my beta um3. Indeed the screws stuff, but that also wasn't the problem on my umo.

    Anyhow I would ask for a fix since you got a new product and you paid for it good money :)


    Yeah, I'm in contact with the UM support and my seller. But I would like to see a definite solution, not sending 10 plates around the continent and luckily maybe there is one that is ok. Like I said, at least the one seller I had in my driving range yesterday, had his store full of warped glasses.

    • Like 1
  8. @neotko I had a quick look on your posting and I've already found it when searching here for warped glass. But I discarded it as my solution as on the UM3 there is no screw between the glass and the bed anymore. You can see the new construction in the 2nd of 3 pictures in my last posting with the ruler picture first. Also the glass is much more bent than the one in your posting inside the printer as well outside the printer.

    Also it should not be able to bend borosilicate glass ;)


    [i don't think that the little aluminium sheet is capable of bending the glass back in shape ;)Also my printing results show that this is not happening.


    Yeah your problem might not be that. On my beta um3 I fixed the deformation points by pressing the points on the metal that where curved and now my first layers are perfect. Ofc that isn't a method I would suggest to a new buyer. But I did that also long ago on one of my umo+


    Ofc the problem to do it correctly on um3 is that you don't have easy access to the gcode to see where the nozzle goes up/down on a banana bed. I did it using the command_util.py that's on the linux of the um3. To access it you would need to ssh root@yourip then pass ultimaker, and cd ../usr/share/griffin then 'python3 command _util.py'

    With this command util you can sendgcode G28 to home all, and G0 Z15 moves the bed near the cores. Then you can bit by bit getting closer to the nozzle. This way moving the printhead you can actually check the points where it's curved.

    Ofc, maybe your problem isn't the plate but just the glass. Also check that the ruler is really straight comparing it, for example, to a x/y shaft straightness.


    Thank you for sharing your experience on that issue :) I will work through it later that day and have a look what I can do on my UM3.

    I've already checked my ruler and used a different ruler, that is not the issue. And the glasses from my seller hat a warping that is even visible to the naked eye. If you want I do a short video on both glasses I currently have.

    But the thing I habe with a new calibration is, even if I have the exact shape of the warping, will the UM3 firmware be able to handle it correctly? At least the standard leveling with 3 (or 2x 3) points is not even mathematically able to compensate warping. 3 way calibration is only able to compensate misaligned bed screws in the plane.

  10. Did you upgrade your printer to firmware v3.5 already? It contains an improvement to the Active Leveling software, which in some situations will give a more accurate result.

    Beware: Active leveling has also been made a little quicker, by starting higher. If you have manually leveled your bed more than 2mm higher than nominally, it fails.

    Nominal position: 14mm distance between top of base plate and bottom of heaterplate.



    I have the most current one the UM3 offers me to update to. But the only change I saw is that now the active leveling also does 3 points for the second print core. But as my glass is warped this is not enough for an accurate calibration.

    After manual leveliing the nozzle at least doesn't scratch on the glass, but the prints are useless.



    As already seen with a different part (the pink one above) this part shows the same behavior. In the front right (where the glass is flat) everything is fine. In the back left the nozzle is too far away and in the middle it nearly scratches the glass again. Yes I am printing very near to the clip, but the area where you see the lines not touching is away from the clip where the head doesn't touch it.

    And I've written before, yesterday I was at a local seller - all four glasses he had in stock were warped, some even more than mine. He tested them himself and I testet them also.




  11. I believe the new active leveling only takes 4 points on the glass (plus 2 with the second core).  So I don't think it will take care of warped glass.  Just non-level glass.

    I was able to bend my aluminum which corrected for any bend in the glass.  Although actually most of the error was in the aluminum in the first place but I think it's worth a shot.  It only took one minute.


    I don't think that the little aluminium sheet is capable of bending the glass back in shape ;) Also my printing results show that this is not happening.

  12. I went out to buy a new original Ultimaker glass and I am shocked. All four glasses at the seller are warped O_O And, not a little bit. 2 of them are actually far more warped (>2 mm) than my current one. Now I've bought the one that has the least warp but this is still causing issues.

    Today I also updated to the current firmware with the new calibration mechanism but this changed nothing. The machine still assumes a flat bed.

    Is there any unwarped/unwarpable alternative to the glass for the UM3?

  13. This feature is built into Cura. I believe from another post that you use another slicer, right? Maybe you can build this "forbidden zone" into your slicer.

    But of course I recommend to switch to Cura. Do you print with PVA as well? Cura contains smart features with nozzle temperature control which is essential for PVA reliability.


    I've used Cura as well and I see the dead zone there. Most of the time I use Simplify3D and yes, I could specify a dead zone in the software. But in my opinion from a software developer view this should not be a software issue, but a firmware issue. Firmware should always prevent hardware damages caused by external interfaces (in this case the gcode) which should be always treated untrustworthy - even for the own tools.

  14. Wow ! I was looking for an issue with the rods :(. I will try with the manual leveling.


    Actually it's a really nice feature. Is your active leveling not working? I would not manual level until it's necessary or you want to experiment. Or is it the sound that annoys you?

  15. I would be surprised if the glass indeed got bent due to heat. Unless you have secretly been printing on 500º/600ºC? ;)

    Could it be that you have been printing smaller objects and you just did not notice?

    I would indeed contact your reseller and see if you can replace it if it is indeed bent.

    What is the thickness of the first layer you are trying to print?


    I hope not ;):p



    I don't think that the glass was warped already. Yesterday everything was fine and now this. When I look at the nozzle while building the first layer (140% height, 100% width, base height 0,2mm using Simplify3D) I can clearly see that the nozzle is very close to the bed on one side, and on the other side it is nearly 1mm away from the bed.



  16. IMG_0730.thumb.JPG.4608a28dc24e991c28e1634943c5e154.JPG

    Little update. I made a video and some photos to prove the warp and then cleaned the glass, turned it over and did a new manual calibration. Then I started the same print as yesterday and no scratching and (at least in my opinion) a nice finish. The stringing and brim blobs are "okay" for now as the nozzle is a little too hot and I didn't apply hair spray as I do it normally (just for this test).

    Do you think this is a good calibration or should it be higher/lower?

    While observing the behavior of the turned over glass I will contact my reseller and share his response with you.


  17. At least one of my magnets is not glued in (or it went loose) and sticked to the rod after the first print core insertion. I just put it back in and no problem so far. I have a stock released UM3 extended and my side blocks are white(ish/cream), not black. Have you modified them?

  18. You could place your finger lightly on the middle Z rod (the one with the thread) while its making the sounds. You will feel a very very tiny movement during the sound, first in one direction, then in the other (moving the bed up and down again a very very little bit). If you feel nothing different to normal movement then it probably is a different problem.

  19. I recently removed the spool holder and then my UM3 startet to make a very nasty noise when the head moved from front to back (left to right and Z were ok).

    After I've checked all screws, axles, cleaned & lubricated everything (originally to find another issue), I realized that I still had the NFC cable blend on the UM3. As the spool holder is not mounted, this little beast resonated with the print head movement -.- After removing it, it's quiet again :)

    Just a little hint if anybody who has not mounted the spool holder has the same issue.



  20. Hi,

    I have a very strange problem since yesterday. From one print to the next the nozzle scratched on the glass very badly - but only in the middle part of the bed.


    After checking all screws, axles, etc. I removed the glass plate and after checking it with 2 different rulers I found out that it is warped from the front to the back. But this must be happened just from one print to the next because yesterday morning everything was fine and the print I started in the evening was messed up.

    Is this normal (to compensate the heated bed, or wear)? Does it go away with time if I flip the glass? Is this a warranty case?

    Best regards,



  21. Okay, thank you very much for sharing your experience! :)

    I've ordered a super cheap china printer for printing crap parts like clips etc. to have my UM3 baby free for the "real stuff" that need its precision and neat tricks ;) This one just feeds 1,75. I'll see if my filament collection will get fragmented but at the moment I think I will just print one spool in 1,75 and don't share them between crap and grail ;)

  22. IMG_0703.thumb.jpg.221b7f88131e6e6742b796ac879f1e88.jpg

    I rearranged my build so it fits and doesn't collide with the hook. You see the arm in the front, if I rotate it so that the long part is in the back and the "pointy" part in the front - same position - the print core switch collides unstopped by the system with the hook in the back right. Similar problem when I add to this model a 5 layer skirt with 1mm distance to the model the head will scrubb (far less violently than colliding with the hook) against the right rail. So the right 1,5-2cm (and around 3-5cm in the back) of the build plate is essentially dead. On the left side there is absolutely no problem.

    Don't understand me wrong, in my opinion this is absolutely okay due to the print head design (which is still really smart and reduces the need of an actuator to retract the second core). But the firmware should cap the coordinates in order to prevent such bad crashes regardless what the gcode tells the machine to do. It would be a really simple fix (when relying on a calibrated X/Y axis, which we should be allowed to).


  • Create New...