30-01-2011 00:15

LLogan

Re: pic: FRC488 Octocanum.

I have been waiting for somebody to create this.

Looks like you have the best of both worlds. Where are you going to place the piston? I'm having a hard time visualizing where you'll be mounting it.

What configurations are you going to have your wheels in? I.e. four traction only, four mecanum only, 4 traction + 4 mecanum, etc.

30-01-2011 00:17

xSAWxBLADEx

Re: pic: FRC488 Octocanum.

isnt meccum wheels designed to use counter friction...more friction the slower the wheels will go sideways if at all, i just see no advantage to this design...im not saying get rid of it at all im saying prove me wrong

30-01-2011 00:21

AdamHeard

Re: pic: FRC488 Octocanum.

I like it, what is the speed difference?

30-01-2011 00:21

bladetech932

Re: pic: FRC488 Octocanum.

Only one is down at a time traction for power then you switch to only mecanum for maneuverability.

30-01-2011 00:22

xSAWxBLADEx

Re: pic: FRC488 Octocanum.

maybe where the ziptie is?

30-01-2011 00:22

spiffyspleen

Re: pic: FRC488 Octocanum.

That looks amazing,are you guys going to be at the Portland Regional this year?

30-01-2011 00:23

xSAWxBLADEx

Re: pic: FRC488 Octocanum.

i see it now

30-01-2011 00:38

Andrew Remmers

Re: pic: FRC488 Octocanum.

I like it!

However a few questions

Why did you chain all the drive wheels, instead of power one of the directly and derive power off that wheel?

How much does it weigh? (without a super structure)

Where does the piston mount?

30-01-2011 10:20

Bjenks548

Re: pic: FRC488 Octocanum.

That is going to be a very fun drive train to watch, just like 217 and 148 last year. How much does it weigh? and Why did you put the mecanum wheels on the outside?

30-01-2011 12:28

Nikhil Bajaj

Re: pic: FRC488 Octocanum.

The 488 version looks awesome. 461 prototyped this drivetrain over the last year and a half and summer, here is a link to that thread.

http://www.chiefdelphi.com/forums/sh...ad.php?t=86171

Hope that helps you see how this sort of system works.

30-01-2011 13:33

s1900ahon

Re: pic: FRC488 Octocanum.

Nice.

You should loop the Jaguar's PWM cable under the front tab on the Jaguar.

30-01-2011 13:36

Katie_UPS

Re: pic: FRC488 Octocanum.

We were joking about making that this year! I'm glad to see that someone did!

30-01-2011 13:41

Madison

Re: pic: FRC488 Octocanum.

The wheels are going to be actuated by two pistons -- one along each side of the robot. It's hard to visualize because all of the hardware associated with that system is still missing. We'll add a vertically mounted arm to each wheel pod that the piston will push or pull upon. This creates a four-bar linkage where none of the links are grounded. The 1x1 cross piece that the pods are currently zip-tied to will act as a stop for the pod rotation and, effectively, ground one link of the four bar.

In retrospect, it would've made more sense to actuate each wheel pod individually, but the system we're implementing is a bit lighter. As a result, we can only run on all mecanums or all traction wheels.

There is an additional 22:32 reduction between mecanum and traction right now, in addition to the difference in wheel size -- 4" vs. 6". We may experiment with this some more.


Yep!


I tried to make direct drive of the mecanum wheel a reality, but AM's "long" output shafts were not quite long enough to make it plausible. It would've simplified a lot of things, certainly.

I am not certain of its weight. We'll try to weight it today, but I'd guess it's around 50/55 lbs.



Mecanum wheels are outside because they're geared faster than the traction wheels and we wanted the widest wheelbase possible to accommodate rapid changes in direction. Putting the traction wheels on the inside also makes it easier for the robot to turn while they're in place.


Here's some brief video of the first test of the mecanum wheel set. It's nothing too earth-shattering, but I understand that, by the time they left last night, our programming group has field-oriented drive running well.
http://www.youtube.com/watch?v=RA-jd48tPFM

30-01-2011 14:08

Ether

Re: pic: FRC488 Octocanum.

Could you elaborate a bit on that please.. How does making the wheelbase longer accommodate rapid changes in direction?

30-01-2011 14:22

Vikesrock

Re: pic: FRC488 Octocanum.

Rapid direction changes + Short wheel base + tall manipulators = Tipping

The longer the wheel base, the less likely you are to tip when accelerating hard or coming to a hard stop.

30-01-2011 14:47

Andrew Schreiber

Re: pic: FRC488 Octocanum.

However, mecanum is not at all like full blown swerve. The bot tends to drift a little bit as the rollers spin which should help remove some of the problems typically associated with rapid direction changes and tall manipulators. That being said, I still would not change direction at full speed (or even drop those traction wheels) if I had any sort of large mass up top. But knowing what I know about 488 they will have thought that out already (this ain't their first rodeo with mecanums).

Cool system, glad to see someone went with it, hopefully the ability to strafe will be beneficial.

30-01-2011 17:37

BJC

Re: pic: FRC488 Octocanum.

I'm very curious as too why you put the traction wheels on the inside as to opposed to the outside. As of now it appears that you are not really taking full advantage of switching wheel base sizes. Why not put the traction wheels in the longer wheel base configuration to make it that much more difficult to turn as opposed to the mecanum wheels which could be made to preform equally well on the shorter (inner) wheel base?

Otherwise, really cool.

30-01-2011 23:15

Dad1279

Re: pic: FRC488 Octocanum.

We prefer to call them über-mecanums.....

31-01-2011 09:12

dmaldonado

Re: pic: FRC488 Octocanum.

Well team 1708 did a very simalar design

http://www.chiefdelphi.com/media/photos/34890

31-01-2011 09:55

Racer26

Re: pic: FRC488 Octocanum.

I like the idea. If the execution can be made perfectly reliable, it should be a formidable drivetrain. The trick comes in ensuring theres no problems with throwing chains etc.

31-01-2011 11:23

1708xMr.Roboto

Re: pic: FRC488 Octocanum.

Team 1708 actually did something like this last year..

Check it,
http://www.chiefdelphi.com/media/photos/34890

31-01-2011 12:25

Madison

Re: pic: FRC488 Octocanum.

We anticipate using the mecanum wheel set most of the time, so we wanted that to be our most stable configuration. The traction wheels are not really intended to prevent others from turning us in place, but to kindly remove defense from our path when the mecanum wheels can't get us around it.

31-01-2011 12:33

thefro526

Re: pic: FRC488 Octocanum.

Well said. Never know when you'll need to gently escort someone across the field.

I really like this drive, I feel like it negates the majority of the disadvantages of a Mecanum Drive in a relatively simple and elegant method.

31-01-2011 12:38

Ether

Re: pic: FRC488 Octocanum.

When you guys assembled those mecs, did you do anything special, like grease the washers at the end of each roller and/or replace the washers with nickel plated ones?

31-01-2011 12:47

Madison

Re: pic: FRC488 Octocanum.

We checked that each of the rollers is spinning freely, but did not do anything otherwise special to them. As we run the chassis in, we'll see how they behave and make adjustments as necessary.

31-01-2011 13:35

JVN

Re: pic: FRC488 Octocanum.

Hi Madison,
Looks good. The Robowranglers built an octocanum platform this fall. Do you mind my asking -- what are you geared at for the mecanum drive?

-John

31-01-2011 13:39

Madison

Re: pic: FRC488 Octocanum.

It's geared at 8.9:1 right now.

31-01-2011 14:13

JVN

Re: pic: FRC488 Octocanum.

Interesting. That is much faster than I expected.
I look forward to seeing video of it in action so I can recalibrate my mental picture of performance.

-John

31-01-2011 14:15

Madison

Re: pic: FRC488 Octocanum.

I think it's a bit faster than I expected, too. We're going to get a bit more weight onto it soon and test it again under more realistic conditions to get a sense of whether or not it's something we can handle.

We've got rookie drivers this year and haven't found an area large enough to let it run at full speed for more than about a second, so necessity may require that we slow it down.

31-01-2011 16:09

Racer26

Re: pic: FRC488 Octocanum.

Robowranglers building more non-conventional drivetrains? 2010 was Nonadrive, I wonder what 2011 holds. Eep!

31-01-2011 16:28

Jonathan Norris

Re: pic: FRC488 Octocanum.

Nice to see someone try this, I've been thinking about this kinda of drive system for a year or so. We developed a CAD of this system ourselves during the offseason, but never got the time to build it. The main concern that I had with this system was the side loading of the 'swing arms' that the traction wheels are mounted on. I would be concerned with the loading on the swing arm when trying to turn with the traction wheels down. We had designed for delrin blocks on either side of the 'swing arm' to transfer those loads to the frame. From my understanding 148 had similar problems with this last year, I'm sure JVN could elaborate more on that. Good luck with it!

06-02-2011 00:37

Madison

Re: pic: FRC488 Octocanum.

A new picture of the entire drive is pending in the gallery, but here's a video of our first test of the articulation mechanism. Our cylinders finally arrived, so we were able to get this hooked up and tested today.

http://www.youtube.com/watch?v=nQgri...e_gdata_player

06-02-2011 00:57

JVN

Re: pic: FRC488 Octocanum.

I really like your articulation setup. I'm admiring the elegance and simplicity of the long connecting link between the two up on top (using the height of the module as your torque arm).

I struggled with something similar in the pre-season, and just ended up using 2 cylinders (I got too caught up in using the cylinder without a connecting link somewhere inbetween the modules, but the force vectors never worked out. This is a much better way to solve the problem!

Kudos!
-John

06-02-2011 01:02

coldfusion1279

Re: pic: FRC488 Octocanum.

Looks like a great mechanism for raising and lowering the wheels. May I ask, with such a short traction wheel base, are you worried about the front wheels 'rearing' off the ground in a pushing match? Maybe the center of gravity is low enough that this wont be much of a problem?

10-02-2011 12:54

Madison

Re: pic: FRC488 Octocanum.

Thanks for the compliments on how we've arranged drive articulation.

In the interest of keeping everything in one place, here's another brief video of the drive. The programming team has implemented field-oriented drive reliably enough to drive the robot freely.

http://www.youtube.com/watch?v=sM8cixsE5fo

I believe this is still at some fraction of top speed, but I'm not sure how it's being scaled. I also believe, but am not sure, that we're using two gyros to determine orientation. This stuff isn't my department.

10-02-2011 13:00

thefro526

Re: pic: FRC488 Octocanum.

That is the most impressive Mecanum Demonstration I've ever seen.

Wow.

I'm assuming a movement like that would be incredibly difficult to do without field oriented control?

10-02-2011 13:07

Chris is me

Re: pic: FRC488 Octocanum.

Hey look, a mecanum drive doing what it's supposed to. Cool!

I'm willing to bet you'll have one of the top 10 mecanum robots this year.

10-02-2011 13:48

robodude03

Re: pic: FRC488 Octocanum.

Absolutely agree with you Chris. This is an amazing system and I love the simplicity in design (referring to the actuation of the drive wheels.

10-02-2011 16:11

Arjun Namineni

Re: pic: FRC488 Octocanum.

Why did you guys decide for two pistons? Wouldn't one work just the same, but with less weight and less air loss?

10-02-2011 16:19

Madison

Re: pic: FRC488 Octocanum.

One cylinder provides enough force at 60 psi to lift the entire robot, but we imagined that added complexity of the mechanical system needs to actuate all four wheels with a single piston offset any weight saved by eliminating the second piston. The second piston is only about 2 lbs. and it means that our actuation system can operate reliably with as little as 30 psi or so.

10-02-2011 16:35

PaW

Re: pic: FRC488 Octocanum.

I'm going to sneak over to the "Seattle Cascade" field from the "Seattle Olympic" field just to watch this.



Your new drive team is gonna have some fun this year...

11-02-2011 12:45

Madison

Re: pic: FRC488 Octocanum.

More video: http://www.youtube.com/user/teamxbot.../1/5_wZ58rzxg4

This time, we were testing to see how the robot would behave if we deployed the traction wheels while strafing. We wanted to test side-loading on our wheel pods and stability of the entire robot. It looks like things will be fine.

The most confusing aspect of the change is that the robot, right now, automatically shifts from field-oriented drive to standard two-stick tank drive. That's why, whenever the wheels are shifted on the fly, the robot immediately starts to move "sideways." We're looking at options for making the switch more intuitive, but it's probably something that a bit of practice could overcome well enough.

16-02-2011 14:13

Madison

Re: pic: FRC488 Octocanum.

This is kinda like blogging my season, right?

Last night, our programming team brought our "follow wheels" on line. These are non-powered wheels with encoders attached that relay actual distance traveled back to cRio. We use three on this robot because of its holonomic capabilities.
You can then do neat things like this: http://www.youtube.com/watch?v=Jr-eglZBAHQ