What tolerances are you expecting? At a minimum I wouldn’t expect anything less than the nozzle diameter, plus a bit. Absolute values, not percentages.
On stock settings I am getting -0.45mm off on a 100mm length and -0.12mm off on a 20mm length. In my world a half mm tolerance is sloppy. I would be oaky with +/-0.15mm and happy at +/-0.1mm. But even then that’s pushing it for making small and/or fitted parts. I feel this is a reasonable ask from a $1000 machine that advertises its usefulness in engineering settings.
1 Like
Fair enough. I see similar results on stock settings. Yet when I calibrate my filaments for thermal shrinkage, I get results well inside the deviations of my means of measurement plus user error. I get those results with 20, 50, 100 and 200mm test pieces. That leaves me to believe that it’s not a systematic error as it would grow with the part. Yet, it doesn’t.
The problem here is - IMO - to identify the true reason for those deviations as that’ll determine what countermeasures you’d take…
1 Like
Fair enough. What is your world? CNC machining?
Those tolerances sound perfectly reasonable for a hobby machine… to me. Just because it says it can print engineering grade filaments doesn’t imply anything about precision. I think expectations are exceeding reality here, and from the previous post, it sounds like proper calibration might help.
I have a Syil X1 mill that cost 3k, and can cut the same material my 15k Tormach can. I don’t expect the same tolerances out of them though, they’re in two different categories completely.
That is 0.4% too short in one direction and 0.6% in the other direction. I wouldn’t trust a measurement of a few 1/100mm compared to 20mm, that would be very sensitive to measurement errors. E.g. if it is 0.15 instead of 0.12mm which is a ridiculously small difference, the second length is short by 0.45%, so almost the same as the first direction.
I don’t say you are wrong. But a part 100x20 mm is not the best to find out. Better use a square 100x100 or put the 100x20 bar twice on the plate, rotated 90° once and measure the 100mm on both. It might be a good idea to print them again rotated by 45° because of the CoreXY.
If those are different in length, then I fully agree with you that something is wrong with your printer.
Btw, the OP didn’t complain about differences between X and Y but about generally too small dimensions. So it seems like different issues to me.
1 Like
Appreciate the replies and data points. I feel this machine is thoroughly mechanically calibrated (at least to the extent Bambu allows) and I’ve tried my own filament settings as well as factory just to make sure.
Calibration model has been PigeonGiraffe’s 100x100x100 which has 20mm increments as well as various other dimensional tests and all-in-ones. Printer performs great except for the consistent XY size issue. I have checked for skew and found none. At this point I’m either going to write a script to automatically resize all models or start replacing belts and checking parts for bends.
My background isn’t machining but aerospace so I work with machinists often without having detailed knowledge of the machines. Maybe my expectations are too high but I believe Bambu’s machines are capable of achieving the level of accuracy I desire. If only they would open up their software a bit more to allow more end-user freedom. Indeed I could be having a separate issue from OP. Once again, thanks.
Modifying the value for the flow ratio from the default of 0.95 should achieve what you are missing with setting the E-steps.
Have you tried to move the printhead with onscreen controls and check the distance it travels?
Actually, I haven’t checked dimensions on my printer yet. Test print is running.
I think you nailed it with expectations being a bit too high. There’s so many more variables than on standard CNC machines. Minimum line widths, thermal issues with the filament, filament properties themselves…it’s a lot more involved than just making a cut at a certain spot
Id give a 0.2 nozzle a try, see if that helps with the tolerances. In theory it should cut the line width in half.
There’s also some slicer settings related to how it calculated edges, I’m not at home or else I would give them to you. I’m assuming you’ve done the whole outside wall first thing. I usually start there when clearances are important for me
1 Like
Lots of replies here. My X1C is within acceptable limits. There are so many factors here. I have been machining for over 40 years, and had to deal with expansion of metals to meet specs. Aluminum for example expands at about 22 millionths of a deg per inch. Ideal degree common around the world for inspection is 68 deg, (20c) when measuring, so if you machine a part that’s10" in diameter, and your shop is 80 deg, then when you go to measure it will be approx. 9.997" @68 deg. Plastic is terrible at expansion. Typically nylon is the worst to keep stable, at a expansion rate of 80 millionths of a deg per inch, that same part would be 9.990" @68 deg. If you want your 3d printer to be consistent all the time, you would need a temperature controlled room, controlled humidity as well. At that point you should get consistent results, then adjust the size of your part accordingly before printing for best results. The X1C is not an industrial printer. Those start at about $70k and go up from there.
Good luck!
I believe that last statement illustrates the issue here that is all too often overlooked in these discussions. Way too many folks buy a consumer product and then expect it to output NASA-class precision. If I may point out, we’re not making gyroscopes here to thread a space capsule through the eye of a needle 4.84 x 10^8 miles away. 
4 Likes
Simple e-step calibration would be nice. Not sure why they chose to reinvent the wheel.
2 Likes
The most generous thing I can say is the OP has unrealistic expectations.
As an Engineer, I have designed traditionally produced injection molded parts and extrusions off and on over the last 40 years.
Typical tolerances are +/- 0.4mm (+/- 0.015")
For super critical dimensions, you can pay extra and specify +/- 0.125 (+/- 0.005")
For reference 0.1mm is the thickness of a sheet of printer paper.
The absolute error on these printers is very, very good. And you can calibrate it in the slicer to improve it. But there is only so much you can do.
Plastic shrinks when it cools. It absorbs moisture in the air and expands. It cools too quickly and it warps. You look at it funny and its off.
Holes shrink. Flash (or brim) happens.
These printers, finicky as they are (and believe me Bambu is the least finicky so far) are AWESOME.
The OPs opening statement “These printers do not print dimensionally accurate parts” expresses a degree of Engineering naivety and intellectual immaturity that I would normally attribute to a troll. But in the spirit of the community forum, I will instead suggest they adjust their expectations.
Engineers have been designing functional parts and assemblies despite manufacturing inaccuracies for hundreds of years. They think about what happens if something is produced that is a little too big or a little too small. And they think about where that part has to fit - and make the space a little bigger or a little smaller accordingly.
OP, I don’t know what you’re printing or building, but consider the possibility that you might be doing it wrong.
9 Likes
Hey all,
Among many valid points here, I don’t see anyone addressing something I’ve recently noticed, and may help with the calibration issues OP is experiencing. Bambu Studio and Orca Slicer both use radius as their metrics for X-Y Hole and Contour compensation; When OP adjusted by .1mm, the diameter measured .2mm larger, which is exactly what it is designed to do. This also applies to contour compensation; .1mm adjustment on a square will move both opposing walls out by .1mm, thus changing the overall outer dimension by .2mm. It may not be what OP was concerned about, and it had caused issues with my prints until I did some digging and learned that the slicers were using a radial compensation.
1 Like
I suggest you click on your own profile and read your own past posts. You help 0% of the time. I mean, how could you be helpful when you think bambu is purre perfection on a ssite meant for helping people troubleshoot problems. Most of your replies are to me and only jonraymond seems to like all your comments. Seems odd. Youve recieved 30ish likes while im well over 200. Youre in the top 10 bambu fanboys though. Your likes will always come from the top 10
“yet he still proclaims to have bought their printer,”
4 Likes
Great post! Thanks for explaining the this in such detail
I’m glad you agree with me about StreetSports.
I guess you missed the subtle hint. I do not agree with you. Hopefully this will clear up any ambiguity.
Personal attacks are a violation of this forum’s rules. Or for that matter, any forum I’ve ever been on. If you have a beef, use the private message system. But don’t PM me because I’ve placed your posts on block which means I won’s be able to see your posts or receive your PMs. Based on your lack of contributing to this forum, I presume I wouldn’t miss anything.
6 Likes
This topic is temporarily closed for at least 4 hours due to a large number of community flags.
This topic was automatically opened after 14 days.