pic: Mars Game 2018 (Work in Progress)



After demoing our robot at the county fair today, I was totally hyped about FRC and wanted to do something… but I couldn’t think of anything useful. So I walked over to the TV and there was a documentary on about Curiosity, so… yeah. This happened.

This is the right side of a 6WD Swank (swerve/tank) drivetrain with a Rocker-Bogie suspension system. Each side will have 3 independently driven swerve modules, which may or may not max out the motor ports on a PDP (I have no idea how many ports there are). I designed it for fun and practice, and over the course of an afternoon, so it

A) is absurdly expensive and time-consuming to manufacture,
and,
B) has a few issues that will take a bit of time to work out.

Anyways, although it’s not really a serious design, tell me what you think!

Looks totally serious to me.

Assuming 2018 rules include something like R1, R22, R23, R24 from this year, you’ll need to think about how to get a legal set of bumpers on it. The MER designers had many challenges, but bumper rules were not among them.

This is very cool!

What do you have the drive geared for?

Any reason you have the drive motors vertical instead of parallel to the wheel axle? Would probably eliminate some inefficiency and allow the modules to turn 360*.

Also why 775pros for drive motors over CIMs? 6 775pros (with voltage/current limits) is barely enough power to drive a full weight robot and not burn up. 6 CIMs is still legal, requires less reduction (smaller gearbox), and is a lot more resistant to high torques, which is likely in a rocker-bogie suspension system. If you think 6 CIMs is too much power, you can use 6 miniCIMs, which is about the same amount of power as 4 CIMs.

FYI, there are 16 slots in the PDP (8 40A max, 8 30A max). Your system uses 12 of those (assuming 6x 30A and 6x 40A), so you would have 2x 30A slots and 2x 40A slots left for manipulators. That combined with pneumatics and servos is enough to have a decent manipulator for pretty much any game since at least 2010.

This is absolutely nuts. Please build this, at least a small one, simply to say you did. I mean, how many people can say “I used a rover drivetrain design for a robotics competition… and it worked.”

Legal… who knows with bumpers. Awesome? Well, that’s a given.

1 Like

I love it. Crazy enough, there is actually more gearing required in this to make it work. You have to have a differential gear box set in between the left and right side on the front pivot so that as one side goes up the other goes down to compensate on rough terrain. I had to build one of these with banebots gearboxes for a lunar mining competition. It does get heavy quick and there are some other things you have to do to keep the chassis level but it worked swell at Kennedy Space Center.

**Challenge accepted.
**

We seriously considered building something similar in 2016. A rocker-bogie suspension at least, not one that slapped swerve drive on top of it.

As to making bumpers work… in 2016 we did build an actuating drive train to help us clear defenses. One drive rail was an integral part of the frame. Then the frame was lifted a couple inches on opposite side. The second drive rail was mounted on a pivot underneath the frame. The bumpers were mounted to the frame and the frame was deemed by inspectors to be non-articulating (the drive rail articulated) as such the bumpers were non actuating as well.

I expect you could design a similar system for the bogie suspension. You mount both suspensions to a solid frame and mount bumpers to it. The question becomes, if both halves of the drive train articulate does the frame articulate?

Yeah… I haven’t really figured out what to do about bumpers. Mounting them at normal height would also make the suspension pretty useless, so I have no clue what to do.

I haven’t selected a gearing ratio yet, but the drive motors have 2-stage VP’s on them, allowing for a range of different ratios.

My plan for getting the frame to stay put is to have a differential bar running across it. This way, when the left side goes down it rotates the bar, causing the right side to go up and cancel out its motion. Because of this opposite reaction, the frame will be able to hold itself up mechanically and without the use of motors or weights.

Yeah, the amount of gearing on this thing is insane. Instead of using a gearbox for the differential, though, I plan on using a bar that will act as a differential. This way, when the left rotates down, the right will rotate up proportionally and cancel out any tilt, allowing the bot to stay upright without the use of motors or weights.

Best of luck to you.

I’d really love to build this, but… it costs like $2000 just for the drivetrain, and I’m not even sure if it would work. I’m prototyping a swerve drive using a 3D printer, but even that requires a lot of components. I think it would be cool to 3D print a little model of this for my desk, though…

I chose to put the 775’s vertical mostly just for the aesthetics, to be honest. Curiosity has some kind of hubbed (or very small and hard to see) motor for its wheels, so I wanted to try to imitate that. Also, I heard that my team had a really bad experience in the past using VP’s horizontally directly to the wheels.

Also, how would you recommend utilizing CIM’s for drive in this design? As far as I’m aware of, CIM’s can’t be mounted to VP’s and creating a custom gearbox for this would be either too big or an absolute nightmare.

Thanks a lot!

CIMs absolutely can be mounted to VersaPlanetaries using a CIM Adapter. You have to pay attention to the load rating guide to make sure you don’t use them above their rated torque, but other than that it’s no problem. You should also pay attention to that guide with the 775pros, because I would think you are starting to approach that limit with (what look like) 6" wheels driven 1:1 off the VP.

As far as directly connecting the VP to the wheels, you’re right that that’s not such a good idea. VPs don’t like much axial or radial force on the output shaft, and direct driven wheels tend to give both of those (especially if not well-constrained). You could, however, use a small spur gear or belt or chain reduction between the VP output shaft and the wheel shaft so those forces aren’t transferred. All of those options are more efficient than bevel gears, and would allow you to rotate the modules 360*. That would also allow you to lower the reduction in the VP, which will bring you further from the max rated torque and maybe even make your VPs smaller.

If you’re not set on using VPs, you could use CIMs with a Toughbox Micro or Nano. They’ll give you plenty of reduction and no worries about forces on the output shafts. The hardest part would be integrating them into your design, but they aren’t too big so you should be able to find a place for them if you so choose.

Yes CIMs can be mounted to the VPs if you’re willing and able to cut the hardened shaft of the CIM motor.

It really isn’t that hard to cut. My team has cut a number of them over the years with just basic hand tools.

  1. Put the CIM in a vice
  2. Mark the spot you want to cut with a scribe (marker will work if you don’t have a scribe)
  3. Plug the CIM into a battery to turn it on
  4. Slowly, using some cutting fluid and all the proper safety precautions, press a hacksaw against the marked line to cut. Stop when there’s ~1/8" diameter left
  5. Unplug the CIM and let it come to a stop
  6. Cut the rest of the way through
  7. Finish off the end with a file

This method hasn’t failed is yet, and it’s basic enough that any team can do it. If you’re really worried about safety, you can cut the shaft without turning on the motor using a good amount of cutting fluid and elbow grease. It really doesn’t matter if the end isn’t perfectly flat for this application.