Best Drivetrain

[EricH]

Originally Posted by Peter Matteson

You also need to take tread wear/how frequently you change tread into account when setting the drop.

If 330's tread wears too far, they change their wheel.

That's one other thing. There are these devices that McMaster sells called pneumatic casters. Take the wheel out and it's pretty high traction. You'll need your own hub, but that isn't too hard.

[Pat Roche]

One solution that hasn't been mentioned yet is using six wheel drive without the drop wheel and a reduced friction wheel on the outsides. We used this setup this past year and it worked out quite well (that combined with inset corner wheels) after some practice. For the "scrub wheels" we used IFI's with worn out rough top flipped over so the backing was contacting the floor. I found this to be ultimately superior to omni wheels for two reasons. The first is that if you are tipped forward or back they offer more traction that an omni wheel does. The second is while turning it adds just enough scrub to slow the turning down to a reasonable speed (reducing the "squirrellyness"). It does however reduce your overall pushing force per wheel due to your weight force being divided over 6 wheels vs. 4 and also due to the lower friction on the corner wheels. That said, we deemed that trade-off acceptable for the maneuverability and over power loss of the system.

[Nikhil Bajaj]

Originally Posted by Lil' Lavery

FIRST Champions:
2005:
330- 6WD, higher traction center wheel (possibly dropped)
67- Swerve
503- 2WD w/ omnis (correct me if I'm wrong)

2006:
296- 2WD w/ casters
217- 6WD, lowered center wheel
522- Treads

2007:
190- 6WD, lowered center
987- 6WD, lowered center
177- 6WD, lowered center

2008:
1114- 6WD, lowered center wheel
217- 6WD, lowered center wheel
148- Swerve drive


correct me if any are incorrect

No corrections to be made, but I WOULD like to point out that taking the fact that there are more 6WD drive, etc. teams on this list as proof that it is a better way to drive and that there is no benefit to more exotic types is somewhat suspect logic and silly generalization.

The reason I say this is that lets say that in the data for 2005, there are what, 800 or 900 robots? Sure, hypothetically. But lets say there were 30 robots with crab drive built that season. Well, in that case, one could say that swerve drive is over-represented, and that obviously, building a swerve drive will afford you a better chance of doing well and getting in the finals, because while only 30 teams had swerve drives, one of them made it (1/30 odds) while then there were probably around 200 robots with 6WD, only two of them made it (1/100 odds.) I'm not saying this data doesn't have ANY merit, but come on guys, lets be responsible in our generalizations. In order to use this kind of analysis, you'd want to make it more statistically valid by either opening it up to all the regional and championship division winners of a particular year.

I've designed three drive trains in my tenure in FIRST, 1 4WD tank with omni (6-motor), 1 crab drive (4 pods, 2x2 chained together), and this year's 6WD with AM Supershifters. I have also given a presentation on drivetrains with Bill Beatty.

But what I would recommend to ANY team is that you shouldn't say "OH THIS DRIVETRAIN IS THE BEST 6WD 4EVAR" and be done with it. You really need to take some time and think about your design requirements--what you want the robot to be able to do, and then build a drivetrain to those capabilities.

The other thing you need to consider very closely (and this is echoed in the ANSI#25 vs. ANSI#35 debate) is your manufacturing and designing capability.

If you are a team that works in a garage with a hacksaw and a couple power drills, you are simply not going to be able to build a lot of really exotic drive systems without outside help. The reason is that many of these systems require tight tolerances (if you want to build your own gear reductions effectively) and if you are off by even a minuscule fraction, you can bind the whole thing up. If you have the capability to do this kind of stuff and build exotic systems, I actually very much encourage you to do so as long as you feel it is in line with what you want to do with your robot.

In terms of DESIGNING capability, if your team has lots of experience building drivetrains, have at it, do what you want. If you are a rookie team, I might encourage taking it easy for perhaps the first year and going with a simpler drivetrain and focusing on manipulator design and also programming. You can do a lot of amazing feats programming even a simple skid-steer robot. But once you have been in the community for a season, even a single regional, you will start to see all the stuff that has been done and is being done--you'll be more familiar with the "state of the art" and be that much more experienced, and ready to go for the next year. Then build something you think will be neat like a swerve or a linkage drive (woo Winnovation!) in the off-season, make sure it works like a charm, and then implement it on your season robot. This way, if you don't get it to work in the off-season, you can still just build the robot with the kit-drive or something else that you KNOW will work. I encourage this methodology for veteran teams too! Plus, you can have a second, better iteration of the design for the season with all of the bugs corrected.

Finally, whatever drivetrain you build or choose to build, learn it inside and out. Learn EVERYTHING about it, how it feels when you drive it, etc. Build SENSORS into the design. Try to do some modeling (mathematical, or build a little replica out of wood, or even a little Vex robot) so that you know how it can move and you can think about how you will control it.

Autonomous driving is only going to get more important guys, lets not kid ourselves.

So I guess, I can sum up my comments as follows:
1.) There is not necessarily a BEST year-after-year drivetrain.
2.) Decide on your drivetrain including your robot goals, manufacturing capabilities, and design experience as parts of the equation.
3.) It is awesome to be innovative and unique (I LOVE INNOVATIVE AND UNIQUE) but please do it in a safe fashion so that what hits the field isn't a janky prototype but a second-iteration. This will help your team, all other teams, the spectators, etc.
4.) If you're a rookie, you need to take a GOOD LONG LOOK at your capabilities, and it may be wiser to take it slow and learn the ropes, and build something phenomenal in the next season.
5.) Whatever you do, think about the design carefully and take controls into the deepest consideration.

[chessking132]

Personally I like a 4 wheel tank drive. With two cims motors going in too a gear box with direct drive shafts. This may not be the best but it is my personal favorite.

Matthew Simpson
Team 75 Driver

[R.C.]

What is the advantage of a direct drive?

[AustinSchuh]

If your chains break, the wheel that you have direct driven still has power. And at least for us, it means one less chain to tension and design for. With that in mind, if you are going to go with direct drive and then run chains to the rest of the wheels you want power to in the drivetrain, drive the one that will be transmitting the most power to the floor, and/or will yield simpler chain paths, and/or will be the one you want to still be connected to your drive motors if your chains fail.

[EricH]

Originally Posted by rc_cola1323

What is the advantage of a direct drive?

One wheel is guaranteed to work.

The problem is that the wheel sends its load straight to the gearbox. It's also a little trickier to do. It can be done, but it takes a little doing. I think 254 has done it in the past. You might want to talk to them a bit.

[AustinSchuh]

Originally Posted by EricH

I think 254 has done it in the past.

I know they have been doing it at least every year since 2004. I don't know about before then though.

[Peter Matteson]

Originally Posted by EricH

One wheel is guaranteed to work.

The problem is that the wheel sends its load straight to the gearbox. It's also a little trickier to do. It can be done, but it takes a little doing. I think 254 has done it in the past. You might want to talk to them a bit.

The impact loading on the gear teeth isn't reduced by a final chain reduction or spread out across a chain wrap around the sprocket. Theoretically you can fracture gear teeth easier with this type of loading.

I've never run the calcs for a drive train to what the loading for this is, but I have seen it some of the other gearboxes we've built that don't have a clutch or other slip mechanism in them.

[Brandon Holley]

Originally Posted by EricH

One wheel is guaranteed to work.

The problem is that the wheel sends its load straight to the gearbox. It's also a little trickier to do. It can be done, but it takes a little doing. I think 254 has done it in the past. You might want to talk to them a bit.

125 did it this year too.

We extended the center wheel shaft into the gearbox where it was coupled via gears to the rest of the cluster. It worked out really well, although it was a bit tricky to pull off.

[XXShadowXX]

if your using chain drive and something hasen't yet broke, something soon will..

[MrForbes]

Originally Posted by XXShadowXX

if your using chain drive and something hasen't yet broke, something soon will..

Not if you designed, built, and maintained it properly.

[Lil' Lavery]

Originally Posted by Nikhil Bajaj

No corrections to be made, but I WOULD like to point out that taking the fact that there are more 6WD drive, etc. teams on this list as proof that it is a better way to drive and that there is no benefit to more exotic types is somewhat suspect logic and silly generalization.

The reason I say this is that lets say that in the data for 2005, there are what, 800 or 900 robots? Sure, hypothetically. But lets say there were 30 robots with crab drive built that season. Well, in that case, one could say that swerve drive is over-represented, and that obviously, building a swerve drive will afford you a better chance of doing well and getting in the finals, because while only 30 teams had swerve drives, one of them made it (1/30 odds) while then there were probably around 200 robots with 6WD, only two of them made it (1/100 odds.) I'm not saying this data doesn't have ANY merit, but come on guys, lets be responsible in our generalizations. In order to use this kind of analysis, you'd want to make it more statistically valid by either opening it up to all the regional and championship division winners of a particular year.

I was not trying to suggest that any drive-train was superior, I was just supplying facts. Notice how I didn't even post any conclusion, just raw data. Some may have drawn erroneous conclusions from this, and I apologize for that, but that was not my intention. It would be silly to judge this data without full numbers on the amount of teams, teams with each drive, and other data. But, sadly, most of that data either doesn't exist, or would be next to impossible to compile, and to make any assumptions about it could and would lead to equally erroneous conclusions. It would be plain stupid to ignore the manipulators, software, electrical reliability, drivers, strategy, and alliance formations of each of the champions as well.

Take it with a grain of salt. It isn't showing the relative superiority of any drive method, rather showing the drives that a handful of successful teams have selected and done well with, for whatever reason, since 2005. Strong teams with strong engineering principles selected these drive-trains for various reasons, that's all I'm saying.

[M. Mellott]

Originally Posted by EricH

One wheel is guaranteed to work.

The problem is that the wheel sends its load straight to the gearbox. It's also a little trickier to do. It can be done, but it takes a little doing. I think 254 has done it in the past. You might want to talk to them a bit.

We also used direct drive this year for the first time, going to a pair of half-tracks centered along the sides for added stability in tight high-speed turns. Also, we wanted to try something different using our success in past years with tank treads.

I agree with Peter, though--a chain to the drive shaft does offer that protection from impact loads, especially with the higher-speed gearbox that we went to after our first regional. Those stresses may have transferred to the tread belts and may have been the cause of a couple stress fractures. We were fortunate that we never lost one during a match--they were found during pre-match checks.

[JesseK]

After a few days of watching TBA and reading some of these posts, I'll make perhaps the only valid & incontravertible sweeping generalisation for FRC:

The best drive train is the one that works every match.

This includes all of the subsystems such as electrical wiring, pneumatics (if you shift or have linkages), voltage of the battery at the start of the match, and the control system the drivers use. These concepts should be drilled into rookie teams just as much as any other drive train concept. They are just as important since (well, without voodoo magic) the robot can't move without them.