I decided to try to design a gearbox housing that could be 3D printed for a two speed gearbox that I designed a while ago. I did my best to design it in a way that the material properties should not have to be superb, but I’m not an expert in 3D printing, and I’m looking for feedback on the design. Do you think it could be adequately printed on a prusa i3 mk2? what material and print settings would you recommend? On a chain in tube WCD with 3.5’’ colsons, do you think it could survive a season?
Please take a look at the CAD before providing specific mechanical critiques.
Here is a link to the CAD. Native Inventor and STEP files are in the folder.
I won’t comment on your design, but I may be able to help with material selection. The Mk2 has a maximum printing temperature of 300c, so your options are pretty open. I’d recommend Nylon and Polycarbonate, but both of those can be a pain to print even after getting your settings dialed in and quality spools are still expensive. Nylon also has a nasty problem with fume emission and rapidly absorbing moisture, which can make it unprintable without the right precautions. If you want something that’s inexpensive and easy to work with, I’d go with PET or PETG. It’s the same stuff plastic bottles are made from and it’s become quite well known for fantastic layer adhesion and high strength in general. If you’re willing to spend the time to get PC to work with your machine, it might be a better option.
Obviously, this goes out the window if you have access to something like a Markforged or Stratasys machine.
Thanks for the info. Team 2471 doesn’t have a Markforged or Stratasys, but I’m curious what you would recommend on those. I’m also wondering how ABS would work in this applictaion.
In my imagination (disclaimer: my imagination has no correlation to the real world), the first failure would be the weight of the CIMs bending the entire gearbox, especially as the plastic softens from the heat. I thought of several ways to help this issue:
Use a material with a higher melting temperature to lessen the risk of it softening (so, not PLA).
Printing extra bits that directly support the bottom of the CIM
do what 5012 did earlier and print with the “grain” in a good direction (tbh I’m not really sure what that means or what that does but I remembered it was a thing).
Putting more vertical distance between the mounting bolts.
Lightly press metal spacers into the gearbox’s mounting holes and bolt that tightly to the drivetrain.
Btw, assuming this is for a WCD setup, where do you plan to place the chain?
The Markforged machines are designed to handle exotic materials like PC and Nylon, and the higher end machines can lay a continuous strand of CF or similar material into each layer.
Stratasys has a massive material catalog, but one of their Ultem variations would work well for your application.
Besides being a pain to print, ABS has tolerance and warping issues. Complex parts shrink in interesting ways and could require several iterations of the model to be printed before everything works. It’s also not particularly strong compared to even the mid-range materials available now.
All in all, this looks like a pretty great design for a 3d printed gearbox, though you might be able to make it a little bit better.
First off, what’s driving the spacing on the four main mounting bolts? There’s less than a half inch of vertical distance between those bolts, and any fraction of inch there will make the whole assembly considerably stiffer.
Secondly, is there any gap between the halves of the housing? Getting the tight fit you want can be a little tricky without a lot of familiarity with your printer or a test print or two. Orrrr, you can possibly cheat and print in a groove for an o-ring; that’d let you get away with a less exacting tolerances and still have a grease-tight gearbox.
Thirdly, I’m not a huge fan of the pneumatic mounting. I’ve used a very similar tactic with printing drop-in slots for hex nuts. It’s one of those solutions that gives you a nasty feeling in the pit of your stomach, but sometimes it’s the best solution you’ve got. But I think that you could use heat-set threaded inserts here with success. And if you’re not comfortable with that, you might be able to use hex head bolts so that you have something a little easier to assemble and keep together.
Let me know if I missed anything or if any of this advice is absolutely unfounded!
For 3D printing, there’s a few major “families” of technology:
FDM/FFF:Fused Deposition Modeling or Fused Filament Fabrication is the most common, “hobby” level method of additive manufacturing. Imagine taking a hot glue gun, and putting it on a CNC gantry to control X/Y/Z. Now you can build up a part by depositing plastic in a “layered” fashion. Imagine taking your 3D file, and “slicing” it into layers, then pathing the gantry to deposit plastic in a material saving manner, so each layer may not be 100% solid. The “grain” means the boundary between each layer. FDM/FFF tends to have weakness in interlayer adhesion. Orienting a part so that the layer lines aren’t planar with any major forces can yield a strong component, but printing in these nonplanar orientation requires nicer supports, like the chemically soluble ones available on the Stratasys machines.
SLS:Selective Laser Sintering isn’t as common. Imagine the same method of making a part layer by layer, but instead lay a thin layer of powder down, then sinter it with a laser to make each layer. The upside is there’s generally very little interlayer adhesion issues due to the really high temperature. Chances are good you won’t run into these parts in FRC, considering the cost of materials is so high (~$5k USD/20kg). Another positive factor of SLS is the features that can be printed, such as living hinges, snaps, and nearly final form nylon components, to include high temp and impact resistant nylon.
SLA:Stereo-lithographyis more common than SLS (Thanks FormLabs!). The basic mechanic for this kind of additive manufacturing is that you take a bath of resin (photopolymer resin), and cure a layer at a time using a UV laser at a specific wavelength. The perk of using this method is that you can cure extremely thin layers (0.025mm/layer), which tends to be below the layer differentiation that your skin can feel, and sometimes below what your eye can see. SLA also has little to no issues with interlayer adhesion, due to the long-chain polymer structure of the photopolymers used. Upsides of SLA is the wide range of materials (anything from 80A rubber to nearly ballistic polycarbonate). Downsides of SLA is the higher cost.
Basically, there’s lots of neat types of Additvie MFG out there. Layers are generally an issue to be considered when designing for AM, but they can be mitigated.
Note: I chose not to go into DMLS or the newer metal printing methods (see the Markforged Metal X, it uses a sacraficial plastic binder to hold the ‘green’ part together before final baking/curing) due to the massive cost. Chances are good FRC won’t encounter them.
In response to OP, I’ll offer a suggestion: You’re designing for additive manufacturing here. You’re no longer limited by traditional machining rules, like internal radii, external features, fixturing, etc. Go a little more wild with the design: Add ribs, use lofts, add organic curvature, and get really familiar with mesh editing tools (Fusion360). You can make geometry that’s impossible to machine and will outperform traditional FRC components if you’re willing to get creative.’
