Anycubic Photon

[narrowminded]

Can you raise one a few thousandths?  It seems like it's seeing the same plane across the top face and trying to join them somehow.  Rotating them in Z wouldn't change that.  So, maybe if those planes were different?

[AlwaysSolutions]

egads!  That appears to be around the same spot as the original orientation around layer 175? (If I'm reading that other picture correctly)  Are there any individual parts on your model that start around that height?  I'm wondering if maybe an unclosed solid, or some other artifact is messing with the slicer.  I really have no idea so at this point it's all exploratory.

[AlwaysSolutions]

Have you settled on any rules of thumb when choosing a print angle?

I think you'll find it's going to be different for each project.  The idea is to minimize the number of lit up pixels on your curing screen at any given time.  So let's say you're printing a chassis for a boxcar - a nice rectangular geometry.  If you print it without rotation you'll have several layers of large, solid rectangles lighting up to cure the resin.  This puts a lot of strain on both the FEP and supports (or if you're printing without supports, the areas connected to your build plate.  This can cause unwelcome pull-off from support mounts and a disfigured print.  As you rotate, those cross sections get smaller and smaller until you're printing vertically, which may be ideal to maximize quality, but nobody wants to spend 12 hours printing something.  So you balance it out.  20 degrees may be enough to keep the model from ripping off its supports while producing "decent" quality, 45 degrees is nice because the step lines are masked a little better.

One thing I'll add as a thought for the print-time component - don't be afraid to use thicker layers to draft-mode something.  I've used .08 mm layers to speed up print time just to get something in my hands where steps weren't important.  It speeds up print times by quite a bit!  Then take it all the way down to .025 for a crazy nice finish when you're ready.

[VonRyan]

The results you guys are getting with these machines are fantastic!
I might have to contract out to one of you guys for a print whenever @Lemosteam finishes drawing up CDJRC "Phoenix".

[Chris333]

I searched "triangles" on the Facebooks and found a post were they said you can use a Photon validator and take the slice with the triangles out and fix in MS Paint, the replace it.

This is all I got looking for the program:
https://github.com/Photonsters/PhotonFileValidator

What in that link is the program?

Found it here:
https://github.com/Photonsters/PhotonFileValidator/releases

[rodsup9000]

 I'm guessing this what your looking for.

https://github.com/Photonsters/PhotonFileValidator/releases

Let us know if it works.

[Chris333]

Needs Java to run 

[John]

OK so I tried it. 2 boxcar bodies side by side, but offset. And each of them was spun 1 degree. One +1 degree, the other -1 degree.

You can see the stepping that would happen.

But then I found this:

So, what does that mean?   Being completely clueless on this stuff --

[Chris333]

The white outlines on the right are what will be exposed on the screen and that image is what the UV light will shine through. The UV cures the resin.

The big white blob is only one slice (0.05mm thick), but it will screw up the finish.

Each slice (layer) is different, but that blob of triangle shouldn't be there. Both layers in the 2 photos should look pretty much exactly the same.

[John]

The white outlines on the right are what will be exposed on the screen and that image is what the UV light will shine through. The UV cures the resin.

The big white blob is only one slice (0.05mm thick), but it will screw up the finish.

Each slice (layer) is different, but that blob of triangle shouldn't be there. Both layers in the 2 photos should look pretty much exactly the same.

Thanks ..that makes sense .. I think

[DKS]

Although I'm in no position to speak from experience, I suspect that if you change the z-position of one of the two renderings by a few thousandths, it might solve the problem.

[carlso]

I am like John and have no clue about what is being written. Chris, in the one photo above you mention that we would see the stepping that would occur. Just what is the stepping that you refer to?

Inquiring minds, dontcha know,
Carl

[Lemosteam]

@carlso I know you asked Chris.  The screen that is used to cure the resing is based on pixles, not lines.  then you angle something, or design a curve, the edge becomes a series of shorter lines that "step".  Where a line would be a straight line from endpoint to endpoint, the pixellated screen cannot duplicate this.
                                                                     __________________________
                              _____________________|
________________|

Another example:

[Chris333]

Exactly what John said. The one one the left is not rotated at all, right is 5 degrees.


The stepping is the resolution of the LCD screen pixels.

In this close up the whole model is square, but the diagonal brace isn't so you see stepping. It is pretty much hidden and is OK by me, but if it was along most of the major surfaces it would not be fine.


Would end up something like this:

[nickelplate759]

That jagged approximation of the desired edge is called "aliasing".   In computer graphics and digital audio there are techniques to improve this, which are called (surprise!) "antialiasing".  I'm not sure how that could be implemented in a 3D printer though.