Swerve Comments

[Team1710]

Just got a basic idea down in cad. Kinda based on Wildstang's mounting on the bottom. I'm going to use V bearings though. Would like some input on the design. I'm a fan of constructive criticism so go ahead. I know its not completely done. Still gotta work out some of the details.

[MattLi]

Consider the following:
1. That trailer's going to be a problem - If you use a swerve drive and drive to the left or right with your robot facing forward, that trailer's not going to align how you want it to.
2. Simplicity - Our team built a simple tank drive and yet we were able to outmaneuver a vast majority of the robots at our first regional (videos at www.adambots.com)
3. Looks good!

[big1boom]

Those CAD's look good, but I personally would want a little big more hand room to fix chain's, change wheel's, and such.

[Team1710]

Yeah this drive isn't intended for this year but rather for next year. I doubt we will be using trailors. Big1boom I agree. There could be some more room. I was making it tight to conserve some weight. Swerves are pretty heavy. I think I will change that though.

[big1boom]

To cut down on weight for our Swerve this year, we used 1/8" plate, and a lot of angle bracket. (but then we gained a lot of weight and used all #35 chain.)

http://www.chiefdelphi.com/forums/sh...ad.php?t=75407

[Team1710]

Yeah we are planning on using 25 Chain next year. I hear it saves a lot of weight. Your swerve looks cool. Just wondering how you managed to run it off of 1 cim though.

[Joe_Widen]

I like to make my swerves with the cims in the wheel caseing and gears only. Try gettin your fingers in there to put in a master link =] But other then that it looks good.

I'm biased I just don't like the whole coaxial idea.

[big1boom]

I tend to lean more toward coaxial swerve for a few reasons

1) full rotation, not limited by electrical.
2) lighter, because of the possibility of less CIM's
3) less CIM's

Coaxial is harder though, because the bevel gears have to be kept perfect.

[Aren_Hill]

One of the easiest things to do to keep a swerve drive easy to work on, is have the sprockets for the module drive on the outside of the module casing.http://www.chiefdelphi.com/media/photos/31077

That is the design we used in 2008, and i'm betting you saw that module in Colorado.


By moving the sprockets to the outside you can make the module much easier to work on, yet keep the overall package quite small.

Anyone can make a huge swerve module thats easy to work on, when you get good you can made tiny ones that are even easier to work on. And when you get really good, you never have to work on it lol.

(08 swerves track record, 1 fix after Chicago, Milwaukee, Colorado, Champs, then finally IRI where i had to spend 5 min and it was back in full form.)

[Al Skierkiewicz]

Looks very nice. Imitation is the sincerest form of flattery. The ring at the bottom that we use takes side loads from turning and robot interaction and transfers them directly to the robot frame so that the top bearing does not bind. We incorporate the CIM motors directly into the module and make the module removable from the bottom of the robot. I think a little research will show that the bevel gears are inefficient and may not be suitable for all designs. When you get it built next year, I would like to see it. Please find me if we are at a regional together.

[Team1710]

I"m going against the idea of integrating the cim motor. By doing so, we would be limiting the number of other motors we could be using on other parts of the robot. Our team has had bad experiences with right angle gear boxes. The only other thing I could think of using would be beveled gears.

Edit:
Also I would like to ask how you can keep the shafts within the modules from moving. Looking through swerves I haven't found a solid idea.

[Al Skierkiewicz]

There is no right angle drive in our crab module. The motor is on the module and the transmission is integral to the module, all gears. Shoulder bearings pressed into the vertical (module) frame members with shoulders on the shafts keep things from walking side to side. We have used two CIM and two FP in the past or four CIM or really long ago, two CIM and two drill motors. We use Globe motors for steering either two Globes for all four wheels or one for front wheels and one for back. The modules pull out with four screws and the power connector.

[Bob Steele]

This is similar to our Skunkswerve this year. One thing you might think about is how you will achieve alignment when you are adjusting your pots...
We steer front and rear the same... our modules have holes drilled in them for alignmnet rods to be placed so all four wheels can be put exactly in alignment with each other so we can align our pots for control...

it is a simple thing to put the holes in during the manufacturing process... much harder to do it after the fact...

thanks to our design team... we put them in early... along with a hole for a chain tensioner... (but we didn't need it..)

Our drive is really a hybrid between crab and swerve... similar to what Al was talking about...

[Team1710]

Originally Posted by Bob Steele

This is similar to our Skunkswerve this year. One thing you might think about is how you will achieve alignment when you are adjusting your pots...
We steer front and rear the same... our modules have holes drilled in them for alignmnet rods to be placed so all four wheels can be put exactly in alignment with each other so we can align our pots for control...

it is a simple thing to put the holes in during the manufacturing process... much harder to do it after the fact...

thanks to our design team... we put them in early... along with a hole for a chain tensioner... (but we didn't need it..)

Our drive is really a hybrid between crab and swerve... similar to what Al was talking about...

What do you mean by alignment rods?

[Bob Steele]

When we align the potentiometers that we use for the control algorithm we need to make sure all wheels are pointing exactly in the same direction. We do this by putting steel rods through holes in the swerve modules. These holes are drilled perpendicular to the module in two directions. We put 3 rods in and all the modules are lined up squarely with each other....all pointing the same direction... then we set the pot values to the same for each of them.

After this alignment we now have the same pot value for each wheel so that when the robot interface gets this feedback from the wheel it can keep each wheel facing in the same direction..

We also can, therefore use the calibration of the pots to change all wheels to face a different direction... or each wheel to face a different direction.

The alignment allows us to have an initial frame of reference for all the wheels..

I hope this helps you see what we are doing...