Swerve Drive Modules

[Ether]

Originally Posted by msimon785

Yes. It fixed the problem, at least for the most part. When it gets down to the root of the problem, though, we had welding displacement and inaccuracies which inevitably caused a misalignment of chain, and thus an unnecessary torsion force on the bearings.

I'm a bit puzzled then by your earlier statement: "I'm going to do my best to get our team away from Mecanum drive". Since it wasn't the mecanum's fault, why do you say this?

[msimon785]

Final-Draft for the CAD! I added an extra bearing to hold the drive shaft. The sides are now riveted on, as opposed to bent from one piece. After creating a BOM, I found that 4 modules will end up costing about $750.

The entire module, from the top of the Timing-Belt pulley to the ground is exactly 7" tall. it is 2.65" wide, and 3.5" long.

[Madison]

You are still likely to see misalignment between the bearing holes on either side of the module that will cause grief for you in the bevel gearset.

If your team isn't yet capable of building a 4WD frame that doesn't wreck ball bearings, you should not expect to have success with a swerve drive yet. Focus on improving your deficiencies instead of replacing them with all new ones and you'll find it to be a valuable experience.

[msimon785]

Originally Posted by Madison

You are still likely to see misalignment between the bearing holes on either side of the module that will cause grief for you in the bevel gearset.

If your team isn't yet capable of building a 4WD frame that doesn't wreck ball bearings, you should not expect to have success with a swerve drive yet. Focus on improving your deficiencies instead of replacing them with all new ones and you'll find it to be a valuable experience.

Again, this is an offseason project. It is to test whether or not we will use Swerve on the 2012 robot. It is very possible that we will put it off for a year and use it on our 2013 bot. We are simply building this to test our abilities and form a relationship with a SM shop.

[Redo91]

It looks like a nice module. Do you have a weight?

Also, I see the gear for a belt to rotate the module, but how are you actually going to attach it to the robot?
From personal experience, that was the toughest part to get working right.

[Clem1640]

We've used Martin Gears' HM1616 (16 teeth, 16 Diametrical Pitch) miter gears for the last 2 years, but will probably switch to M1616 (not hardened) if we do this again. Gears have 3/8" unfinished bores. We cut our own keyways (finished bore gears are too expensive). We've had good experience with these.

Here's a view of our 2011 Pivot Module. Our Module includes the motors and angle sensor:

http://wiki.team1640.com/index.php?t..._annotated.jpg

Mass is 9.1 lb including motors, gearbox & sensor.

[42!]

I would have doubts on the ability of the single output shaft to be able to take the side force from impacts regularly incurred in FRC competition, also how would you be taking thrust load from the weight of the robot? In general these are both really important factors for swerve drive. I would recommend looking into Delrin and other bearing systems to attach it to the robot.

[BJC]

Just a good thing to keep in mind when building drivetrains:

The best drivetrain is the one that gets you where you need to go, when you need to be there.

Ask yourself:
-Does a 6wd get you there? In a timely mannor? Consistantly?
-Does a swerve get you there faster than a 6wd? Every time, Consistantly?
-Is the difference in performance between a 6wd and a swerve enough to outweigh the potential driver practice that comes with finishing more quickly?
-Could the effort put into the drivetrain be better placed in another part of the robot?

Just trying to keep you honest , Bryan

[JVN]

Originally Posted by BJC

Just a good thing to keep in mind when building drivetrains:

The best drivetrain is the one that gets you where you need to go, when you need to be there.

Ask yourself:
-Does a 6wd get you there? In a timely mannor? Consistantly?
-Does a swerve get you there faster than a 6wd? Every time, Consistantly?
-Is the difference in performance between a 6wd and a swerve enough to outweigh the potential driver practice that comes with finishing more quickly?
-Could the effort put into the drivetrain be better placed in another part of the robot?

Just trying to keep you honest , Bryan

Love it! Yeah... what Bryan said.

-John

[Chris is me]

Originally Posted by BJC

The best drivetrain is the one that gets you where you need to go, when you need to be there.

The best drivetrain never has issues. If there's one part to make reliable, make it your drive.

If you have issues with non-swerve drivetrains, don't waste your time with a swerve prototype. You won't be able to pull it off. You have to learn to walk before you learn to run.

[Hawiian Cadder]

for 750 per set, why not just buy wild swerve modules from andymark. if you want to go further, get the revolution modules.

[msimon785]

Originally Posted by Hawiian Cadder

for 750 per set, why not just buy wild swerve modules from andymark. if you want to go further, get the revolution modules.

I resourced a bunch of the parts, and we're going to be paying $640-$660 now. The revolution costs $990. And again, we're not building this necessarily for the season. This is an offseason project designed to test our abilities, teach freshmen/newcomers, and form a relationship with a SM shop.

All in all, I think that the value of the project outweighs the cost. However, that is subjective, and is my personal opinion. To those who promote 6WD or 4WD, that's more than likely what we will use in the 2012 season. It depends both on the game and whether or not the swerve is extremely successful and reliable.

[Siri]

Originally Posted by msimon785

All in all, I think that the value of the project outweighs the cost. However, that is subjective, and is my personal opinion. To those who promote 6WD or 4WD, that's more than likely what we will use in the 2012 season. It depends both on the game and whether or not the swerve is extremely successful and reliable.

Test the living daylights out of it. Weigh it down, run it hard, long (lower voltage), and against defense. Complex components have a nasty habit of seeming really great until you actually need them to be.

And, just to clarify what I think a lot of people have been saying (I think you realize it): Don't build a swerve to test if you can do it. Build it to learn the skills you need to build great drivetrains: testing, troubleshooting, consistency, accuracy, robustness. (not exhaustive) Emphasize those lessons over the result of the actual device. They're more important, for FIRST and beyond.

[Taylor]

Originally Posted by Madison

If your team isn't yet capable of building a 4WD frame that doesn't wreck ball bearings, you should not expect to have success with a swerve drive yet. Focus on improving your deficiencies instead of replacing them with all new ones and you'll find it to be a valuable experience.

Ding ding! I have to spread some rep around before giving Ms. Madison any more, but she deserves it.

In 2005, 2006, 2007, and a little bit in 2008, we struggled with throwing chains on our drives. At the single most inopportune moment, the robot would decide to drop a chain and do the robopeepee dance. in 2009, we put a priority on chain routing - the routes were clean, properly lined, and shortened - we had a great year with no drive problems. In 2010, our chassis flexed in ways we didn't realize until after the FRC season had ended, causing us to throw what we thought were properly aligned and tensioned chains; we put on structural supports on before the offseason events and ran them smoothly.
However, the notion around the team became "chains are nothing but trouble" - the truth is chains are fine. We just weren't implementing them correctly.
As I mentioned, this year we went to direct-drive mecanums right out of the gearboxes, with absolutely no complaints. This year taught us that simplicity brings success. I applaud you for pursuing the swerve drive as an offseason project - I think your team will learn a lot about iterative design, programming, and machining. However, I'd be careful about creating these with the 2012 or 2013 game in mind - as others have said, let the robot fit the game, don't force the game around your idea of what a robot should be.

[buildmaster5000]

Others have said this and I will echo their sentiments: if you do swerve drive, it needs to be reliable, and built like a tank! We spend many hours at our (only) competition fixing little things that kept creeping up on our swerve drive system. Also, I have trouble believing that the module you disigned will hold up against any amount of side load (imagine a 200 lb robot traveling 20 fps raming you at a dead stop. If you are moving, that net speed goes up, and bumpers will not do everything.) without torquing or just sheering the ouput shaft. Personally, it was really a nice feeling to have the bottom assembly of the WildSwerve modules there, where we could not sheer the output shaft.

In addition, decide on a control scheme and programming strategy, and if at all possible, have the driver be the programmer. I both programmed and drove my team's swerve drive this past season, and I was able to quickly and easily diagnose software problems just because I knew the program so well.

If you are interested, I can point you to the team code from this year, as well as answer any questions you have via PM.