New Phrozen Printer "Sonic Mighty Revo"

[Peachymike]

Thanks for the photo Mike, but it is taken nowhere near close enough to be able to judge the surface quality.  I guess I would have to see it in person.  In any case, thanks for the explanation.

OK tried again with as much zoom as I can and as close as I can. Really need to dig the DSLR out  This is the boiler, in n scale. The diameter is about 0.5 inches and overall length is about 2 inches.
[ Guests cannot view attachments ] [ Guests cannot view attachments ]

[robert3985]

A bit of thread drift....

I've got a little app on my old Google Pixel 2 called "Magnifier" that gets right down on top of the small stuff.  You might want to see if you've got it, or if you can install it for your model close-ups. 

Here's a bell and bracket I made a couple of days ago which I photographed using "Magnifier"....



I very seldom use my Nikon and my Nikkor Micro Lens any more.  The phone is too handy and does a great job.

Cheerio!
Bob Gilmore

[Chris333]

If you get close enough in a photo you can see the layers:


This is barely visible to the eye in person and will be even less once painted. But it is there.

[peteski]

Yes Chris' photo is  exactly what I was talking about.

[robert3985]

If you get close enough in a photo you can see the layers:


This is barely visible to the eye in person and will be even less once painted. But it is there.

This is an N-scale engine??  These are "layer" lines and are common on complex, curved surfaces when 3D printed in just about all different types of printers.

These are alleviated by fiddling with "blur" and "Anti-Aliasing" on both MSLA and DLP printers, as well as making the layers as fine as possible.

Each of these types of 3D resin printers will soften the harsh edges of the layers, while at the same time lessening the sharpness of other details on the print.  MSLA printers are getting very adept at minimizing these layer lines, while DLP printers, having much better light collimating capability have trouble getting rid of these lines...especially the first consumer model, the Anycubic Photon Ultra, which although printed noticeably "sharper" than any MSLA printer at the time, was plagued with the layer line problem because of an inadequate blur/AA algorithm.

So, this means that it's up to the user to determine how much blur and AA to apply, because it's a compromise between sharpness (more definite layer lines, but sharper corners, rivets, screw-heads, square holes, etc.) and smoothness of complex, curved surfaces, but slightly rounded edges, corners, less detailed rivets and screw-heads.

Locomotives are especially difficult because of the complex, curved surfaces that abound on them as can be seen in Chris's photo.

Another thing to keep in mind is Z-axis accuracy (up and down).  Slicers allow the user to determine the layer height, and some resin 3D printers are built with finer, more precise Z-axis accuracy than others.  A common minimum Z-axis accuracy is .01mm or 10microns (0.000394") using a "lead" screw for up & down build plate movements.  Premium machines with more expensive, more precise double ball-screw Z-axis actuating screws commonly have a .005mm or 5microns (0.0001968") Z-axis accuracy, which is twice the accuracy of the less expensive machines. 

So, between messing with blur, and AA, and layer height, with MSLA 3D resin printers, the near elimination of the these pesky "layer lines" should be possible...and because of the sharpness of Chris's layer lines on his print, I am going to guess he's running an Anycubic Photon Ultra...the first consumer-grade DLP printer...which was extremely sharp, but notorious for having a substandard blur/AA algorithm, making sharp layer lines quite prevalent.

Cheerio!
Bob Gilmore

[Chris333]

That loco is HOn30 so about N scale in size. Think it was printed on the Sonic Mini 8K (22 microns). Not sure AA works with that printer using Chitubox. I could blur them out a bit, but really I'm fine with the fact my eyes can't see it.

EDIT: I print everything at 0.05mm layers. While this mostly round steam loco is printed at an angle, I usually print everything perfectly flat. And 0.05mm layers keeps sagging away from larger openings like windows.

[haasmarc]

AA works on the mini 8k in Chitubox.

What layer height are you using?  I use 0.02mm and it works well for me.

[Chris333]

I just edited my post. A bunch of Youtube videos say that you can set AA, but it does nothing with certain printers.


NEW EDIT: https://www.reddit.com/r/Phrozen/comments/mijsnc/sonic_mini_anti_aliasing_doesnt_work/

[Peachymike]

If you get close enough in a photo you can see the layers:


This is barely visible to the eye in person and will be even less once painted. But it is there.

Definitely don’t get anything like that on mine. Angling a print like this one was does not always help with layering, you need to calculate the arctan based on the size of the pixel verses the layer height. The idea is to limit stepping so it occurs in sync with the layers. As I said before I use 25 micron layers, but the pixel size is 28.5 micron. Doing the calculation you end up with an angle of 41.6 degrees. If I were to arbitrarily use 45 degrees as many do, I would likely end up with layer lines as in this photo. If you are still getting lines at that point, try rotating the print on the plate as well, very often this causes the lines to diffuse somewhat. I don’t have a formula for this, it is somewhat subjective to the shape of the object, but as a rule of thumb 15 degrees is a good starting point. Rarely I will tilt the object as well, as I have done with a coal gondola that was dealing me fits with its tapered sides. Again no formula, that was just a matter of experimentation to find the optimal angle. Note that this was on a Mars Pro 2, I only need the first two angles with the Saturn 2.
Sorry for the drift, but I am hoping that the information will assist some of the novices and reduce their frustration with results. I love to tinker with my technique just to see what is possible. 
Mike

[peteski]

Coming up with a sweet-spot angle to minimize the visible layering is very useful for flat surfaces, but will not do much for complex curved surfaces (like Chris' loco), that is because some parts of the curved surfaces will be printed with less favorable angle.  That is where the size of the voxel makes a difference.  The smaller it is, the smoother the surface will be (to the point where the stepping can be successfully smoothed with a layer of primer.

[Lemosteam]

@peteski is right- There is no escape...   Regardless of the calculations you make or the angle you set.

Assume you've heard the adage: If you only take 1/2 step toward the wall, it does not matter how many steps you take, you will never arrive to touch the wall.

Such is the digital world we weave in 3D printing.

Muwahahahahahaha!

[Chris333]

Yeah the stack shaft  and main part of the sand dome on my steam loco have no lines at all. So there is no ideal angle.

[robert3985]

Yeah the stack shaft  and main part of the sand dome on my steam loco have no lines at all. So there is no ideal angle.

Sure there is.  It's the angle that gives you the least layer lines...or the least layer lines where you can most readily see them.

The object here is not to eliminate layer lines, although that would be nice, but to MINIMIZE them.

The whole world is composed of particles of one kind or another, cells, grains, fibers, molecules, atoms, subatomic particles...and the closer you get the more evident it becomes that there is no surface that is 100% "even".  There is always granulization, pixelization, voxelization, etc., in our quantum reality...which, if I understand it correctly, means that most things exist (at least part of the time) as particles.

Since the goal here in 3D resin printing for model railroading is to minimize layer lines to the point they are unnoticeable to the unaided human eye...or even in close-up photos...some point of reference might help us determine just how "good" a printer needs to be before layer lines disappear from our everyday notice...and from our close-up photos.

Just exactly what is the close-up visual acuity of an average 20/20 human eye?  Wellll....there's no consensus to exactly what that limit is and varies in opinion from 200-400 microns, 100 microns, down to 58 to 78 microns.  The problem is how to calculate it and quantify it since the human eye is much different than an LCD screen.

One small explanation I read that seemed very logical and well thought out, puts "normal" human eye visual acuity at 100mm focus distance, at 6 - 29 microns.  You can read it here if you're interested: https://biology.stackexchange.com/questions/26189/whats-the-smallest-size-a-human-eye-can-see

Sooo, assuming that "normal" visual acuity is somewhere between 6 and 29 microns, I'll just put it right in the middle at 18 microns...which is 0.000708", or slightly less than an N-scale 1/8 of an inch.

right now, the very best MSLA printers (the Elegoo Mars 4's) have LCD's that have 18 micron square pixels...right at the limit of human eye discernment if my references are correct...and produce in the resin vat voxels that are probably around 45 to 48 microns, and that would be SHARP voxels, with sharp corners and edges...which MSLA LCD screened printers don't do...they produce blurry voxels, with rounded edges and corners.

This means that although the diagram that @Lemosteam has provided us with illustrates the point that there will always be layer lines & voxelization, the ACTUAL printed voxels in the resin are not nearly as sharp-cornered (especially with MSLA printers) as is illustrated.

Increasing blur and AA will further degrade the sharp edges and corners of the printed voxels to where they may actually be 99% unnoticeable...in fact, this assumption is based on the material I researched and my interpretations...so, if "normal" human visual acuity is less, then we may already be to the point where with proper blur/AA, small print layers and print tilt, 3D resin prints will have better visible surfaces than injection-molded models.

On the other hand, layer lines will never be completely eliminated, but I am confident we are getting very close to where they will become invisible to the naked human eye and to our close-up photography too.

Cheerio!
Bob Gilmore

[Peachymike]

Man you guys are a tough audience! OK I thought to compare apples to apples I put a light coat of Tamiya primer on it. Took the picture with my wifes phone which has a much better camera, at a similar angle. Lighting is a little dark on the smoke box front. The boiler is almost exactly 0.5 inches across. The whole boiler looks like this, nary a line to be seen. Proof is in the pudding so to speak... [ Guests cannot view attachments ]
Mike

[robert3985]

Man you guys are a tough audience! OK I thought to compare apples to apples I put a light coat of Tamiya primer on it. Took the picture with my wifes phone which has a much better camera, at a similar angle. Lighting is a little dark on the smoke box front. The boiler is almost exactly 0.5 inches across. The whole boiler looks like this, nary a line to be seen. Proof is in the pudding so to speak... (Attachment Link)
Mike

THAT looks like we're already there.  I don't see any layer lines...and neither does the close-up shot.

EXCELLENT!

Cheerio!
Bob Gilmore