Globe gear box

[Karthik]

Originally Posted by Joe Johnson

Of late, I have given up building my own custom gearboxes, swerve drive, shift on the fly, and any number of other complexities...

...perhaps I should start a 12 step program for multi-motor, shift on the fly, steerable drivetrain addicts:
Step 1 We admitted we were powerless over fancy FIRST drivetrains - that our FIRST designs have become unmanageable
Step 2...

Joe,

I'd be very interested to hear your reasoning on reverting back to simpler drive trains. Is it just that the advantages gained are too small to warrant the effort and complexity, or are there other reasons as well?

To me, I think teams just need to recognize their limitations. For example, there's no point in strafing, if results in your robot being a cardboard box that can strafe, or a robot that can strafe but is uncontrollable and unreliable. Your drive system is the most important part of your robot. Here are my guidelines on choosing and building a FIRST drive system.

---

1.You're better off with a simple, 99.1114% reliable drive system, than a complex 50% reliable drive system.

1a. A robot that doesn't move will not win many matches.

2. If it's going to take 6 weeks to perfect your complex drive system it may not be the best idea.

2a. It's really not the best idea if you don't have enough people to work in parallel with the rest of the robot.

3. Remember, this robot will be driven by young, pubescent and agressive High School students. As a result...

i. It will be driven into walls, robots and the ground. Make your drive system durable.

ii. Your drivers will be inexperienced. Not many people drive robots in their spare time. Make the controls as user friendly and intuitive as possible. No matter how much extra coding this takes, it is worthwhile.

4. If you do build a complex drive system, expect for it not to work at first.

5. Looking at #4, hints strongly towards off-season prototyping.

6. Evaulate your abilities and resources in an honest and realistic way.

7. There needs to be a good reason to go complex. Don't strafe or shift just for the sake of doing so. Look at the game, evaluate your strategic priorities, and ask yourself, "is it really neccesary?", "could we better use our resources somewhere else?"

---

I like these guidelines a lot. But then again, this is coming from a guy whose team used six motors and four 6" pnuematic cylinders to drive the 2004 robot...

[Joe Johnson]

Originally Posted by Karthik

...
I'd be very interested to hear your reasoning on reverting back to simpler drive trains. Is it just that the advantages gained are too small to warrant the effort and complexity, or are there other reasons as well?
...

I am a recovering complex-drivetrain-aholic.

We were one of the first teams to really push the limits on what we call "swerve" (steerable drives).

• 1998: 2 wheel drive with 1 motor parallel steering (a.k.a. "2 by 1 Swerve")
• 1999: 4 wheel drive with 1 motor parallel steering (a.k.a. "4 by 1 Swerve")
• 2000: 2 wheel drive with 2 independent steering motors (a.k.a. "2 by 2 Swerve")
• 2001: 4 wheel drive with 4 independent steering motors (a.k.a. "4 by 4 Swerve")
• 2002: 4 wheel drive with 2 independent steering motors (a.k.a. "4 by 2 Swerve")*
  

But, in 2003, the "K-3 Rule" more or less caused us to hit rock bottom. We just could not make the complex drives and meet the $3500 limit on material and labor (at the time we had no mill or lathe at our school and we had limited access to our internal shops, forcing us to farm out our manufacturing).

So... ... we decided to go back to basics. For the last 2 years we have had more or less pretty straightforward 6 wheel drive systems (lowered middle wheels of course) with servos shifting drill tranmissions.

I can honestly say that I don't miss the more complex drives a bit. In fact, quite the opposite. I feel we have had robust drive systems that have gotten the job done when we needed it.

Swerve is a real resource hog. We used 8 motors just for our drive system when we had 4 by 4 Swerve!!! It also requires feedback on the steering angles so that the control system can contain the beast. Plus it uses up extra programming time to make it work. Plus it drives complexity into the operator interface. Plus it uses up valuable drive time for your drivers to make it look easy to drive. Etc. etc. etc....

Will Chief Delphi EVER go back to swerve? Never say never, but I, for one, will be fighting hard for the simplest drive system that gets the job done. If that is a swerve drive, fine, but if a 2 wheel drive system can do the job, that is what we are going for.



Simple, robust drivetrains let you spend your resources on other important tasks, like climbing stairs, sucking up and spitting out small balls, capping goals and hanging from bars.

That is were we came down this year anyway...

Joe J.

*By my naming convention this year, WildStang 2004 had "2 by 1 plus 2 Swerve" I think they had 2 wheels in "front" that steered but always in parallel with eachother plus 2 driven omni wheels that did not steer at all. I believe their 2003 machine had "4 by 2 Swerve" 4 drive wheels that turned but only 2 independent steering motors (the left drive wheels stayed parallel to eachother as did the right drive wheels).

If we ever build my dream of 4 wheel drive, 4 wheel independent steer, with 3 gears on each drive wheel, I will call it "4 by 4 by 3 Swerve." Yes, we can do it, even with the K-3 Rule (did I mention Chief Delphi bought an old Bridgeport and a WWI vintage lathe?), but so far the game has not needed it...

...maybe next year...
...did I tell you I also have ideas on how to give it 4 wheel independent suspension? -- I haven't thought of a good naming convention for that yet...
...also, I THINK I can make it climb stairs too (even better than Wildstang's 2004 cool drive system)...

...Man! That would be cool!...

...no no no... ...stop me before I design again!....

...I am a recovering complex-drivetrain-aholic...
.................I am a RECOVERING complex-drivetrain-aholic...
...............................I am a RECOVERING complex-drivetrain-aholic...
................................................I am a RECOVERING complex-drivetrain-aholic...

[Pat Roche]

I understand your reasoning of at different load points i actually used it at 85% of the original rpm. I also understand you when you mention simple drive. We have never used more than a single speed tranny. The bad part of a single speed tranny is you have to choose speed over power. We are looking to have both.

-Thanks for all the advice

-Pat


*edit*WOOT* My 134th post, that is the best number ever!*edit*

[rswsmay]

Well I don't necessarily agree that one has to choose speed over power. Its all a matter of the design parameters. One can always choose speed over power (Torque) if the game and your strategy requires it. You can also make the choice to the other extreme if it suits the game. There is also the middle ground when you design the drive train to deliver a balance between speed and torque. My view is that there is a 3rd parameter that usually drives the thought processes that others use to develop multi-speed drivetrains. That is control/manuverability.

In any FIRST game, there may be a need for the driver execute tasks, manipulate objects, or navigate a course that will require some amount of precision. This may be hard to accomplish with a bot whose top speed is 15fps and has a hair trigger joystick. Even with compensating through programming, there is only so much you can do without an appropriate mechanical design.

True a bot that doesn't move isn't worth much on the playing field as evidenced by our rookie year performance. Hence, in our second year we put a lot of effort in to making sure the drive train worked and worked as designed. So unless FIRST comes up with a really unique game, I don't really see a compelling need for more than 2 speeds in the games I have been involved in.

Still... its great to see the teams with the resources to do so, jump in feet first and develop these really unique drivetrains and their companion control systems. The ones I've seen this year really inspired me to think on a different level for the upcoming season. I still don't think you'll see our team with a 6 motor, CVTranny anytime soon. But ya never know.

Originally Posted by T134guy

I understand your reasoning of at different load points i actually used it at 85% of the original rpm. I also understand you when you mention simple drive. We have never used more than a single speed tranny. The bad part of a single speed tranny is you have to choose speed over power. We are looking to have both.

-Thanks for all the advice

-Pat


*edit*WOOT* My 134th post, that is the best number ever!*edit*

[Pat Roche]

I think you raise a good point in your post, however my team has been striving for the ideal drivetrain for us and very adaptable for the game. For the past three years we have been creating newer and more powerful drivetrains and we feel this is out next stage. We are at a point now that we havent had to do any work on our drivetrain during a competition which is a very big plus, we are now ready to take it to the next level. Definately a good post w/ many arguements in it.

-Pat