I would assume that shifting would get you finer control over the position in low gear, but a very fast high gear, which may or may not be necessary depending on your strategy.
Fast is good if you are trying to get the cans in the middle quickly, for example.
Fast is good, yes, but that doesn’t always mean a higher top speed. If you’re starting in between the auto and landfill zone, going for the center of the step, you’re driving five feet. At distances this short, you can easily gear your robot so quickly that you actually lose the race to a slower top speed robot factoring in acceleration. Even if it’s just a few milliseconds slower, you’re giving up low speed precision to have a high top speed you’ll almost never actually reach. Acceleration models are available on CD to help verify this sort of thing, just keep in mind that the constants people use in these models vary.
What I’m trying to say is that you can’t really spout out old “rules of thumb” as fact this year, in a game that isn’t open field and for a different drivetrain configuration. Especially when you have never built a mecanum drivetrain. If you give advice without qualifying your lack of experience, teams may be led to make decisions they otherwise wouldn’t that could end up hurting them. I’ve made this mistake before and i’m sure teams have made missteps as a result, hopefully minor.
This might be totally made up, but I’ve also heard that strafing performance decreases at higher speed gearing. This is probably a function of roller efficiency, if this is the case. I lack experience in this specific aspect of mecanum drives, though.
I was responding to GeeTwo’s implication that reduced top speed gearing requires using a shifting gearbox. For this game, that proposition is arguable. See Chris’ post.
[EDIT: The below comments are referring to variable-speed gearboxes, not gearboxes in general. -Thromgord]
For this year’s challenge, gearing systems are completely unneeded to control speed. Last year, our team had quite a bit of success by keying buttons on our controller to ‘reduce speed by 50%’ and ‘increase speed by 50%.’ It allowed us to pick up hypothetical* exercise balls with great precision, speed to the other end of the field really quickly, and then get to the precise place to shoot the hypothetical* balls into the goal. It wasn’t even difficult to code!
*Our robot never actually did this, which is why I’m considering this “hypothetical.”
It wasn’t difficult to program, and while it added no extra torque, no extra torque was needed that year. Except for when we were getting pushed around backwards. You won’t require much torque at all this year, though, so it shouldn’t be a problem.
By the way, one comment I haven’t heard is that mecanum wheels don’t work as well under very light loads; we built a mecanum 'bot in the off-season two years ago, and without putting about 40lbs of weight on it (initial weight was about 60lbs), the mecanum wheels wouldn’t work properly (it had difficulty going in even the vague direction we wanted to, and after blaming the programmer like we always do, we added more weight and it helped tremendously). I know that you’ll have no trouble making a heavy-enough robot, but just be sure to keep this in mind.
I just realized that our team wasn’t all that great last year. Ah well, we’re off to a good start so far… 
I am factoring in acceleration. I use a drivetrain calculator that was on CD a year or two ago to determine the effectiveness of different speeds. Of course, the “speed loss constant” is somewhat of a fudge factor (80%), but that’s hard to avoid. If somebody could supply the Kro and Krv to use, that would be pretty useful, as right now I’m just using the stock values.
Five feet is more than enough for a 15fps bot to gain the advantage over a 10fps bot, and as the distance gets larger the 15fps bot gets faster. The crossover point is at ~2.2ft, when both a 10.5fps and 15.0fps bot take ~.41 seconds. Of course my constants are probably throwing it off a little, but if you could supply your own calculated numbers that would be really helpful.
Because we will likely need to drive quickly at the start of the match/ auton, I feel like the minor time advantage we get at distances lower than 2ft is not worth it, especially because at those distance we probably don’t want to go too fast anyway.
Our robot last year weighed in at 75 lbs with the battery and bumpers installed, and had no issue whatsoever with driving in any desired direction. Was your robot properly balanced so that there was even weight distribution between all 4 wheels? Did you use a gyro for field-centric drive?
It almost sounds like you are advocating driving the wheel at the same rotational speed as the motor.
Is that really what you meant?
I am well aware that this can be done.
The question I was asking Thromgord is whether his post was to be construed as advocating doing such a thing for the 2015 game.
Yes for your first question, no for the second. If I remember correctly (which I likely don’t), the wheels were spinning in place a lot, and when sideways translation was tried, the robot still just had its wheels spinning in place. Maybe if the wheels were turning slower, or maybe if our carpeting was different… 
No, not at all. I was intending to state that a variable-speed gearbox wasn’t needed. I personally thought that I was being reasonably clear, but then again, I’m new here. Changing my first post to be a little more clear, now.
Actually, I was presuming that since Hoover put the throttle control on the Z slider, that it was going to be adjusted during the match, so that you could switch (or actually grade) from “full speed” to “parking maneuvers”. To reduce speed through gearing while allowing the high speed would require a shifter, no?
Yeah, ya know, I am not sure we are seeing how small of space we are in this year. First off we have half the court. Then quite a bit of the court is occupied by playing pieces. And then sharing this space with two other full size robots.
Then there will be things that will be detrimental to a team
- Out of control driving into objects they don’t want to hit.
- Out of control steering into objects they don’t want to hit.
- Not having the speed to get to objects
- Not having the finesse to manipulate objects
- Competition for space or objects by other robots on their own ‘alliance’.
All this shows that the game goes fast, slow, fast, slow…
The ideal robot should be able to move in all directions either fast or slow. If mecanum is used then does that mean that shifters are also needed for all 4 transmissions. What a nightmare if just one of the fails, then 2 above is probably going to happen to that team. Shifters are out of the question for out team anyway so my solution then is voltage regulation, i.e. limiting top voltage in a way convenient for the driver (to be determined).
What does “high speed” mean to you, in the context of this game? How many feet per second?
The point I was making (for the benefit of rookie teams reading this thread) was simply that reducing top speed by limiting motor voltage does not provide the same kind of fine control of slow small motions that is obtained by reducing the top speed of the bot via gearing.
Gearing a bot for a top speed of 18 fps and limiting the voltage to 6 volts does not provide the same fine control of slow motions as gearing the bot for 9 fps.
I suspect that most of this game will be played at speeds of less than 1fps. In this context, 2-4fps is high speed - what you’d want to use to carry that stack of totes over to the scoring platform - though you may make a dry run or a single game piece trip faster. Except in the first 30 seconds of the match, anything over 8fps is likely to cause more un-scoring than scoring.
Granted, a shifter is better at fine maneuvers in low gear, but its real purpose is to improve acceleration at low speeds.
I’m having trouble parsing your post.
Let me rephrase my question: For a robot with a single-speed gearbox, what do you think would be the correct top speed gearing for Recycle Rush?
Granted, a shifter is better at fine maneuvers in low gear, but its real purpose is to improve acceleration at low speeds.
For Recycle Rush, improving fine control of slow motions is a “real purpose” of gearing for low top speed… even arguably a more important purpose than acceleration.
More than I’ve seen in a number of years, ideal top speed is dependent on what a team’s strategy is. Our strategy will require getting to a couple places on the floor quickly, so we have geared for a “high” top speed, which is still significantly slower (11 fps) than last year (17 fps); we are also forgoing the shifter. If another strategic approach is taken, 8 or 9 fps maybe plenty or even too much. There are legitimate strategies this year that require 0 fps.
To answer Hoover’squestion about the Toughboxes, I don’t think you will have any problem using them with mecanum. Good for you and your team, for using the resources you have in a smart and creative way.
The AM14U2 is a nice package. But for FRC purposes what I would really like to see is a “toughbox multi”. It would be similar to the mini but would come with alternate gears to give a team a choice of speeds.
When I began this reply I was going to say that I wish this chassis came with shifting transmissions but I know that is more expensive.
In either case, changeable gearing is a great teaching tool. The first year our mechanics mentor drew gearing on the board and asked us which would be for speed and which would be for power, more than half of us (mentors included) got it wrong.
Of course, it depends on your game strategy. For a robot that stacks on the field and carries the stack to the platform, and does not plan a bunch of trips back and forth from the alliance wall to the landfill, I think 6fps is about right. We plan to either carry a bunch of crates from the human station to the nearest platform, or clear the landfill (at least of the upright totes; we’ll work on a tote-flipper after we get the forklift functioning well enough to tune the software practice driving). That said, we’re gearing at about twice this speed (13fps), because this is our team’s fourth season, and we’ve mis-analyzed game play three times so far. We want the flexibility to adapt to the game as it turns out to be played as opposed to being optimal for what we expect.