To Shift or Not To Shift, That is the Question

[lukevanoort]

Often you will put out the same amount of maximum pushing force in low and high gear because generally your wheel/tread material stays the same. (shifters with designs like 67's in 2004 being the exception)

The benefits of shifting generally are that:
1) The motors are loaded less in low gear
2) You can go very fast for covering ground
3) You can go very slow for accurate driving

Ideally, your wheels would slip at about 40A current draw from each motor so that you don't pop a circuit breaker and lose mobility; thus, with the most popular tread material (CoF 1.0) this means you (assuming a four chiaphua drive system) must go at most about 7-8 ft/s. 7-8 ft/s happens to work pretty well as a middle ground speed, but if you have higher traction tread such as as SBR Roughtop incline conveyor belting, that number becomes something more like 5-6ft/s. That is too slow for many teams; thus, they shift. The shifting then frees you up to have both ideal current draw in low gear and a blazin' 16ft/s high gear. As you can see building an effective two-speed drivetrain doesn't only consist of making (or buying) a good shifter, but also choosing the correct ratios for your drive system. (arguably, the latter is the hardest of the two issues)

[Warren Boudreau]

Here is a rule of thumb that I have developed over the years.

You want to gear your drivetrain to go as fast as you can but still spin the wheels/treads when you are up against a wall.

<EDIT> If you want to go faster, you need to a multi-speed gearbox.

[AndyB]

I think the benefit of a two speed transmission was best exemplified by teams 1114 and 330 last year (I'm sure there were others as well). Both used multi speed transmissions and both were very great robots. They used speed to maneuver around the field and beat defenders who were playing closely. They used low gear to push around robots that played closer to the rack.

A two speed transmission will make it easier for the driver... I am not saying that it will do the trick by itself, but if you have a driver that can handle it, you can really work it to your advantage.

[Richard McClellan]

Has anyone ever experimented with having a sort of automatic transmission system? So, when the driver moves the joystick between neutral and halfpower, it's in low gear, but then it will automatically switch to high gear if you put the joystick to full forward? Or does everyone use a manual transmission controlled by a switch?

[Cory]

Originally Posted by richardmcc2

Has anyone ever experimented with having a sort of automatic transmission system? So, when the driver moves the joystick between neutral and halfpower, it's in low gear, but then it will automatically switch to high gear if you put the joystick to full forward? Or does everyone use a manual transmission controlled by a switch?

Team 33 did this with their 4 speed transmission in 2004. I think team 368 used a similar design in 2005. You might want to search for the threads about 33's transmission, as i believe they discussed the autoshifting in depth

[AndyB]

I'm not a driver, so keep that in mind, but it seems intuitive for me that it would be nice on the joystick to have it default to low gear and then when the trigger or a button was held, it would shift to second gear, when it is released, back to first... kind of like a turbo button.

I say this because in my opinion, low gear is for precise driving and motion and high gear is for moving quickly between two points.

[Dan Petrovic]

You don't necessarily need the low gear for precise movements.

Last year we had a single-speed transmission that was always running at 75% power, unless we pulled the trigger, then we put out 100% power.

Same idea, a little bit simpler, however, you do lose some pushing power with that 75%.

In the past year, it became quite clear to me that you don't necessarily need to be able to push, you just need to be able to stay still when you want to. If you can get enough momentum, then a single-speed transmission can knock someone off target. You do have abide by the rules of no "long distance, high speed ramming".

[AdamHeard]

Originally Posted by InfernoX14

You don't necessarily need the low gear for precise movements.

That's true, but it sure helps....

As a former driver, based on experience with our previous years robots, the jerkiness of trying to make precise turns with a drive geared at 7-8fps (or higher) was kind of annoying. Now, I know a better designed drive (or one with less traction.....) could have helped, but that isn't all ways as easy as just bolting on a 2-speed.

[wilmo]

we had a 2 wheel drive with the AM Gen-2 Servos. We tried shifting once, and we almost fell over from the sudden burst of speed, wasn't worth the weight for the 07 game, rather have more tanks for air

[DonRotolo]

Originally Posted by AdamC

that finds a balance between power and speed.

You mean a compromise, not a balance, because that's what it is. A shifter eliminates much of the compromise...but that might not be important to your strategy.

[Daniel_LaFleur]

The answer to this question has to do with your strategy and the design of the game.

Last year Our strategy was to defend the opposite side of the rack and to come back to open our ramp / platform at the end of the match.

We were not too concerned with speed because all the scoring was in the middle of the field. That meant that the scorers had to come to us.

Our solution was tank steering, 4 wheel drive, 1 small CIM per wheel, Banebots single speed transmission with 16:1 conversion, and soft 7" pneumatic tires (CoF ~1.3). This gave us a speed of ~ 6 fps and ~ 175 lbs pushing force ((120 lb robot + bumpers + battery) *CoF). We adjusted the pressure in our tires so that the tires would just break free at stall current.

In all of our compitition matches (25) we only met one robot that we could not push (checkmate, team 40).

I'll say it again, don't put in a multi-speed transmission just to have a multi-speed transmission. Design in what you need to be successful in the game using the strategy you chose.

[Ian Curtis]

Originally Posted by Daniel_LaFleur

The answer to this question has to do with your strategy and the design of the game.

Last year Our strategy was to defend the opposite side of the rack and to come back to open our ramp / platform at the end of the match.

We were not too concerned with speed because all the scoring was in the middle of the field. That meant that the scorers had to come to us.

Our solution was tank steering, 4 wheel drive, 1 small CIM per wheel, Banebots single speed transmission with 16:1 conversion, and soft 7" pneumatic tires (CoF ~1.3). This gave us a speed of ~ 6 fps and ~ 175 lbs pushing force ((120 lb robot + bumpers + battery) *CoF). We adjusted the pressure in our tires so that the tires would just break free at stall current.

In all of our compitition matches (25) we only met one robot that we could not push (checkmate, team 40).

I'll say it again, don't put in a multi-speed transmission just to have a multi-speed transmission. Design in what you need to be successful in the game using the strategy you chose.

Trinity learned their lesson about being pushed around in 2006. You can have the best darned scoring mechanism on the field, but if you can't stay in scoring position, then it doesn't matter.

Also, the shifting debate really does come down to *gasp* robot strategy. In my experience, the better you are at scoring, the faster you need to go, for two reasons. If you could be scoring, why waste time plodding around the field? Second, if you're good, and people know it, you will be a target, and its much easier to rush out of their way than to have to fight through them. Because even if they don't manage to stop you, if you spend time fighting through them, you're wasting time you could be using to score.

Take Buzz (175) at BAE for an example. During the final practice matches on Thursday, they were putting up over five tubes with ease. Obviously, this got noticed, and during the course of 8ish seeding matches they managed to only score 15 tubes. They had an excellent drive team as well, but they were just pounded non stop, and couldn't score well through all the defense. Of course, when alliance time came around, their AC made sure that defense wasn't a problem, and they essentially walked through the elimination tournament.

On the other end of the spectrum, don't make a useless gear for the game. Last year in our top gear (we had 3 speeds) we could cross the field in about three seconds. Since the rack was in the middle of the field however, it wasn't really useful, and just resulted in a banged up grabber.

Most games require precision however. Let's review the last 4 games, shall we?

2004: Line up with 13" playground balls. Line up with 3'(ish) wide chute to deposit them.

2005: With a 5 lb object swinging wildly through the air, line up about a 2' square area over a tetrahedron.

2006: Line up with a 4' hole in wall

2007: Line up a circle with a foot radius over the end of pole.

Games require precision and speed. They typically lie at either ends of a spectrum. Shifters are one way to get there. There are other ways to get relative precision in code however, which is less of an investment in time and money than buying/making two speeds, and requires a smaller time investment than shifters.

[Billfred]

I have to echo the ideas above: it's all in your strategy. We wanted to be a fast arm robot, but I've had a hand in far too many flying bricks over the years for the robot to be unable to push. AndyMark Shifters allowed us to hit that sweet spot (ours, as I recall, were set for about 12 and 5.5 fps through sprocket and wheel choice).

What we do for next season will depend on a lot of things, at least on my end:

1) The game next year. Obviously, a robot that's going rock-crawling will not be using Uppercut's drivetrain.
2) Durability of the 2008 KOP gearboxes. If we can do the job with the kit gearboxes, I'll be more than happy to put them on. But if we had another year with gearboxes showing problems, I'd start investigating alternatives fast.
3) Resources available to us. If we need the $500 difference between AM Shifters and AM Gearboxes for other parts of the robot, decisions have to be made, particularly if item #2 is in question. (Ideally, of course, this would not factor into the equation.)
4) Familiarity. Given the choice, I'd much prefer to use a system someone else has had all the frustration fun of debugging already. ("Someone else" can be us.)