Are cutting 2.75’ wheels good for odometry, do they get bumpy?
people often slice them in half for low-profile tracking wheels, so they should work fine.
although you could cut the wheel, i would recommend using the wheel as it is because it’s really annoying to cut the wheel.
1 Like
You only need to cut the wheel if you have no space, or only one wheel (or two wheels depending on the set up). If it’s the latter, than I suggest you invest in more wheels (especially if you want to do an advanced programming scheme like odometry).
2 Likes
They will not be containing the ground always because of the gaps between the rollers if only half of the wheel exists. This is bad. This is a game where it’s fine to have a wide chassis, so better to make room and use whole wheels.
2 Likes
Sure there is some friction loss but an ideal wheel is supposed to free spin with minimal friction on the axle and maximum friction between the wheel and the ground. The static friction between the foam field tiles and the rubber rollers is more than enough force to cause the tracking wheel to spin.
Not to mention the tracking wheel should be pushed down into the ground with elastics further increasing the force of friction that causes the wheel to spin.
Here’s a diagram that I think will be useful:
4 Likes
The issue is that with a half cut omni, it is no longer a true wheel. There are significant gaps where the wheel diameter is smaller, and will absolutely result in slip that leads to noticeable drift in the odometry.
2 Likes
No, even with “gaps” in the wheel diameter, with rubber banding, a part of the wheel will be contacting the ground at all times. This means that there will not be any “slip” in the wheels.
Alright, now I’m just curious what evidence you have to substantiate these claims. I personally have not cut omni wheels, but I have used the single-roller omni wheels that were discontinued that are functionally the same as a cut omni wheel (arguably even less surface area). I have never experienced or noticed any issues relating to wheel slippage for my tracking wheels.
I think the point is that the rollers in each half are offset alternatingleft and right. With just half, you have large alternate gaps where the other halfs roller would have been and so has a lot more flat sides. The single style dont suffer this issue - they have something like 14 rollers around the circumference. With the 2.75 cut in half, you end up with 7. When it is still whole, the offset rollers still give 14 round the circumference.
2 Likes
2.75 inch omnis have staggered rollers such that there is always a contact point on the ground. Half of these are on each half of the wheel. Therefore, by removing one half, you remove close to half of the wheel’s ability to contact the ground, creating the issues described.
Yes, I know that, but I’m just trying to understand the rationale behind your drift claims. I’ve seen many teams cut the omni wheels in half and none reported drift issues. Here’s a notable team with the configuration (4001a):
The issue you pointed out is virtually eliminated by the tensioning on the wheel. Remember the only thing causing the wheel to turn is the force of friction between the ground.
Even with half the amount of rollers, the force of friction on the plastic is enough to turn the wheel.
2 Likes
Slip must lead to drift. It’s just how it works dude, I don’t see why you think it needs to be backed up. It can be managed by ramming into the wall and resetting, but with full wheels it doesn’t need to be done as often, as the accuracy will drift more slowly.
since the wheel is rubber banded into the ground, it physically can’t slip.
1 Like
Everything theoretically works, but no one knows until it’s actually tested
We used halfcut 2.75” encoder wheels on our 24” robot last season; we had no issues with drift.
3 Likes
It doesn’t theoretically work, it works more poorly than a full wheel. Sure it works, but this is about what’s better, and a full wheel is undoubtably better. It’s really not that hard to fit them.
This isn’t true. The rollers on the 2.75 halfcuts are quite large, meaning that the divots between them are quite deep. This means that even with bands, the areas where there is no roller will not touch the ground, causing the issues I mentioned.
The issue you’re mentioning is so minimal it shouldn’t even be considered as a factor. If your wheel needs that much friction with the ground then you’re probably doing something wrong on your end. The wheels should free spin with the slightest amount of force. And because the wheel is tensioned down, there is ALWAYS a friction force between the ground and the wheel. (This is mainly because the wheels and the foam can “squish”) The only discontinuity would be if the robot physically tips away from the ground.
2 Likes