Skip to main content Link Search Menu Expand Document (external link)

Building Clank(s)

build

Variants

I’ve (accidentally) developed two version of this thing. Both use the same toolchanger, so the notion is that both can share end effectors. However, we sometimes want to machine things that are wide-and-flat (sheet stock) and sometimes things that are more-like-squares (billet parts, additive mfg). So, you’ll want to pick one of these - probably you already have one in mind, if you’ve landed on this page:

Clank-Stretch

Clank-FXY

Clank-CoreXY

Default Volume is 620 x 420 x 30mm (24 x 16.5 x 1")
(z-tools can retract above the build volume)
I suspect this design can be built out to 4x8'
Default volume is 255 x 255 x 255 mm (10" cube)
Max build out to 500 x 500 x 750mm
WIP WIP: default ~ 200 x 180 x 20 mm | made to mill PCBs
Stretch is hardware stable, but needs end effectors. Flying-XY is very nearly hardware stable. The hot newness, we are building and testing these during <a href=https://fab.cba.mit.edu/classes/863.22/ target="new">HTMAA 2022</a>

Click on one of the variants above to access CAD for either variant.

Required Tools

You don’t need much to start building one of these. Here’s a list of everything that is strictly necessary to build either Clank:

  • any FDM 3D printer; the Ender 3 Pro (cheap) and Prusa i3 (better) are great options, costing $300 and $900 respectively
  • a set of metric allen keys
  • tweezers (for one tricky washer install) and needle nose pliers

There are a few things that make this easier - but aren’t required:

  • a deburring tool (to take brims off of printed parts)
  • a pliers-style wrench, I use Knipex 86 03 180 - besides being one of my favourite tools of all time, it’s wonderful when inserting nylon hex nuts into printed parts. there’s a lot of that.
  • a set of M3, M4 and M5 taps

I typically keep a stock of extrusion in 2 meter lengths, then cut it to size. But Misumi is great in that you can specify for delivery of cut-to-length parts; you can even ask them to tap the ends for you. For example the misumi part number HFS5-4060-735-TPW specifies at 40x60 extrusion profile (5-series) that’s cut at 735mm. Their precision is pretty good, I’ve checked (to 0.1mm is reasonable to expect). The TPW here specifies that ends are fully tapped; this is how I spec them in the BOM to save you some time.

If you want to cut your own extrusions, you’ll also need:

  • a miter saw (!) any carpentry saw will do, go slowly. you can get non-ferrous metals blades if you prefer being fancy

That’s about it. Done wisely, my bet is that you could bootstrap yourself a whole tiny shop (4x8’ multipurpose clank, fdm 3d printer, hardware kit, basic tools) for about $1500 all-told, not a bad start.

Printing Parts

There is admittedly a fair amount of printing required to make these machines, maybe ~ 60 hours or so each. Every part has at least one solid planar face that should be oriented towards the printer’s bed; hopefully that’s easy to identify. You’ll notice some Nylock inserts have a chamfered bottom, that chamfer is there to eliminate overhang: its presence is a good indicator of which side should be oriented towards the bed.

I print in PLA, though it (shouldn’t) matter too much. I often use “Tough PLA” which has improved impact strength, but normal PLA will do just as well (and is maybe even stiffer), and ABS should work as well bar the normal warping problems.

If you want a stiff machine, print everything solid. For milling, mass is your friend: if you want to whip around, lighten things up: that’s your call. I normally choose solid since speed is overrated and plastic needs all the stiffness help it can get, but lately I have been just printing enough perimeters (4-5) and ~ 50% infill for speed, ~ 90% infill when I want to maximize stiffness.

Here are the settings I use, some of them are important:

Setting Parameter Notes
Nozzle Diameter 0.4mm Standard
Layer Height 0.2mm Structures are fairly low-res (blocky) so resolution isn’t too important. This increases speed
Combine Infill Every 2 Layers This speeds things up, I would suggest more but the nozzle is only 0.4mm in diameter, so here’s the limit
Number of Perimeters 4-5 Defaults are normally 2-3. This slows things down, but I prefer the integrity.
External Perimeters First Yes, turn this feature on Slic3r lets you specify this in expert mode in the layers and perimeters tab. This helps the printer actually hit the geometry specified in the file; it biases polymer squish to the inside of a part, to the detriment of overhangs (that’s why it’s off by default)

I’ll bet you can have success with almost any print settings, but I thought I would include my setup anyways.

There are some things with GT2 teeth printed in; for those I turn the external perimeter speed way down, like 10mm/s, this improves the fidelity of the profile and makes belts grip better.

W/R/T bed adhesion, do whatever you think is best: I use a “Layerneer” adhesion fluid on a powder-coated Prusa bed, washing it every ~ 2-3 prints. A little prep goes a long way, as they say.

Assembly

My strategy here is to publish very thorough CAD models and then leave order-of-assembly up to the reader. I see that’s a little bit rude, but I think it’s also better than nothing. I hope to link to others’ build logs in the future.

If that has you stuck in the water, for reference check out the clank-lz build log (incl. videos!): the design patterns there, and here, are mostly the same

Build Feedback

I put these machines in a few public repositories on github;

Clank-Stretch on GitHub

Clank-FXY on GitHub

Hotplate on GitHub

Posting an issue there is the best way to get in touch. If I’ve made any blaring omissions, unacceptable design decisions, etc (or if you like this stuff and want to say hi), please do so.

Contributing

Along the same lines, if you like the project and want to contribute, there are a few ways to do so:

  • build one! let me know how it goes. write a build log to answer others’ questions while you answer yours
  • design a new tool! I love the hotplate toolchanger but barely have any stable end effectors for it. check the tools page here and whip up a new design! there’s lots of processes to get after

Table of contents