The Ultimaker-cura slicer can figure the compensations automatically. It’s free and supports 3d printers from a lot of different manufacturers, but, unfortunately, bambu labs printers aren’t among them, at least not yet.
So, in an effort to find a solution, I reached out to support with all my measurements that I posted above along with the analysis of the situation. It took a really long time - 24 days - to get a reply, and here it is:
Hello,
Thank you for contacting Bambu Lab support. I am sorry to see that you are experiencing problems with your printer, but rest assured we will provide support to solve them as soon as possible. In this case, the problem seems to be related to the print quality.
To address this issue further, can you please start by performing a bed leveling and belt tensioning procedure? Then print some calibration cubes and compare the tolerances.
Additionally, once complete, leave the printer idle for 10 minutes or so before exporting the print logs for further review.
Belt tensioning: Belt tensioning guide Bed leveling: Bed leveling guide Log upload: Log upload instructions
Please let me know if you have any questions or concerns.
Best Regards, Miles Bambu Labs Support
In good faith, I’m going to do all the steps that they have asked, and then I will take my measurements again and respond. I’ll update this thread with both my results and their response once I receive it. I’m expecting them to take another three weeks once I’ve replied, but we’ll see.
2 Likes
@PEZ3D interesting issue.
I regard that the printer “setup” is focused on aesthetics, which, from what I read from experts, conflicts with tight accuracy or tolerances.
I value accuracy and tolerance more than good looks, as I use the printer primarily for functional prints.
I can say that with a 0.4mm nozzle, I never notice.
Yet, after the typical flow rate calibration procedure for each filament spool, I start a fine-tuning procedure using tolerance tests until I get a 0.05mm fitting.
Also, I set my outer wall print speed to 60 mm/s when tolerance is critical. With a 0.6mm, I had some difficulties getting accurate dimensions in the three axes; it seems to have improved in the last months, maybe by software and firmware updates.
Yet, after reading about others’ experiences, I need to recheck to be sure I am still achieving such an outcome.
I read about a similar issue months ago; in that case, the BL support team solved the issue by providing accurate instructions to fine adjust the belt tension. If I find it, I will share it.
@RPMiller, I look forward to knowing more about BL support solutions for your trouble ticket.
For fixing skew on printers like the X1, CNCKitchen gives a couple workaround options until there’s a more permanent solution:
TL;DR: octoprint or g-code post processing
I don’t see any instructions related to X1C. Just a quick mention of the Prusa. Did I miss it?
back it up a few seconds. He says if you can’t change your firmware like on the new bambu machines or the prusas…
Right, but now instructions of what to actually do.
Hello,
new here, running Mk3 for some time and recently X1C carbon. On X1C also having troubles with dimensional accuracy, tracked it down to skew:
- Califlower calibration test show => skew -0,19° (pretty bad for my needs)
Based on good discussion in this thread I see 3 possible skew remedies:
- manually distort print model (step) to compensate for printer skew prior slicing – not easy, not convenient, but seems feasible
- Boneyard labs video recommend to shim-out the x/y skew – based on my tests, this seems to work in short term. But I actually don’t fully understand the core XY kinematics & belt system and I am a bit concerned this may cause some undesired effect/belt tension/wear in long term?
- CNC kitchen Stefan’s video seems to indicate GoSkew G-code postprocessor - I don’t have any experience with such technique and was not able to find any reference/discussion anybody using it for Bambu or Orcaslicer
Does anybody here have relevant experience to share which of the solutions actually work and is most reliable/robust/convenient in long term please?
1 Like
Belt tension comes to mind as a 4th point. But you have probably already done that.
Goskew gives directions on what to do and how to do it:
In case you didn’t see it, have a look at:
Progress!
1 Like
Thank you pointing me to this! This is definitely significant and I agree it should just be in the official wiki. I’m still waiting to hear back from support, but it will be interesting if they suggest this same thing.
Just to a little add to this. I’m coming from 30 years mechanical engineering and 10 years of personal CNC. Quite a noobie in 3D printing at 2 month.
When having prototype parts made, you give a model the machinist who will adjust offset based on the material for materials. (at prototype you may use a different material for cost) This similar to filament types.
Many times in subtractive the cutting tools are measured to 4 decimal points and that dimension is put into the CNC machining equipment.
Lastly for threads, if they were designed and made exactly the same (line to line) they would thread together, some clearance is necessary. Same goes for mating part with slots.
Steve.
The problem you describe seems to be related to the printer not calibrating the printer correctly in steps per millimeter along the X and Y axes. Adjusting the hole and contour compensation settings in the cutter may temporarily correct this problem, but will not completely solve it, especially for round and square shapes.
1 Like
As already mentioned further up, the printer mechanics and electronics are rarely the root cause for too small parts. So adjusting steps/mm does not solve the root cause. Instead, you should look out for shrinking of the material when cooling down. Accordingly, that should be compensated in a material profile, separately for each material you use.
Skew is a different story, where calibration of the machine absolutely makes sense.
Going to give a try to this firmware fix, seems almost too good to be truth! 
To provide update:
“mechanical” tape fix is difficult to fine tune and does not seem to be fully stable over time (probbably change in belt tesion?).
Spend quite some time getting GoSkew integrating in Orca Slicer - finally got it to work in terms of “skewing” the g-code but also generates weird artefacts in the gcode trajectories.
Next step - give to try to firmware enabled fix as pointed out by NeverDie!
Thanks, and yep, just belt tension re-set did not help.
Have you see this post about the skew compensation built into Bambu firmware?
4 Likes
That does indeed look odd. Maybe the model you’re working with is too “coarse”? How was it created?
In the past I would “print” my fusion360 CAD model to a .stl file. When doing that, I could specify the degree of “refinement” in the output file. Default was “medium”.
Is that how you’re making your model? If so, maybe try the “High” setting or something custom that’s more refined. The goal would be to export a mesh that has, say, 10x as many triangles in it.
Anyhow, I no longer do it that way myself. Now I export .step files, because I’ve heard it is more geometrically precise. More like vector graphics than pixel graphics, if I’m getting the gist of it right. So, maybe try a .step file.
Unfortunately, I’m out of my depth and these are both just wild guesses, but it seems somewhat plausible and very easy to try, so maybe you want to see if it makes any difference.