Six Wheel Drive Question

[Mr.Twister]

When making a six wheel drive, how much do you have to offset the middle wheel Down?

[Joe G.]

1/16" to 3/16" is generally the standard range, with 1/8" probably being the most common drop. The shorter your wheelbase and harder your wheels, the smaller a drop you can get away with.

[Michael Hill]

Originally Posted by Mr.Twister

When making a six wheel drive, how much do you have to offset the middle wheel Down?

There have been several threads about this, and I know people are going to gripe about searching. However, those posts are rarely useful...

Anyway....like everything in FRC, it depends. It depends on your drivetrain flexibility, it depends on your drivebase configuration (short, long, square, etc.). If you have a shorter drivebase, you generally need a smaller drop, and just the opposite for a longer drivebase. And the more flexible (less rigid) your drivebase is, the more you'll need. I've heard everything from 0.1" to 3/16". In 2014, we had a square drivebase and used 5/32, and it worked out fine for us. The whole goal of dropping the center wheel is to turn a 6 wheel base into 2 short. 4 wheel bases. This helps significantly in skid steer (all traction wheels) turning.

[GeeTwo]

And in case it isn't obvious, for most robots you want the drop to be as small as possible while achieving a short wheelbase for turning. The greater the drop, the more the robot will pitch up as you accelerate and down as you brake, resulting in manipulators moving along with the chassis.

If your manipulators are relatively insensitive to pitching or you have sensors that reduce this sensitivity, and the maneuverability is essential, you may opt to add a bit more drop as a safety margin.

On the other hand, if you have a long arm that reaches forward and back to pick things off the floor, you probably want to minimize the drop. With that long arm, you may want to consider an 8 wheel configuration so that while you're doing low speed/acceleration pickup/placement maneuvers, your chassis will be horizontal. Another solution would be to put your CoG intentionally ahead or behind your center axle so that you're at least at a known attitude.

Note also that "stiffness" in this sense is primarily in the plate/bar/channel in which each track of wheels runs; the required drop is not nearly as dependent on flexibility in other axes. Using pneumatic wheels also counts as reduced stiffness for this purpose.

Another alternative to actually dropping the center is to turn down the corner wheels on a lathe. This is not usually the best solution, because the wheels are usually constrained by chain or belt to rotate at the same angular speed, which does not result in the same linear speed with uneven wheel sizes. However, if you scaled the number of teeth on the sprockets to be proportional to the diameter, this would not be a problem. (E.g. 14 teeth on a 3 1/2" wheel and 16 teeth on a 4" wheel, and with holes on the same level, you'd have a 1/4" drop).

[asid61]

We tried 1/8" drop in 2014, it rocked a little too much for me. We moved to 3/32 IIRC for an offseason chassis.

[Chris is me]

For what it's worth, keep in mind that the 1/8"-ish number became a pseudo-standard in FRC back when drivetrains were 38 long by 28 wide. Excluding this year, in 2013 and 2014 shorter chassis shapes were the norm, and honestly I think people stuck to 1/8" drop on those chassis sizes just out of habit. I don't think it's necessarily perfectly optimal, but the marginal benefit one would get from figuring out the "ideal" amount of drop for a drivetrain of a particular size / wheel type / tread type / etc is probably not worth the effort of finding it out. 1/8" works.

The more your wheel tread material compresses, the more drop you'll want. 1/8" is good for roughtop tread. Works for Colsons too, but you could also run a smidge less (3/32?) for them. Pneumatic wheels may want 1/4" drop. Your mileage may vary.

Note also that "stiffness" in this sense is primarily in the plate/bar/channel in which each track of wheels runs; the required drop is not nearly as dependent on flexibility in other axes.

Not exactly; how stiff your drive sides are relative to each *other* matters too. If opposite corners flex and both make contact, you'll have far more turning scrub.

[Michael Hill]

Originally Posted by Chris is me

Not exactly; how stiff your drive sides are relative to each *other* matters too. If opposite corners flex and both make contact, you'll have far more turning scrub.

Yeah, this is what I was referring to. Doing the whole hold 3 corners down and lift up on the other.

[Mr.Twister]

Thanks!! Your replies really helped!

[GeeTwo]

Originally Posted by Chris is me

For what it's worth, keep in mind that the 1/8"-ish number became a pseudo-standard in FRC back when drivetrains were 38 long by 28 wide. Excluding this year, in 2013 and 2014 shorter chassis shapes were the norm, and honestly I think people stuck to 1/8" drop on those chassis sizes just out of habit.

At the same time that chassis were getting shorter (due mostly to rules changes), frames were getting stiffer. The 2012/13 KoP chassis used 1.25" x 1.125" x .125" c-chanel. The 2014/15 KoP chassis side plates are also .125" thick aluminum, but they're 3.8" tall - many times stiffer in terms of support, and the construction is also stiffer in terms of the joins. Vex 2" x 1" tubing, nanotube, and other similar tall tubing has become increasingly popular. While I haven't built one (other than our nanotube air cannon, but we use angle as cross members) , my guess is that these would be nearly as stiff in loading and probably stiffer in torsion than the latest KoP chassis. Custom chassis designs seem to be moving towards "taller" and therefore (other things equal) stiffer structures.

I recall seeing at least one robot reveal this year that boasted a 1/16" drop.

[fargus111111111]

The opperating surface is also something to consider. The typical carpet in frc applications does compress some but not much and produces a lot of resistance if all 6 wheels touch. Due to these factors some extra clearence is not bad. Someting smoother and harder like the reoglith floor from 2009 does not compress and produces much less resistace and so a much lower clearence can be used. I assume you are using all traction wheels but putting omnis on at least one end eliminates the need for a drop at all.

[Loose Screw]

Is there any benefit to having 6 traction with a drop over 2 traction and 4 omni without a drop?

[bigbeezy]

Originally Posted by Loose Screw

Is there any benefit to having 6 traction with a drop over 2 traction and 4 omni without a drop?

What makes it easier for you to turn (omni's) makes it easier for others to turn you. That setup would have little defense against a defensive robot. However things like speed and driver control/ability could mitigate that.

[Joe G.]

Originally Posted by Loose Screw

Is there any benefit to having 6 traction with a drop over 2 traction and 4 omni without a drop?

Yes.

Good traction wheels have more forward/backwards traction than an omniwheel, so for maximum traction, you don't want to compromise by having some of your weight on omnis. Four corner omnis can also lead to unpredictable behavior over non-level surfaces, as there is a chance that the central traction wheels would be lifted off the ground, leaving you on all omnis. Four omnis provides even less turning resistance, enough that it's very easy for others to turn you. The main reason to use a flat 6+wheel drive with some omnis and some traction is to arrange your omnis in such a way that you have a non-central point of rotation, to help generate more favorable turning characteristics for lining up manipulators.

[Loose Screw]

Cool, thank you guys.

[Kevin Leonard]

Originally Posted by Loose Screw

Is there any benefit to having 6 traction with a drop over 2 traction and 4 omni without a drop?

Not having to deal with omni wheels. It also becomes harder for other robots to rotate you (although that can be mediated by running 4 traction, 2 omni). It can also depend on what point you want the robot to rotate about (which was a drivetrain consideration for many teams in 2015).

Interesting case studies in 2014:
1625 ran a 2 traction, 2 omni setup, which made them move about the field in interesting ways sometimes. I can't say whether that made them better, worse, or neither. They were good because their catapult was remarkably consistent and their drivers were very well practiced with that machine.

33 ran a 4 omni wheel setup, which made them extremely slippery for defenders. Again, I don't know whether it made them better or worse (I have my own opinions on it, but thats not for this thread). It worked because their drivers were phenomenal and their catapult had a giant sweet spot.

20 ran a 6 wheel drop center with colsons, while 340 ran 8 wheel, 4 traction, 4 omni. From experience at four different 2014 events that both 340 and 20 attended, the wheels on the ground made no difference in performance between the two machines, and as long as the team understands how to use their drivetrain, has well practiced drivers, and the other parts of the robot are also consistent, you should do well enough with either.