Swerve and Steer Module

[ArthurA]

So my team is considering buying four Andymark Swerve and Steer modules modules so we can develop a swerve drive platform that can be adadapted to different games due to a modular design. Does anybody have any experience with this platform? If so, my main questions are:
1. Is it any good as a swerve drive?
2. Do parts break easily?
3. Are parts easy to replace? If one breaks, do you have to buy a new one?
4. Any other infro you have on this, or videos of it in action?

Thanks in advance

[Peyton Yeung]

Originally Posted by ArthurA

So my team is considering buying four Andymark Swerve and Steer modules modules so we can develop a swerve drive platform that can be adapted to different games due to a modular design. Does anybody have any experience with this platform? If so, my main questions are:
1. Is it any good as a swerve drive?
2. Do parts break easily?
3. Are parts easy to replace? If one breaks, do you have to buy a new one?
4. Any other info you have on this, or videos of it in action?

Thanks in advance

1. I like the fact that the module is all in one. I bought 3 for testing this summer and I quite enjoy them.
2. I only had 1 part break break on me and that was due to my assembly error with the arbor press and a bearing.
3. Many of the components of the module can be purchased individually. For example I broke the housing in the press and was able to buy a new one for pretty cheap.
4. A couple of teams found success this year with a swerve and steer platform.

[InFlight]

The one downside is the absolute encoder is not 1:1 on the steering module. There are several ways to deal with this, a quick search here should yield some potentional solutions to this issue.

[Knufire]

Originally Posted by InFlight

The one downside is the absolute encoder is not 1:1 on the steering module. There are several ways to deal with this, a quick search here should yield some potentional solutions to this issue.

There is an alignment feature; a hole in the top plate that is cocentric with a hole in the top plate of the rotating part of the module with the thrust washer. This allows you to stick an allen wrench through both for an easy way to align the wheels in the same position. Making sure they're aligned whenever you reboot the robot should let you use incremental quadrature encoders without much of an issue.

[bmammen]

Originally Posted by InFlight

The one downside is the absolute encoder is not 1:1 on the steering module. There are several ways to deal with this, a quick search here should yield some potentional solutions to this issue.

AndyMark mentioned they would have a fix for this sometime in September, but haven't heard if they're on track with that or not.

[Peyton Yeung]

Originally Posted by InFlight

The one downside is the absolute encoder is not 1:1 on the steering module. There are several ways to deal with this, a quick search here should yield some potentional solutions to this issue.

I made a small flat plate from aluminum (2" x 1.5") to mount an MA3 encoder and a 20t gear to make the steering 1:1.

[bmammen]

Originally Posted by Peyton Yeung

I made a small flat plate from aluminum (2" x 1.5") to mount an MA3 encoder and a 20t gear to make the steering 1:1.

Do you happen to have any pictures of the plate with the encoder/gear mounted to it? We're still trying to figure out how we want to handle the absolute encoders.

[Peyton Yeung]

Originally Posted by bmammen

Do you happen to have any pictures of the plate with the encoder/gear mounted to it? We're still trying to figure out how we want to handle the absolute encoders.

Here are some pictures of the setup.

The encoder has a 1/4" output so I took a 3/8" hex shaft, drilled a 1/4" hole in one end, slotted the hole with a band saw, and slid a shaft collar over the slot and voila. The encoder now has a 3/8" hex shaft which I coupled with a 3/8" hex gear with the same number of teeth so the feedback is 1:1. The whole plate has just 5 holes in it.

[Mark McLeod]

We also added a 1:1 encoder gearing, but 3d printed the gears. We were looking for a smaller footprint.

[sraque]

Thanks for mentioning us in your post Peyton! Yes, we had quite a bit of success with these modules this year. We also used the US Digital MA3 absolute encoders for wheel direction. The mechanical solution mentioned earlier is fine if you are OK with always setting up in the same orientation and don't care about putting one more task on the drive team. We chose not to do that, and our different auto routines ran best with different initial wheel orientations.

My only observation on these modules is that after a bit, the gears begin to chatter. The shaft with bevel gear can flex away from its mating gear on the wheel and result in slipped gears. We still did well with it. It also allowed us to get going with swerve very quickly. But I must also shout out to Apple Pi. Without their Labview swerve code, we would have had a really rough time.

[page2067]

Originally Posted by sraque

Thanks for mentioning us in your post Peyton! Yes, we had quite a bit of success with these modules this year. We also used the US Digital MA3 absolute encoders for wheel direction. The mechanical solution mentioned earlier is fine if you are OK with always setting up in the same orientation and don't care about putting one more task on the drive team. We chose not to do that, and our different auto routines ran best with different initial wheel orientations.

My only observation on these modules is that after a bit, the gears begin to chatter. The shaft with bevel gear can flex away from its mating gear on the wheel and result in slipped gears. We still did well with it. It also allowed us to get going with swerve very quickly. But I must also shout out to Apple Pi. Without their Labview swerve code, we would have had a really rough time.

We were psyched to hear that a team that did as well as you guys used our code as a basis.
Congratulations on a great year!
Our programming team will be posting our past year swerve update later this fall. Our code and some powerpoint that describes the basis for it are on our resource tab of our website.

We appreciate hearing from teams that find it useful.

[bmammen]

Originally Posted by Peyton Yeung

Here are some pictures of the setup.

The encoder has a 1/4" output so I took a 3/8" hex shaft, drilled a 1/4" hole in one end, slotted the hole with a band saw, and slid a shaft collar over the slot and voila. The encoder now has a 3/8" hex shaft which I coupled with a 3/8" hex gear with the same number of teeth so the feedback is 1:1. The whole plate has just 5 holes in it.

That may be the simplest solution I've seen so far, thank you so much for the information and pictures! We would also like to give a shout out to Apple PI for the great LabView code they created for Swerve, our programmers were able to use it for the base code on our off season bot.

[bmammen]

Originally Posted by page2067

We were psyched to hear that a team that did as well as you guys used our code as a basis.
Congratulations on a great year!
Our programming team will be posting our past year swerve update later this fall. Our code and some powerpoint that describes the basis for it are on our resource tab of our website.

We appreciate hearing from teams that find it useful.

It was very useful to our team! We look forward to checking out the update code when you get it posted as well. Thanks

[page2067]

Originally Posted by Mark McLeod

We also added a 1:1 encoder gearing, but 3d printed the gears. We were looking for a smaller footprint.

Mark:
Curious - how much backlash or wiggle does the steering have before the steering encoder feels it?

[Mark McLeod]

Originally Posted by page2067

Mark:
Curious - how much backlash or wiggle does the steering have before the steering encoder feels it?

There is nearly none between the steering motor gear through the plastic encoder gearing. There is some normal backlash between the metal-to-metal steering gear and the wheel module gear. I should mention that these swerve units were constructed based on the AndyMark design, but were not purchased units, so play in through the mechanism may be different than in the commercial units.

I'd need to test again for more exact measurement. I don't remember the original numbers, but they aren't anything that affects the operation.