Team 1533 Strange Swerve 4.1 CAD Release

This is the module used by Team 1533 Triple Strange for the 2018 Power Up Season!

GrabCAD model

There are two main differences between the module as shown and what was on our robot. The drive motor was changed to a Mini-CIM with a 3:1 Versaplanetary reduction. Also, the wheel shaft was not held in with one bolt but rather two bolts on either end and the hex shaft was tapped.

If you have any specific questions that you would like to keep away from ChiefDelphi posters, feel free to email us:

1 Like

Looks awesome! Thanks for sharing! The swerve looked great at Championships, and I can’t wait to dig deeper into the CAD model! Do you guys by any chance have a CAW for the parts that were purchased off the shelf for each module? Thanks again for sharing!

Here are the COTS parts that we used this year (some of these parts are available from other sources):

Colson Performa (4" x 0.875", 1/2" Hex Bore) 217-4049 Qty 1 $9.99 ea
Flanged Bearing 13.75mm (1/2"ThunderHex ID) x 1.125"OD x 0.313" WD 217-4006 Qty 2 $3.99 ea
Flanged Bearing 0.375" ID x 0.875" OD x 0.280" WD 217-2733 Qty 2 $2.99 ea
64t Aluminum Spur Gear (20 DP, 1/2" Hex Bore) 217-3575 Qty 1 $14.99 ea
32t Aluminum Spur Gear (20 DP, 3/8" Hex Bore) 217-5455 Qty 1 $8.99 ea
30t Aluminum Spur Gear (20 DP, 3/8" Hex Bore) 217-3210 Qty 1 $8.99 ea
15t Steel Bevel Gear (12DP, 3/8" Hex Bore) 217-3344 Qty 2 $14.99 ea

Versa Planetary Gearbox 7:1 ratio + 9:1 ratio 3/8” output shaft Qty 1 $61.96 ea

Mini CIM Motor 217-3371 Qty 1 $24.99 ea
Versa Planetary Gearbox 4:1 ratio, 3/8” output shaft, CIM adapter Qty 1 $50.97 ea

Steering Motor
9015 Motor (am-0912) am-0912 Qty 1 $14.00 ea

Designatronics (SDP/SI):
100T 48DP Acetal - Brass insert A 1T 2-Y48100 Qty 1 $7.98 ea
30T 48DP Acetal - Brass insert A 1T 2-Y48030 Qty 1 $5.18 ea

Mouser Electronics:
Potentiometer TT Electronics 4.5 - 5.5 Volts 858-6127V1A360L.5FS Qty 1 $12.32 ea

Bearing Ball Store:
1/4” Delrin Bearing Balls Pack of 300 Qty 1 $28.00 ea
These Delrin bearing balls are cheaper if you buy them in higher qunatities. We bought a case of 1000 this year which was enough for 10 modules when we added it to what we had left over from last year.

I did not include the hex shaft or Churo shaft stock to the COTS list. I also did not include the nuts and bolts. I assume you can handle that.

Including the motors, this works out to about $290 per module.

A few notes:

  1. The CAD model shows using a 775 motor. We used a miniCIM with a 3:1 geabox for the drive motor this year instead of the 775s. We used 775s last year and we experienced some issues so we switched to the miniCIMs this year. The price list above reflects that switch.
  2. We also added encoders to the drive gearboxes this year, but we had a bunch of those laying around, so we did not have to buy those and I don’t have a receipt. Since these are optional, I left those off the list.
  3. We added powerpole connectors to the motor wires. I did not include the price for those.
  4. you will need a PWM cable to connect the steering encoder (potentiometer) to the RoboRIO. I did not include that cost as the length will be dependent on the robot design.

The rest of the parts shown in the CAD are custom waterjet cut parts. I did not attempt to cost the raw material needed for this.

Let me know if you have any questions.

As seen in the attached image, the CAD indicates that you used what appears to be a Churro as a standoff (blue), which can get tapped to 1/4-20. However, you used a 1/4-20 bolt (red) all the way through the assembly and then into a lock nut (pink) on the other side. What led you to do this instead of just using a bolt on either side and tapping into the churro? Also, how was this accomplished? Did you tap the churro and then thread into the nut in addition (is that even possible?) or did you drill out the churro?
Thank you.

There’s so much I love about the execution of this design! Well done!!

The original CAD had the churro bore opened to 1/4" with a thru bolt, but at build execution, the team decided to simply tap both ends of the churro and then use a 7/8" 1/4-20 on each end. That allowed them to quickly remove what we call the spider (drive motor mount) from top to change or repair. We used blue loc-tite on bolts and had zero issues.

I noticed that the wheel is not centered in the module. Have you had problems with this like in auton navigation?

The team had discussed shortening the shim on the shaft and then adding a small shim between the drive gear and the wheel in order to center the wheel. Honestly, I thought that had happened, but looking at the pictures, it looks like it did not happen. Obviously, it would be very easy to adjust the shims on the axle to center the wheel.

It did not cause any noticeable problems with the way that the robot behaved during auton or teleop. With swerve, our switch auto was basically driving in a straight line diagonally to the switch, so the offset would not affect that. After dropping the pre-loaded cube, the robot did turn and collect additional cubes from the pile and drop them in the switch, and this seemed pretty consistent, so the offset did not appear to affect that. The scale auto involved more turning and navigating. It seemed pretty consistent, so it did not appear to affect that either. I don’t believe we did any deliberate tests where we changed the starting orientation of one wheel by 180 degrees to see if the consistency was affected.

It would be interesting to re-shim the wheels to be centered and then run our autos again and see if the consistency changes. Maybe a good summer project. We can call it Strange Swerve 4.1.1. :cool:

Thank you for posting this, after seeing your swerve in action on Einstein it’s obvious this is a great model for many teams to dissect and learn from.

I have a question. In the 4.1 model the top ring, bottom ring and mounting plate don’t have a raceway for your ball bearings. Is this an oversight or are the exactly the same as in the 4.0 version?

Same as 4.0

For the sake of my sanity I have included a stp file of the module with the bearing grooves in the grabCAD post. I hope that this oversight hasn’t caused any confusion for those who have already looked at the module.

I really like your implementation of the azimuth encoder indexing spur gear. Well done, this will find its way onto my space saving prototypes for sure.

How much does it weigh and can you put a CIM on it

With a miniCIM, it is 6.8 lbs.

A CIM is 0.64 lbs heavier than a miniCIM, so with a CIM it would be 7.44 lbs.

I’ve not done the math for what it would weigh with a 775 instead of the miniCIM.

We had the CIM vs miniCIM power vs weight tradeoff debate a couple of times now for the swerve drive motors, and the miniCIM always wins out. But the answer may depend on the situation. Once you have configured for one, it is relatively easy to switch to the other.

Excellent job, everyone! These modules look GREAT! :slight_smile: