Photoshop of course might be an good option to modify the files. But for me it is unclear how you want to put these back into the toolchain later on for printing. Beside that I don't like the G-Code approach to control all the motors and stuff on the machine. I like a more comprehensive approach but that is probably also because I come from embedded development background
Your idea with a kind of 'snap' mechanics to lock the projector into position sounds good. With that one could use low-cost mechanics while still having great repeatability. The only downside: this mechanism needs to be controlled too e.g. with a motor/solenoid or something else. I'll consider that...
Beside programming my stuff I also design the electronics. Not that hard at all - just takes a lot of time
Thanks,
radi2k
Statistics: Posted by radi2k — Mon Apr 11, 2016 6:47 am
]]>
I'll bet one could use Photoshop with a Photoshop action script to slice up the slice images for experimenting around. I've used Photoshop before to do something like that.
As for the mechanics, the best way to do it would be to just move the light coming from the projector. One would need to find a way to turn the light coming from the projector into a beam so that it could easily be directed around. Failing that, one would be left with moving the projector or the build tray. For this, one way to get around the repeatability problem would be to use adjustable detents. So, the projector or build tray would be positioned using regular means and then when close to the projection position it would stop there and a solenoid would active a locator pin that would insert into a locator hole and then the projector would project the image for that position. The locator holes would be made to be adjustable, such that they could all be positioned at the exact locations needed for perfect alignment of the projected image every time.
That's pretty cool that you're creating your own slicer! I wish I knew computer programming. That's the one thing I was never good at when it came to building stuff.
Statistics: Posted by James — Sun Apr 10, 2016 7:35 am
]]>
my projector and reservoir both were fixed. I just wanted to do some tests before the machine is redesigned in total e.g. additional moving mechanics. Yes, I use my own software for stepper control and the UI + projector image generation. I generate my slices using the free version of netfabb - a custom slicer is work in progress. Thus I don't use CW. I'm a software developer on my own and like to do my own stuff
I have big plans but I'll keep that on my own for the moment.For the test setup:
My projector has 1280x800 resolution. I split the image in two parts: Left and right, both exposing after each other and then the tilt move. That's why the repeatability is so good. Regarding mechanics to move the projector later on I don't have plans yet. I'll probably consolut my brother who is a really good mechanical engineer. But in general I do invest good amounts of money...
And yes, repeatability for the projector moves is critical. If the projector has 50µm resolution the mechanics should be much better so there is no chance for gaps as you mentioned. I'd like to avoid special post processing stuff since that's kind of bodge
Thanks,
radi2k
Statistics: Posted by radi2k — Sun Apr 10, 2016 5:24 am
]]>
Also, I reread your first post too and found, "However, no results yet - software not fully ready yet." Are you coding a completely new slicer or are you using CW and post processing the output images in some way?
Statistics: Posted by James — Fri Apr 08, 2016 4:29 pm
]]>
One idea I thought of while writing this just now is the linear exposure positions could be spaced slightly during the build. Once finished the resin would be held in place by capillary action, which could then be post process cured to "weld" everything together.
Not germane, but that brings up something else. If that worked, one could use that to alleviate shrinkage issues for regular SLA printing, maybe.
That's cool Steve that you're willing to work on this! Definitely a way to get more people interested in CW since a working machine and software combination would generate quite a bit of interest on AM news sites and maker news sites.
Statistics: Posted by James — Fri Apr 08, 2016 4:13 pm
]]>
First exposed one side and then the other. In the light reflections you can see it but structure is good. Also inspected it under a microscope. Looks like one piece.
Almost forgot it: no special pixel blending on the overlapping area was used to get that result. Probably also because I have very short exposure time of only 150ms (FunToDo Standard black)
However, I didn't move the projector for that. Now after that test I can redesign my whole machine!
Thanks,
radi2k
Statistics: Posted by radi2k — Fri Apr 08, 2016 3:20 pm
]]>
Statistics: Posted by PacManFan — Thu Apr 07, 2016 11:27 pm
]]>
]]>
Let's see if I can contribute some efforts making this home-ready. This evening I'll do some tests...
Thanks,
radi2k
Statistics: Posted by radi2k — Thu Apr 07, 2016 10:50 am
]]>
I wonder if Steve (the owner of this forum and the creator of Creation Workshop) has ever had a chance to work on this.
Statistics: Posted by James — Thu Apr 07, 2016 8:07 am
]]>
Even the commercial machines only provide poor printing area with DLP based systems. I want to take the benefit of DLP precision but scale build volume up.
Thanks,
radi2k
Statistics: Posted by radi2k — Thu Apr 07, 2016 6:30 am
]]>
]]>
]]>
]]>