Re: CVT, which teams are planning on using them for next year?
 

CVTs would nead a pretty amazing program to utilize it properly. any ideas of how a program would guide it?

Re: CVT, which teams are planning on using them for next year?
 

I see what you mean, Stonefan. I was going to say: what if you needed to change ratio's really fast? but then again, I guess you could finally put use to that Z-axis on the joystick, or just make your own custom switch with potentiometers.

I sort of had that question myself. I know that in cars, CVT's start to change around their reduction ratios once the engine starts to reach the peak of its torque curve -- but in the case of an electrical motor, that torque peak is at stall, and stalling motors isn't a very good thing. I think the basic algorithm would go: shift up if current is increasing and wheel speed is less than some constant X, which is what the wheel speed should be when that amount of current is going through the motors, so you shift up while accelerating. If the current is increasing and wheel speed is decreasing, then shift down.
And there's more for when the robot stops, is slowing down, etc...
Sounds like a fun project for a programmer.

Re: CVT, which teams are planning on using them for next year?
 

I was thinking for the controls if you did tank style steering and had the joystick 1 for a side, that you could possibly adjust the ratio with a pedal that you controlled with your foot. The harder you push the higher your belt goes up the pulleys.

Re: CVT, which teams are planning on using them for next year?
 

That would actually be a pretty cool way to program it. Might take some time and math to get it well though, but that would be cool and seems reliable, now that I think about it, it would be easier for a novice driver. It would be totally automatic and so he could focus on driving around.And ya, also that would be pretty neat to instead of having to use the buttons to change the gear ratios, using the 2nd axis of the joystick that the driver isn't using. That would be even easier if he had to use manual. But I was wondering, is there a way using some kind of sensor or with programming to detect current, or would you just go with shaft encoders to detect speed?

But ya, keep the ideas rolling.

Re: CVT, which teams are planning on using them for next year?
 

I believe that current sensors were shipped in the kit, or at least they were in 2004, I can't remember if they did in 2005 (I never bothered to look into the electronics, the kit bot was interesting). I'd go with both current and shaft encoders, I'd think you'd need to at least have two different sensor values to compare so the program can give the right gear ratio.
Also, when I meant Z-axis, I meant the 3rd axis, which is that little throttle thing that looks like a calibration wheel, but isn't. I don't think I've seen any team use it; on the new joysticks the thing is really small and hard to quickly manipulate, but on those old joysticks, the thing was pretty huge.

Re: CVT, which teams are planning on using them for next year?
 

That was true a couple years ago, but as others have said, not anymore.

A CVT is the only automatic transmission option on a MINI Cooper. It's a MINI Cooper - not a big car, not a big engine, but the transmission is really small (not light though...). I tinkered with the idea of buying broken ones from BMW for a robot but decided that it's not the least bit necessary. The company that manufactures the transmission and sells it to BMW for MINIs has a really really nice whitepaper on every aspect of their transmission and the 1.1L MINI engine combo.

The CVT on the MINI is really the only way to go if you like feeling your left leg after driving a bit - friggin' clutches :rolleyes: (the other option is a 6-speed manual). The belt I believe is entirely metal and there's not the least bit of "rubber-band syndrome" found in early CVT cars. You can select between 3 different shifting algorithms while driving, and if you don't value your warrantee, you could hack it and come up with some yourself (I would, but it's not my car...). I could extoll other aspects of the MINI too (now can use regular instead of premium because of a software upgrade :ahh: :D ), but most of them are true of many modern cars.

George

Re: CVT, which teams are planning on using them for next year?
 

On a different topic, if one was using a CVT with electric motors (itself a proposition only necessary because of voltage limits in FIRST - look at the majority of electric vehicles), a wise algorithm to use would not let the driver have any control at all... Merely, joystick position would correspond to a specific wattage input (measure with ammeters), with accelerometers indicating your speed so that the RC (or a suitable likeness) can decide the best gear ratio and throttle position to take the fullest advantage of the power. This would work in many scenarios, whether the robot is shoving or accelerating down the field.

not that I've thought about it for the past 4 months or am a programmer or an intelligent person or anything ;). Course, you need integrators and voltage-dropping circuits because have you ever tried getting a PIC18 to integrate at 400hz realtime even with interrupts (which are hard to implement anyway on the adcs)? coz I have. That was painful. and whatnot but hey, what do ICs cost nowadays? lol, integrating integrating integrated circuits into an integral platform to measure the integrity of the integrals from your integration of the integrating integrated circuits... ahh tonguetwisters and the many definitions of integral...

Hope that helps. ;). PM if interested.
George