Can an stl file for 3D printing made with Rhino cause Z height failure?

[Steve1]

Hi,
V8.
I have made in V8 a high end calibration file SR6 to test super fine detail printing for scale modelling purposes (not dwarfs and warhammer stuff) but serious scale modelling.
Its for what is called a R_E_R_F file test, and gets printed lying flat. 2mm thick with 1mm tall test details on top, thus 3mm in all.

However fed into the R_E_R_F file test for an AnyCubic Mono X 6K the 1mm tall test details on top are only 0.4mm tall ! Thats anti aliasing (AA) off and 15 microns layer thickness. 8 tiles starting at 0.75secs. 0.25secs increments.

I have now received the R_E_R_F file tests from Elegoo , printed on the Saturn Ultra 12K, with AA on and AA off (so 8 tiles per test, 16 in all) AND THOSE TEST ITEMS ARE 0.3 mm tall !

same SR6 printed on a Phrozen were 1mm tall, CORRECT HEIGHT ! though printed as a single print and at an angle.

Is this an issue with Rhino3D made items ? I MUST KNOW AS I AM WASTING FOLKS TIME PRINTING IT IF SO.

Its passed all the checks in Rhino I was told to do and also then in Prusa.

How can this be even from the HQ of the best printer out there (pending test results comparison !)

Can anyone with an AnyCubic Mono X6Ks (note the s) print it with resin suited for fine detail items, at 15microns as a R_E_R_F file test. I will pay for resin, postage etc.

SR6 attached, free for use, just say in any reviews its source, here ! , (I need to add it to my website, a WIP)
SR6 arch chamber supports 3D print calibration Rhino8 PrusaOK has 5pnt5mmdiaDome F.stl (4.5 MB)

Cheers

Steve

[wim (Wim Dekeyser)]

Hi Steve -

@Steve1

2mm thick with 1mm tall test details on top, thus 3mm in all.

When I import your STL file and set it’s units to be in mm, this is what I get:

STL files are unitless. What other systems do with those files is up to those systems (and, possibly, their users).
-wim

[Steve1]

Hi Wim,
yes thats correct, 1mm tall test details.
so why is it that AnyCubic fails to make them 1mm tall, and now Elegoo HQ has failed.

Itching to find an AnyCubic 6Ks user to see what happens. and just how good is the 6Ks at detail.

The whole point of a R_E_R_F test is to see what print is best, 1mm details made 0.3mm tall sort of fails that, though the XY accuracy is assessable. (new word ?)

So there is no chance that a Rhino made file can be the cause ?

How can two completely different independently designed machines and software have the same issues with that and R_E_R_F tests ?

I see youtube tests with the infant XP2 file and no one says oooh eck its failed in the 1mm area !

Steve

[theoutside (Kyle Houchens)]

fwiw- the mesh on this part is pretty coarse.

lots of very long triangles and such… some printers don’t like that.

is it possible to see the original rhino file not meshed?

I’d like to see this meshed at a much finer/more even/higher poly count for my tastes.

I’m wondering if that is contributing to the failure.

[Steve1]

Hi,
at the time, I asked of what tests and checks to do to make sure irs ok for 3D printing, and carried them all out, it also was then passed fit by Prusa.

It was exported as stl using all the defaults, as I havent a clue on what extra settings to use, and trust its set up for us to be ‘good to go’

Its 23.2MB how do get it to you ?

Steve

[theoutside (Kyle Houchens)]

you can use our uploader-

fwiw- here are some videos explaining the mesh settings-

02 - Basic Mesh Settings for 3D printing

03 - More Mesh Settings for 3D printing

[Steve1]

Hi,
sent.
I need to know ALL the things I need to do when making an stl file, I cant keep finding out more after the event.
I need to be 100% sure what I am sending is correct.
I am sure I am not the only one that trsusts in what defaults we are fiven, assuming thet are to allow us to get things right with normall accepted settings, and stl is the file type for 3D priinting.
Maybe there should be a button called stl for
3D PRINTING
WHICH WOULD CHOOSE THOSE SETTINGS FOR US, ON THE STL OPTIONS PANEL.
marked resin fine dtail, or fdm etc.
That would be an EXCELLENT IDEA .

please fwd that the boffins at McNeel.

As I am sending to a bureau service, who are spending a lot of time trying to solve the loss of 1mm characters, I have to be sure the file sent is 1000% correct.

and now there are stl settings to consider.

I did ask at the time what needed doing.

we all pour our heart and soul and burn many candles after the midnight hour, to get things done. so to have some options box and a can of worms called settings to have to overcome as well. please make it easy for us.

Steve

[theoutside (Kyle Houchens)]

replied via email- with a file to download via dropbox

I’d mesh it with these settings, then run triangulatemesh to detail it a bit further (some printers don’t like quads)

[ShynnSup]

It’s basic knowledge that higher detail requires a higher polygon resolution.

I expect any software generating an STL file to provide a dialog with resolution settings.

That’s exactly what Rhino does.

[Steve1]

Hi Kyle,
@theoutside

@theoutside

I’d mesh it with these settings,

Got the dropbox, meshed it with your settings received in the email, sent it off again.
There was no other instruction as such.

but now see a further step to do, not in the email.

@theoutside

I’d mesh it with these settings, then run triangulatemesh to detail it a bit further (some printers don’t like quads)

So I need to open the stl back into Rhino, then run TriangulateMesh

** followed video steps and used Kyles settings, no playing around with Min Initial grid quads setting.**

I have to get this right and form an entire checklist and export method, that will be pinned to the wall for every object being sent from now on. I have to have a finite list.

and anybody else sending for 3D print can also refer to this thread.

At least the loss of 1mm tall detail cannot be then blamed on Rhino3D and my settings.
Bureau have found its 1mm at 50microns, and it gets less after 35 microns, on a machine capable of 10microns, what the F is going on ?

but then failure.see next post.

Steve

[Steve1]

Hi,
@theoutside
ok sussed that, was in video.

Watched the videos,

Selected my object (which passed all the tests I was told to do in v8, and then was declared ok in Prusa.)
and again after this below i run zoomNaked, finds nothing on the object. but finds 122 on the mesh.

followed the settings EXACTLY of Kyle

then ran triangulateMesh,

then selected it and export selected, as Binary, following video.

sent it off to Bureau.

Then I opened it in Prusa as they also did at same time.

And declared faulty

122 naked edges

They also say their prusa says:-

exposition time is out of printer profile bounds.

As well as 122 naked edges.

Now as said this was bullet proof passed all tests in Rhinov8, has been used by Elegoo, no one spotted anything, Prusa was OK.

and now he tells me …Fusion 360 reports that the “Mesh is not closed. Mesh does not have a positive volume”.

As said it was declared 100% ok by V8. before being meshed, and 100% healthy after my stl export. I also did a selbadobjects on the mesh I just made.
I also ran Check, and all read ok.

I see also Kyles is with 122 naked edges.
I redo the mesh and all is ok, no naked edges.
I do the triangulation and get naked edges on a simple 90 deg edge.

thats where it goes wrong. Triangulation !

THIS IS VERY ANNOYING INDEED.

what has happened with the mesh settings to do this now ?

embarrasing to send it off saying its even better.

Steve

[David53 (David)]

What Rhino NURBS thinks and what whatever mesh is produced can be two very different things.

Here are several areas of suspicion for me, that clearly could be liable to failure if the accuracy is poor, or if they are so close, it comes down to computational accuracy and luck that it is closed or not.

[HugoIII (Pascal Heußner)]

Maybe one try is to use an 3mf instead of a stl file
The 3mf is newer and better for kind of imperfect (not watertight) meshes
Just an idea

[Steve1]

Hi, David,
I revisited those almost touching areas as I did have them touching, and touching is a no no.
so separated them out so as to fit within the tolerances.
and all is good with normal stl export, also Kyles settings.
all until I do the triangulation .

why should those areas nowhere near the simple long edge, cause failure for triangulation ?
I have sent it off yet again, without triangulation , meanwhile, so we can get some printing done.

I dont think the failure to get anything higher than 0.4mm when details are 1mm tall in a R_E_R_F test is down to triangles not existing throughout.
It works at 50microns, and at 30 then progressivley fails down to 15mics.
why ?
I printed ok on a Phrozen at an angle.

Hi Pascal,
I did try for 3mf and the resulting mesh was a mess. it was then suggested stay with .stl

it comes down to computational accuracy and luck that it is closed or not.

I am starting to wonder if owning Rhino is like owning a Ford Escort , forever at the garage, and thought this and that forever going wrong was part and parcel of owning a car, until I got a Mazda 323 Fastback, then no visits.

Steve

[Zaxxon (Zaxxon77)]

Hi Steve -

I have a bit of experience doing fairly dimensionaly demanding prints on Formlabs Form 3+'s with Rhino output - if it would be at all helpful and if you are able to share the .3dm file - I can run it on one of these at our shop and report the results.

It isn’t necessarily the identical workflow (different printer/slicer) - but maybe it could help shake out which at which stage issues are being introduced? Can only imagine how frustrating it must be trying to diagnose this issue remotely through the game-of-telephone through the bureau service - there are so many variables.

I do remember early on messing a bit with the settings in the STL meshing - but once they were dialed in I haven’t paid attention to these, and in that ecosystem - things have been decently stable, (at least on the model side - printer maintenance is another story).

[HugoIII (Pascal Heußner)]

My workflow is to use the mesh command in rhino to generate the mesh itself first (with precise personal settings) - and only then I use the „export selected“ command- for my Cura / Ultimaker Workflow I usually have no Problems with my meshes. But I know it hardly depends on the precise model / situation / slicer / settings / printer (and print Technologie)

[Steve1]

Hi Zaxxon,
Thats most good of you.
sent you a PM for email and then I can send the stl files old and new and the .3dm

@HugoIII

My workflow is to use the mesh command in rhino to generate the mesh itself first (with precise personal settings) - and only then I use the „export selected“ command- for my Cura / Ultimaker Workflow I usually have no Problems with my meshes.

I followed the two videos, ran the mesh command, used Kyles settings, checked it and all is good, Prusa likes it as well.
then triangulaemesh and it gets naked edges along the most simple edge. so why ?
If you want to take a look send me a PM as well.

Steve

[Zaxxon (Zaxxon77)]

Hi Steve -

Here are the results that I got using our typical workflow - i.e. my standard settings out of Rhino →.STL→Preform (the formlabs slicer/job-manager)

I made two prints in the Tough 2000 material @0.05mm layers. One was printed at an angle using supports and the other directly on the platform.

(don’t mind the unsightly sticky residue the IPA bath is a bit depleted, that’s the part they don’t tell about doing this stuff, you need to have IPA on subscription!)

(angled print left, flat print right)

3mm features (flat print)

3mm features (angled print)

The bottom of the angled print definitely shows some of the “bloating” effect that undersides of suspended prints exhibit.

[Zaxxon (Zaxxon77)]

[Steve1]

Hi,
extracts from PM sent.

I see you used 50 microns. Its for 15microns. It shows what can be achieved detail wise at 15microns, or 20 microns.
For all my figures and other items I intend to have 3D printed, 15 microns (or 20 at most) is essential.
50 microns was shown to be useless for such ultra fine details, after being sent such for 5 years, faces at 1/32 scale 1/48 scale etc
I had folk trying to emulate the Envisiontech Perfactory printers wth 50 mics, no chance !

Lying flat in a R_E_R_F test…
50microns gives 1mm tall details.
We are finding 30microns also 1mm tall, but then it tails off down to 0.4mm by the time we are at 15 microns
This when doing a R_E_R_F test on the AnyCubic Mono X 6K.
and Elegoo HQ on their Saturn Ultra 12K , theirs is 0.3mm tall for the 1mm tall details, so they also got afflicted by whatever is the cause.

Are you able to run the .stl you made with your settings at 15 microns ?, lying flat
Does the printer you have do a R_E_R_F test ? its in that test it goes flat.

if poss do that before sending in the post. I need to be comparing 15micron prints.

I see your settings differ from those Kyle used.

I will email you the results from Elegoo and AnyCubic, taken with a dSLR macro lens 60mm Canon EFL. I am just assembling them right now, and adding text.

You will find the results INTERESTING !

You have had some wierd happening on the ‘female’ sunken in ‘sea urchin’. compare yours to what you see these other printers have done. You have the .3dm and the .stl to see what they should be like.

I am dissapointed that the sunken text on Elegoo and AnyCubic fails. I have not seen it appear as it is in the original file.
I do have a mystery print, dont know who sent it, but it was printed at a 45deg angle, the sunken text is there, BUT ITS GOT ROUNDED EDGES TO EVERYTHING !
the base is smooth, no build plate pattern, somewhere is a box it came in then I will find out who did it ! the base has a raised lip with the file name running round it.

Steve