Dimensionally Inaccurate Parts Being Produced by X1C

[dc3D]

Hi,

I can totally confirm the issue…
Old post I started 1 year ago:

X / y dimensions of prints differ

Hi, I noticed slight differences in the dimension of prints: Y: should be 100mm, reads 99,93 → OK for me [1674291748353] X: should be 100mm, reads 99,78 → not OK, should be closer to Y Dimension at least [1674291748348] Problem scales with larger prints, and is independent from material (green is PLA, black is PETG). Also X dimension, desired size should be 150mm: [1674291873878] Any ideas how to fix that? I always dialed in my other printers via e-steps calibration, but what can I …

I have years of experience and always could dial in (with patience ) my printers to +/- 0,05 on 100mm via E-Steps.

The only way that works on the X1 is compensate via scaling X/Y dimensions.
My sweet spot for PLA: Scale X and Y to 100.2%
My sweet spot for ABS: Scale X and Y to 100.5%

By the way, there is also still an error in Bambu Studio in visualizing the X,Y,Z directions via the colored arrows after rotating the part depending which value you are changing (% or the the actual dimension in mm)

Object was imported and rotated 90deg in X, than:

Scaling tool and % selected → Y arrows are shown up and down

Scaling tool and mm selected → Y arrows are shown front and back

Similar issues with X and Z arrows.

[bignazpwns]

First things first. If you wanna talk about precision You need to ditch The uncalibrated random Amazon calipers and get get a real instrument. Not only are those cheap ones inaccurate but they have drastic swings and measurements depending on the temperature.

[user_514568738]

That comment is irrelevant when you use the same caliper for two samples of printed object in the same minute.
I don’t need to know exact dimensions in this specific case. I need to know they differ by 0.2mm, which is perfectly achievable with basically any caliper available on market.
Please don’t get this comment wrong, I don’t question that in general people should use as precise tools as possible. Only it is not needed in this specific case.

[FaithCD]

Usually i calibrate my filaments regardless of brand. Every single one. I do not have such a problem so far.

[teilkuh]

at least the material shrinks after printing and cool down. if I want dimensional good results i upscale the model to 100,4 %. this is a value I found after printing some test models in different sizes in petg. at least it is 3D printig and you measuring with a calliper. There are limits to both.

[duane777]

All plastics shrink, I’ve had PETG shrink 1mm over 180mm. I only use ABS now so I deal with it before printing. I upscale everything to 100.6% when i remember or if it needs to mate with other parts that aren’t printed. I’m still waiting for Bambu to add shrinkage like Orca has had forever already. I’m not sure if orca scales in X&Y only or all axis. I need all axis. Also don’t trust auto flowrate calibration, I’ve had black and red be over 100% and when printing black it was over extruding a lot.

[Viking27]

Bonjour.
Personnellement, quandj’ai besoin de précision, j’imprime la forme sur 1 mm d’épaisseur, je verifie mes cotes et je modifie sur Solidworks.

[bignazpwns]

Just like any measurements because I can take a print and make it show whatever I want. Combine that with trash calipers with junk for reliability lol

[bignazpwns]

Even with a good calibrate accurate set if you don’t know how to use them it’s not accurate.

[Chip_Wallace]

You are being a bit stubborn to people trying to help and explain things to you, and for seemingly no good reason. When more people respond in kind, you basically state you no longer have time to deal with it because you’re not getting the response you thought you would get or wanted to hear.

Please don’t do that. Be humble, admit you don’t know everything, or maybe at least that you don’t know everything you thought you knew. This is how we learn things, and sometimes, relearn things due to our own misconceptions. Everyone here is literally trying to help you.

The base issue here is that you are expecting dimensional accuracy in the thousandths of an inch from technology that is not made for that purpose.

I too have been 3D printing for a long time, almost 10 years since I got my first MakerBot. Among other things am a mechanical engineer as well as a master machinist. I deal with all types of manufacture and fabrication every single day. I run 3d Printers, FDM, SLA, and SLS, CNC routers, lathes and mills, plasma cutters, and Laser Cutter/engravers.

Plainly put, FDM technology has too many variables to consistently get such tight tolerances. Even if you dialed it in perfectly this afternoon, you may get different results due to ambient temperature or humidity tomorrow morning. Change the filament, nozzle temp, bed temp, speed, etc… you would have to recalibrate every time. Your just chasing a ghost that is impossible to catch.

If your trying to fabricate something where the tolerances are that crucial, you should not be doing it on an FDM printer.

As someone already mentioned, it’s more about knowing how to design for 3D printing, as the printer just will not produce your design down to the micrometer.

If you need a part to be that exact, then something like a CNC manufacturing is the way to go. However, there are also variables to this as well. Machine a piece of aluminum to spec, then measure it outside in 98 degree heat and it will be larger. Aluminum will expand up to .024mm per meter for every 1 degree Celsius of temperature increase. So when designing parts we accommodate for this by designing for it, the same way you need to design for FDM printing.

Try to keep an open mind, be thankful there are such great online resources, forums, and people to learn from. I’ll be fifty this month and I am still learning every day, and you can learn something new from everyone.

[JonRaymond]

Solid first post. Welcome to the forum!

[Chip_Wallace]

Exactly.

Not too mention, how dimensionally stable is the part? Does it flex at all? Expansion under ambient heat or humidity? Far too many variables with FDM, your lucky if you can reliably stay with .00, let alone .000.

I print mechanical parts all the time and I never go to .000mm when designing. Which as a machinist makes my teeth itch, but it is what it is.

[Chip_Wallace]

Thank you!

Glad to be here

[PEZ3D]

I’m sorry if you took it that way, but that is not the case at all. You might understand more once you make a post of an issue you are having. For someone chiming in, you can do so and be done at any time, but it isn’t that simple for whoever made the post (look how many posts there are in such a short time). I keep getting the notifications of more and more replies and there is no possible way I can maintain a constant interaction or response for everyone who replies to the post on top of everything else going on. I once had a 3 month support ticket with Bambu, while also trying to attend one of my posts, and it was just too much to handle. Had I known I could not delete this post, I would not have created it and opened a support ticket at the same time all over again.

I also thanked everyone for taking the time to share their thoughts and feedback, so you are assuming things that just aren’t true. If I had the time, I would gladly reply to everyone, and do many tests to show my findings, but I simply dont have time to respond to everyone and attend this like I want to which is why I said I would not be responding anymore after realizing I could not delete the post and open it later on when I had more time.

I do appreciate everyone’s feedback (like I said multiple times on this post), and I am not always right which I will happily admit, however, this is a topic I have done many tests with and I would not have posted this unless I knew for certain that something is just off with the machine… I also noticed my resonance and leveling data is wiped or not saved (or not used?) if I calibrate from the printer LCD, but it is if I calibrate from the slicer, or when I select reprint and tick the bed leveling before the print. Something is just not right with the printer saving or calibrating, I just dont know what it is or if it is related to the dimensional accuracy. I could go on and on, but again, I just dont have the time to attend this like I want to.

Again, thank you all for your input, your experiences, your suggestions, your critiques, and even mild insults, but I just dont have time to respond to everyone… Maybe once this support ticket is worked out and some other stuff clears out, I can then try to come back to this and catch up. Right now I just need to put it on hold till I can put more time and focus into it.

[simonthompson1965]

Is the solution as simple as decoupling the X and Y axis compensation, so they can be adjusted individually?

On my various Ender’s I always measure the actual movement of X, Y, Z axis and compensate inaccuracy by adjusting the stepper movement, and also do similar for my E steps, so I effectively remove the all variables of different filament from actual movement.

I understand that core XY systems use two steppers in unison for X and Y motion, so you can’t decouple the actual X and Y stepper rotations to travel a given distance, but you could add compensation in the slicer at a machine level (ideally not at model level, so you only need to do this once) to adjust movement in all axis to compensate the actual movement measured vs the movement command.

Ideally E, X, Y and Z compensation should be available as separate adjustments, the you can be confident the bed, tool head and extruder all give as near as possible accurate movement. After this you can then refine flow etc on a per filament basis for final fine tuning.

So, just to clarify, I do not print a model say 40mm square, measure that and then adjust X,Y and Z steps. I attach a set of callipers to each axis in turn, send a command to move say 100mm, measure the actual distance travelled, re-run the test several times to ensure consistency, if necessary take an average, then modify X, Y, Z steps and re-test until the actual distance travelled is as close as possible to the movement command sent. Once I’ve done all that, then I will run calibration prints to check flow etc, and if necessary and any further compensation on a per filament basis to fine tune dimension of printed parts.

Should this should be achievable and fairly simple in Bambu slicer etc?

[Zenith1]

This thread is bizarre. I’m having the same issues as the OP. Multiple filaments, including ones that advertise minimal/no shrinkage and on stock settings the XY dimensions are simply wrong. I’ve confirmed this with multiple calibration prints, on multiple tuned filaments, and checked every belt and screw. I shouldn’t have to manually resize every part knowing that my printer is ~0.3% to 1% off. I am 100% sure it is the slicer or the printer’s firmware. The printer should arrive with good dimensional accuracy OR have a way for the user to manually adjust. The XY hole/contour fields are inadequate when used on complex parts and manually resizing is just unneeded tedium.

Yes, FDM does not have the tightest tolerances but we are talking about accuracy. People are receiving inaccurate machines and there needs to be a fix.

[dan.pinvidic]

Interesting…every post on this thread states valid information.
HOWEVER everyone seems to ignore the following PEZ3D post

“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”

When you have this situation it seem to me to really point to the machine.

[Olias]

@dan.pinvidic

HOWEVER everyone seems to ignore the following PEZ3D post

“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 don’t see this as anyone ignoring the original topic. We have a statement made by the OP without any details. Given that, in what manner can one respond to such a statement? It’s a statement devoid of any factual data. We have no idea what the printers or slicers being used are.

I would also add that the conclusion drawn here does not follow a logical progression.

@dan.pinvidic

When you have this situation it seem to me to really point to the machine.

Sticking to the closest principle of scientific method, what proof do we have that apples are not being compared to oranges?

The statement

@PEZ3D

I have printed same test models on multiple printers

On the surface that statement does not imply causality between the 3d models being printed and the printer. Unless I missed it and then if so, please point this out, there are two glaring omissions in this statement are:

• Which printers were they printed on?
• What slicer was used on those printers?

So if one were to conduct a purely scientific approach, one would have to first create a control group with the least amount of variables. Fortunately, Bambu Studio was based on Cura and it inherited support for multiple printers. So duplicating a scietifically valid experiment should be possible.

Simple experiment.

• Use the same slicer(Bambu Studio)
• Use the same filament
• Use the same test model.
• Use the same profile
• Measure and ensure that the ambient temperature inside the enclosure remains as stable as possible in order to rule out temp influences.
• Repeat each test on each printer.

Now this would be much closer to a scientifically valid experiment and therefore would lend force to the statement

@PEZ3D

I have printed same test models on multiple printers

and your subsequent conclusion

@dan.pinvidic

When you have this situation it seem to me to really point to the machine.

[Alex_vG]

@dan.pinvidic

and all of my opensource printers will print accurately since I can set the steps per mm on them”

I might be wrong, but this to me smells a lot like he actually calibrated the material shrinkage by adjusting the steps/mm on his older machines. The measured error of 0,3% to 0,6% is exactly in the range of what filament manufacturers state as shrinkage for e.g. PLA or PETG.

He also confirmed that the mechanics travel the correct distance, so I really don’t see any hint towards a faulty machine.

[Zenith1]

I did another calibration with eSun PLA+ today and got parts on average 0.35% small in the X and 0.8% small in the Y. That is not material shrinkage. That is the same inaccuracy my printer has had since day 1. It is a Bambu issue. This is after doing another round of belt tension and calibration. This time fully moving each axis thru it’s full motion over 12 times. Furthermore, I’ve seen people get +/- 0.01mm accuracy on Bambu machines without compensating for shrinkage. It’s just that not every printer that Bambu ships is accurate.

I was close to pulling the trigger on a P1S as a second printer but seeing other people complain about this very important and very solvable issue only to be met with handwaving instead of solutions is turning me off. It’s not the filament. It’s not the user. It’s not unreasonable expectations. It’s the printer.