Are there any parts shops for SLA/DLP/resin printers yet? There are several shops selling kits, but none of them seem to be selling any replacement parts for them. I’m considering doing a custom experimental build, but I’d like to find a place where I can get a bunch of the parts along with some resin. Maybe a circular vat/petri dish like the LittleRP and DropLit use, plus the non-stick coating, a build stage of some kind, all parts that someone using one of these machines might eventually need/want to replace. Has nobody set up a shop for this type of stuff yet, or am I just not finding it?
I haven’t either, although I haven’t looked that hard. It’s been mostly “get this here, that there, the other thing at some third place”.
I have just been buying the parts here and there. I wish there was a New Egg for 3D printers where you can buy hardware for 3D printer experimenters.
Most of the stuff is self source-able. I am using EGW15CCZ0 / EGR15Txxxx for rails from automation4less, robotdigg threaded rod / anti backlash nut, and I cut my own mirrors from kaleidoscope glass. The larger acrylic petris are nice but glass corning are nice and flat just shallower. I have 1000 acrylic petris coming thursday if want some. MLsolar.com is the cheapest for the silicone coating. The goal yes is to have a webstore up at least selling fastener kits, petris, coatings, sample resins etc. I am open to suggestion of what to carry.
Cool Brad I hope you do well with the store. I think the mirrors are special type I forget what is special about them.
I found this pretty interesting. Testing on build platforms and vat coating. http://www.khwelling.nl/3d.dlp_printer.testing.php
There’s info on sources in the article.
I’d be most interested in finding a suitable LCD. DLP seems a little overkill (even a pico). As I’m interested in making a tiny dirt cheap petri dish sized SLA with FDM printable parts.
@Liam_Jackson is that possible? I thought they used dlp for the uv content of the light to speed layer exposure time?
@Ben_Delarre there have been a couple of designs recently using an LCD over a UV LED array, but you really want all of the light beams to either be parallel or converging on a single point and evenly distributed. I haven’t been impressed with the output of any of these machines that I’ve seen, though to be fair most are still in the experimental phase.
@Whosa_whatsis I guess a uv led array with a large collimator lens would do the trick. Probably expensive though.
I know that the guys at mUVe sell a lot of the individual components of their SLA printers.
We are making an assembled bot but would consider a kit. If we go ahead with the kit, we would sell all the key parts. We investigated the LCD method but it is still immature, although shows promise if someone made a cheap high res transparent screen. We are using the same dlp as B9 - $500-800 retail. Since it includes the optics, light source, hdmi input… It sidesteps headaches. We are making and coating our own vats but it’s a pain and I have to believe there is room for a manufacturer to perfect this and do well selling replacements. Perhaps a standardized vat size across a few manufacturers would help. Some just use (rip off) form labs cat design… A plus if you are ok just buying them from form labs. They are expensive though.
If you want to collaborate, I’ll release all the files- we will open source them but I tend to be slow if not pushed. Welcoming all the cooks into the kitchen doesn’t always produce a better recipe… But you are a good cook.
We use linear rail which I have a ridiculously cheap source for - direct manufacturer in China. We use 2 Nema 17 (one for blocking the light in the chamber). We also use our printrboard because we can make them cheaper than an arduino+ shield. Off the shelf open source software… Forget the name of it. Made Solid resin.
Due to the other kits out there not really taking off yet, I think assembled is the way to go for us, but I love those die-hard builders out there and have a hard time saying no to like-minded enthusiasts. Total cost to me will be around $800, so assembled or kit would be shy of $1499. Around 4x3x6" build volume - give or take. … Haven’t decided on the z… It’s almost free to go to 8".
Let me know if you want to share knowledge. I might even have a dlp knocking around that I could let go.
The downside is we are two designers with strong opinions that may want to take different approaches ;). Often, we all go our separate ways and wait for the public to cite with their dollars. Frankly, I may be at an advantage since I have a strong r&d budget and lots of equipment to manufacture. Chances are I can manufacture cheaper than most and it makes it hard to compete with me in the market. I hate that this fact squeezes out some small startups, but hopefully publishing the plans opens doors for clever folk who are welcome to jump past me with their own innovation.
Regardless, my offer to share stands. I’d love to see your take on a resin printer.
Brook
Actually, I’ve been thinking that the thing to do for a controller for a DLP machine is to just make a daughterboard for a Raspberry Pi to carry a single pololu-style driver (or equivalent in a single board) for Z. Running the printer as a second monitor on a non-dedicated computer is just asking for trouble, but a Pi can drive the HDMI, and without the need for the coordination and high speeds and accelerations you have in the 4-axis machines we’re used to, a Raspberry Pi’s GPIO should be fine for running the Z axis without handing-off the real-time functions to a microcontroller. All of the current implementations I’m aware of have several layers of complexity that serve no purpose other than creating confusion and points of failure.
Files to print could be transferred over ethernet/wifi or supplied from a flash drive, and the reduced complexity of slicing for DLP means that supplying files in STL format should be fine as long as they are already oriented and scaled with any necessary supports. If we have a good automatic support algorithm (that we can trust without inspection), the Pi could probably run that too. A small display connected to the Pi’s SPI (Adafruit has a bunch that would fit the bill) should provide plenty of interface in this configuration, so that the HDMI never has to be used for anything but curing resin, and a web interface could be used for additional configuration and manual control if necessary.
I don’t imagine such a system (aside from the automatic support stuff) would be difficult to implement, but for now, I just have a bunch of mechanical ideas for a machine that I want to try out for a minimalistic machine (which would look like an attachment for the projector, rather than vice versa).
That’s pretty plausible. There isn’t a lot of fancy timing or complex calculation, but you do need something that has a display and isn’t subject to the occasionally interesting behavior of multi-head desktops. Ideally throw a servo control on there too for a shutter, rather than the stepper (for the cost of level-shifting you can just plonk on a a Tiny84 or something, btw).
I’ve got a 10-year-old VIA board and a Teensy doing just that, and it’s really nice to not have to worry about things going all wonky if my desktop machine reboots. (The teensy does all the work and tells the VIA what to do by HID, but that’s just stupid if you have purpose-built hardware.)
@Whosa_whatsis Octoprint based? Because there is a OctoLaser Fork from @Mr_Beam afaik.
@Brook_Drumm you need to supply black labcoats with the Printrbot logo on em for all your crazy RnD 
@Nils_Hitze there might be some networking and interface code that would be worth borrowing, but for a machine that doesn’t run gcode, doesn’t connect to a microcontroller, doesn’t have heaters, etc., it seems like starting from octoprint would be more work than starting from scratch. Mr Beam is still a multi-axis machine, this is not.
@Whosa_whatsis true
I love the Pi idea. oh, if I could write code! I made an led turn on and off tonight so I can’t help there. I know tons of awesome programmers but the good ones are always busy!
I like the octoprint idea for a few reasons…
We wouldn’t have to rewrite all the web server stuff, wifi stuff, UI stuff, or figure out how to integrate Slic3r (it slices for dlp)… That’s all done. Gina is constantly developing it and is committed to open source. Stripping out unneeded things is sure easier than writing from scratch.
Getting some fast progress and tapping some active Pi folks would be likely.
She also has a plugin system that lends itself to stuff like this
Thoughts?
Brook
I haven’t really looked under the metaphorical hood of octoprint, but it really doesn’t seem like it would be appropriate. Like I said, there are a few things that could be reused, but the core of it is something completely different. What it needs is something closer to the web implementations for Smoothie and the Ormerod. The Pi is, in this case, serving an equivalent function to one of those controller boards, not to a printer host like octoprint is designed to be.
It seems likely that the ideal distro to run such a machine on wouldn’t be Raspbian, but rather Arch or another more stripped-down version of linux. We’re essentially using it as a microcontroller that happens to be able to drive an HDMI display and do USB host and network stuff, not the stripped-down desktop machine that Raspbian is intended to be.
BWT, on the subject of Slic3r integration, I was thinking of implementing OpenGL-based rendering of cross-sections directly to the frame buffer in real time. I’d be surprised if separate slicing software turned out to be necessary.
I’d kinda like to see some kind of slicing option, just from a design point of view. You could probably save a lot of resin (and curing time) and get some interesting visual effects from having internal patterns. And it doesn’t cost you anything in printing time the way it does for FDM. (In addition, for the resins that still have significant shrinkage, non-solid prints might well improve quality.)
Obviously you could do this at the design stage instead, but why?