Masking Troubleshooting & NanoDLP Host?

Forum dedicated to discussion and technical support centered around the LittleRP affordable, open source 3D resin printer.
[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1266: count(): Parameter must be an array or an object that implements Countable
Head-Zonk Studios
Posts: 25
Joined: Tue May 24, 2016 6:14 pm
[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1266: count(): Parameter must be an array or an object that implements Countable

Masking Troubleshooting & NanoDLP Host?

Postby Head-Zonk Studios » Mon Jan 09, 2017 5:38 pm

What I'm working with:
- LittleRP 2 Beta
- Fun-to-Do Deep Black resin
- ViewSonic PJD7820HD with
- See-Me-CNC Universal Flex Vat with 5 mil FEP sheeting

Quick recap: which (I think) was mostly due to the Z Lift being too fast. After that, things started coming out more cleanly, but had severe light bleed.
The exposure time was twice what it should be FTD resins are crazy sensitive.

Subsequent test prints revealed a fairly serious light imbalance with my projector. I created a new test print model to test how severe things are and where. Feel free to use it. You may have to scale it down by 10% or so.
16x9Grid.zip
(58.17 KiB) Downloaded 791 times


TestGrid_01.JPG

TestGrid_02.JPG


The one on the left is my first attempt. As you can see, about a third of it failed to build, even the base grid. I then made a mask, which resulted in the right one. It improved, but it wasn't quite enough. There's still a fair amount unformed and peeling.

Here's my situation in trying get this thing working:
1. I don't have the means to purchase and make the whole UV sensor set up I have to do this with trial and error, creating masks in GIMP.

2. I'm wondering if I may have screwed in the lamp slightly off kilter when I moved the color wheel. It didn't exactly have a LEGO tight fit in there.

3. I'm thinking sticking is still a problem. When the second test grid was printing, it was peeling up awfully hard. Got kinda loud at random points. One idea suggested by goopyplastic was using a different Flex Vat. The sheeting on SMCNC model is drum-head tight, where as HTL's model has more slack, which should allow for easier peeling.

4. When I printed with the mask I made, I also doubled the exposure time (500ms -> 1000ms). I think this contributed to the extreme sticking, but there was also some little squares of cured resin stuck to the vat. They didn't cover the whole underexposed area, but they were there. I've made a new mask that covers more and is significantly darker than the last. Since the weak areas were still underexposed, I'm wondering if I should increase the exposure time more.

5. I'm questioning the settings of the projector. With the way it was originally, just the ambient light was curing the new FTD resin (wound up with a cured sheet the size of the whole build area), so I had to turn down the brightness and up the contrast by quite a bit. I'm wondering if more light for less time would work better than less light for more time. Again, I'm fighting/avoiding light bleed here.

6. I have done nothing to the GCode. I've seen people on here writing things in for timing, speed, distance and whatever else. Is there something I could do that would help any of this?

7. Even the fully printed pegs on the test grid still had a ribbed surface. Goopyplastic said something about the high tension Flex Vat causing lines on the prints, but I'm not sure if that's what he meant. If that's not it, then it might have something to do with the reactivity of the resin changing, or the exposure settings, or the Z Lift speed... I don't really know.


I've also been looking at the new NanoDLP host software. It seems to have many advantages over CreativeWorkshop, including it still being free and updating.
The spread sheet for the LRP2 said it was an option (among other things), but it's made to work with Raspberry Pi, while I'm still working with Audrino. I don't know if it could work with Audrino, I don't know enough to convert or patch it in if I had to, and I do not have the money to throw at a Raspberry Pi set up.


TL;DR: My prints are still sticking to the vat, the lighting is grossly uneven, I'm having a hard time judging what to do with making the masks, and there's just so many variables to work with that all interact with one another.
I am at a loss for what direction to go in, and it seems like it's been taking a long time for people to get back to me.

I need some help and advice. I've had this thing for almost a year and I have yet to get a usable print out of it.

goopyplastic
Posts: 236
Joined: Sat Mar 29, 2014 6:16 pm
[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1266: count(): Parameter must be an array or an object that implements Countable

Re: Masking Troubleshooting & NanoDLP Host?

Postby goopyplastic » Mon Jan 09, 2017 6:03 pm

I think going the approach you have been is a good way to do it. UV meters are not easy to get good results from for a number of reasons. There is also a photographic method: http://robotsinthesun.org/dlp-printer-b ... tribution/

You will have problems with cure times as fast as you have, timing from the software is likely to vary a bit and 500ms is really just too fast imho. Additionally quicker cures results in more shrinkage, I would say you want to at least be in the approaching 2 second range.

You removed your color wheel but not sure that you were addressing a problem, you potentially have made things harder. The PJD7820HD is already a pretty hot projector in terms of uv. The LittleRP has a smaller projection window that a lot of printers and thus the uv / ir / visible light levels are higher given a certain area, resins will react differently in these circumstances. Fun-to-do deep black is not something I have tested on the LIttleRP so I can't vouch that it will even work, it should be workable though as the LittleRP is hardly any different in function than the titan1, b9 , solus etc.. I have a bottle so I will put testing in the queue.

It is possible that fun-to-do black will have a bleed at any exposure level given these light levels. Adding back the color wheel, putting the projector into eco mode, lowering projector brightness, changing projection color to a grey etc.. are options that can increase the cure time.

The PJD7820HD like the P1500 and H6510BD have image focus issues around the borders, so you are fighting more than just uneven light distribution, but areas out of focus may require more cure time as well. Typically resin cures enough to fully form a voxel or it doesn't, additional cure time beyond that makes the voxel harder or more fully cured. Border areas that are blurred (perimeter of projected object) will not be fully white but a gradient or grey if you will and won't receive as much uv seconds of light as an area in full focus... in turn they don't receive enough uv to fully form and are not cured.

So not only is it ideal that you need to correct for light distribution levels, but areas that are blurred will need to be over-cured to be fully cured up to the object border.

Most people with the PJD7820HD , P1500 and H6510BD will avoid the particularly bad areas of it's projection. It isn't as common that someone needs to print something that fills the entire build area ... in fact back when the LittleRP came with petris there was quite a limitation on how much surface area your print could occupy of around 25x25mm.

Going forward the PJD7820HD , P1500 and H6510BD will not be recommended as flexvats have allowed for larger and larger prints. That isn't to say the PJD7820HD , P1500 and H6510BD are not usable as at least half of my example prints were using one.

Head-Zonk Studios
Posts: 25
Joined: Tue May 24, 2016 6:14 pm
[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1266: count(): Parameter must be an array or an object that implements Countable

Re: Masking Troubleshooting & NanoDLP Host?

Postby Head-Zonk Studios » Sun Jan 15, 2017 3:17 pm

That light distribution tutorial is spot on. Their pictures even look a lot like the masks and prints I made. I think the distribution is so severe in this model is because the bulb is at such an angle. The light gets redirected at least twice before it comes out. I imagine those other models are built similarly.

FTD Deep Black is made to use for high resolutions with no bleed. According to the manufacturer, .030 is the maximum slice thickness. They say you could add more pigment if bleeding persists, so I may do that. They also say that it's made to work with DIY DLP printers like the LRP, and it's supposed to cure at 500ms with an unmodified projector.
So I think you're right, the wheel-less projector is too intense for the FTD resins, and the time is just too short to control.

I moved the color wheel to speed things up with MakerJuice SF Red for LRP. It was taking somewhere around 9 seconds a layer at .025 slice thickness. I had to plan half my day around one single print, which sucks when you have no idea if it'll come out right. Although now that I think about it, I may have really been fighting the uneven lighting the whole time. Increasing the exposure time overcame the uneven light but created other problems in the process.
Fortunately, I made things so I could put the wheel back.

I haven't been trying to build to the very edge of the build area. I figured you couldn't really do that, especially since the build plate isn't that big. Seeing what that tutorial said about closer mounted projectors having focus issues and your larger LRP2 model got me thinking. I could rearrange things so that the projector is mounted behind the machine, giving a little more distance for better focus, and maybe help with the curing issues. Not to mention getting the projector out from under dripping resin. But, again, that's another experiment for later.

I'm not sure why being able to make larger prints would make those three projectors undesirable, especially considering it seems like they're the only true HD projectors that don't cost thousands. When you're working with something at any size the higher resolution the better. I know you're compressing an image that's meant to be clear at 30-60+ inches to about as big as a credit card, but still.

Alright then... I'm going to put the color wheel back to help with the bleed and exposure (and maybe with the sticking and uneven distribution), then I'll rework the mask(s) to balance out the lighting. I also await the new integrated Flex Vat, thank you for that.

That leaves the NanoDLP host program. Your said that would be an option for the new printer. Would I need the Raspberry Pi board to work with that?

Mattasmack
Posts: 25
Joined: Thu Dec 17, 2015 1:11 pm
[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1266: count(): Parameter must be an array or an object that implements Countable

Re: Masking Troubleshooting & NanoDLP Host?

Postby Mattasmack » Mon Jan 16, 2017 5:42 pm

Hi Head-Zonk,
You mentioned a ribbed surface; what did that look like? I had problems with layer-to-layer banding at one point (I posted a couple photos in this thread: viewtopic.php?f=19&t=5062 ) that turned out to be due to inconsistent exposure times. If you're using very short times that makes this sort of problem more likely.

You might want to try slowing your lift waaaay down, if only for a test-print or two. I'm using B9 Emerald in an HTL flexvat, and I use a lift speed of 40 mm/min. If you have a flexvat with higher tension you could slow down even more; not having to lift as far should help compensate in terms of printing time. (I might be lifting more slowly than needed, but I had a lot of trouble with parts sticking to the flexvat and once I found a setting that worked I didn't have much motivation to keep experimenting, even at the cost of somewhat slower than necessary prints.)

If you're looking to move away from CW and/or untether your printer from a PC, another option is running Photonic 3D (nee Creation Workshop Host) on a Raspberry Pi. It doesn't control the stepper driver directly as NanoDLP does; you would keep your current Arduino-based motor controller and just plug its USB into the Pi. That's what I'm using currently and it's working fine on a Pi version 3 (it did not run well on an original Pi). If you go this route, having a Z endstop switch on the printer is a big quality of life improvement, but maybe the LRP2 has that already?

goopyplastic,
Are there particular projectors you do recommend going forward? I'm looking to upgrade the old one in my current LittleRP, and getting one that will also work in the LittleRP 2 would be preferable since I do plan to upgrade once the new printer is available.

goopyplastic
Posts: 236
Joined: Sat Mar 29, 2014 6:16 pm
[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1266: count(): Parameter must be an array or an object that implements Countable

Re: Masking Troubleshooting & NanoDLP Host?

Postby goopyplastic » Wed Jan 18, 2017 1:35 am

Mattasmack,

The x152h seems to be what I will steer people towards for 1080p it has superb optiocs and is priced aggressively, I also have been testing the hd37 like the solus uses and it works fine as well although it isn't a candidate for shimming and needs diopters that will induce barrel distortion ... ok for some things like jewelry, not ok for others like printing boxes. I imagine I will start putting emphasis on diopters and how to use them, as well as using projectors without the fixed projector base to encourage people to use whatever projector they can find cheap / used. It is a little silly to buy a 900 dollar projector for these purposes :) Although the x152h is a really good value the downside is that newer projectors seem to have more uv filtered and removing the filters has some sporadic reports of damaging the light tunnel. I will be testing a hot filter you can use on the x152h to decrease cure times but protect the internals, I imagine this is similar to what b9 has the vivitek factory do for them. Solus just runs the hd37 without a filter, I don't see any reports of damaged projectors....

Low uv output isn't that bad at LittleRP image sizes, going up it starts to make you wish you had something quicker.
LittleRP2 is getting close, I am just putting the final testing on the larger build area option for it.

Mattasmack
Posts: 25
Joined: Thu Dec 17, 2015 1:11 pm
[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1266: count(): Parameter must be an array or an object that implements Countable

Re: Masking Troubleshooting & NanoDLP Host?

Postby Mattasmack » Sat Jan 21, 2017 1:11 am

Thanks for all the info!
Oh jeez, I tried using a closeup lens on my current projector and I just could not figure out the focus and distortion problems. I eventually decided I didn't know what I was doing and shimmed it instead. Even though I mostly print jewelry, I don't like losing a portion of the print area to focus problems or distortion, so the x152h sounds ideal.


[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1266: count(): Parameter must be an array or an object that implements Countable
[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1266: count(): Parameter must be an array or an object that implements Countable