I think that if you put only one driven mecanum on your chassis, say on the back left, and the other three wheels omnis, it would do the same thing as you describe in your test above. You would likely see some diagonal movement, especially while accelerating or pushing.
But if you tried putting mecanums on the back left and back right (or in fact all four wheels), the roller are now trying to roll in opposite directions when driving forward, so they effectively cancel each other out and behave much like omnis. At least this is my understanding.
I didn’t even need to take the video. They moved in virtually lock-step along either side of the bump. Analyzing data that obvious would have been a waste of time.
You can rebuild a base in ~4 hours tops. I’ve done mine three different ways this season already. You could easily have it done by this weekend, even, if you really needed it.
The vector of a single mecanum wheel should be just this:
If you want to test it, your idea should work (although there is likely to be extra friction from the omni wheels, etc.) because the mecanum wheels on the diagonals of the base have the same roller orientation.
From what I can tell, your image depicts a 45-degree angle. However, it has been hypothesized that the angle when strafing is between 45 and 90, and this causes mecanums to strafe more slowly than they move forward/backward (and slower than an X-drive).
You are correct, the angle is actually more like 60. I believe this is caused by friction in between the sides of the rollers and the wheel hub, but I may be wrong here.
I think the info on vectors and speeds has been posted before, but I’ll just say it again. If you break the vector into components, it should be pretty easy to see why strafing is slower.
Ok, maybe this picture is kind of bad, but you can sort of see that because of the (~)60 degree angle, the X component is smaller than the Y component. So when driving forward, the X components cancel each other out and your left with the Y components, which are about the same as normal wheels (a little less I believe). But if you strafe, the Y components cancel out and your left with the X’s, which are noticeably smaller than the Y’s, thus causing the slower strafing speed.
With this info, a few measurements, and a bit of math you could probably find the exact angle if you wanted to.
From what I understood, the rollers lock while driving forward due to friction created by the forces, meaning that the wheel no longer is visualized as an omni wheel, but rather a high traction wheel. Because it doesn’t have that sliding movement, it doesn’t get the speed boost
I followed that same conclusion that cancelling vector forces would result in it behaving like omni’s. But if it did behave like omnis in X drive, then it should of had the speed boost. But according to their results, the speed boost didn’t occur
