Dimensions acurate

[Nacho_abad]

Hello, I have a new Saturn for one month ago. Is my first resin printer, in fact, my first 3d printer. I'm Mechanical engineer so i am used to work with other manofacturer machines, CNC codes, etc...

I am impressed with the quality optained on my first printed pieces. I am using water washable resin, and I have noticed that succesfull prints need proper temeperture and proper supports.

Now im starnting measuring printed pieces dimensions. I show dimensions deviations betwen designed and obtanied. In the interior holes,  where I need  best accurate I have obtained the worst one. Maybe I didnt clean the piece well enought on the hole interior? I thought it could be antialiasing paramter, but I would had obtained same deviation on all dimension.
What do you think about it?

https://1drv.ms/u/s!ArXk-yRXcPgZgp5oYNuj...g?e=nhNjdT
https://1drv.ms/u/s!ArXk-yRXcPgZgp5q6EDc...w?e=FGQUWH
https://1drv.ms/u/s!ArXk-yRXcPgZgp5pmoCN...Q?e=g7OXSu

[cliffyk]

You need to keep in mind that these are inexpensive hobbyist level machines. Intended primarily for printing figurines (dolls) and the like where utmost precision is not needed.

The X-Y axis resolution of the machine is 0.05 mm (0.002 in.) I.e. an "accuracy" of no better than ±0.05 mm.

Also note this is the resolution of the LCD shutter, not the overall process--which I estimate to be at best ±0.5 mm, with "accuracies" in the range of ±0.5% (Industrial quality SLA printers "max out" at ±0.15% or so).

Small holes are worst affected by the process as the resin pools up in the bore and is cured via diffusion of the curing light in the media (which is clear with suspended pigment) and physical "overlap" of the curing process to unexposed areas.

I have found that attaining precision and tolerances in the range of even relatively low standard machining practices (I.e. ±0.001") requires post printing/curing machining of SLA printed parts.

Here's a good article from FormLabs re: accuracy and precision.

THe good news is the process has great repeatability--I have developed a set of empirical correction factors for various shapes to be applied as the component warrants--I perhaps should document these, they now exist only as "figments" in my imagination and "gut feelings'...

[Nacho_abad]

Thanks for the info. If you need help with your documentation, or make some experiments, let me know.

[cliffyk]

My general rule of thumb is that for male components of (threaded fasteners, rods, pins and the like) I decrease outside diameter by 0.5% to 1.0 %--and for female components (bores, nuts, etc.) I increased the inside diameter by the same. I have found trial, error, and experimental fitting up to be the best solution. I sometimes print only sections of components specifically as experiments prior to printing an entire part.

Like you I had hoped for better precision--thankfully the good repeatability makes the process viable...

[Nacho_abad]

I am experiencing the expected precision. But I think I can have more if I understand the origin of the issue. Once you work with any manofactring machine you know you have to work hard against it to have good results. All has its capacities and limitations, but i think i am far from the limitations of this machine. Maybe I can find some correlations and formulas to build a postprocesor between CAD and chitubox. This simply rule you had just sayed, can be easily implemented.

[cliffyk]

I wish you good fortune--I can offer some oversight and field testing if needed, however I have Parkinson's Disease and it sometimes affects my ability to concentrate...

[nighteagle]

I am experiencing the expected precision. But I think I can have more if I understand the origin of the issue. Once you work with any manofactring machine you know you have to work hard against it to have good results. All has its capacities and limitations, but i think i am far from the limitations of this machine. Maybe I can find some correlations and formulas to build a postprocesor between CAD and chitubox. This simply rule you had just sayed, can be easily implemented.

Have you seen my post for correct calibrating the printer - i don't mean the levleing-thing i mean the tesprints and measure and change the parameter on the chitubox slicer to become perfect dimensions in x and y - but not in z!..

https://www.elegoomars.com/forum/showthread.php?tid=699

The same for the Saturn - i have mine since three days and have printed now 0,5kg for calibration. Now x and y are very accurate about 0,02mm but z-Height the same prolem like the Mars Pro 2 and  elegoo-service are incompetent and ignore the problem by saying print in another direction. So the the model is fine accurate but not in the dimesion the z-Axis is.
I have calibrate alos the parameter for  Z-Height but the results different - looks like is some mechanical backlash problem - i can only modify the cad.model after first printing to add offsets in z-Height to have absolute correct dimensions. But it is a workaorund - not the best solution.

[cliffyk]

0.02 mm (±0.01 mm) is about as good as you are going to get. Some tweaking might pull that in a bit, however we need to keep in mind that he pixel size of the LCD shutter is 0.05 x 0.05 mm--so our "printing" consists of piling up stacks of mushy 0.05 x 0.05 x (whatever layer height) mm "bricks" that arealternately squashed and stretched during the printing process.

(See my detailed post here).

One must also consider the vagaries of the resin printing process (light scatter in the resin and that the light only "initiates" curing; further curing beyond the exposed area is continued by molecular action ).

The "Printing" process consists of piling up stacks of mushy little 0.05 x 0.05 x (whatever layer height) mm partially cured polymer"bricks" that are then alternately squashed and stretched during the remainder of the  printing process. Given all that a ±0.01 mm accurate actual output seems pretty good.

It might improve if the vertical surfaces of prime interest could be placed precisely at inter-pixel boundaries--but that would be a daunting task.

My brief research on the web seems to indicate--somewhat counter-intuitively--that our "happy-homeowner" hobbyists level machines may operate at or near the limits of the technology.

[essorrac]

I'm kind of relieved of at last finding this thread. I'm the recent owner of an elegoo mars 3, coming from the FDM world, and I'm trying to print mold masters as my first actual non-tests or random mini fig prints. I honestly drove me mad this week-end

The main problem is a z-axis discrepancy between my two mold masters, which should be almost identical as their model is basically sliced in the middle, with some minor adjustement. Both are mostly a big flat rectangle with the actual mold object in the center, yet I have big differences between one half whose flat rectangle base is thinner, and the other one that is a bit thicker to  accomodate hollow half spheres, and that in addition to that I did print hollowed.




The thinner half (half spheres printed, the two on the right) is missing at least 1mm of details, as if sunk into the flat base, with the same printing parameters. I did try to print it twice, reslicing anew, for a similar result. All the slicing were done with chitubox.

From what I'm reading here, it seems that the main problem is that I did print my flat rectangles flat on the bed, on supports, not tilted at all, and that there is some major lack of precision on this configuration when you're printing flat details above a big chunk of resin.

It really looks similar to the staircase example described in one of the previous posts : https://blog.honzamrazek.cz/2019/08/test...s-problem/ , the object is kind of squished just after the flat surface. i can't really explain why it's less noticeable and inaccurate on the first print, on the left of the picture, just that the flat base if hollowed there, and probably less thick just below the surface.

Did anyone finally figure out and know how to mitigate that, apart form just printing the pieces very titled on supports (which will take a hell lof or additional time but well...) ?

[billtodd]

I am experiencing the expected precision. But I think I can have more if I understand the origin of the issue. Once you work with any manofactring machine you know you have to work hard against it to have good results. All has its capacities and limitations, but i think i am far from the limitations of this machine. Maybe I can find some correlations and formulas to build a postprocesor between CAD and chitubox. This simply rule you had just sayed, can be easily implemented.

Arrrgh Chtibox just crashed and wiped out my reply... once more with feeling:

I find I have to treat the process like casting, allowing for and compensating where possible for shrinkage and warping.

I have seen, but not tried, suggestions such as heating the whole build to 30+ 'C to reduce the heat induced shrinkage, however , there is also some shrinkage from polymerisation , so it remains necessary to adjust the model.

There is a beta Tolerance compensation option in ChituBox but I've not found this useful (OK it's crap - it distorts small objects)

The standard scaling option works OK for some things , but most of the time a tweak of the model post print in necessary.

I now tend to use designed-in supports where I think the model will distort and add sacrificial material on hanging edges to avoid rough finishes .

Ultimately, it may be necessary to add machining tolerances so the part can be milled/ground to size.

Attached is a printed extruder  including the stepper motor housing , to get a reliable fit I had to machine the stator mating surfaces and bearing fits , but after three iterations it is still working