Author Topic: Anycubic Photon  (Read 38101 times)

0 Members and 1 Guest are viewing this topic.

Chris333

  • Crew
  • *
  • Posts: 13081
  • Respect: +2430
Re: Anycubic Photon
« Reply #495 on: December 12, 2018, 03:47:22 AM »
0
No chance it was an errant drip from above when the reservoir was out?  Maybe during a color change?  I guess it wouldn't matter anyway as the film would already have a flaw in/ on it that could distort the curing light. :|

Naa there was a mark on the FEP in the same spot and the spot on the glass looked like the last layer of what I printed.



Chris333

  • Crew
  • *
  • Posts: 13081
  • Respect: +2430
Re: Anycubic Photon
« Reply #496 on: December 12, 2018, 03:49:33 AM »
+2
I posted my boxcar to a HOn30 Facebook page and they all wanted one. Asked if I could just put it up on Shapeways. Just for fun I uploaded it and my cost to have them print it in FUD was $43.82   :o 

Just a guess that I can get 10-15 boxcars out of a $30 bottle of resin.

Lemosteam

  • Crew
  • *
  • Posts: 3910
  • Gender: Male
  • PRR, The Standard Railroad of my World
  • Respect: +1187
    • Designer at Keystone Details
Re: Anycubic Photon
« Reply #497 on: December 12, 2018, 05:19:22 AM »
0
@narrowminded , so you have the Z dimension postulated well I think, and you have to understand one thing.  Not only is the screen curing the last section as a set thickness of resin atop the previous layer, it is curing that TO the previous layer, hence the extra time to do it. Of course all of this is relatively infinitesimal, but you being a miniaturist machinist understand that infinitesimal effect.

I do disagree slightly on one of your points though, that the displacement of the fluid has no effect on previous layers, there must be some, particularly on thin walled parts, much like filling a cavity with waterm there will be deflection of the walls (as I noted with the tender shell waviness of the sides, which can also be addressed play below.

Here is what I want to know.  What is the accuracy of the screen's projected section (layer, slice, or whatever one wants to call it)? This accuracy (inaccuracy?) has much more effect on the dimensional stability of printing the nominal design.  Much like an old time projector surely there must be some way to adjust the X and Y slice dimensions to calibrate its accuracy?

This is fine if one is printing a virtual blob like a human figure or a stagecoach (not to pick on you @Mark W ), but when fitting one part to another accurately, these dimensions and the tolerance on them matter.

If not, this is a major flaw in the design or software of the printer, IMHO. This will also be exacerbated when slicing at an angle because now the dimension is no longer lateral to the plane of the screen but foreshortened by the angle when sliced and printed. I mention this because the tender shell that Mark so kindly printed and mailed to me, unfortunately would not fit the mating, sintered metal printing of its chassis floor., critical to my kit.

Can you find anything about that? No one should have to redesign a part or scale it in either direction for dimensional accuracy.

In the early days of SLA, each layer of film was controlled by a wiper bar that wiped across the previous layer and then a round laser was used to trace the outline of each slice (how accurate were the stepper motors then?) to trap the resin and then the internal area of the slice was cross hatched to sinter the last layer to the previous. Thankfully this is all done in one DLP flash which is why parts can be printed in hours instead of day, I digress.
« Last Edit: December 12, 2018, 05:25:15 AM by Lemosteam »

peteski

  • Crew
  • *
  • Posts: 20363
  • Gender: Male
  • Respect: +1783
    • Coming (not so) soon...
Re: Anycubic Photon
« Reply #498 on: December 12, 2018, 05:37:27 AM »
0

In the early days of SLA, each layer of film was controlled by a wiper bar that wiped across the previous layer and then a round laser was used to trace the outline of each slice (how accurate were the stepper motors then?) to trap the resin and then the internal area of the slice was cross hatched to sinter the last layer to the previous. Thankfully this is all done in one DLP flash which is why parts can be printed in hours instead of day, I digress.

As I understand the X and Y dimensions are permanently fixed since the UV light is simply projected through pixels of a screen (like screens used in smart phones) sitting directly under the clear FEP film.  There are no XY stepper motors - the screen is stationary.
--- Peteski de Snarkski
--- Honorary Resident Curmudgeon

narrowminded

  • Crew
  • *
  • Posts: 1519
  • Respect: +466
Re: Anycubic Photon
« Reply #499 on: December 12, 2018, 06:40:20 AM »
0
As I understand the X and Y dimensions are permanently fixed since the UV light is simply projected through pixels of a screen (like screens used in smart phones) sitting directly under the clear FEP film.  There are no XY stepper motors - the screen is stationary.

Yes.  And there's a fixed percentage adjustment to the program, unlikely to be needed as they are set up from the factory, that can be made if needed to correct those axis if for some reason you were consistently printing with an error in any of the planes. 

And John, as far as forces from the Z axis movement, while something is surely there, it is so low that it doesn't really matter to the process.  If you look at the pictures of the tipper chassis, take note of the saddle coming off of the frame and the two guides that project upwards from that saddle.  Those were printed with no supports whatsoever, hanging out in the breeze.  What forces exist aren't enough to hurt those parts where the largest dimension in any plane is .03", then .015" for the tee with both of them .008" thick.  The two vertical supports are .028" wide and it all printed VERY nicely.  I was amazed to see that too but the original test pieces and every one in the twelve piece group after that printed flawlessly.  That was part of what made me think this through further, coming to the realization that the print process was as I understood and now observed first hand.  There was no real force applied to those parts to make them NOT print straight and true.  100% success rate so far, tipper hoppers and chassis.  No rejects after the very first test print, the lower section of the chassis only.  That print was done with their automatic support placements and the failure was anticipated just from the review of the support placement, with zero experience, the first print ever, and utilizing my first ever 3D drawing, done in Sketchup. 

One thing I have seen in these threads was a problem with a very broad face printed flat with full surface area printed on each layer.  Those had a problem remaining stuck to the build plate but I suspect that was due to the z axis upward move with that much area effectively sealed to the bottom of the vat by the viscous resin, much as two flat faces will stick momentarily with an oil or grease layer on their face.  The resin creates a seal around the perimeter so the atmospheric pressure keeps force against its removal.  That might be able to be fixed by adding some holes to an inconspicuous place if that were possible so it would afford more opportunity for the resin to flow through, breaking that seal easier.  Otherwise, just slowing down the Z upward movement might help but that's not an adjustable parameter, at least on the menu.  But the conventional fix is to print the part at an angle, not engaging the whole face at once.  BTW, my chassis prints were tipped 30 degrees and 15 degrees in X and Y, not recalling which was which.  The hopper was tipped 15 degrees in only the X axis.  The chassis is open enough that it might print just fine flat but I don't know that.  Afterall, I'm brand new at this with much to be learned. :D
« Last Edit: December 12, 2018, 06:47:03 AM by narrowminded »
Mark G.

Chris333

  • Crew
  • *
  • Posts: 13081
  • Respect: +2430
Re: Anycubic Photon
« Reply #500 on: December 12, 2018, 07:05:20 AM »
0
My boxcar ended up printing flat with no tilt. The roof is peaked so the whole thing doesn't stick all at once. But it does leave a boat to hold resin so when done printing I tip it for a while to drain.

I drew up a similar gondola, but haven't tried it yet because the deck is flat. Maybe some heavy supports will keep it from ripping off?  I don't want to print at an angle because it comes out soooo much better flat. Some surfaces look machined.


narrowminded

  • Crew
  • *
  • Posts: 1519
  • Respect: +466
Re: Anycubic Photon
« Reply #501 on: December 12, 2018, 07:34:43 AM »
0
My boxcar ended up printing flat with no tilt. The roof is peaked so the whole thing doesn't stick all at once. But it does leave a boat to hold resin so when done printing I tip it for a while to drain.

I drew up a similar gondola, but haven't tried it yet because the deck is flat. Maybe some heavy supports will keep it from ripping off?  I don't want to print at an angle because it comes out soooo much better flat. Some surfaces look machined.

Well, that's incentive enough to change both of my parts to flat.  The hopper, due to its length, had all wall faces engaged at once for a goodly portion of the print time and before I printed it that caused me some concern so I just added extra supports down the center of the bottom, hanging it there with the open end facing down.  After seeing how it printed with zero troubles I am pretty sure that the whole thing would print as you described and for the reasons you described.  I'm beginning to think that the supports and issues aren't as difficult, sensitive, nor as mysterious as much of my reading suggested.  The process is beautifully simple and beautifully clever and I think many of the day to day issues that might be encountered will be satisfied with a minimal understanding of the gory details but with a good logical understanding of the process and then good logical review of the problems that do surface.  8)
Mark G.

Chris333

  • Crew
  • *
  • Posts: 13081
  • Respect: +2430
Re: Anycubic Photon
« Reply #502 on: December 12, 2018, 07:38:27 AM »
+2
Have you messed with that ChiTuBox slicer yet?  Once sliced you can play it like a video and can even zoom in on each slice to look for problems.

Some screen shots:


Here you can see at one point the roof is solid.


And the zoom in on the door:

« Last Edit: December 12, 2018, 07:45:52 AM by Chris333 »

narrowminded

  • Crew
  • *
  • Posts: 1519
  • Respect: +466
Re: Anycubic Photon
« Reply #503 on: December 12, 2018, 08:02:18 AM »
0
Have you messed with that ChiTuBox slicer yet?  Once sliced you can play it like a video and can even zoom in on each slice to look for problems.

No I haven't.  The only reason I looked at that at all was because it was mentioned on the facebook blog in the thread about my video card issue in my old computer.  Once I got that all I've done is use the supplied slicer and placed my own supports after seeing the results of theirs, testing out the function of everything this far, including my drawing and all of the peripheral tools like the wash rigs, curing lamp and turntable, stuff like that.  It's only been four or five cycles running parts.  And right now I have some other things occupying my time so the printer is cleaned up and put away.  I'm thoroughly satisfied that it's a great addition to the bag of tricks and beyond anything I would have predicted just a month ago.  It's a huge change for me.

I take it you like that slicer.  I suspect that it's not the same one I mentioned because that one seemed to be identical to the one that came from Anycubic and when I looked at it on a wild chance that it might work on my old computer I didn't have to do any signing up or offering up of my firstborn.  Is it a truly free download or just a trial with fees to follow if you keep it?
Mark G.

Chris333

  • Crew
  • *
  • Posts: 13081
  • Respect: +2430
Re: Anycubic Photon
« Reply #504 on: December 12, 2018, 08:11:35 AM »
0
It is free , but you had to sign up to download it. It lets you save supported files as an .stl. And after you add supports it will let you move the supports around.  Besides that it go rid of the couple problems spots I was having:
https://www.therailwire.net/forum/index.php?topic=45736.msg598043#msg598043

narrowminded

  • Crew
  • *
  • Posts: 1519
  • Respect: +466
Re: Anycubic Photon
« Reply #505 on: December 12, 2018, 08:45:22 AM »
0
Sounds good.  Easy to use too, I take it. 

I'm okay with signing up as I've gotten pretty quick with the delete key. ;)
Mark G.

rodsup9000

  • Crew
  • *
  • Posts: 480
  • Gender: Male
  • Respect: +162
Re: Anycubic Photon
« Reply #506 on: December 12, 2018, 12:33:19 PM »
0
@narrowminded


Here is what I want to know.  What is the accuracy of the screen's projected section (layer, slice, or whatever one wants to call it)? This accuracy (inaccuracy?) has much more effect on the dimensional stability of printing the nominal design.  Much like an old time projector surely there must be some way to adjust the X and Y slice dimensions to calibrate its accuracy?

If not, this is a major flaw in the design or software of the printer, IMHO. This will also be exacerbated when slicing at an angle because now the dimension is no longer lateral to the plane of the screen but foreshortened by the angle when sliced and printed. I mention this because the tender shell that Mark so kindly printed and mailed to me, unfortunately would not fit the mating, sintered metal printing of its chassis floor., critical to my kit.

Can you find anything about that? No one should have to redesign a part or scale it in either direction for dimensional accuracy.



 The way I understand the calibration of a SLA printer like the Photon, is the cure time. There is a calibration plate on thingiverse that if you take the time to do it right, you can get the dimensional accuracy I think you looking for.


 The way to do the tender shell is once you have a print, find out what needs to be adjusted and adjust. Then print again. Do this till you get the fit you want. I've had to do this several times when printing parts (on the FDM printer) for things like a gearbox with bearings. I keep making the adjustments to get the bearings to fit a slight press fit. For this type of adjustment, I just used the scale tool in the slicer. With your tender shell, you may have to go into CAD and make the adjustments till you get the fit you want. It just takes time and a little wasted resin, but I think what you want is achievable with the Photon.
Rodney

My Feather River Canyon in N-scale
http://www.therailwire.net/forum/index.php?topic=31585.0

Lemosteam

  • Crew
  • *
  • Posts: 3910
  • Gender: Male
  • PRR, The Standard Railroad of my World
  • Respect: +1187
    • Designer at Keystone Details
Re: Anycubic Photon
« Reply #507 on: December 12, 2018, 12:56:16 PM »
0
Yes.  And there's a fixed percentage adjustment to the program, unlikely to be needed as they are set up from the factory, that can be made if needed to correct those axis if for some reason you were consistently printing with an error in any of the planes. 

And John, as far as forces from the Z axis movement, while something is surely there, it is so low that it doesn't really matter to the process.  If you look at the pictures of the tipper chassis, take note of the saddle coming off of the frame and the two guides that project upwards from that saddle.  Those were printed with no supports whatsoever, hanging out in the breeze.  What forces exist aren't enough to hurt those parts where the largest dimension in any plane is .03", then .015" for the tee with both of them .008" thick.  The two vertical supports are .028" wide and it all printed VERY nicely.  I was amazed to see that too but the original test pieces and every one in the twelve piece group after that printed flawlessly.  That was part of what made me think this through further, coming to the realization that the print process was as I understood and now observed first hand.  There was no real force applied to those parts to make them NOT print straight and true.  100% success rate so far, tipper hoppers and chassis.  No rejects after the very first test print, the lower section of the chassis only.  That print was done with their automatic support placements and the failure was anticipated just from the review of the support placement, with zero experience, the first print ever, and utilizing my first ever 3D drawing, done in Sketchup. 

One thing I have seen in these threads was a problem with a very broad face printed flat with full surface area printed on each layer.  Those had a problem remaining stuck to the build plate but I suspect that was due to the z axis upward move with that much area effectively sealed to the bottom of the vat by the viscous resin, much as two flat faces will stick momentarily with an oil or grease layer on their face.  The resin creates a seal around the perimeter so the atmospheric pressure keeps force against its removal.  That might be able to be fixed by adding some holes to an inconspicuous place if that were possible so it would afford more opportunity for the resin to flow through, breaking that seal easier.  Otherwise, just slowing down the Z upward movement might help but that's not an adjustable parameter, at least on the menu.  But the conventional fix is to print the part at an angle, not engaging the whole face at once.  BTW, my chassis prints were tipped 30 degrees and 15 degrees in X and Y, not recalling which was which.  The hopper was tipped 15 degrees in only the X axis.  The chassis is open enough that it might print just fine flat but I don't know that.  Afterall, I'm brand new at this with much to be learned. :D

Mark, you might want to measure some of your parts before you make that claim on your first point.  Factory set or not, the shell below is nearly 0.31mm too small at the 11.08mm dimension which measures 10.77 on the printed part, let's us assume this was the Y dimension.  The 18.38mm dimension on the part measures in two places 17.97mm near the 11.08mm dimension (0.41mm too small), and  17.75mm respectively at the rear of the shell which is 0.63mm too small, and the reason why the shell cracked back there when I tried to install the tender frame, the front was a slight press fit without cracking. The 71.62 (image is wrong) came in at a very close 71.68, only a 0.06mm deviation, so the shell fix in the X dimension nicely, although I do not like the bow on the back wall. 

I was not referring to the Z axis force.  I was referring to the pressure the fluid is under when an open cavity is compressed on viscous fluid with nowhere to exit.  It may be imperceptible on on shallow cavities, but when a shell that is 20mm deep with 0.5mm thick walls is compressed you can get this, which may or may not be a function of the supports having difficulty keeping the wall where it should be during printing:


While not exorbitant, it IS warped and that will be noticeable after paint. 

@peteski, I think you missed my point.  In the old days a TV had H an V adjustments to fill the picture tube properly. I am not referring to resolution, I am referring to which row of "pixels" (this is DLP after all hence the quotation marks) the screen puts a straight line in.  Adjustment could move the line to the next pixel "row" effectively moving the printed wall.

Thanks @rodsup9000 , you replied as I was typing this.  Yes a trial and error is always required to establish a dimensional baseline, but please look above.  I have dimensions that should be parallel and off by roughly the same amount, but the clearly are not.  This leads me to believe that in addition to dimensional adjustment, there is something else at play.

rodsup9000

  • Crew
  • *
  • Posts: 480
  • Gender: Male
  • Respect: +162
Re: Anycubic Photon
« Reply #508 on: December 12, 2018, 01:49:36 PM »
0


 Yes a trial and error is always required to establish a dimensional baseline, but please look above.  I have dimensions that should be parallel and off by roughly the same amount, but the clearly are not.  This leads me to believe that in addition to dimensional adjustment, there is something else at play.


 One thing that is it could be is the way the shielded light (pixels) is bounced off the mirror to the plate. Another possibility is something is happening in the slicer and making the dimensions that much off.

 Just a couple of my thoughts.

Rodney

My Feather River Canyon in N-scale
http://www.therailwire.net/forum/index.php?topic=31585.0

narrowminded

  • Crew
  • *
  • Posts: 1519
  • Respect: +466
Re: Anycubic Photon
« Reply #509 on: December 12, 2018, 02:41:07 PM »
+1
John, I'm new at this and don't have the definitive answer to your specific question about this machine printing to your specified number as it sits.  I don't have that experience yet.  I'm also not proposing that your part should be made in this way or any other way.  What I do know is that the parameters are adjustable both in the machine primary settings as well as in the slicer, adjusted per individual print.  As I run this more I will know what those capabilities are for certain.

But from what I DO know, speaking not just for this machine but from parts manufacturing in general, all I need to make properly dimensioned parts is a consistent, repeatable process and that is the SINGLE thing that will matter, all measurement numbers be damned, metric, English, or cubits.  All else is adjustment from the start point whether in the machine basic parameters, a secondary adjustable parameter, (the most convenient ways) or even if the dimension that is inserted from the drawing isn't the result, BUT it is, ALWAYS some other very exact number, change that number.  (Pattern makers do exactly that as a key tool of their trade.)  And that's ALL I need to cover that requirement, a repeatable process. 

With the repeatability requirement in mind, as I measure these pieces, 16 of one and 20 of the other, all I have made over a few runs, most measure comparatively within a half thousandth in x, y, and z, splitting the dial divisions on a set of 6" dial calipers, and the two excursions are no worse than +/- .002".  That's ALL I NEED to repeatedly make good mating parts as your example requires.

What I will do with running time is confirm the details of what's happening over the range of part dimensions in each axis.  If adjustments are warranted they will be made.  If it's something that varies with different resins, that will be recorded for future compensation and as required.  And if it's a single thickness adjustment, that too will be made when required.  Point is, if the process is repeatable, so CAN the parts be that are produced by that method.  I will know with experience but the small sample this far suggests that the process is VERY repeatable with various options to compensate as needed.  I hope that helps clear up my claim and if not, sorry, it's all I've got at this point. 
Mark G.