Shifting X-Drive

Yet another transmission system, this one allows you to shift from normal 4-wheel tank drive to 45 degree x-drive with the power of pneumatics!


Read the full write-up (including close-up pictures) of the drivetrain here:

So what’s the point of having an x-drive? Apart from being able to strafe, it goes faster than a tank drive too! Don’t believe us? Read this article here - “Why is X-Drive Faster?”:

Got any questions? Feel free to ask (or discuss) below!](


Very cool :slight_smile: I really like this concept.

As for X-drives being faster, I’ve heard it called in fancier terms the Velocity Augmentation Factor. But I appreciate that in the paper you guys explained it in an easy to understand way. Nice job. :slight_smile:

Our team tried to come up with this for our wallie, 4886c and 4886b, so we could expand then switch to tank drive. The problem was trying to using only 2 pistons to shift.

Ummmmmm Last time I checked Holos were slower than tanks due to canceling vectors…

You might want to check your math there…

  • Andrew

To be fair, that was what most of us said too. And then we proved it both theoretically and empirically that x-holo is FASTER in the “forwards” direction than a tank drive. TooMuchStrategy says watch the video - if necessary we might do a slightly more scientific version of the video with tape measures and timers etc.

I think that this actually works.

When I draw the vectors for both theories, the one on the right (where vectors cancel) seems to be incorrect.

Of course, if it’s been proven empirically, then our work is done here, isn’t it? Unless someone comes up with a weird exception that doesn’t make sense with our current model of understanding.
X holo drive vectors.jpg

If you watch the video starting at 12s, you can see that it slows down noticeably on switching into tank:

Hmmm… I’m seriously considering trying this, it seems pretty awesome, there aren’t many downsides…

Only problem is I only have access to two pneumatic cylinders. Do you think its worth trying to chain/gear the front and back together to get them to both unfold from the action of only a single cylinder?

EDIT: Also, if you were to use this on a competition robot, would you use the vex hinges, or would you mount the wheel “pods” on a shaft? Does the hinge work well in this application or not?

The way I see it, it’s a swerve drive, only it has omni wheels and the turning is controlled by pneumatics rather than a motor. Because of this, I would think it’s very possible to actually have all 4 pods turn with the action of just one piston if you gear them together correctly. The piston could be attached to a rack that rotates a gear roughly 45 degrees. Also if you do this, you would likely have to put the pods on shafts instead of hinges.

That’s just my opinion anyway.

We’ve found that hinges are generally stiffer and less wobbly than using shafts in similar applications, not that you couldn’t do it well with shafts.

It is possible to do it on two pistons, just harder to build and probably not as effective (less total force available). Although four pistons definitely means air runs out quickly with just one reservoir.

Also if anyone’s interested, if you make the x drive past 45 degrees you would get more speed in one direction and more torque in the other direction, essentially making it like a shifting drive train without actually having shifting mechanisms.

Indeed a cool way to make an effective CVT (continuously variable transmission). But, I’m a little wary of that design because of the pivot point. Can it support the full weight of the robot?

Other than that though, very nice use of the speed advantage of an X-holonomic drive.

We’ve got most of an arm/lift on that base right now, and it seems okay, although we haven’t made it pick up a gazillion stacks yet. With our testing so far it seems that it should work, although that might change.

On a slightly unrelated side note, 254 borrowed two 60T gears from us at Worlds and we never got them back. :frowning: (it’s okay!)

where’s the continuous part?

If you use a continuous actuator instead of the pneumatic cylinder then you can get the full range of ratios 1 - ∞, not just 1 and 1.4.

We were hoping to be the first team on the forums to point that out, in our writeup of the offset X-drive Previn linked to, but I guess the cat is out of the bag now :p.

How would you calculate speed and torque if the offset is x degrees apart? I tried to figure this out but I don’t know enough about vectors.

I have a draft document on this that I’ll try to upload tonight. I’ll see if I can get our webmaster (bobsalive) to add it as a page on our website tomorrow.

If the offset is X degrees from the wheel being pointed straight ahead, the force component straight ahead is given by cos(X) times the force when X = 0. The maximum speed is given by 1/cos(X) = sec(X) times the maximum speed when X = 0. You can verify this result by multiplying the speed by the force:
cos(X) * 1/cos(X) = 1
which says that you are not losing any energy as you rotate the wheels. You are losing a lot of energy to friction as X --> 90 degrees because the rollers are turning very fast.

Yes, dontworryaboutit is right (as usual). Here is an in depth explanation if you are interested:

The math checks out. An N-motor, Y-ratio “X” drive will move faster than an N-motor, Y-ratio tank drive. However, this is a little bit misleading. For a single speed system, you loose signifigant amounts of torque in the “X” drive. Therefore, the N-motor tank drive can be safely geared faster than the N-motor “X” drive, cancelling out the difference if we’re talking about two separate robots.

Cool shifting concept though! Reminds me of this. I wonder if a similar linkage system could be applied to reduce the number of cylinders/air consumption required.