Drive Train Choices

[cadandcookies]

Originally Posted by tuXguy15

Our team wants to experiment with different drive trains and was wondering which would be the best to play with. What are the advantages and disadvantages of them?

I (and many of the other people here) could probably fill multiple encyclopedia-sized tomes with the advantages and disadvantages of different drive trains and configurations of drive trains. There's an entire world of custom drive trains that opens up the moment you step outside of the appealing and excellently designed kitbot-shaped box.

I'll reiterate for a moment what KrazyCarl said-- before you even start thinking about what kind of custom drive train you're going to build, consider very carefully your resources-- what kind of budget does this have? How much time is your team willing to spend developing it? What kind of sponsors do you have? What are the lead times on parts from those sponsors (we have a waterjetting sponsor that needs us to give them CAD files at least three weeks before they manufacture, which limits their use during build season)?

Once you've answered all those questions, then brainstorming is a good idea-- if it's just for the joy of designing something out of the box, then you might very well consider octocanum or swerve drive or any of the other myriad of cool custom drive trains out there-- if you're developing it for competition, your goals are probably quite different. If you're developing specifically for competition, in fact, you'll probably want to stick to making a 6WD that you can manufacture and start driving super fast, in addition to being as flexible as at all possible. There has never been a year where you couldn't complete the game challenge with a well-designed 6WD (at least in recent memory).

In terms of awesome projects for drive trains, a lot of great ones have already been mentioned-- an octocanum is a very cool project (in fact that's what my team has been working on). Butterfly is, as mentioned, a nice, relatively simple project (considering that 3928 has so kindly used a good deal of COTS parts and posted many pictures). This late, I wouldn't suggest developing swerve-- you might be able to pull it off, but there's a lot of mechanical and programming complexity there.

Another suggestion would be to CAD all of the above drive trains-- it's good practice and you'll learn a lot about your program of choice. And CAD has that nice thing about being free. If you do that, then you can look at all your designs and pick out the one that makes most sense/ you'd like to build the most.

[tuXguy15]

We just want to try something new so we have some experience during off season instead of fulling around with it during build. I think its safe to say that my team has just about mastered 4 - 6 Wheel Drive. We want to try new stuff to see if we like it and want to use it in a future year instead of the same 4 - 6 drive every year. We just don't know which one to pick to start with.

[Akash Rastogi]

Originally Posted by tuXguy15

We just want to try something new so we have some experience during off season instead of fulling around with it during build. I think its safe to say that my team has just about mastered 4 - 6 Wheel Drive. We want to try new stuff to see if we like it and want to use it in a future year instead of the same 4 - 6 drive every year. We just don't know which one to pick to start with.

IMO there's still a ton of development your team can do within 4-6-8 wheel drives.

Once you've iterated through those, the only one I'd suggest next would be butterfly or octocanum. But, I'd say continue on the route of further refining your drives. Try belts in a 6 or 8wd, try custom gearboxes, try new wheels, etc...

[efoote868]

Originally Posted by tuXguy15

We just want to try something new so we have some experience during off season instead of fulling around with it during build. I think its safe to say that my team has just about mastered 4 - 6 Wheel Drive. We want to try new stuff to see if we like it and want to use it in a future year instead of the same 4 - 6 drive every year. We just don't know which one to pick to start with.

"Which drivetrain should we prototype in the off-season?" is a whole different question than "what are the advantages of each drivetrain?"

If all you're looking for is a simple answer to the former question, I'd recommend mecanum for a few reasons. One is that it can be mastered in a season. Two is that if all else fails, swap out regular wheels and you'll have a drive that you're familiar with. Three is that it will make a good demo bot all by itself, since most people aren't familiar with mecanum wheels.

Hope this helps.

[LeelandS]

Originally Posted by tuXguy15

We just want to try something new so we have some experience during off season instead of fulling around with it during build. I think its safe to say that my team has just about mastered 4 - 6 Wheel Drive. We want to try new stuff to see if we like it and want to use it in a future year instead of the same 4 - 6 drive every year. We just don't know which one to pick to start with.

I'll echo what everyone else has said about making sure your 6wd is perfect. It's important to have a strong "default" drive system. However, I don't think you came here to read a couple of posts that are not helping you make a decision like you wanted! However, you could still experiment by designing or iterating your own gearbox (or other improvements to a standard Tank system)! My team is currently examining this possibility because we recently acquired new machining resources and it would be a cool project and a valuable resource to have in our pocket.

Hm... Well, Mecanum drive is a good one to prototype during the off-season. It's a challenging exercise for your programming and mechanical subteams (which is a reward of itself). It also opens the door for Octocanum drive in the future (which I've wanted to look into for a while). The wheels and gearboxes are going to be kinda costly but other than that it can be mounted to a standard C-Channel chassis, which is nice if you don't want to mess with a special setup system.

I've wanted to check out Butterfly drive for a bit as well (if you were unaware, Butterfly drive is Octocanum but which omni wheels instead of Mecanum wheels, it has the added maneuverability without the strafing and cost of the wheels). It's a fancy system that will also challenge your mechanical and electrical teams, as the development of a good, reliable wheel-switching system could be a bit of work. This is also a fairly new drive system, so there is a lot of experimentation that can go into it because not much has been done with it already. The machining of the modules might require resources beyond the standard team's availability though, so research into a design is required beforehand.

If you don't want to venture too far away from your Tank drive systems, Drop Center 8 Wheel Drive has been becoming increasingly popular. I'm not sure what the strict advantages are to it over 6wd, but there are some teams now (some of whom are very high-leveled) who are swearing by it now, so there must be something that makes it so desiring.

It all depends on what kind of drive train you want to have up your sleeve. If your team really just has an itch to play around with a drive train, then present a couple ideas to the team and see what sticks! You might find one that really interests all the people on your team. Just remember to consider cost, machining requirements, needed man-power, general complexity, and the support the FRC community can provide in the endeavor!

Looking forward to see what you guys come up with! Feel free to share if you make something cool!

[bardd]

Originally Posted by tuXguy15

We just want to try something new so we have some experience during off season instead of fulling around with it during build. I think its safe to say that my team has just about mastered 4 - 6 Wheel Drive. We want to try new stuff to see if we like it and want to use it in a future year instead of the same 4 - 6 drive every year. We just don't know which one to pick to start with.

As a big fan of tank drives, I'd suggest you try the 8WD, with the two middle wheels lowered. I only know one team that does it, 2630 from Israel, but I'm sure it's a common design, I just never looked into existing designs (since I know people on 2630 that does it so when I have a question I just ask them).
The great advantage of this design is that you have a very small wheel base, even smaller than dropped-wheel 6WD, and since you have two lowered wheels, it only rocks when you have high-speed collisions. The major disadvantage is it's way more complex than 4/6WD, and you can't use the KOP to build it (not enough wheels, you need special c-channels since the KOP ones only have one lowered hole).
2630's drive in 2010, first time they did it. It's kinda hard to see but if you look closely you'll notice the external wheels are a bit higher than the lower wheels.

Did you ever do a west coast drive? If not, you can try that too. It too is harder to make than the standard drive, but it clears up a lot of space inside the chassis, which makes component placement much easier, and in some cases allows for better game mechanisms.

Side note: most of 2630's 8WDs were west coast drives, if not all.

EDIT: Sorry, I didn't see the "different wheels" post. I'm keeping the post though, for anyone else who might be interested in trying something new and looks at this thread.

[Akash Rastogi]

Originally Posted by bardd

The major disadvantage is it's way more complex than 4/6WD, and you can't use the KOP to build it (not enough wheels, you need special c-channels since the KOP ones only have one lowered hole)

.

This is not entirely true. An 8wd is easy to chain together even with the kit setup. You'll obviously need the additional hardware and parts. It is very easy to modify the kit chassis with little blocks of tubing bolted on with 3/8" bolt holes milled into them. Select your drop, select your wheel spacing, mill the blocks, bolt them on and you're good to go. You can also drill/mill new holes right into the kit frame, like 2495 did in 2012, but using blocks allows more room for error. I'm sure plenty of teams do this, but 2495's are the only pictures I can refer you to:

http://i.imgur.com/32N82ta.jpg
http://i.imgur.com/Xey74Gy.jpg
http://i.imgur.com/rUbLYc9.jpg

Is there something specific you find to be complex about 8wd? Perhaps I can explain it better if you have a question.

Other than added stability, 8wd allows for the ability to cross obstacles on the field without needing huge wheels. See the top 2012 robots as examples and 2010 robots. I would highly suggest prototyping 8wd and figuring out the best amount of drop to use, wheel spacing to use, and even which wheels work best.

[bardd]

Originally Posted by Akash Rastogi

This is not entirely true. An 8wd is easy to chain together even with the kit setup. You'll obviously need the additional hardware and parts. It is very easy to modify the kit chassis with little blocks of tubing bolted on with 3/8" bolt holes milled into them. Select your drop, select your wheel spacing, mill the blocks, bolt them on and you're good to go. You can also drill/mill new holes right into the kit frame, like 2495 did in 2012, but using blocks allows more room for error.

I'm assuming not owning a mill. I don't know how common they are in the USA, but I know many people there I talked with don't have such equipment, and in Israel only 3-4 teams have them (that I know of, but knowing the teams in Israel I highly doubt a team I didn't ask has one). If you have a mill, than yes, the job's easier.

What I meant by it being more complex is not being hard to design as much as 4/6WD being very easy. I mean, if you have the drive base parts, you can build a drive base without even having to think about it, and I'm not talking only building the KOP by instructions: we've been using KOP parts for our custom drives every year since we started, excluding 2010 and 2013. None of these drives were ever CADed, or even pre-thought of ("let's put plaction wheels, I heard they're better" was usually the most discussion about it before actually building). Only one of these drives ever failed, and even that was because we got the wheels super-late and didn't have time to place them properly.
With dropped 8WD, you obviously can't do that. Not saying it's a good practice to do that, and I really don't like the concept of doing that (the reason we didn't do that this year), but the fact it can be done shows how much easier it is to get a 4/6WD than 8.

[magnets]

If you are looking to get a more competitive drivetrain over the summer, I would recommend improving a six/eight wheel drive. If you are looking to have fun building a cool looking robot, you could try swerve/mecanum. In the past 4 years, there have been almost no good robots with swerve drive.

To develop a good swerve you need multiple seasons/off seasons of development, people who are skilled with CAD software, lots of money, and lots of CNC and machining equipment. Look at team 148, one of the most competitive teams. They have only done swerve once, and that was in 08, when their robot didn't have any other manipulator/features, it literally just drove in circles! After 08, they stated that swerve is almost never beneficial to have.

As for meccanum, I personally have a strong dislike for these horrible wheels. Many good teams automatically cross off teams with meccanum wheels from their pick list, just because they know they will be inferior to a 6wd.

If you are interested in making your 6wd more competitive, figure out how to make it lighter, cheaper, simpler, and more reliable. Make it a two speed (if you haven't already) and play around with different gear ratios to see which works best. You could also try to implement a more advanced method of control such as a gyro that helps your robot drive in a perfectly straight line, or something like team 254's cheesy drive. There is currently no team in FRC with a perfect 6wd, although the cheesy poofs are getting pretty close.

[Jay O'Donnell]

Originally Posted by magnets

In the past 4 years, there have been almost no good robots with swerve drive.

What about 16 in 2012? And 1717 all of the last 4 years? Those are only two off the top of my head and I know there are others.

[bardd]

Originally Posted by Jay Money 1058

What about 16 in 2012? And 1717 all of the last 4 years? Those are only two off the top of my head and I know there are others.

He did say "almost"... There are good examples of swerve out there, they're just very uncommon. 1640 is the only team that I know of to use swerve in the past 3 years other than those you mentioned.

[Andrew Schreiber]

Originally Posted by Jay Money 1058

What about 16 in 2012? And 1717 all of the last 4 years? Those are only two off the top of my head and I know there are others.

2008 - 16, 148 (World Champs)
2009 - 111 (World Champs)
2010 - 1717
2012 - 16 (World Champs), 1717, 973
2013 - 1717, 1640

16 has been swerving for years and have iterated the crap out of it, same with 1640, 1717, and 111. 973 built several in the off season before running it in season. Of interesting note are the teams that swerved and never went back to it:

1114 - 2004
67 - 2004, 2005
1503 - 2007
469 - 2007, 2008


Of particular interest, the Beatty Curse. They haven't won a World Championship since they started swerve drive. (This is a complete joke)

[magnets]

Originally Posted by Andrew Schreiber

2008 - 16, 148 (World Champs)
2009 - 111 (World Champs)
2010 - 1717
2012 - 16 (World Champs), 1717, 973
2013 - 1717, 1640

16 has been swerving for years and have iterated the crap out of it, same with 1640, 1717, and 111. 973 built several in the off season before running it in season. Of interesting note are the teams that swerved and never went back to it:

1114 - 2004
67 - 2004, 2005
1503 - 2007
469 - 2007, 2008


Of particular interest, the Beatty Curse. They haven't won a World Championship since they started swerve drive. (This is a complete joke)

If you notice, all of the good teams with swerve have been at it for years. (111 has had swerve since at least 03, 1640's had swerve since 09, 1717's had swerve for 4 years, and 16's had swerve for as long as I can remember) The reason 16 was so great in 2012 was not because they had swerve. They were great because of their ability to quickly collect balls, and accurately score them. These teams' success is not caused by their swerve drives, but the swerve drives happen because these teams are so good at building robots.


The only exception to this is 148, who successfully implemented swerve in one season. However, JVN has said that he "wouldn't recommend" swerve, and that there are close to no situations where swerve is as beneficial as people say it is.

My point is not that swerve drive ruins robots, but that swerve is VERY easy to do wrong, and if the original poster was interested in making their robot as competitive as possible, they should work on their 6wd.

[Brandon_L]

Originally Posted by bardd

I'm assuming not owning a mill. I don't know how common they are in the USA, but I know many people there I talked with don't have such equipment, and in Israel only 3-4 teams have them (that I know of, but knowing the teams in Israel I highly doubt a team I didn't ask has one). If you have a mill, than yes, the job's easier.

What I meant by it being more complex is not being hard to design as much as 4/6WD being very easy. I mean, if you have the drive base parts, you can build a drive base without even having to think about it, and I'm not talking only building the KOP by instructions: we've been using KOP parts for our custom drives every year since we started, excluding 2010 and 2013. None of these drives were ever CADed, or even pre-thought of ("let's put plaction wheels, I heard they're better" was usually the most discussion about it before actually building). Only one of these drives ever failed, and even that was because we got the wheels super-late and didn't have time to place them properly.
With dropped 8WD, you obviously can't do that. Not saying it's a good practice to do that, and I really don't like the concept of doing that (the reason we didn't do that this year), but the fact it can be done shows how much easier it is to get a 4/6WD than 8.

We did not use a mill, nor do we own one. Just carefully measured and drill pressed. Not really an exact science and could be done better but we never had an issue with it.

As for more complex, its not at all. If anything maybe even easier because you can just plop the gearbox/motors dead center in your bot. you just double sprocket the two inner wheels as you would a center wheel on a 6wd.

without even having to think about it

wheres the fun in that? The only thought that went into ours was wheel spacing, because we wanted to make sure we could clear the bump with no issues (beaching ourselves on it, breaking chains, ect) which we just did some simple geometric drawings in CAD for, not even a full 3d model. If it werent for that, you can pretty much just evenly space them.

[thefro526]

Originally Posted by Akash Rastogi

This is not entirely true. An 8wd is easy to chain together even with the kit setup. You'll obviously need the additional hardware and parts. It is very easy to modify the kit chassis with little blocks of tubing bolted on with 3/8" bolt holes milled into them. Select your drop, select your wheel spacing, mill the blocks, bolt them on and you're good to go. You can also drill/mill new holes right into the kit frame, like 2495 did in 2012, but using blocks allows more room for error. I'm sure plenty of teams do this, but 2495's are the only pictures I can refer you to:

For what it's worth, there's an even easier way to go about this than the method that Akash described, and that's to use standard COTS pillow blocks and drill the C-Base to match (if need be). Assuming the blocks are bolted to the frame, you can then shim them up or down with washers/shims/etc to tweak the drop or move them around for different configurations.

Or, if you've got an extra C-Base rail or two sitting around, you could cut it into a series of 3-4" long segments and use those as your 'axle' blocks. To be honest, if you're one of those teams that doesn't use the C-Base or has an extra one laying around, I'd suggest modifying it with axle blocks and trying different drive configurations to get a better first hand understanding of the pros and cons of each type.

Regardless of what you decide to pursue, I'd suggest emphasizing reliability over all else until you're confident that you can build a 'zero-failure' drive train. I've seen way too many teams chase performance at the expense of reliability and it's a really disheartening moment when an awesome drive train fails in the middle of a match - and it's usually the 'most important' match of the day...

For example, the 341's drive this year is probably the best drive train I have ever had a hand in over 8 seasons of FRC. It is by no means the lightest, or the fastest, or the coolest, but it's bullet proof. Over the course of 60+ matches on season and hours of off-season use, we've never had a single failure - and have not needed to do any maintenance work other than swapping wheels when they've worn too much.