Saturday, 26 July 2008

Every little helps

The RepRap Darwin design has 10 diagonal tie bars across the corners of all but the top face of the cube, making it very rigid. These are attached by 20 diagonal tie brackets.

The brackets are held onto the protruding 8mm stubs by M5 set screws through a captive nut. The diagonal bars are then held in place by M8 nuts either side of the bracket.

When fitting them I noticed that the set screws and nuts are not necessary. All the holes I make come out a little undersized and stringy so I clean these out with an 8mm drill. This makes them an interference fit onto the M8 rods. The force exerted by the M8 nuts is enough to squeeze the bracket to make it a tight fit. This is the case when they are made from ABS with 25% fill. Other plastics may be too strong or brittle.

This shortcut saves 20 grub screws and nuts and the time to fit them (inserting the nut can be quite fiddly). Not only that, the bracket can be simplified and made smaller because it does not need space for the nut and grub screw. This optimisation is well worth doing because, although these brackets are quite small, there are 20 of them so they are a significant part of the time taken to replicate.

Here is my smaller design which uses 21% less plastic and reduces the time to make 20 from 11.5 hours to 9 hours on my machine :-

I also used a truncated teardrop for the lateral hole. This relies on the fact that filament can span gaps as well as being able to build out at 45°. The drawing below illustrates that, even for an 8mm hole, the difference between a proper circle, which would require support material, and this truncated shape is very little. It also shows where the full teardrop would extend to.

Here is a picture of it installed alongside the old design: -

I think this is a beneficial mutation that will slightly increase the rate at which Darwins reproduce in the wild. The new DNA can be found here.


  1. Cool

    I have just been making some wooden ones and 20 to file up and put grub screws into is a bit of a chore. (Although worth it)

    Something I feel that could also be beneficial is a center cross piece for the tie bars.

    Fastening them together at the point they cross should add a lot of rigidity.

    Whether it needs it or not I can't say as I have'nt go that far yet with the wooden one.

    Most excellent work as always.

  2. I also felt it might be beneficial to tie them together where they cross. My thought process was: it makes the tie bars stronger against compression by stopping them bow. It doesn't alter the strength against extension. Since any skew force on the frame tries to extend one tie and compress the other I dont think it will add much stiffness.

    I have to say it already plenty stiff enough for FFF and the frame is not the weakest part of the system.

  3. Nophead, your post made me think about an idea I had earlier to replace more structural parts. The easiest place to start would be the diagonals, since the strength against extension would be mostly important (as opposed to compression).

    Be sure to have a look at my forum post.

  4. x,y,z, axis work productivity and area of operation can also be 'expanded' i.m.h.o.
    Think of 3 cubes, on top of each other, ' glued with flexible tape' on only 1 of the now 4 vertical sides.
    This 'print head' can be put under, side and above and be attached on only 1 rod.(middle or other cube attached on a ehhh knob-joint , don't know english word for it.)
    It can work in corners with 2 cubes under and 1 a-side, or 2 a-side and 1 under, as a U vorm from 3 sides at same time for i.e.. a 'floor' etc.

    b vd berg

  5. b,
    I am sorry but I don't understand what you are suggesting. Perhaps a diagram would help.

  6. Nophead,

    I'm from Melbourne Australia, going to be in Leeds UK for a few weeks in Sep. I would very much like to see your work on Reprap, so if you're interested, email me on

  7. What if you slotted (either using a hacksaw or in the printed part itself) from the smooth hole all the way through? Then the compression from the nuts on either side of the allthread would clamp the bracket on the smooth rod, even if you were printing it from stuff that doesn't do the interference fit properly. I don't think it would be harder to set up, either.

  8. Cutting it with a hacksaw would not be a good idea because it is sparse filled so it would leave open edges with the criss cross infill exposed.

    A slot can easily be added if anybody needs it by editing the STL file in AOI.

    It would only become necessary if the hole came out too big for some reason, in which case the hole could just be made smaller instead.

  9. Slot patch is indeed a good idea; we've put it here:

  10. I can see that might be necessary with PLA as holes come out a little bigger and it is harder to squeeze.