Hacker News Reader

Comment by namirez | Hacker News Reader

I don't understand the benefit of turning the whole wheel into a triangular track; would it not be sufficient to make only the bottom third flat?

View on HN

Here's a video from Mattracks (2014) where the geometry changes like you've suggested: https://www.youtube.com/watch?v=6zTxMwYM1tI#t=20s - that should start at 20 seconds in.

It doesn't look like their whole wheel rotates though.

If it wasn't symmetrical it would be far more difficult to balance properly for when it's rotating in "wheel mode"

Yah the real trick is seeing that thing spin perfectly at highway speed.

and then smashing into a pothole. That is A LOT of heavy complicated unsprung mass.

The vehicle in the video was already wobbling around madly at low speed, as it is it’d be completely uncontrollable at highway speed and at whatever frequency the wheel and suspension resonate.

We add grams to balance wheels to prevent damage to suspensions etc, this thing seems just a dream for vehicle applications.

Note that the whole assembly rotates when in wheel mode. Making it radially symmetric (i.e. a triangle) in track mode means that it can switch modes with at most a 120 degree rotation rather than having to rotate through up to 360 degrees to get the flat bottom into position.

I didn't say anything about the design of it, which is by the way, very cool. I don't understand the point of the triangle. The same mechanism could be used to change the geometry of the wheel only at the bottom.

I would guess that it has to do with the tread tension. A rounded top with a flat bottom may leave the tread too loose. The triangle keeps it taught.

Note: taut (pulled tightly) and taught (past tense of teach) are homophones

Considering that this is a rather a rounded triangle, I doubt it has anything to do with tightness, and more with convenience, esthetic. A symmetrical displacement is also slightly less constraining in term of keeping the overall circumference/periphery (i.e. less deformation).

A better design would have been to flatten the wheel into an more oblong or even an egg shape, like snowmobiles or tanks have. However, it requires more mechanical displacement and adjustements.

"Better" is meaningless in isolation. The design you propose is better in some aspects, but likely not in the areas the designers of this wheel are optimizing for (simplicity, reliability, etc)

I suspect it may apply torque from the drivetrain to the mechanism with the most optimal geometric leverage against both the mechanisms internal frame and the contact patch on the ground during transition into and out of tread mode.

Changing it only at the bottom adds complexity: Which side is at the bottom right now? How can we know this; how can we know this in a way that will survive the tough conditions we expect this to be driven in? And then we need to have three separate actuator paths to transform just one part of the wheel.

We would also need to deal with the fact that just flattening one side of the wheel would make for a shorter perimeter than that of the circle; the tread is a loop of a constant length, and it needs to be kept fairly taut to function in both wheel and tread mode.

The triangle solution absolutely does need to keep track of which side is at the bottom, or you'll end up with all the weight on one vertice.

Agree with the perimeter consideration.

It's a symmetrical wheel -- there is no bottom.

I'm guessing the triangular track mode is the default mode. If the track gets damaged the whole thing can be switched to the wheel mode and continue working that way.

Reply on news.ycombinator.com