Quote:
Originally Posted by JVN
We refer to the 5-omni setup as something called a "Slide Drive". When you combine a slide drive with 4 drop down traction wheels you get what we're calling "Nonadrive" (not Nano Drive as Mike Copioli seems to think...)
Why do we use a slide drive?
Here is the quantitative part:
1. This drive allows for the front and rear wheels to be "linked" without needing a ball differential or anything like that. If one wheel comes off the ground, you still have the full power available on the other wheel.
2. We have 4 motors worth of power pointing forward/backward at all times. With a Mecanum drive, you only get part of this. (The mecanum drive has better side-side power than our 1-CIM, but we don't care.
3. The "slide drive" part of the Nonadrive is 100% intuitive to control, with ZERO programming. This is not meant to discount the efforts of our incredible programmers (148 and 217 have some GREAT ones), but the slide drive can be fully utilized with default code, and its control would be identical to Halo, Call of Duty or any other FPS videogame.
Here is the "not so quantitative" part:
This will be a controversial statement...
In all of my competition robotics experience I have never encountered any drivetrain (swerve, mecanum, or omni) that drives as well as our slide drives. Maybe it has something to do with a lack of programming. Maybe it has something to do with a driver's mental block. I don't know. What I do know is that you put a slide drive in the hands of a good FRC driver, and they'll be almost immediately doing maneuvers that make your head spin.
Before you knee-jerk and reply, remember who you're talking to. We understand design tradeoffs. We understand your value propositions may be different than ours.
Honestly, I expected more people to ask us "Why the heck did you do THAT instead of just doing a swerve or mecanum drive?" My answer is... "If you built one and drove it, you'd understand."
|
I think you may have misunderstood what I was asking, though the more I read your response the more it seems like you might have gotten it.
Many, including you, have seen my personal feelings about [pure] mecanum drives (you even commented on my facebook rant). I wasn't suggesting comparing the nonadrive to a simple mecanum.
But rather I was comparing the "slide drive" portion of it to a mecanum drive. You could replace the four omni-wheels with mecanums and still be able to accomplish any strafing with equal results.
You are also correct about a mecanum drive losing power when wheels lose contact with the ground (barring, as you mentioned, differentials of some sort). This isn't a major issue with chain/gear-linked "sides" of drive trains.
Granted, I was operating under the (apparently incorrect) assumption that you would be operating in the traction wheels drive configuration for most forward/reverse driving (though even a "2+2" also applies to my assumption). The "slide drive" does, in fact, move in forward/reverse more efficiently than a mecanum drive.
Operating under that assumption, I was basically considering the two drives in terms of strafing and rotational ability, in both of which the mecanum drive is advantageous compared to the slide drive.
Knowing the characteristics of the traction wheel set-up, it really comes up to how you want to drive the machine in order to justify any engineering trade-offs. Given the way you stated you wanted to drive it, I absolutely understand the reasoning behind your design choices.
Also note, I never was attempting to call into question your design choices, but rather just seeking more information about why you made the ones you did. And there were indeed some factors I didn't take into consideration, and learned more about your design philosophy and your strategy for playing the game.
Quote:
Originally Posted by JVN
Nope. The slide wheel is positioned near enough to the CG that she moves sideways almost straight as an arrow. The same can be said for forward and reverse, even without the traction wheels down. No complicated control needed. For testing (before we had any of the other electronics wired up) I took a VEX PIC Microcontroller from inventory, hooked the Victors up and we drove it with default code. Really.
|
I was more wondering about diagonal motion. With different total torques in either cardinal direction, an input of a "45 degree motion" isn't going to immediately produce a 45 degree motion from the robot because the acceleration in each cardinal direction will be different. Granted, this could be accomplished by using different gearing on the wheels, but then the total speeds would have to be reduced (on the faster geared wheels only, obviously) in order to produce a 45 degree motion.
This really isn't a big concern at all, I was just curious if you felt it was an issue. It's easily adaptable by a skilled driver (though we all know that kid down there can't drive his way out of a paper bag, so maybe he might need programming help
).
As a side note, how is she going to be controlled? Given the amount of options available, I'd assume you're using some sort of game controller in order to have enough buttons readily available to the driver.