Wednesday, August 24, 2011

Delta Robot 3D Printer. Sounds cool doesn't it?


Alright, I've been meaning to post an update for a while now, I was just hoping to have finished the rebuild by now... Anyway, I have been working on a UV/Visible cure resin 3D printer using a 405nm laser. I started off thinking if I could mix a Delta Robot and a resin printer in one it would be ultra double awesome. While it was pretty darn cool and fun to build, it just didn't work great.

I first started with a dynamixel servo based delta robot design thinking they should have a decent resolution.



Then after I built it, I found that was not the case, at least for 3D printing purposes. Also the controller started resetting every couple of moves so I probably fried it somehow.

At that point I did some more research and found some people working on a delta robot for use as a reprap printer. Following their design I used steppers and gears. The gear ratio was chosen mainly for what would be easy to cut on the taig mill. All worked out on paper it looked like it would have a decent printing resolution. My brother drew it up and we got to work cutting it.



We made the Z axis pretty much out of scraps and it didn't work the best. (linear servo... what was I thinking?) It was good enough to test the theory though.


After assembly and mangling the reprap firmware into a usable shape for a delta robot we were ready to go.



Yeah, it printed an awesome lump! It was supposed to be a small hexagon.




So, what went wrong? First off I think that shooting a laser at a shiny aluminum plate is going to make the laser reflect all over... Yup, also I think dye needs to be added to the resin to help with the reflections and diffraction. The delta robot worked alright but I just didn't feel it would make the best quality prints. There was a little too much backlash and the inverse kinematics is not my thing.

So now we have redesigned, again, but this time back to the tried and true (but boring) Cartesian robot. Here is the X and Y axis, still working on the Z axis redesign but hopefully soon we will have more acrylic lumps cool 3D stuff printed.


8 comments:

  1. If you have access to a rather powerful IR laser ( a laser lab that has a femtosecond laser) you can print using the 2-photon effect ( effectively only curing material in the focal point of the laser( 1 um sphere )...

    I'm working on a more economical way of doing this, as not many poeple can access a femtosecond module...

    This is very cool.. :D

    congrats

    ReplyDelete
  2. Seems to me that you haven't optimized your process parameters, it looks like you are applying too much laser for too long so diffraction is an issue. In stereolithography the idea is to apply just enough laser to cure only a small portion on the top. So increase your speed or decrease your laser power.


    Though using a cartesian or delta robot is a bit overkill if you want to do traditional stereolithography, one could just use galvanometers or a DLP projector. But since you have already have 4 axis manipulator you might consider adding an extra axis and doing layerless stereolithography where you move an optical fiber hooked up to a UV source through the photopolymer like a CNC machine un-milling something.

    See more here:
    www-bcf.usc.edu/~yongchen/Research/CNCAccu_Web.pdf

    Nice work on the blob, blobs are the first step in any additive manufacturing process!

    ReplyDelete
  3. Have you thought of using 2 lasers. I don't know if it is practical with the curing characteristics of the resin, but if you are curing resin at the intersection of the 2 beams, you no longer need to worry about the exact focal point of a single beam.

    ReplyDelete
  4. @Gene Hacker

    Thanks for the info, the usc.edu paper is very cool, I hadn't seen that before.

    @Donnie Darko

    I don't think that would help too much with this resin.
    I have thought of adding a second laser with a wider beam focus as a second tool for filling in layers. Then use the small focus laser for detail work.

    ReplyDelete
  5. Great work, RobH. Successive approximation will integrate to ultimate coolness.

    ReplyDelete
  6. Hello RobH, what is the highest resolution you get??

    ReplyDelete