Building a new drivetrain?

Technical Technical Discussion
#1

Our team is in its second season, we were rookie in 2017. We collectively decided that we would use the AM14U3 chassis this year.
Some of our team members are now suggesting that we modify the drivetrain to work with omni wheels or the likes, allowing better control of movement about the field. An example being helping with accurate positioning against a fixture like the exchange or portal. They recall the difficulties we had lining up to the gear peg last year.
I fear that this would tangle us up taking away from other much-needed functionality, but I’d like to get the thoughts of much more experienced teams.

thank You

#2

If you replace the time spent modifying your chassis with drive practice, you will overall be much better off. You might think that omnis, 2+2 drive, or mecanum will be a magic bullet that makes driving easier, but it’s not. If anything it makes it harder. Instead of modifying your chassis and pushing back the finish of that a few days or weeks, build it now or soon and practice every day for an hour or two. Even if you don’t have any manipulators, just drive.

#3

If all you’re doing is changing from traction wheels to omni wheels, I don’t see this taking up a significant amount of time or requiring all that much modification at all. The omni wheels offered by Andymark interface to the provided hardware the same way as the traction wheels do.

My only caution would be that I’m not 100% sure both wheel types are the same thickness, so you may have to alter the spacing on the shafts slightly if they are different.

#4

We went from kit tank to Omni H our second year and subsequently…it took over half our build season (4-weeks) to figure it out mainly working on the suspension. But was worth it for Recycle Rush and worked pretty well last year in SteamWorks.

So depends you can drop Omnis in certainly or consider going beyond that and add things like a suspension, be careful only 6-weeks and you need drive time.

Whats nice is once you figure out a decent drive base you then can always have that as an option under your belt for future game consideration.

Good luck in your second year… go for it.

#5

Andymark has many configurations and add ons for the AM14U3 There is a Mecanum option that my team did last year and worked well but we had amazing programing but it took weeks to get it right but it does cost 700$+ The cheaper Option is Omni wheels costing 200$ http://www.andymark.com/product-p/am-14u3-du6.htm
And Mecanum http://www.andymark.com/product-p/am-14u3-mk6sr.htm

Some people have strong opinions on Mecanum just do a quick search on the form

#6

Also the AM14u3 Works very well for this game being a tank drive and you do not have to be that precise as the scoring platforms are relatively large

#7

Getting the cube into the hole is going to be different than getting a gear on a peg.

Driving practice definitely helps.

And as usual, robot design can make some difficult tasks easier. Spend a lot of time figuring out how, and you’ll do better.

#8

Honestly mecanum wheels will not be as effective this year, as compared to last year. This year is going to be all about getting around the field as quickly as possible without being destroyed by ntor robot that has more torque.

#9

If you are sticking with the AM14U3, then you are really talking about Mecanum and not Omni, as Omni would require a non-trivial redesign of your base. Be aware that using Mecanum usually translates to slower robots, but not as slow as many with no holonomic experience would want you to believe.

Take going forward at 100% throttle. With a traditional tank drive, you get 100% forward throttle. With Mecanum, the vector math will only give you 70.7% forward throttle. In theory, An Omni-wheel robot would still be traveling at exactly the same speed as the conventionally wheeled robot. Unfortunately for Mecanum, with identical motors and gearing it will only go at 70.7% of the speed. If an enterprising programmer were to throw out the typical vector normalization and let the motors run to the input maximum, it will result in a 40% speed up for forward/reverse and strafe left/right directions for Mecanum, matching the tank drive speed in two different directions! The robot would still be the same slower speed on diagonals. This non-linear behavior is frustratingly bad for control, so the default normalization is what we tend to stick with.

This is one of the main reasons Mecanum is perceived as slow. It could be fixed with a 40% lower gear ratio (each wheel has its own motor), but friction losses and moving into a worse part of the power curve make this a difficult solution to fully realize. A good balance is a modest drop of 20-25% on the gear ratio.

Our team does a holonomic drive more often than not, and it can really do some nifty things. We are currently designing a 6-wheel full Omni robot for this year’s game. Even with the magic of holonomic, there really is no substitute for lots of drive practice, just like conventional tank drive.

Not knowing how big your core team is, I would recommend going with what you know, and practicing a lot. If you want to tackle holonomic, work on it as an off-season project. Stop by our pit at Northern Arizona, and we’ll tell you far more than you ever wanted to know about holonomic drives.

#10

You need to drive for 100% of your matches and you need to be able to drive to score points, andymark builds a drive base that works and there is nothing wrong with a tank drive to get around the field, focus your time onto your upper parts of your robot and you will be rewarded, good drive train doesn’t win you anything, but a bad drive train that doesn’t function definitely could make you lose, my advice is leave the drive base because you are absolutely certain that it works and focus on the parts of the robot that score you points and differentiate you from everyone else.

#11

Majority of the time, spending time on a more complex mechanism, such as omni, that has fewer pros than west coast or tank is a waste of time. There is a reason that most successful robots end up going with some variation of west coast drive or tank.

#12