omni drive transmissions

[FourPenguins]

Well there's a few point to consider. You want a four-wheel omni drive. What kind of four wheel omni drive?
Are you referring to a true omnidirectional drive (strafing ability) or simply a tank-drive with omniwheels?
If you're looking at a tank-drive with omniwheels, I advise against it. You'll get shoved sideways very easily and a well thought-out tank drive with regular wheels can turn just as well. In short, the advantages just aren't there.
If you're looking at a true omni drive, please elaborate. Crab? Omni wheels? Mecanums?
As for the kit, only time will tell. However, you can plan based on the Andymark boxes, and if they aren't in the kit, either adapt or buy some.

[Thermal]

Last year, my team (1293, D5robotics) used the 8" mecanum wheels from AM along with 4 BB transmissions, one for each CIM. Mechanically the bot worked fine last year, but we had some very very strange coding issues with the accellerometer that caused the robot to be highly uncontrollable at Palmetto.

http://d5robotics.org/Content/View.php?id=550

That is a video of our bot during build season, that is the first night we had it running. Here is an inventor rendering of the bot....

http://d5robotics.org/Content/View.php?id=548

We had no issues with the single CIM banebot planetary transmissions, we found them to be delightful (moreso than the dual CIM BB gearboxes, which shattered on us at Palmetto in 07, but that could be attributed to the fact that they were in the front of the bot, unprotected, and took a direct hit as we plowed into another bot at full speed)

The "white" chassis is made from UHMW, a nice and flexible plastic but very strong. We took a 1 1/4" (I think) sheet of it and water jetted the chassis from it. It offered a kind of makeshift suspension system that is vital with mecanum drives. It kept all the wheels on the ground since it was able to flex and absorb the vibrations caused by the wheels. The chassis UHMW costs 200 dollars for a slab that large.

8752K518 is the mcmaster product number

[gorrilla]

i thinking 4 omni wheels at 45 degrees and i was assuming the toughboxes will come in the kit again seeing as how they worked great


well mecanums are 330 dollars so that kinda limits them there


and for some reason we have always had a problem with weight and being slow i dont know why? but UHMV wouldent that be likley to kinda bend and twist ad stuff and not be signifigently better than the kitbot chassis

[AJ R]

last year 1714 used an omni drive. We used the Andymark tough boxes without a problem. I would also recomend using the molded plastic omni wheel. We tried to use the coolie dualie wheels at first, but they would bend.

[FourPenguins]

If you're looking to build that sort of omni drive, I see no reason your plan (4 tough boxes, 4 CIMs) won't work. Be careful to:
a) Do the math on the speed/torque this combination will yield (searching CD will get you lots of help with this, and many in this thread can assist).
b) Take into account that some of your speed/torque will be lost to wheel scrub (the wheels will be driving against each other as well as forward/backward). I'm sure someone here knows more about this than I do.

I'd say your first step is to define your goals. Speed? Pushing power? Terrain ability? (this type of DT works best only on flat ground)
From there, work back to your desired gear ratio. See if the toughboxes are within your needs. If they aren't, you can modify them pretty easily by swapping out a stage or two, or adding a stage after the gearbox. You could also design and manufacture your own boxes, but I gather you don't have the resources for this.

[AdamHeard]

If you are not directly driving the wheels with the gearbox output shaft, 1-2 stackerboxes and a chain reduction seems to be a very light and compact option. You could probably direct drive from a stackerbox, but I would modify them to replace the output shaft with a 1/2".

Originally Posted by Thermal

Last year, my team (1293, D5robotics) used the 8" mecanum wheels from AM along with 4 BB transmissions, one for each CIM. Mechanically the bot worked fine last year, but we had some very very strange coding issues with the accellerometer that caused the robot to be highly uncontrollable at Palmetto.

http://d5robotics.org/Content/View.php?id=550

That is a video of our bot during build season, that is the first night we had it running. Here is an inventor rendering of the bot....

http://d5robotics.org/Content/View.php?id=548

We had no issues with the single CIM banebot planetary transmissions, we found them to be delightful (moreso than the dual CIM BB gearboxes, which shattered on us at Palmetto in 07, but that could be attributed to the fact that they were in the front of the bot, unprotected, and took a direct hit as we plowed into another bot at full speed)

The "white" chassis is made from UHMW, a nice and flexible plastic but very strong. We took a 1 1/4" (I think) sheet of it and water jetted the chassis from it. It offered a kind of makeshift suspension system that is vital with mecanum drives. It kept all the wheels on the ground since it was able to flex and absorb the vibrations caused by the wheels. The chassis UHMW I believe costs about 460 dollars to buy for a chunk that large.

If it was $460, your robot was illegal.

[gorrilla]

i know there will be wheel scrub


the stacker boxes look great but but those are more for arms and stuff right?

we do have the tools and machine to make our own but we would have to purchase gears and deal with shipping and its just a pain and would probobly not as reliable




i would have gone with the plastic molded omnis anyways because they are the cheapest


and i was planning on making my final decision after the kickoff to see what terrain is involved something tells me it will be flat and require lots of movement again

[AdamHeard]

Originally Posted by gorrilla

the stacker boxes look great but but those are more for arms and stuff right?

A gearbox is a gearbox; A toughbox is just two sets of the gears in the stackerbox (different bores/shafts... same ratios) in a larger package with the ability to mount two motors. If you don't need to mount two, as is the case, I'd say the stackerbox is a much better option as it is much lighter and more compact.

[gorrilla]

so its got a 3-57-1 ratio and with two that would be very fast and the toughboxes are 12-75-1 so it would be faster(tourqe is not a concern since its omni wheels)


and they are cheaper than the toughboxes and the two from this year have been dismantled and pieces lost!

[AdamHeard]

Originally Posted by gorrilla

so its got a 3-57-1 ratio and with two that would be very fast and the toughboxes are 12-75-1 so it would be faster(tourqe is not a concern since its omni wheels)


and they are cheaper than the toughboxes and the two from this year have been dismantled and pieces lost!

Two would yield a ratio of 12.75:1, just like the toughboxes (both are made up of 14:50 reductions).

Torque/speed is still very much a concern. If you gear it too high, accelerating will take forever and low speed movements will be hard to control.

I have never built an omni-wheeled drivetrain, but I imagine directly driving off of two stackerboxes two 6" omni wheels would yield a decently fast drive, probably around 9 fps (assuming that at free speed, it'd be just shy of 11 fps).

[gorrilla]

Originally Posted by AdamHeard

Two would yield a ratio of 12.75:1, just like the toughboxes (both are made up of 14:50 reductions).

Torque/speed is still very much a concern. If you gear it too high, accelerating will take forever and low speed movements will be hard to control.

I have never built an omni-wheeled drivetrain, but I imagine directly driving off of two stackerboxes two 6" omni wheels would yield a decently fast drive, probably around 9 fps (assuming that at free speed, it'd be just shy of 11 fps).

oh i must have missed that day in math class



so your saying that two stackers is just like one toughbox in that case it should be around 12fps depending on how much wheel slip there is


and i was under the impression that if you have it geared high it would have less torque(therefore less wheel slipping) but a higher final speed or am i backwards? or is it like this because there omni wheels?

[FourPenguins]

Higher torque=faster acceleration, more power at low speeds, easier turning(this last point is less of a concern with omnis)
Higher speed=higher top speed
Since most FIRST bots don't have much room to move or much time to accelerate, a top speed over about 11ft/sec is usually excessive.
Wheel slipping has to do with how "sticky" your wheels are. If you have more torque than your wheels' "stickiness" can transfer to the floor, and then your wheels will slip. Though lower torque reduces the risk of slipping, it also reduces maneuverability, pushing power, etc. I don't know much about the AM omni wheels, but an omni wheel should be designed to be very "sticky". (Regular wheels must slide sideways to turn, omnis don't.) Perhaps someone could enlighten us as to the sliding characteristics of the AM omni wheels.

[gorrilla]

also is there a way to make them stickier

[Thermal]

Originally Posted by AdamHeard

If you are not directly driving the wheels with the gearbox output shaft, 1-2 stackerboxes and a chain reduction seems to be a very light and compact option. You could probably direct drive from a stackerbox, but I would modify them to replace the output shaft with a 1/2".



If it was $460, your robot was illegal.

I'm sorry, I misquoted the price, the actual price of a slab of that size is 200. Here is the mcmaster-carr product number for the sheet we got.

8752K518

Sorry for the confusion.

Also back on topic, we originally went with the 45° omni design in cad, but decided to go with the mecanums. Here is the inventor render of the omniwheel bot. Also, if you have access to a waterjet (or even if you don't) you can try and setup a way to get those mecanums cheaper. We contacted AM and just got the sideplates for 180 dollars, and we fabricated our own rollers from brass tubing bought off of mcmaster and sheets of rubber which we then jetted out the rollers from. It ended up saving us about 100 dollars for the set overall.

http://www.chiefdelphi.com/media/photos/29534

Also theres a bonus, a very lengthy discussion on the design for you to read through :-D