Team 2517 Swerve Now and future.

[Tyler2517]

This year was my teams first year attempting a swerve drive and was by far the most complex, challenging and rigorous thing we have done.

The rest of our robot may not have turned out as well as it should have. I guess this is living proof of the resource draw on us. Even with that this has been our best year with out robot. We won our first engineering awards this year. We also placed the best we have ever as well.

We did not have the strongest of design and engineering so we knew we could not attempt our own fully custom swerve so we used Team 1640s as our launching ground with a few improvements to decrees machining and decrees weight.

Our biggest improvement was using a in line gear box with the drive belt going around the steering motor. Something that allowed us to make it out of a single piece of 2/3 extrusion drastically increasing decreasing weight and increasing the strength much the same way the bends in sheet metal help it stay strong and ridge.
To save even more weight we milled off 1/8 of the .25 thick extrusion on 3 sides allowing the bottom to remain strong to take the hits and such the this game dished out.

All of the upper modules were manually machined holding 2 thousands on all holes and the boring was held at half a thousandth. All laything work was done manual as well holding 1 thousandth all around.

For our prototyping we had printed everything on a 3D printer allowing us to make sure that the gears and such were in line and belt lengths were right.
http://imgur.com/a/Rx370 - Assortment of image's from the assemble and testing phases.

Our biggest problems and their fixes.
1. Power we sat still for most of the first district event due to constantly browning out the jag's thought can bus. Everytime we would try to start steering motors and the drive motors at the same time the resulting voltage drop would take out our can bus. We had 2 fixes for this one was to buy new battery something we had no done for several years. (running 4ish year old battery that are in horrible disrepair on the field was a bad idea). We also devised a system that would not offset every other module by 50-100 milaseconds allowing them to ramp up their electric fields and overcome stall at different times drastically saving our voltage drops.

2. Encoder slipping. After much post season testing and a bit of disassemble we found that this was due to the press fits that were holding tge coaxle to the pivot top were slipping after awhile this would stop thought so their was no mid season fix. We account this to poor material/poor maching. After post season testing we will be testing them and most lickly new modules will be welded or some other feature.
https://www.youtube.com/watch?v=P_aGnqAmhiw - testing

For programming we used C++ all of our code was custom at least to my knowledge. We did not use field centered control. It was pretty basic but with some driver skill would hold its own. Standard right stick for rotation and left stick for translation.

We did have a really cool feature that would allow us to change our center of rotation on the fly to anywhere with in our robot or out side of our robot allowing us to pull some pretty neat on the fly maneuvering. http://youtu.be/1oVNrp2L1EQ?t=2m26s -best example that i can find on the top of my head.

Even with our limited successes we are moving forward with our development of this drive system. We are mainly attempting to improve our mechanical efficacy and our weight. To do so we are removing the transfer axial from the swerve unit this includes not using chain. http://imgur.com/5fuaJQs
We are hoping that his will improve mechanical efficacy drastically decrees the size of the unit and how tall it is along with how heavy it is.
http://imgur.com/a/H6Qcm

Any criticism is welcome along with comments to improve design.
Our cad for this years and our designs for next year. https://drive.google.com/folderview?...sp=dri ve_web

[Bryce2471]

Hey Tyler, I'm curious what kind of wheel you plan in using for the new lower castor. Or if they will be custom, how you plan on making them, and what material you will use for the tread.
I like the concept of only having one gear stage in the castor.

[seanthompson]

Sweet design! What speed are you geared for? Also, how long did it take to machine the gearboxes? What was the final weight of your drive train?
I am super interested in your design as our team is planning on creating a swerve drive over the summer and into next preseason.

[Tyler2517]

For the lower wheel we are planing on making them our selves. Probably out of 6061 both sides are identical in regards to the wheel. We designed them to be ran on our tiny cnc machine one operation and then cut from the block on a table top layth. Our biggest concern is that our steering motors wont have the power to turn under static load. Most likly will be using blue nytrox treed from mcmaster carr.

We were geared for 12.5-14 fps we changed several times within the season for different reasons. For the new one we are hoping at gearing at about 14.5-15 fps with the efficiency improvements but that is still in the air.

Defense is not a problem for us in terms of pushing we will play passive defense by turning our wheels against their primary pushing force so they cant back drive out motors and have to brute force over our large surface area wheels. And if we have to give a bump we can do accelerate pretty fast to deture anything.

[Tyler2517]

*

Originally Posted by seanthompson

Sweet design! What speed are you geared for? Also, how long did it take to machine the gearboxes? What was the final weight of your drive train?
I am super interested in your design as our team is planning on creating a swerve drive over the summer and into next preseason.

Me and my sister managed to turn out a swerve module and a half a week. We had very little machining experience. Our lead machinists left along with our couch left the morning of kick off. The biggest thing that hurt us is that it got super cold causing a pipe to burst in the machine shop shutting it down for 2 and a half weeks. This caused problems.

They weighed in at something like 7 pounds 8 ounces? We were not very comfortable with bushing and our machining so they were made much heavier they they needed to be. Full bearings other then one bushing.

[asid61]

Is the large bushing on the swerve module aluminum or plastic?

[Tyler2517]

On the new one or the old one? the old one was a brass/plastic on the top. On all of them we are running bearings on the bottom Right above the yoke.

[Jared]

These are really cool. It looks to me like there's a thrust bearing on the bevel gear shaft. Did you guys find it necessary to add this, or did you just add it to be safe?

[cjl2625]

Why didn't you guys try field centric control?
I can't think of any reasons not to, aside from handling gyro drift. But especially with a bit of driver practice, I find it to be pretty manageable to deal with.

Also how does the driver change the center of rotation during a match?
I recently wrote some code to change the rotation point, but I don't really know how to elegantly integrate it in the driver controls.

[Tyler2517]

We ended up not running the thrust bearing on the top of the bevel gear due to running out of time there was a lot of wear on the top swerve yoke. I would have preferred to use one.

We use a standard ps2 controller and the bumpers on each side would correspond to a corner. To change we would press down on the button and as long as you were holding down it would be using said corner as the point of rotation using the left stick to move.
We did not have the experiences to start the coding on field central drive when we started to build and program. When we did get it it was to hard for the drivers to switch to with our getting a few mess ups our code also seemed very slow as we could conserve more momentum using the robot central control.