Swerve Module Mounting Bearing

Technical Technical Discussion
1

I am pursuing designing a new coaxial swerve/crab/swivel drive train for my team. If possible, could those who have designed and used similar swerve modules please leave your feedback? I know Craig Hickman has designed several coaxial swerve modules, which seem to use similar mounting types. Also, I know the 118 swerve modules use similar bearings, so I’d particularly appreciate hearing from these two groups; however, many others have similar designs/insights!

Two pictures are attached in the zip folder. Fyi, the two semi-transparent plates are made of 3/16" thick aluminum, but I temporarily made them semi-transparent for the sake of the picture.

All elements are made of aluminum, except for the two pale blue/gray rings, which are Silverthin Bearings (part no. SAA15XLO, SAA 3/16in X 3/16in Thin Section Bearings | SilverThin). The tall central hub would be bolted onto the roof of the pod. The two semi-transparent plates would be bolted around two aluminum box rails. The two bearings would be pressed into these plates. The ring with the holes in its perimeter at the top of the hub would be screwed onto the hub, and removed during installation.

To install, the two bearings would be pressed onto their respective plates, the bottom plate would be slid down the hub from above, bearing-side down; the upper plate would be slid down the hub from above, bearing-side up. After these have been slid on, the upper ring would slide down from above, and be fastened with 4 10 bolts that are tapped into the hub.

The idea is that the upper plate/bearing takes the majority of the weight of the robot under normal functioning, the lower plate/bearing prevents the pod from sliding out from under the robot when it’s lifted, and they together share the radial and moment loads. Obviously, that’s a simplified explanation of the loads… :slight_smile:

I have never been quite clear on how other teams support pods that are only supported from above… Is this similar to others’ designs? Would this be able to handle the radial, thrust, and moment loads? To me, this seems functional; however, I have little experience with swerve modules. Thank you!

Nathan

Screenshots.zip (261 KB)


Screenshots.zip (261 KB)

2

I’ve always liked Wildstang’s mounting system for their modules. So, when we made a swerve last summer, that’s the type we used, more or less. I’ll see if I can describe what we did.

Up top, we had a thrust bearing for the weight of the robot, and a bushing for horizontal loads. To take the module out, all we had to do was remove the monster snap ring that held it all together. And then, at the bottom of the module, we had a round sheet metal ring with delrin rubbing on it. The ring was mounted to the robot frame, and the delrin was mounted to the module. By holding the module laterally in two spots, we didn’t have to worry about moments up top in the bushing, which made things a lot easier. We were originally going to just use a bushing to hold the weight of the robot, but we found a $6 thrust bearing on McMaster that fit our needs, and it was hard to say no after that.

If you find a picture of Wildstang’s module on CD-Media, you’ll see them holding the module in two spots like we did.

EDIT: or you can look at the picture that Akash Rastogi posted…

Good luck!

3

Our coaxial module was entirely supported on top by a big 1" bronze bushing about 1.25" long, and it sat on two thrust bearings that were just 1/16" thick bronze disks.

Not as efficient as bearings, but sure as heck a lot cheaper.

4

Here’s a picture of what Austin’s referring to. Its definitely the cleanest mount I’ve seen.

Adam, are these the same bushings you used on the King Krab?

5

Our coaxial module was entirely supported on top by a big 1" bronze bushing about 1.25" long, and it sat on two thrust bearings that were just 1/16" thick bronze disks.

Not as efficient as bearings, but sure as heck a lot cheaper.

Bushings are more than sufficient for the application.

Silverthin Bearings (part no. SAA15XLO, http://www.silverthin.com/4saa.htm).

Though frc118 did deploy silverthin bearings successfully for many seasons they would be the first to tell you that these bearings are often the weak point in their setup.

Fee free to review our co-axial swerve design or our Wildstang clone. www.team221.com

Good luck!

6

Maybe, maybe not.

We just used bronze bushings for a 1" shaft, all sorts of options on mcmaster. Paid about $7 each iirc.

7

This year 1625 had a 1" long 1.125" ID roller bearing supporting the top of the module, 14$ a pop, got replaced with another bearing due to hard anodize wear but thats a material issue not a bearing issue.

8

Team 1983 used two oilite flange bushings from Mcmaster in addition to one needle roller bearing. A shaft collar was used to constrain the module vertically. The steering sprocket was bolted to this shaft collar to rotate the module.

A cross section of the module can been seen here.

9

This is quite similar to the 1986 module mount. Two inexpensive flanged bronze bushings supporting the pivot stem, captured with a collar. A roller thrust bearing took the load above the steering sprocket. The stem weldment was steel and quite stong, no lower support required. This mount was very simple and robust. We could change out a module in a few minutes. Here’s some pics.

DSCN2703.JPG
DSCN2703.JPG607×873 80.8 KB

DSCN2706.JPG
DSCN2706.JPG766×1093 83.8 KB

DSCN2746.JPG
DSCN2746.JPG858×782 77.1 KB



DSCN2703.JPG
DSCN2703.JPG607×873 80.8 KB

DSCN2706.JPG
DSCN2706.JPG766×1093 83.8 KB

DSCN2746.JPG
DSCN2746.JPG858×782 77.1 KB

10

Just out of curiosity, were you considering doing an infinite rotation coaxial design, or were you planning on having the motor inside the module?

11

I will post pictures of the swerve module I designed as part of my senior design project at Northeastern when I get them off my other hard drive…however here are my thoughts until then…

The silverthin bearings are a really nice solution, however they do have some issues, and they are ridiculously expensive (on order of ~$78 a piece)

I used the most simple/cost effective solution I could and figured I would start adding complexity and price as needed. My approach was to start from the ground up and see what was the simplest solution I could use.

I ended up taking a stock bronze bushing from mcmaster and bore-ing it out, and parting several smaller bushings from it. I pressed these into the plates that mounted the module to frame. Because I turned the bushings myself I was able to get a very precise fit on all of them. This resulted in a perfectly workable solution.

Had that not worked as expected, my plan of attack was to do a large sleeve like Adam described, and then use thrust bearings (w/ ball bearings) to alleviate the thrust load.

My swerve can be seen in action here: http://www.youtube.com/watch?v=mlla1xahkrk

pics to follow…

12

Thanks for the input, everyone! I’ve taken what you guys have been saying into account, and am finishing up an alteration of the what I’ll call “bronze sleeve bearing with thrust bearing” design. I’ll post some pics of that and the pod itself.

To answer your question, Dave, this module would be a true coaxial swerve module… the steering and drive for the module are located off the pod. I plan on having independent steering for each of the four modules, and shared drive for the right side with shared drive for the left side. This allows for any of the swerve drives capabilities… strafing (crabing), spinning in place, and pseudo-ackermann steering… although it sometimes requires rotating a pod more than 180 degrees to switch orientations.

13

Are you going to be supporting the Module at both the top and bottom like the classic 111 swerve, or only at the top like 118 used to do?

14

I plan on supporting the pod exclusively at the top - I have always been willing to take the extra design complications of the 118-esque swerve module to reduce the frame complications and to free up area at the base of the robot.

15

I also have a preference for the top supported modules. I’d advise trying to make the distance between the point of contact of the wheel and the lowest supported point on the outside axle as small as possible to minimize the moments that you will encounter.

16

There are a couple of pics for your reference.

Roy

971swerve2.jpg
971Swerve-module.jpg


971swerve2.jpg
971Swerve-module.jpg

17

The 221 Wildswerve is an excellent example of Bushings only. All my my later designs revolved around only bushings mounted inside of either a block, or two plates. I usually use a 1-2" pivot shave that has the coaxial drive shaft running inside. That pivot shaft sits in two bronze bushings. We built a non-coaxial with the same setup on 114 (here), and that system still runs to this day. In fact, it was strong enough that we’ve had two of the modules on a little person cart for the last two or so years, and it has yet to fail.

As Anthony Lapp said above, Bushings will be more than sufficient for this use. Bearings may seem tempting, but the cost is not worth the advantage for a FIRST system.

I’ll invite you to check out this latest CAD of a crab system (and by latest I mean early 20008…): http://picasaweb.google.com/ironspork/CrabDrive#

The system uses two bushings inside of a milled block that anchors to a bar of 2x1. This is the system we used in 2005 on our crab, and it never broke. Still hasn’t to this day. Keep in mind the chassis I linked isn’t legal anymore due to the constraints of the newer bumper rules. It would take some creativity to use that style of frame legally.

I’ve also done a crab out of punched sheet that was supported by bronze bushings. It wasn’t coaxial, and didn’t allow full rotation… But it was on a production robot that weighs ~200 lbs. If I had the rights to, I would post an image.

Basically, bushings are fine.

Feel free and email me at craig AT hickmanfamily DOT com if you have any other Crab questions.

18

The key difference between bushings and bearings is cost as everyone has said, and then efficiency.

Efficiency might matter for your design, it might not, it depends on what motors you’re using to steer.

We had two globes steering 4 modules on bushings, and it worked fine; with such an overpowered system we actually ADDED friction to help tune the control loops.

If we were trying to do a crab steered by window motors, I’m unsure what we’d do, I imagine bushings would still work depending on your reduction.

19

We did giant needle bearings, largely because we needed to steer 6 modules with 2 window motors, bushings would’ve killed us.

Some people switched their 4wd swerves away from windows on steering because the windows were overheating quickly.

20

228 used large Delrin bushings for both the lower and upper supports on our 4WD independently steered GUSwerve this year.

Since each pod was steered by a single Denso motor, we had some initial issues with them overheating. Once we correctly shimmed the Delrin plates using .010" Delrin shims and set the Jaguars to coast instead of brake, we haven’t really had any issues with the motors overheating.

http://www.team228.org/gallery/125/slideshow/build-season-week-five_1e2a3-1e2a3.jpg](http://www.team228.org/media/pictures/view/6133)

http://www.team228.org/gallery/126/slideshow/img_1a141-8a64e.jpg](http://www.team228.org/media/pictures/view/6182)