Anycubic Photon

[u18b]

So here is something strange. I printed 2 boxcar bodies side by side. No more triangles since I've been using the pixel editor. Right after the print was done I scraped it off the plate and loaded up a different file to print. Went to scrub the boxcars and found one was deformed, it is the top on here.


Once I see it I think there must be something on the FEP and go pause the machine. Nothing on the FEP, continue printing. That print came out fine. So I try the boxcars again. Same exact file, and get the bottom boxcar deformed even more. Each time the "other" boxcar printed fine.

I checked this file before and after and there is nothing wrong with it. I just now did the screen test and it is working fine. And remember I did a print between the 2 failed prints that was fine.

Any ideas?   

I think these would look good heavily weathered in a dead line.

[timwatson]

I purchased some Elegoo Gray since Anycubic was out. It prints really nicely. This is at .04mm layer height. 11sec exposure time (probably on the high side for exposure).

N scale unittrack mated to my Ttrak riser for code 70 flex track.

[ Guests cannot view attachments ]

[BuffaloJohn]

Looking at the bodies, the interesting thing I see is that there are parts of the print showing on both that are correct, but skewed. Since the body is created from the bottom of the body toward the top and the features are bent, that seems to indicate that area had a support failure of some sort, either a blowout (unlikely since the bottom of the body is open), a support detachment (might add a few more supports in that area, or, the FEP is flexing and not returning to flat (time for new FEP or need more retraction & time to ensure resin flows into area before lowering for next layer.

My thought of testing this would be to print only one of the bodies, but in the same location to see if it comes out ok. I don't think it would fail, but if the FEP is flexing too much, there would be distortion on the side that fails with two.

So here is something strange. I printed 2 boxcar bodies side by side. No more triangles since I've been using the pixel editor. Right after the print was done I scraped it off the plate and loaded up a different file to print. Went to scrub the boxcars and found one was deformed, it is the top on here.

_snip images_

Once I see it I think there must be something on the FEP and go pause the machine. Nothing on the FEP, continue printing. That print came out fine. So I try the boxcars again. Same exact file, and get the bottom boxcar deformed even more. Each time the "other" boxcar printed fine.

I checked this file before and after and there is nothing wrong with it. I just now did the screen test and it is working fine. And remember I did a print between the 2 failed prints that was fine.

Any ideas?   

[wvgca]

how far can you move the boxcars away from each other ?? that way you -might- reduce the tendancy of the two supports to interfere with each other ??

[narrowminded]

I just compared the specs between Elegoo resin and Anycubic.  They are identical specs, to the decimal.  Sounds like the same stuff.  

Not saying one's better or one's worse as I don't have that experience in use but just that they are the same in Amazon's listing for them.  Without anything to compare to at this point I can say I've been happy with the prints I've gotten using Anycubic resin.  I'm also inclined to look harder at durometer and higher elongation for break than tensile strength, expecting those to indicate toughness and ability to withstand some abuse while I'm handling the parts, the worst load I expect they will ever see, but again, still don't have enough experience with either to make a strong recommendation for one over the other.

I did the same comparison a few days ago between Nova and Anycubic because the packaging appeared identical but the specs were different.  My recollection is that the Nova had a slightly higher tensile and was slightly harder and less elongation along a break.  I read that to be slightly stronger but more brittle, exactly what I don't want/ need.

Edit add:  I guess specs depend on which seller is selling it.   Never knew the same thing could have two different specs.  Disturbing.  

This link: https://www.amazon.com/dp/B07FCM5V7M/ref=biss_dp_sa2  states the specs that are identical to Anycubic's listed specs.  Here's the site with those specs from Anycubic: http://www.anycubic3d.com/products/show/1273.html  They read identical.

Then there's this set of specs from the same brand:  https://www.amazon.com/ELEGOO-UV-Curing-Standard-Photopolymer-Printing/dp/B07FCNHQDJ/ref=sr_1_1_sspa?ie=UTF8&qid=1547492036&sr=8-1-spons&keywords=elegoo+resin&psc=1 


Two different sets of specs.      I guess you pick the one you like and then say the other one doesn't mean it.  Or are they the same and represent a vendor change?  If it's confusing, it should be... but shouldn't be.   See what I mean?

[narrowminded]

@Chris333  That failure almost looks like a substantial piece of debris got caught up in the motion and just started mangling the parts with each "Z" move.  The inside face of that angled failure, the face of the break, does that look to be broken or is it possible it was printing that way with a minor void in the part face that shouldn't have been there.  Almost like a removal of the original triangle line somehow dropped out the pixels for the line, causing a minor void that was unsupported and then finally broke loose sometime late in the print.  Was there any substantial debris in the vat?

What is striking is that it almost had to be attached early as those properly printed artifacts are present at the bottom of the scrunched up remnant, bottom face, boards, and all, albeit all scrunched up.  (Scrunched is the scientific name for "fubared".)

[narrowminded]

I am starting to get some predictable dimensions on surfaces using gray resin.  Not sure if this will hold for other colors but if not, will probably be close. 

"Z" axis is basically within .001"+ and isn't a surprise based on how that axis functions.  Both "X" and "Y" seem to grow between .002" and .003" per face.  That's not a growth or percentage error over a distance but a straight up addition to the face, suspected to be from the light reflected through the resin, curing a little more than intended.  This means that on a square part, two opposing faces, it will measure .004" to .006" over the drawn dimension but the location, or the C/L remains as designed.  Everywhere that I used that correction the part came out within .002" of the intended dimension, usually better. 

That factor was arrived at by measuring a number of parts from various runs (finally sat still and did it) and then utilized on a new design and to my total satisfaction.  That design was making Code 40 Nn3 flex ties to receive code 40 rail, complete with tie plates gauging the rail.  The rail guides held .001" (no surprise, utilizing "Z" axis) and the gauge came out exactly as drawn, at .257", the minimum tolerance (also "Z" axis).  I may add .001" or .002" to the guide plates for paint allowance but we'll see as the paint I just applied to a set of ties will need at least 24 hours to cure/ shrink to finish thickness.  I just sprayed it a half hour ago.

Really small pockets or holes (>.01") are... who knows.  Keep in mind the .003" per side then the viscosity of the resin and you see what you're fighting.  I will play more with that as I go but expect to always drill the small holes using the designed in hole as a locator and pilot hole.  They drill very easily, nicely aligned, and accurately located with just a pin vise. 

I used a trick based on this type of info when designing the rail foot guides in the ties.  The concern was any radius that might occur at the transition from vertical to the tie bed as that is the guide for the rail foot.  With only .009" and another raised section (spike) in the center of that tie plate topping out in the center at .012" I decided to relieve the vertical guides at their inner base, the two sides that capture the rail, to assure a true vertical guide.  The relief was a pocket, designed extending the inside vertical face into the tie at .004"  deep x .005" wide, .003" wider than the guide face, leaving a flat foot for the rail base at .030".  (This was what the design said, never intending the final part to show any of that.)  This gave somewhere for the over cured material to go, based on the behavior of the material noted above, and the result, first time, was a very straight, sharp corner, vertical guide for the rail foot.  I was pleased.  That is the repeatability thing at work and work it did!   That is all for now but I will add things as I go.  Pics maybe later.   It really worked well.

[Chris333]

Need rust or grimey black colored resin 

[Lemosteam]

I am starting to get some predictable dimensions on surfaces using gray resin.  Not sure if this will hold for other colors but if not, will probably be close. 

"Z" axis is basically within .001"+ and isn't a surprise based on how that axis functions.  Both "X" and "Y" seem to grow between .002" and .003" per face.  That's not a growth or percentage error over a distance but a straight up addition to the face, suspected to be from the light reflected through the resin, curing a little more than intended.  This means that on a square part, two opposing faces, it will measure .004" to .006" over the drawn dimension but the location, or the C/L remains as designed.  Everywhere that I used that correction the part came out within .002" of the intended dimension, usually better. 

That factor was arrived at by measuring a number of parts from various runs (finally sat still and did it) and then utilized on a new design and to my total satisfaction.  That design was making Code 40 Nn3 flex ties to receive code 40 rail, complete with tie plates gauging the rail.  The rail guides held .001" (no surprise, utilizing "Z" axis) and the gauge came out exactly as drawn, at .257", the minimum tolerance (also "Z" axis).  I may add .001" or .002" to the guide plates for paint allowance but we'll see as the paint I just applied to a set of ties will need at least 24 hours to cure/ shrink to finish thickness.  I just sprayed it a half hour ago.

Really small pockets or holes (>.01") are... who knows.  Keep in mind the .003" per side then the viscosity of the resin and you see what you're fighting.  I will play more with that as I go but expect to always drill the small holes using the designed in hole as a locator and pilot hole.  They drill very easily, nicely aligned, and accurately located with just a pin vise. 

I used a trick based on this type of info when designing the rail foot guides in the ties.  The concern was any radius that might occur at the transition from vertical to the tie bed as that is the guide for the rail foot.  With only .009" and another raised section (spike) in the center of that tie plate topping out in the center at .012" I decided to relieve the vertical guides at their inner base, the two sides that capture the rail, to assure a true vertical guide.  The relief was a pocket, designed extending the inside vertical face into the tie at .004"  deep x .005" wide, .003" wider than the guide face, leaving a flat foot for the rail base at .030".  (This was what the design said, never intending the final part to show any of that.)  This gave somewhere for the over cured material to go, based on the behavior of the material noted above, and the result, first time, was a very straight, sharp corner, vertical guide for the rail foot.  I was pleased.  That is the repeatability thing at work and work it did!   That is all for now but I will add things as I go.  Pics maybe later.   It really worked well.

It would seem that the x and y, I'll call it growth from light bleed, is a function of the Z in the sense that the more Z you print, the wider and wider it gets?  So a set of ties shows infinitesimal growth, but a 19mm high boxcar would measure noticeable growth, and the height would be dead-on.

So with a known deviation over Z would it be possible to modify the slicer algorithms to compensate for that growth in terms of pixels per layer?

It would seem that printing an entire layer at a time saves print time, at the expense of X & Y accuracy?  A single point laser SLA machine would not have that issue.

[narrowminded]

John, each of these errors appear to be a one time event at the face of the cured resin.  It's a one time event because it all is happening in a small, controlled, repeating window.  The X and Y are straight up light bleed and the Z is mechanically limited by the computer controlled absolute Z travel.  For Z to be different it would have to jack up the whole damn screw, frame, screen and anything else mechanically in the way.  For X and Y it's the light bleed at each step.  There's is nothing else to stop it.  The X and Y bleed occurs at each step or it WOULD do what you decribed but it doesn't.  The .002"/ .003" face increase occurs pretty uniformly so must be happening in the immediate area of the section under light being cured.  It doesn't keep growing because the light doesn't.  It goes as far as it can, curing everything in its path... which is apparently about .002"/ .003".   And because it's all  repeatable, so is the correction.  And in fact, that's what I just described and just did.  Compensated for the repeatable and made good, accurate parts to my specified directions (drawing).   The beauty of machinery.  It doesn't operate on mood or opinion.

[Lemosteam]

John, each of these errors appear to be a one time event at the face of the cured resin.  It's a one time event because it all is happening in a small, controlled, repeating window.  The X and Y are straight up light bleed and the Z is mechanically limited by the computer controlled absolute Z travel.  For Z to be different it would have to jack up the whole damn screw, frame, screen and anything else mechanically in the way.  For X and Y it's the light bleed at each step.  There's is nothing else to stop it.  The X and Y bleed occurs at each step or it WOULD do what you decribed but it doesn't.  The .002"/ .003" face increase occurs pretty uniformly so must be happening in the immediate area of the section under light being cured.  It doesn't keep growing because the light doesn't.  It goes as far as it can, curing everything in its path... which is apparently about .002"/ .003".   And because it's all  repeatable, so is the correction.  And in fact, that's what I just described and just did.  Compensated for the repeatable and made good, accurate parts to my specified directions (drawing).   The beauty of machinery.  It doesn't operate on mood or opinion.

Then I misunderstood, I took from your post that the variation increased as the print grew from the first layer. Did exposure time also have an effect?

[DKS]

Would the resin fail if small amounts of dye were added, like the dyes used to color casting resin?

 

[Mark W]

That shouldn't cause any issues.   Some DLP Resin manufacturers even offer pigment sets to allow users to mix their own color.  At most, one may need to adjust exposure settings depending on the color, just like is necessary with different resins/colors already. 

[narrowminded]

Would the resin fail if small amounts of dye were added, like the dyes used to color casting resin?

As Mark said, you can color.  But I don't think you can color too much beyond translucent because light still has to pass through to cure the resin.  This is evident in the additional cure times needed to cure colored resin over clear (2-3x?) and they are just translucent, especially evident in thin sections.  When you think about the process it makes sense.

[peteski]

As Mark said, you can color.  But I don't think you can color too much beyond an opaque because light still has to pass through to cure the resin.  This is evident in the additional cure times needed to cure colored resin over clear (2-3x?) and they are just opaque, especially evident in thin sections.  When you think about the process it makes sense.

Maybe I'm being pedantic Mark, but to me opaque means "fully light blocking" as in "not allowing any light to pass".  The same definition is at https://www.dictionary.com/browse/opaque :
not transparent or translucent; impenetrable to light; not allowing light to pass through.

I think what you mean is that even the gray resin is translucent (but not transparent like the clear or green resins are).  Translucent material allows some light to pass through.