Re: Questions about crab/swerve drive.
 

http://usdigital.com/products/ma3/ ? It was the first hit on Google for "Ma3 abolute analog encoder"

Re: Questions about crab/swerve drive.
 

Thanks, multitasking off my ipod touch and internet is super slow.

Thanks,

-RC

Re: Questions about crab/swerve drive.
 

Since this thread has taken a lot of tangents about crab drive, let me throw this out.. In a coaxial system, where the power is transferred to all wheels via chain, would it be better to use a belt instead of a chain in case you get into a pushing match on a friction surface so you don't slip or lose the drive chain? This creates more side load since the belt only works with tension but if preserving your drive is the main point, would a belt drive be a better solution?
( I ask because I noticed a belt "upgrade" for my milling machine. It is currently geared and the belt preserves the system in a stall situation presumably by slipping. )


Steve

Re: Questions about crab/swerve drive.
 

I'm not going to claim expertise on the subject, but I have seen teams to use chain for this purpose and as far as I know it works very well. I have never seen a team use strictly belt to power a coax, but somebody has probably done it. 1625 and 118 would be great to teams to ask about this.

If you want some intersting info about belts vs. chains check out team 234's white paper about chains and belts

Re: Questions about crab/swerve drive.
 

Re: Questions about crab/swerve drive.
 

The difference is that in using a belt to control the turning a timing belt would have the same effect as a chain essentially, but if there was slippage, then the entire steering system would be uncalibrated, the wheels would be in different places then the program thinks, a logistical nightmare. That is assuming you monitor the turning from the motor. If you do not measure from the motor then you are complicating the modular designing of your wheels.

And if you are running with multiple modules and wheels tied together then you complicate things if your wheels become out of sink.

Re: Questions about crab/swerve drive.
 

I didn't mean for the steering, the steering should be chained. I meant for the drive system that supplies power to the wheels. Even if you settle on a system that has a motor dedicated to each wheel, should you belt drive the wheels so the belt acts like a virtual clutch to avoid over stressing the transmission on a friction surface (like carpet)?

Re: Questions about crab/swerve drive.
 

I didn't mean for the steering, the steering should be chained or electronically synced. I meant for the drive system that supplies power to the wheels. Even if you settle on a system that has a motor dedicated to each wheel, should you belt drive the wheels so the belt acts like a virtual clutch to avoid over stressing the transmission on a friction surface (like carpet)?

Re: Questions about crab/swerve drive.
 

Thats a good point. Just wondering, how do teams like 118 keep the modules alligned? Is it just tensioning out the chains correctly, or more complicated than that?

Re: Questions about crab/swerve drive.
 

Chain tensioning but their wheels are generally within 2-5 degrees off from each other. I don't think I've ever seen them perfectly parallel.

Re: Questions about crab/swerve drive.
 

This year we bound our wheels F/B. The way we did this was with chain tensioning, and physically rotating the rotation sprocket

We didn't line up our wheelboxes completely, but I would estimate that we were between 1 and 5 degrees toed-out for each wheel

Re: Questions about crab/swerve drive.
 

Hmm, here's a way to eliminate the misaligned wheels. If teams limit the total rotation to, say, 720 degrees, a couple of in-line turn-buckle style chain tensioners could be added to a coaxial crab system. 720 degrees on a 3" sprocket radius is a total ~12" of chain run needed to rotate the modules. Hence, there should be plenty of room to add two tensioners to each chain run, one on either side of the sprocket, centered between the sprockets when the wheels are approximately in the center of their rotation range. With some adjustments, the wheels could become perfectly parallel to each other.

This might not be worth the added complexity, but if toe-in is really a concern for a team it may be a start.

Re: Questions about crab/swerve drive.
 

Steve, the issue of overworking a gearbox is not an issue unique to crab drive. Every FRC robot has to take this into account during design regardless of drivetrain to make sure that they don't stall the CIMs. Our systems are a bit different than your mill in that our systems are current limited by breakers well below the stall current of the motors typically used in the drivetrain. A stall condition will trip the breakers and your robot will just stop moving for a little bit. This is still an undesirable situation so most teams try to avoid it.

The most common way to do this is to gear your system so that the motors have enough torque to make the wheels slip. Having belts that slip somewhere in your power transmission path would be another potential way to attack the problem. As a mechanical ignoramus I will avoid commenting on the pros/cons of using belt slip to address the stall condition.

Re: Questions about crab/swerve drive.
 

I had an in depth conversation with people from 118 the gears on the modules are actually a friction fit, so they just turn the modules to realign, instead of turning the motors and chains that position them.

Re: Questions about crab/swerve drive.
 

Steering with chain works great, but I wouldn't say it should be chained. I've seen teams cable. I believe Wilstang did in 2008, I didn't get to see their robot this year up close, but they may have used it again. My team used cog belt this year and it worked just as well as the chain we used the year before, the only thing about belt is that it needs more wrap around its pulleys/sprockets than chain does. I have also seen teams use a linkage, if you look at 141's robot this year it is a good example of this.