I recently saw an old CD post about an 8x mecanum drive train for FTC (FTC 8-Wheel Mecanum Drive - NX PowerPlay) down below, the user says “Let me state the obvious: this drivetrain is impractical, and this writeup is a huge meme” The post has been since archived so I am unable to ask why this is impractical.
So I was wondering why is it so impractical. dose it mitigate the effectiveness of mecanum drive? I want to know why an 8x mecanum drive doesn’t it work? and on the flip side could it work?
Pleases understand i am asking this with the mind set of “no dumb questions”
8wd mecanum does not provide a significant traction increase while increasing weight and complexity and decreasing efficiency.
In an ideal world, traction does not increase with surface area, as increasing surface proportionally decreases normal force resulting in a net 0 gain. In practice, additional complexities such as tile sink can result in slight traction gains, however its not very significant.
The downsides are fairly obvious, more parts means more weigh and more frictional loses and more things to break.
Theoretically, the kinematics get a little funky. The inside wheels won’t want to move at quite the same speed as the outside wheels, so you’re introducing some friction into the drive. However, this friction will only be there when rotating the robot, and it should be way less than what you see on a turning skid steer. Additionally, there may be traction/weight distribution benefits to the doubled contact patch, though I’m not really certain of that at a glance (and it’d be inferior to a properly balanced or suspended drivetrain regardless).
I’m not familiar with FTC’s constraints, so I can’t speak to that. For FRC, it looks to be a fair bit of extra weight and space for an edge case trade-off, as well as all the caveats about mecanum generally in a post-COTS swerve world.
I’d love to hear what drove this decision for 1986, and if its impact over a traditional 4 wheel mecanum was measurable. At a glance, those look to be 4" mecanum wheels, which are smaller than what I remember most FRC mecanum bots using.
Team 1986 did a Behind the Bumpers interview in 2017, so you can at least hear the reasoning they had for their 8x mecanum design used in FRC. Quick summary from the video below: more consistent strafing, good weight distribution, more even wear on wheels.
In some situations, motors can provide so much instantaneous torque, that the wheels will slip in place when trying to strafe. This results in slower strafing speeds. If you’ve ever seen an FRC bot that had a mecanum drive which seemed to turn and drive in straight lines faster than it could strafe left and right - this is what was happening. Due to a variety of factors (mostly friction/rollers binding) it is harder for wheels to slip going forward / backward as it is for them to slip while strafing.
The benefits of 8wd mecanum aren’t really about pushing power or anything. Adding extra mecanum wheels is one of several things that can be done to limit wheel slip. With two points of contact per corner, it’s harder for that instantaneous slip to occur. This is highly dependent on the exact wheels, motors, structure, weight, etc. of the robot though and is not universally necessary or better - if anything it is probably a niche benefit in specific circumstances.
There are other ways to limit wheel slip which may work even better with less mechanical complexity and weight. Closed loop speed control of each wheel, traction control, etc. is not difficult at all to implement and can similarly improve mecanum performance, by reducing power to motors that are spinning a wheel “too fast” after it has broken traction with the ground. I would much sooner implement control like this before going to 8 wheels. Make sure the problem you need to solve is actually a problem first, you know?
I have no context for FTC specifically, but my guess would be that good software control of mecanum wheels would make using 8 of them unnecessary.
FTC actually has good mecanum wheels that use ball bearings instead of bushing which allows them to strafe extreme well with minimal binding and roller losses. All the mecanums that are popular in FRC are bad because they use bushings.
