pic: Swerve Design

[z_beeblebrox]

loading photo ...

[Chris is me]

Awesome design. Using a bike fork for the module top is very clever - I've been looking for an excuse to use one in FRC for years but we never build anything dealing with that kind of loading. I do have to ask, how are you handing the headset compression?

[MrBasse]

Can we get a view of one of the modules that would show the interior components and how they interface with each other? We've never done swerve before and I would just like to get mind head around the design process a little.

[z_beeblebrox]

Here you are:

[yash101]

Great job, 4183! Is this your 3rd or 4th year in FRC?!?! Keep up the great work!

[Bryce2471]

Cool design, I like how you chose to use stand-offs instead of front and back plates on your lower wheel modules. Although, this may just be how I'm looking at the render. It seems like you could cut down on the foot print size taken up by the drive system.
The first time I designed a swerve module, it took up a 7" by 7" area in the corner of the robot. Upon redesign, I got it down to a 6" by 6" foot print. Then I redesigned again and got the size down to a 3" by 9" box.
I've noticed that each time I make my design smaller, it seams gets stronger and lighter as well. so there's multiple benefits of optimizing for size.
Also, make sure you have accurate and robust encoders on the steering of each module, they are important.

[cmrnpizzo14]

That looks light, how much should it weigh?

[z_beeblebrox]

Originally Posted by cmrnpizzo14

That looks light, how much should it weigh?

Solidworks estimates 36lbs, which seems pretty reasonable.

The top and bottom plates and the module sides are .25" aluminum. Would any pocketing on those be useful?

[theCADguy]

Awesome design! Nice work.

Here is a similar design by 221 Robotic Systems. It's a customizable research platform, and has an interesting setup. Steering is accomplished using one motor to rotate two swerve modules. Another motor rotates the rear swerve modules. It seems as if the modules can be turned more than 360°, as the CIMs driving the wheels are located on the center of the drive train, and connect to the modules using a sprocket and chain.

http://www.team221.com/viewproduct.php?id=128

It seem very similar to your design.

[z_beeblebrox]

Originally Posted by theCADguy

Awesome design! Nice work.

Here is a similar design by 221 Robotic Systems. It's a customizable research platform, and has an interesting setup. Steering is accomplished using one motor to rotate two swerve modules. Another motor rotates the rear swerve modules. It seems as if the modules can be turned more than 360°, as the CIMs driving the wheels are located on the center of the drive train, and connect to the modules using a sprocket and chain.

http://www.team221.com/viewproduct.php?id=128

It seem very similar to your design.

221's module design was one of the things we looked at while designing ours.

[magnets]

Originally Posted by z_beeblebrox

Solidworks estimates 36lbs, which seems pretty reasonable.

The top and bottom plates and the module sides are .25" aluminum. Would any pocketing on those be useful?

If you've got machining resources (you probably do if you're going with swerve), then it's probably worth it to pocket the plates. 0.25" aluminum is really strong, so you can make big pockets.

A few ideas for weight savings

You can shrink the module plate by making the whole thing a big triangle and pocketing the center, like this. Your current design looks like you're making them wider so you can add the little support thing. You shouldn't need one because the lower axle with the wheel on it will make the thing stiff enough.

The big green plates mounted to the frame could also get smaller by making them curved, and you could probably mount the steering motor directly to the frame, with no need to make the green plates go out that far.

The idea for using the bike parts for the bearing is really neat.

[z_beeblebrox]

Based on the suggestions given, I've made some changes to the design. I've removed the spacers from the module and added pockets. I also mounted the steering motor to the frame rails and reduced the size of the gearbox plates to what was actually needed. It's now about 38.5lbs. (For comparison, the 2013 KOP chassis with 4 CIMs weighs about 32.)

I also noticed that the first iteration was several pounds too light; I hadn't counted the correct mass for the the motors and gearboxes.

Edit: I added some pocketing to the gearbox plates and took off another .5lbs.

[yash101]

That is quite lightweight! If you want to get even lighter, I notice that there is some material that won't necessarily add any structural value, so there are many places to drill holes at to reduce weight! Hopefully you use this in 2014 so I get to check it out at the competitions!

[z_beeblebrox]

Originally Posted by yash101

Hopefully you use this in 2014 so I get to check it out at the competitions!

I'm almost certain we won't run a swerve drive in 2014. It's heavy and machining and software intensive for little performance gain over a simple 6WD. This season, we want to build a simple, elegant effective robot quickly and have plenty of time to work on the practice robot, software and driver practice. It would take our team about a week extra to make a swerve drive, a week we can't afford to lose. We might build a prototype the next off-season, though.

[s_forbes]

Originally Posted by z_beeblebrox

The top and bottom plates and the module sides are .25" aluminum. Would any pocketing on those be useful?

Depends entirely on the resources at your disposal and the rest of the design of the robot... If the other mechanisms require every ounce of spare weight you can sacrifice, then it would be useful in those cases (our 2013 robot fell into this category - not a fun situation!). If it's not necessary to cut drivetrain material in order to meet the weight limit, I would leave it solid. Make the manipulator light instead.

Weight down low is a good thing. Weight up high is not.