I have the issue - even after all calibration with the vision encoder as well, the holes are .1 to .3 mm smaller than they are designed. With my P1s, the sizes are 0.005mm off. I use on both printers the same spool of Bambulab Basic PLA.
I could try the hole compensation, but it seems not the right solution if all my other printers do not have this issue.
Does anyone has an idea or is it a bug?
With your P1S the sizes are only off by .005?? Are you sure you wrote that correctly? Maybe you mean .05? Which is really good accuracy for a 3d printer. .005 would be insanely good accuracy. My P1S is accurate to .1mm. When I make a 6x3mm hole for a magnet I have to make it 6.15x3.1. Although I haven’t tested my H2D, judging by the even better prints alone it must be as accurate or better.
For an average printer I’d say .1 to .3 was pretty good but it is not good for the H2D especially with the vision encoder and its claims. I don’t know what could be wrong but you may find it helpful to look under the shrinkage settings. I have never played with those settings but maybe someone else could chime in to help.
Hahaha, yea 5 microns, measured with what, a ten dollar caliper 
Anyway, yea compensation is often required machine to machine.
yes sorry I meant 0.05 
I also see the same thing on the H2D, all models containing e.g. M2, M3 screw holes are way too thight and I can’t get the screws in. My A1 mini and X1C print these just fine.
Yeah there is certainly an issue here. Im unsure if it affects all models or just a few. Holes that are supposed to be 3.2mm wide come out to 2.5mm-2.6mm for me. Applying hole compensation “fixes” the issue but it increases all internal geometry by that compensation amount. I believe there was a reddit post about this issue as well.
It’s a universal problem for all FDM machines. It’s not a single point failure. Even if the motion system is 100% accurate, e.g., you’re using Magneto X from Peopoly.
The smaller the hole is the less accurate it become. Flow rate, speed and acceleration, line width, layer height, filament fluidity and dimensional stability (especially shrinkage stability), humidity, all matters here.
If you want good dimensional accuracy, you would definitely use fibre reinforced filaments, or print very very slowly.
Most calipers are only accurate up to 0.1mm, even the quiet expensive mitutoyos (0.001" or ~0.02mm)
The Mitutoyo is definitely a “quiet” caliper. 
But still I sure all the other FDM parameters you mentioned are the cause for the major part of the deviation mentioned above.
That was my reddit post. Everyone says you need to calibrate the machine. But I have the same models printing across prusa mini, 4 different X1’s, and an A1 without being off any noticeable amount. It has to be an issue with the machine slicing to be as extreme as it is. IMO.
I appreciate the response, but this is not the case here and seems to diminish a legitimate issue. Yes it’s true that dimensional inaccuracies do exist due to expansion/motion system etc. But it is not normal for a $2000 machine to be turning 3.2mm holes into 2.5mm holes when a $600 P1S and $400 A1 are able to put out holes of 3.0mm - 3.1mm given the default print profiles and the exact same filament as the H2D.
A deviation between 0.1mm - 0.3mm is acceptable and can be attributed to motion system/expansion.
A deviation of 0.7mm or greater is simply far too much to chalk up to inevitable inaccuracies, especially when cheaper machines produce far better results.
I 100% agree with you. Something isn’t quite right if the deviation is so high on this machine while entry level printers from both Bambu and other manufacturers are able to consistently and accurately reproduce holes within 0.2mm of CAD. Being close to a millimeter off is frankly unacceptable for engineering work.
This guy will dismiss any issue or even suspicion of issue anyone raises. Just a heads up.
More over, can you check your Gantry Leveling? Check your belts and see if they drift up/down on the rollers while moving the toolhead around. If yes it could indicate a bent chassis/missalligned gantry, which could cause this type of issues.
If you wonder why I propose to check that, if your belts are grinding on their rollers, the whole motion system is put under load that it is not expecting, and ofcourse affects the motion accuracy too as any number the printer “knows” to be true, is thrown off because all the variables in the belt path change.
Afaik the filament profiles for H2D have been overhauled due to it’s stronger cooling and heated chamber. Might be that for certain profiles there are mistakes leading to this.
I am sure it is a slicer/profile issue and not the dimensional accuracy of the motion system.
Thanks for the heads up. I’ve noticed that quite a bit on these forums with people being dismissive of issues he/she definitely seemed to be in that group.
As for the belts, unfortunately I’ll be away from my printer for about a month so I won’t be able to check on that for a while. But you are 100% right that could certainly be something that could cause an issue. I saw a recent post from a user by the name of Bringle related to this with belt grinding and that could be something affecting my unit, I’ll have to check when I get back. I have a feeling it’s possibly both a combination of hardware issues like these along with under baked slicer profiles that causing these problems.
This is the thread you are talking about, I posted an analysis there of this issue, with photos from my machine, and couple users already confirmed that their belts show wear signs within a week of use.
In my case its confirmed that the gantry is not level, im just waiting for more people to confirm to see how will act from now on.
Because, on my Voron and other CoreXY’s this is something I can fix, and I expect to do so. On a 2.5k machine, brand new, right out of the box, with not even a wiki page on the matter though ? Hell no. Call me crazy but if this turns out to be a common issue, a recall might be imminent.
Seeing fibers on the side of the belts is not explicitly meaning that the belt is wearing off on the sides.
If the fibers are sticking out, you are right.
fibers by themselves though indicate something is not running smoothly. This is not normal even on your voron. You can see that it grinds, and grinding is throwing off anything in your motion system, more or less. Especially when the wear is inconsistent, indicating a deviation in belt path, which translates LITERALLY to “your nozzle is not where it thinks it is”
This belt is straight out of the package, and has not been on any printer yet!
then it simply does not apply to the discussion. This is not about an unused belt (which honestly yours appear to be in rough shape, ive never had one new that looks mangled, but that might just be the camera)
This is about ones that are in use and get destroyed. FAST. If its in the condition it begun with, its a different matter.
