Pages

▼

Thursday, January 27, 2011

Improved Huxley design


I am working on an improved Huxley design.  This is slower than I'd like, owing to lack of time, but it's coming together.

A few key points:

  1. It's all being done in OpenSCAD - the files are in the RepRap svn repository under huxley.
  2. It uses printed sliding bearings that are screw adjustable, so they always fit exactly.  Further, you can compensate for wear.
  3. Other adjustments are done with shims.  Or small rectangles of paper...
  4. Rhys has found a very high torque NEMA 11 motor which I hope to use for my microextruder for it.
  5. I'm trying to design it for ease of assembly.  In particular, I'm trying to get it so that you can put things together in just about any order, and so that you don't have to take four things off to extract and replace a fifth.
The first version will have a moving Y platform, like Mendel.  But MakerBot have taken Ed Sells' idea of building on a conveyor belt and made it work, so I think replacing the Y motion with a belt will follow soon after.

    at

    13 comments:

    1. whosawhatsis?January 27, 2011 11:19 AM

      I built my Huxley with these: http://www.pololu.com/catalog/product/1209

      1400 g•cm (enough to run a full-sized mendel) vs the 1200 for the one you linked. Just sayin'...

      ReplyDelete
    2. UnknownJanuary 27, 2011 7:31 PM

      The pololu motor it is a better but is more big

      ReplyDelete
    3. whosawhatsis?January 27, 2011 7:40 PM

      Sorry, I was half asleep when I read it the first time.

      1200 g•cm is impressive for a NEMA 11. How long is it, and how much does it weigh? If this is for a carriage-mounted extruder and the difference in weight isn't huge, it might make sense to go with a slightly heavier NEMA 14 with a wider mounting point and its center of mass closer to that mounting point.

      ReplyDelete
    4. Adrian BowyerJanuary 28, 2011 1:03 PM

      The NEMA 11 advantage is entirely in its geometry - it is long and thin. It weighs only 5g less than the NEMA 14s I use, but I hope that it will allow a much more compact extruder design. I'll measure it's length when I get home and add that in a comment.

      ReplyDelete
    5. Rhys JonesJanuary 28, 2011 2:55 PM

      Its 50.5mm long. (its on the datasheet....)

      ReplyDelete
    6. whosawhatsis?January 28, 2011 4:52 PM

      Hah, should have thought to check the datasheet, although I've seen datasheets that don't include that type of information.

      Can't wait to see what you've got in mind. NEMA 11 uses smaller mounting crews too, doesn't it?

      -*checks datasheet*-

      Yes, it does, M2.5. That it adds another type of vitamin, and printing and reaming holes smaller than 3mm would be annoying (why am I such a downer today?)…

      This is a geared extruder, right? It probably goes against your "put it together in any order" idea, but it would be cool to use an internal-tooth gear to get a high gear ratio with a small extruder body.

      It should be noted that Makerbot's conveyor belts do not run during printing. The conveyor belt sits on top of a moving x/y stage, and it has a DC gearmotor inside to run the belt after the print is finished. I recently got one of these, but have yet to have a successful print while using the conveyor belt attachment. The tension on the belt is quite high, and it would probably take at least a high-torque NEMA 17 to run it at a decent printing speed, though mounting the rollers on real bearings might make a big difference.

      I just noticed the diagonal cross-brace on the bottom in the picture. This is a good idea, and something I've been trying to work into a design for a while. That added stiffness might allow you to use a single threaded rod at the front and back instead of two, though this will require a different type of mounting for the Y motor.

      Your vertices look like the prettified parametric ones I made for Huxley a while ago, but you might consider using something more like my idea for smaller vertices, found here: http://www.thingiverse.com/thing:5515

      Also, some of the newer designs like the emaker huxley are mounting the X-bars vertically, and I'm thinking that you might be able to mount a NEMA 11 between them for a compact carriage/extruder combination.

      ReplyDelete
    7. D1plo1dJanuary 28, 2011 5:48 PM

      I like these new adjustable sliders, it makes sense that a printed slider would tend to erode faster. Is that ABS your using on the sliders in the picture? How well do they work?

      ReplyDelete
    8. whosawhatsis?January 28, 2011 5:50 PM

      I'm sure it's PLA. PLA is a better material for them, and I've never seen ABS that white.

      ReplyDelete
    9. D1plo1dJanuary 28, 2011 5:54 PM

      Also really looking forward to your automated build platform. Personally I'm building a makerbot ABP into my hl-reprap on top of a traditional belt-driven y carriage. However a rr printed solution for replacing the y-carriage all-together would almost certainly be considerably cheaper.

      ReplyDelete
    10. D1plo1dJanuary 28, 2011 5:55 PM

      whosawhatsis?: I was guessing ABS because I've never seen PLA that white! :)

      Your probably right though, ABS bushings would be a weird throw-back to one of the worse aspects of Darwin.

      ReplyDelete
    11. UnknownJanuary 29, 2011 12:32 PM

      Adrian:
      Can you publish this model in reprap github?.I can't acces to sourceforge

      ReplyDelete
    12. D1plo1dFebruary 02, 2011 6:49 AM

      some research later: liking the nema11s but surely servos are the direction we should be headed ultimately? A rc mini servo like this (http://www.hobbyking.com/hobbycity/store/uh_viewItem.asp?idProduct=8742 ) can output 2300g*cm with a much smaller foot-print and has the added benefit that with a openservo board it can stall-detect for homing.

      ReplyDelete
    13. markFebruary 02, 2011 9:58 AM

      howabout instead of screw adjustable slider bearings, springloaded bearings. So as they wear out the spring pulls them onto the rail from the side.
      -mlagana

      ReplyDelete