pic: Team 2220 Blue Twilight Octocanum Offseason Project



As an offseason project, my team has been working on creating a new showbot/electronics test platform. We have settled on creating an octocanum drivetrain for this robot for both the experience of designing a drivetrain outside of our norm (we’ve done a kitbot-style 6wd just about every year since we started), but also for the “wow” factor when we’re showing sponsors or for school assemblies.

This is our version two of the actual octocanum module, rendered in with Creo’s built in tool and combined in Photoshop. The plates are 3/16" aluminum, and it uses VexPro mecanum and traction wheels. All axles are dead, and it’s powered off of a gearbox above the frame to the sprockets on the module. We plan to articulate the drive “Neutrino-style,” or, barring that, using the more “traditional” method of attaching the piston to the bar (using a combination of channel and angle stock).

The module is designed to be dropped into a 5" wide channel, sandwiched between two pieces of 3x1 1/8" wall 6061-aluminum tube stock. The top side of the module wall is parallel to the tube stock.

Feedback, critique, and comments are welcome. We’re aiming for production around the end of July, with assembly and powdercoating taking place in August.

Are you using threaded rod as an axle for the traction wheel?

I ask because we’ve tried doing that before, and it didn’t work very well, the axles bent after only a few minutes of driving. Then again, I think we were using pretty small rod, I think it was 3/8", but it might have been 1/2", I can’t remember exactly.

Also, what is keeping your wheels centered on the axles?

The module is looking pretty excellent though! Can’t wait to see the finished product! :smiley:

which wheels do you plan on having at the corners of your drive?
Also what are you planning on for a gearbox? It appears that the sprockets on both wheels are the same size is that correct? Overall looks very nice hopefully i can see it sometime at an off season event when its done.

Based on the fact the double sprockets are on the meccanums I think it’s a safe guess that they have the traction wheels on the outside.

I don’t see this as conclusive evidence.

For the OP,

If you have decided, can you tell us why you decided on the placement of the traction wheels on the inside or outside. I think there are a few obvious pros and cons to each, but would like to hear your view on it.

One comment I would make is with the sprockets chaining the traction wheel and mecanum together, did you pick the wheel separation distance to ensure an even number of chain links? It is a bit harder to put the modules together under tension, but never worrying about throwing a belt or chain is great.

One comment I would make is with the sprockets chaining the traction wheel and mecanum together, did you pick the wheel separation distance to ensure an even number of chain links? It is a bit harder to put the modules together under tension, but never worrying about throwing a belt or chain is great.

Yes! We saw the multitude of posts on that subject and ensured that we were spacing the sprockets at a multiplier of the pitch for #25 chain (.25").

If you have decided, can you tell us why you decided on the placement of the traction wheels on the inside or outside. I think there are a few obvious pros and cons to each, but would like to hear your view on it.

The current plan is to have the traction wheels on the outside. The primary reason has a lot to do with one of out mentors, who is pushing us to develop this for the 2014 season (which may or may not make sense). Our primary rationale was that, in a competition bot, one of the main benefits we wanted from octocanum was the ability to go in a very straight line without too much feedback correction (for autonomous mode). The main thing we were weighing this benefit against was being able to use a single piston on each side to actuate.

Are you using threaded rod as an axle for the traction wheel?

No, we’re using 1/2-20 Grade 8 steel cap screws. We’ve used them on previous robots with much success-- we haven’t managed to bend one in competition yet. We’re still running our 2012 competition bot on the same bolts we used for our two regionals, plus championships and several demos and off-season events.

Fortunately the size of the pistons you need to shift is pretty small, I think ours had a stroke of 1" tops, and a bore of 1 1/16" or so, we also plumbed them all off of 1 or 2 solenoids, I can’t remember what we decided on at the end of the season. You don’t have to use air to pull the traction wheels back up, robot weight should be more than enough. Also, the bot was downright hard to move if it stopped in traction mode.

Yeah, this confirms what we were thinking about the process. I believe Aren mentioned something about “cute little 8lb springs” too.

Could you explain this further?

If you don’t use air to retract the traction wheels, won’t the end result be having both the mecanum wheels and traction wheels touching the ground simultaneously?

Was trying to figure out what the confusion was, then I realized.

We also have a small spring (maybe 1-2lbf tops) that pulls back on a small “finger” sticking up from the plate by the axle lifting the traction wheel the remainder of the way off the ground.

Here is a side profile picture with the “finger” which sticks up through a milled slot in a chassis tube.

Sorry I missed a couple of your questions.

We’re planning on either running them through something like the VexPro double reduction gearbox or a custom solution (we can always use more practice!).

The sprockets are indeed the same size-- and they’ll probably stay that way, unless we find a good design reason to change it. We have a bunch of 32t #25 sprockets lying around, and might as well use them for this!

I hope we get a chance to show it off at an offseason event!

You could gear the traction wheel slower because i would assume in a competition the mecs would be the primary drive mode except during auto and when you encounter defense. changing the speed of one of the wheels would essentially give you a two speed drive train with your high gear being omni directional.

The traction wheel is already smaller, which will make the robot slower already in traction mode. Whether that is slow enough the team needs to test and play with.

I don’t have the exact dimensions off the top of my head but I think neutrino went from a 4" omni to a 2.75-3" traction wheel which provided a substantial speed difference. With the current size difference between their mecanums and traction wheels they should see a similar if not larger effect.

Good point i forgot about the wheel sizes

The way we have it geared currently, by my calculations (using a 1:9.52 gearbox like the the VexPro double reduction), we would be running at ~12 fps adjusted on the mecanums and ~8 fps adjusted on the tractions (using the JVN 2012 design calculator).

To me, those both sound like rather reasonable speeds. There’s also that we have a maximum sprocket size on the 4" wheel (which I believe is something like 44t sprocket). Running that reduction would bring us down to ~6 fps adjusted speed. That’s something we might consider if we were to run this in a competition, but probably won’t deal with on the prototype, at least initially (because of the aforementioned large amount of 32t sprockets lying around). We might also get a variety of sprockets and test out how it feels in different configurations.

It sounds like you are not running a reduction between the gearbox and the wheels. I urge you to do this. If you do, you can switch to a single reduction box which will save some weight and efficiency (especially important on a drive like this).

If you really want to save weight, you could integrate your gearbox into the wheel module.

That is one possible improvement, but regardless of whether the reduction occurs inside of or outside of the gearbox, it will still add inefficiency. A double reduction, direct drive gearbox should have similar efficiency and weight as a single reduction gearbox with an additional chain reduction between the gearbox and the wheel.

Direct drive eliminates one axle, 2 sprockets, and one length of chain per module, while adding one drive hub (and possibly an extended output shaft) per module.
Changing to a single reduction and changing the sprocket ratio removes 2 gears, a stub-shaft, and 2 bearings. One sprocket and the gear box housing are also reduced saving additional weight.

If you really want to save weight, you could integrate your gearbox into the wheel module

This is yet another solution with its own pros and cons to consider. While lighter, it could also be more complex and more difficult to actuate due to the additional weight (not an issue with some methods, but if it is retracted with springs, it might not be worth it).

All possible solutions should be evaluated to determine which is best for the team and its goals.

Imagine moving the pivot outside the two wheels in the module, where ‘o’ is a pivot, ‘X’ is a mecanum wheel and ‘x’ is a traction wheel:


o -- X -- x

This arrangement eliminates the need to lift the modules at all – either by spring or by other actuation. The weight of the robot itself will cause the module to rotate about the pivot point; you simple need a hard stop somewhere after the traction wheel comes off the floor to stop the module from over-rotating and letting your frame hit the ground.

You can make this sort of thing work by either integrating the gearbox entirely or by continuing to use the pivot point as the link between gearing in the fixed gearbox and gearing to make the wheels spin.

I have a CAD model at home of the module I’ve been refining. It’s not quite done yet, but it is based upon this idea and uses an integrated gearbox. If the above description isn’t clear, I’ll post a picture when I get home later tonight.