Swerve Drive and Ball Bearing Questions!

[Shelabot]

This year in Atlanta, I attanded the Swerve Drive Seminar in one of the conference rooms... but cannot remember for the life of me how the actual module was kept in place. I have been looking into ball bearings, and allot of it is frankly over my head. Though after a few hours of researching, I have come to understand that the best way to keep a swerve module constrained to it's axis with no vertical or horizontal movement, are two Angular Contact Bearings, fixed at opposite directions of their... what's the term... perhaps ball angle? I was looking into Radial Contact Bearings, and they seem good too, but they seem to be weaker than the above combination.

Another question in my mind, is how the ball bearings are attatched to the frame and the module. Are they welded on (I have to doubt ) or are they pinned somehow?

If you or your team has any experience or knowledge in Swerve Drives, I would appreciate being thrown a bone here!


( Oi, did I write all that? Two hours ago that would have been complete gibberish to me!)

[Dave McLaughlin]

Are you considering a coaxial or a non-coaxial design?

[Joe_Widen]

When make a crab, you should aim to have two points that support the module, rather far apart near each end. The modules don't normally spin fast enough to warrant a bearing being a necessity. When I had dreams of making a crab, I've always thought of using bushings as the point for the top contant and some sort of teflon pieces with groves for the bottom of the module.

Heres something I toyed with for a while.

http://www.chiefdelphi.com/media/photos/25250

[BornaE]

Took us a while to figure out how to do our 2 wheel swerve this year, after lots of research, here is the proved solution for this.
http://static.howstuffworks.com/gif/...e-bearings.gif
you can get them here
http://www.nashbar.com/bikes/SubCate..._200276_200304

never had a problem with them

[Collin Fultz]

We probably attended the same conference, a very good one by team 118. I'll assume you're talking about doing a coaxial design.

The bearings they use are from Silverthin, part number SAA17. I would think you'd want 4-point contact (X-style). You use the dimensions listed as L1 and L2 to create lips for the bearing to react against.

I have wondered though, if bearings are required for the rotation between the wheel pod and the frame.

[Al Skierkiewicz]

Wildstang uses a bronze bearing at the top of the module. The upper module attachment is machined to fit this bearing. The bearing has a flange as well so that the weight of the robot rests on the flange (thrust bearing) while the remainder of the bearing provides a surface for lateral loads. The bottom of the module is a circle of polished aluminum resting inside a delrin ring, resting inside a fixed aluminum plate attached to the bottom of the robot. The delrin is machined into a "Z" shape so that it self aligns and reguires no additional parts to keep it in place during use. This provides lateral support when the robot is turning or is hit from the side during robot to robot contact.

[IKE]

If it is a high traction game (typically they are), do not underestimate the bending forces going into each module.

We used nylon bushings for the top bearing support, and then made a round delrin plate as the lower support and lateral load reaction. The low traction this year did not require the support as much, but this is a design that should work well with high traction too.

One of the most important things to figure out is what tools you have to make your pods. We went with a Wildstang esque sheet metal fabrication because it is easiest for our kids to make. We also did a sheet metal frame. I am not saying that is the way to go for everyone, but that is what worked for us.

[DonRotolo]

Bearings are often held in place with light friction fits. Some bearings have flanges to hold them in a hole. You can also cover the sides of a bearing with something to hold it down, like a screw with a washer, the edge of the washer holding the edge of the bearing down. For the inner race, you can keep it from sliding along a shaft with a shaft collar, held by setscrews.

Ball bearing races are hardened and cannot easily be machined.

You never want to have a situation where the shaft is rotating but the bearing isn't, or the bearing is rotating in the bigger housing hole. Thus the light friction fit. The idea is that a little friction will hold the bearing in place, and the very low friction of the bearing itself won't put much load on the frictional parts.

If you are using angular bearings, you are expecting some side load on the bearing. You then seen something more than a screw with a washer or a friction fit to hold it in place. If the load is in one direction only, a flanged bearing is great; otherwise you machine a hole that has a lip at the bottom, press the bearing into it, and use a cover (with a hole for the shaft!) to hold it in.

Try this link, read the articles

[daltore]

ausTIN CANs just used cheap little 5" Lazy Susans (something in the range of $5, I think). Due to the low friction this year, we only used one on each module, and we did end up with a little bit of jamming, but for our summer project we're planning to do something very similar, but with 2 Lazy Susans, one on top and one on bottom. Lazy Susans are very nice because they can support a surprising amount of load, and have large plates with mounting holes on them. They require a bit of setup maintenance (bending the plates a bit to align the bearings better and crimping the inner flange that holds the two halves together tighter so it doesn't wobble), but if you have the side loads distributed between two of them, you shouldn't have to align them ever again. And for that price, if you screw one up, it's not a big deal, you don't have to machine another one.