These printers do not print dimensionally accurate parts. I have noticed this issue on the first printer I bought/returned, and this replacement printer suffers from the same issue. It does this with all types of filaments, and all different brands, so this is not an issue with filament shrinkage or expansion.
Only way I can get parts close to accurate is to adjust X-Y Hole and Contour Compensation in the slicer, but this does not translate the same to circles and squares the same way. The other problem with having to adjust the slicer is that it can mess up the geometry of the parts where they have visual defects like countersunk bands on the outer walls (you can see these defects in the slicer preview as well as printed part).
If setting X-Y Hole and Contour Compensation to 0mm, my hole and contour dimensions are undersized by .15mm - .2mm which makes important parts not fit that must be dimensionally accurate. If setting to .1mm on both, the circle dimensions are almost perfect as far as accuracy, but this does not translate the same for squares and other shapes.
Basically these machines are calibrated incorrectly when it comes to the steps per mm on the X/Y axis, and there is no way to fix this. Calibrating the machine or printing the calibration patterns before prints does not make a difference either.
Kudos for posting something clear and concise. It sure would nice if other posts included this level of detail.
Unfortunately, everything you’re experiencing as well as the remedy you’re using with respect to X-Y compensation is exactly how the slicer works. If it’s any consolation, you’re doing all the right things.
The issue is not the printer or the software, it’s simply that results will vary widely based on filament type, moisture in the filament and filament parameters.
On your last note however, I think you may be presuming that calibration at the factory is possible between filaments. As stated above, the laws of physics just make this impossible without extensive sensor technology to provide real time compensation for filament shrinkage/expansion. It just isn’t available yet in Consumer 3D printers just yet. Maybe some day, but not today. The X1 tries to do this with LIDAR. How’s that been working out? According to many X1 user posts, not very good I’m afraid.
I think you’ll find that X-Y compensation works just fine. I might suggest Orca Slicer rather than Bambu Studio because it has baked-in calibration tools that really work nicely once you get the hang of it.
There was a lengthy thread on this topic that you may want to look at. In this case, it was the challenges of getting 3d printed threaded objects to mate. But the issues are the same regarding intermateability between two parts and calibration of each filament.
Are you sure? Have you calibrated your filaments for shrinkage before tinkering with dimenstional compensation? Because these deviations (0.2% at most) are about the same as I get shrinkage with most of my filaments.
I calibrate all my filaments for this (ignoring the manufacturers statement) and I do have good results. Although i know that 0.1mm deviation is totally normal for an FDM process.
Yes, I have calibrated different filaments of all types, and all have roughly the same discrepancy in measurements. I have printed same test models on multiple printers (not Bambu machines) and all of my opensource printers will print accurately since I can set the steps per mm on them.
I have been 3D printing since 2015, and I own around 12 printers running marlin, all of which print accurately regardless of filament shrinkage or expansion from moisture.
In my opinion, such small parts are completely inadequate to evaluate or even calibrate the steps/mm. If at all, you should use a part that makes use of the dimensions of the print bed. And even then, it is not clear, if you really calibrate the mechanics or the filament or probably a combination of both. The mechanics in most 3d printers i have used don’t have any tolerance to speak of. The steps/mm is fully determined by the drivers, motors, GT2 belts and pulleys. So the steps/mm can be fully calculated and don’t need calibration.
If at all, I would measure the distance the printhead travels without any printing.
To check your theory, you could do exactly that: With the on screen controls, move the head completely to the left, stick some magnet to the bed that touches the print head on the left side. Then move the head e.g. 200mm to the right. Put another magnet on the bed, again touching the print head from the left. Now measure the distance of the two magnets.
@gooselk@Alex_vG@Olias Thank you for replying to this topic, but I am going to close this post for now, and I will create a new one once I have more data, images, measurements, tests, etc.
I dont think steps per mm is the issue as I found the movements are accurate if i control travel movements from lcd, but something in the slicer is over or under compensating for whatever reason, so I will have to come back to this at a later date.
@Alex_vG@gooselk@Olias I will wait till you see this message before I close it, but here is proof that it has nothing to do with shrinkage or filament issues.
.4mm single wall done in vase mode using PLA that I have calibrated. Wall is .4mm wide (it flickered between .4 and .39mm), yet the outside of the printed part is .16mm too small when setting X-Y Hole and Contour compensation to 0.
I could go on and on with proof, but I just dont have time to respond to everyone and perform new tests to back it up, on top of dealing with the support ticket I have open, so once you see this final message, I am closing the post.
Really on such a small part with a single wall, slightly reduced diameter is absolutely expected. The point here is not shrinking due to cooling. Instead, for any inner curved wall, the elasticity of the molten plastic will slightly pull it inwards. Imagine the filament coming out of the nozzle like a rubber band. The last few millimeters are fixed by the layer below, while the freshly extruded plastic is still soft enough to slightly collaps to the inside.
Your single wall cylinder effectively is only one inner wall.
I guess that if you print such a cylinder at double diameter, it will still be too small the same 0,1…0,2mm.
I have heard and experienced that everywhere throughout my 3d printing journey. Slicers even have a dedicated setting to compensate for it.
In my opinion, your printer and slicer are perfectly fine and that effect is just something that you have to be prepared for.
All of those figures are within typical 3DP tolerances… and several within what you’d get out of mass-production machine shops. You need to design for 3d printing and remember how the wall generators work. They only seek to be within line_width/2 of the polygon they’re tracking.
Arachne may get you closer, but it carries its own caveats.
I can’t seem to delete the post, so I will just have to stop responding to it if anyone else replies to the post. Thanks for taking the time to share your thoughts and feedback, but I just dont have the time to stay on top of this and continue to respond to replies right now.
The OP said that his multiple open-source printers running Marlin aren’t having the issue, so I don’t think it’s anything to do with him. If he was new to the game and wasn’t already getting the tolerances he needs out of other printers, I’d be the first to suggest that it was user error, but I think in this case it’s down to the machine somehow.
I haven’t used 10s of printers but my Ultimaker 2+ (a Marlin open source printer for 2500€ at the time) and 2 other Marlin printers I use occasionally all do exactly the same thing:
For smaller parts, outer dimensions are typically within 0,1mm accuracy. Larger parts need compensation of shrinking, according to the material datasheet. Inner diameters (actually any concave arcs) are typically 0,15-0,2mm too small in diameter. I increase those already during design by at least 0,2mm. Others have reported the same story here.
So I’m very sure that this is not a problem of his unit or bambu printers but it is a limitation of the manufacturing process.
And I’m sure it’s the problem of the machine itself. I have the same problem with my X1C. Same parts printed from the same material have different dimensions and usually don’t fit.
It’s either dimensional calibration of the printer itself or the way how the slicer prepares gcode. It’s not between Bambu Slicer and Orca Slicer (I use Orca), both provide the same results.
I tried with the same STL and the same spool of PLA on my Prusa MK3 and my X1C (I even tried to slow down X1C so it prints as slowly as MK3).
Nice you have 9+ years experience and xxx + years on marlin machines. If you need tolerances that are better than .03mm as a user as well I wouldn’t use a fdm printer. I would do subtractive manufacturing instead ie a cnc. OR I would stick with the old marlin machines.
IMHO anyone on these forums complaining of a tolerance problem of .03mm using plastic/pla/asa filaments are buying the machines and expecting cnc tolerances.
We aren’t building atomic implosion weapons and that kind of tolerances you’re describing I believe is overrated for ANY fdm machine. THEY ARE NOT RESIN MACHINES @ .03microns. Your nozzle is .4mm and they also make a .2mm so even if you used the .2mm nozzle your literally expecting??? Capabilities.