pic: Swerve 4


  • 6.82 lbs not including mounting hardware
  • 1 CIM @ 5.25:1 reduction (18:42 belt, 24:54 spur gear)
  • 14.40 ft/s free speed, 11.66 ft/s adjusted
  • AM9015 @ 42:1 (10:1 VP, 20:84 spur gear) for ~3rps loaded
  • Major bearing is McMaster Turntable 6031K16
  • VP encoder for rotation, S4T optical encoder for speed (not pictured, to be mounted over CIM via Lexan and surgical tubing)

CAD can be found at https://workbench.grabcad.com/workbench/projects/gcaFW2UsMko62pkyqWgTqwQMsoTccy8VDRyF8lsWVke50S#/space/gcDuRMGRReXyDbsOb86ajMXL0dsFJBqbAJGsXLbWx4BbfX

I would really appreciate any comments, concerns, and criticisms you may have.

Interesting module you got there. I like the use of a lazy susan to take up thrust loads from the module; I don’t think I’ve seen that on here before.

Some questions/comments for you:

  1. Why the use of #6 screws to retain your dead axle? That seems pretty weak compared to using a 1/4"-20 or a #10 screw, especially because you can use unmodified Vex tube axle for the former.

  2. I like how you’ve designed it to be easy to make without much lathe work, with hex shaft everywhere.

  3. It looks like you have only a single gear reduction to the small gear of the module rotation. By my calculations, that means you can only have a maximum ratio of 42:1 to rotate the module, which is a little too low (~370rpm from an RS-550 motor). You may need to add another stage to your VP.

  4. The use of 1/8" plate to support your module is very thin, thinner than any other swerve I have seen. Consider upgrading to sticking it into a 2x1 or just 3/16" plate. You’re lightening very aggressively for 1/8" plate.
    As a general note, keep in mind that when losing weight, it’s usually better to change your overall design rather than focus on lightening patterns. Patterns get it down maybe 1/4lb to 3/8lb on a 5.5lb swerve IME, and cannot lose significant amounts of weight.

  5. You free speed looks to be around 14.4fps, which is solid.

  6. I like the belted first stage. If you plan to make this IRL, I would add some kind of a tensioner just in case it doesn’t turn out 100%.

  7. Why do you have a bearing at the top of your module? It doesn’t seem to do anything.

  8. Why are you riveting the lazy susan on instead of using screws? Just curious, may or may not be an issue.

  9. You shouldn’t rely on the lazy susan to align your bevel gears. It will help a lot if you stick 1 or 2 bearings into the large turning gear to constrain that shaft. Right now, because the large turning gear is constrained only by the lazy susan, it can easily slosh around and screw up your bevel gear alignment.

  10. If you mount the S4 over your CIM via surgical tubing, how will you keep the hex adapter from sliding upwards?

Apart from that, it looks nice. Solid lower body. Good lightening on the large gear. Not too much to make from scratch. Clamps onto 1x1.

To respond to your questions:

Thanks, I haven’t seen it eather on here. I saw it on McMaster and thought “hmm…I wonder if I can make a swerve around this.”

The wheel shaft is 3/8" round. I was hesitant to drill a large bore hole in it, but I could probably step it up to #8.

You’re right on the 42:1. I think I could theoretically step it up to 60:1 by using a 14t 3/8" Hex gear instead of the 20t 1/2" Hex gear I’m using now. According to the JVN calc, an AM9015 through a 42:1 reduction pushing 1/4 of the robot weight at a radius of 1" results in 1140 deg/s at 23A. I thought that would be enough, but if you think it isn’t I can look into increasing it.

I wasn’t sure what was a good thickness plate to use. My next iteration (if I iterate again) will probably use a less-pocketed 3/16" plate.

Probably a good idea. I don’t actually plan on building it because I aged out last year, but I will look into adding a tensioner to the next iteration.

Right now, the two bearings retaining the vertical shaft are on the bottom and top plate. If I move that top bearing to the 84t gear, I will have to shift the vertical miter gear down 1/2" (because right now the miter gear hub fits inside the 84t gear), thereby increasing the module height by 1/2". It might be worth it though to get rid of the top plate and to better define the cocentricity of the vertical shaft.

I chose to use rivets because it meant that I don’t have to rely on tapping the plates or having nuts hidden inside the module that I would need to get to to take it apart. This way, all I need to take apart the module is a clear shot to drill out the rivet. If there is a good reason to use screws instead of rivets, I wouldn’t be strictly opposed to it.

I was planning on using CIM retaining rings above and below the hex adapter to keep it in place. The surgical tubing would just slide over it.

Thanks for all the comments and suggestions.

P.S. - I forgot to put the custom parts list in the original post:

  • (6) waterjet cut sheet aluminum plates
  • (1) waterjet cut gears
  • (3) center bored shafts for retaining screws
  • (1) bent piece of lexan - not modeled

Changes in v2:

  • Plate width changed to .19 (from 1/8")
  • Top plate thinned to 1/16"
  • Bearing added in 84t gear to keep vertical shaft concentric with module
  • VP pinion gear changed to 14t 3/8" hex from 20t 1/2" hex
  • Drive encoder changed to E5T to fit right over the 8mm CIM shaft
  • Small pocket added on module side plate

New weight is 7.06lbs per module. CAD updated on GrabCAD here.

In making changes I realized why I needed rivets and the top bearing. The top bearing is necessary because otherwise the shaft is only constrained by one bearing in the big plate and one in the 84t gear, which can move radially and mis-align the bevel gears. By adding the top bearing, the vertical shaft is now constrained by two bearings relative to the frame and the bearing in the 84t gear will keep the module concentric with the vertical shaft. The rivets are necessary instead of gears because there is not enough room for two screw heads (and certainly not two nuts) to pass be each other between the plates of the lazy susan. The rivet heads, which are shorter, have enough clearance.

That bearing won’t do much if it’s in 16th plate imo. 1/8th might be safer. Looks much better than before, although I would still test to see if it holds up before a seasonal use.

On grabcad I only see the .sldasm, is there another link to download all the .sldprt files?

You should be able to see everything through the built-in GrabCAD 3D model viewing plugin. For good measure though I just added a Pack-and-Go and STEP versions so you can download and view in your own program.