pic: Toltechs drive-train teaser

[greencactus3]

wow thats pretty high up. are you going to be able to suck in the balls that roll in from the sides too?
or better yet from all 4 sides?? if i had holonomic drive thats what id wanna do.
otherwise, i love the spiral. someone has a lot of time lol

[Andrew Schuetze]

Originally Posted by greencactus3

wow thats pretty high up. are you going to be able to suck in the balls that roll in from the sides too?
or better yet from all 4 sides?? if i had holonomic drive thats what id wanna do.
otherwise, i love the spiral. someone has a lot of time lol

I am seriously hoping that this thing drives tomorrow. We are 8 inches off the floor with the frame so that balls can move under from 3 sides and be whisked into our con-VADER. The idea of the omni-drive is that if balls are boucing and rolling over the court, one just swerves to the side and harvests them.

Like I said though, it is untested and we have to put the belt/fabric on the con-VADER and then mount our cannon. Weight is becoming an issue

Thanks for the positive comments and yes the students did describe their ball sweeper design to the mentors as,"You know ... like under a vacumm cleaner"

[Rick TYler]

Originally Posted by Andrew Schuetze

(...)the students did describe their ball sweeper design to the mentors as,"You know ... like under a vacumm cleaner"

You mean you let students work on the design? You rebel.

[greencactus3]

Originally Posted by Andrew Schuetze

I am seriously hoping that this thing drives tomorrow. We are 8 inches off the floor with the frame so that balls can move under from 3 sides and be whisked into our con-VADER. The idea of the omni-drive is that if balls are boucing and rolling over the court, one just swerves to the side and harvests them.

Like I said though, it is untested and we have to put the belt/fabric on the con-VADER and then mount our cannon. Weight is becoming an issue

very nice. once you get it driving, tell us if 8inches of clearance is always good enough .i was wondering if bouncing balls will bounce off the frame that is "too low" like 8in maybe? or if they just get dragged under..

[Rick TYler]

Originally Posted by greencactus3

I was wondering if bouncing balls will bounce off the frame that is "too low" like 8in maybe? or if they just get dragged under..

Maybe eight inches is too high? It's important to remember that Poof balls squish. Here's a thought. A lot of teams are going to do sweepers that have counter-rotating rollers or belts designed to lift balls from the ground. Many of these will be slanted back, or have curved ramps that skim along just above the ground. What if, instead, the lifter was a belt that went up vertically, with a pressure plate about six inches away. The center point of the bottom roller of the conveyor is below the center point of a ball (how about 2-1/2 inches?), and the bottom of the pressure plate is about 5-3/4 inches from the floor. You drive the bottom of the plate over the ball, which, since you are impacting it well above its CG, starts to roll. The robot is moving faster than the ball, so it drives over it. The ball then rolls into the bottom of the conveyor which acts as a wheel, lifting the rear edge of the ball. The ball tries to roll forward, but, since the force from the belt is actually a vector pushing forward AND up, it rises a little. Once it lifts up, it becomes trapped between the belt and the pressure plate, which act just like a single-wheel shooter and a pressure plate or bar, and the ball walks up the pressure plate into a curved sheet, which walks it over into a hopper.

This must have come to me in a dream. I'm sure it's not the design we are using on Top Gun. Incidentally, this is a testimony to prototyping. We would have a sloped ramp with a curved Lexan floor sweeper (which would act just like a bulldozer blade and knocked balls out if we hit them going fast) and two rollers if we hadn't built three prototypes of this. Now we know exactly what geometry will work, and how fast the belt needs to go. Not that this is our design or anything...

[Andrew Schuetze]

Okay for those that said No, you are partially correct. I think we can live with our design. Video of the robot will hopefully be posted this weekend on our team's discussion board.

The robot drove good the first time with the code. When they added in code and control for zero degree spin, that worked the first time as well

We can get up the ramp if a second robot is behind us keeping the robot from spinning and rolling down the ramp when the front wheels are about 3/4 the way up the ramp. I know this because I was able to put my feet on the back end while seated behind the robot which keep the angle of attach perpendicular to the ramp. Once the front tires get over the lip they are back on a level surface and pull the robot over the top. It is a lot of spinning wheels So one of those alliance partner robots with the IFI or AM traction tires gets behind us we are up and on without too much work.


On the plus side, it is way to cool to swerve or strafe

[Andrew Blair]

Originally Posted by Rick TYler

Maybe eight inches is too high? It's important to remember that Poof balls squish. Here's a thought. A lot of teams are going to do sweepers that have counter-rotating rollers or belts designed to lift balls from the ground. Many of these will be slanted back, or have curved ramps that skim along just above the ground. What if, instead, the lifter was a belt that went up vertically, with a pressure plate about six inches away. The center point of the bottom roller of the conveyor is below the center point of a ball (how about 2-1/2 inches?), and the bottom of the pressure plate is about 5-3/4 inches from the floor. You drive the bottom of the plate over the ball, which, since you are impacting it well above its CG, starts to roll. The robot is moving faster than the ball, so it drives over it. The ball then rolls into the bottom of the conveyor which acts as a wheel, lifting the rear edge of the ball. The ball tries to roll forward, but, since the force from the belt is actually a vector pushing forward AND up, it rises a little. Once it lifts up, it becomes trapped between the belt and the pressure plate, which act just like a single-wheel shooter and a pressure plate or bar, and the ball walks up the pressure plate into a curved sheet, which walks it over into a hopper.

This must have come to me in a dream. I'm sure it's not the design we are using on Top Gun. Incidentally, this is a testimony to prototyping. We would have a sloped ramp with a curved Lexan floor sweeper (which would act just like a bulldozer blade and knocked balls out if we hit them going fast) and two rollers if we hadn't built three prototypes of this. Now we know exactly what geometry will work, and how fast the belt needs to go. Not that this is our design or anything...

This is actually the same exact idea that I had towards the beginning of the season. We are not using it either though...

[Rick TYler]

Originally Posted by Andrew Blair

This is actually the same exact idea that I had towards the beginning of the season. We are not using it either though...

Sorry for making humorous jokes and not being clear. This is how we will be lifting balls from the floor. The biggest advantage is that it only takes up 10 inches of length on the robot, leaving the rest of the real estate for the hopper/loader/shooter mechanism.

[Andrew Schuetze]

As promised, mentor D. Sean Kelly did a quick edit and uploaded a .wmv to our site, you can view the first cut of our robot driving in the shop. Not seen in this video is the code addition which allows us to complete zero degree spins in C.W. or C.C.W. via a second controller.

http://www.toltechs.com/Rocket/Uploa..._Holonomic.wmv

[Andrew Schuetze]

Originally Posted by greencactus3

very nice. once you get it driving, tell us if 8inches of clearance is always good enough .i was wondering if bouncing balls will bounce off the frame that is "too low" like 8in maybe? or if they just get dragged under..

With all the excitement, we didn't test it with poof's on the floor much less bouncing around. Maybe Wednesday before it is driving again as the electronics board is being finished. The demo plywood board with the quickclamps is not the final version I'll post another update when we test the drive with poofs.

[Kris Verdeyen]

Love the video - that base looks slick. Do you have closed-loop speed control over the wheels? It looks like the lines it's moving in are way too straight to be from open loop motor commands. Of course, I've never built a base like that, so what do I know.

[Andrew Schuetze]

Originally Posted by Kris Verdeyen

Love the video - that base looks slick. Do you have closed-loop speed control over the wheels? It looks like the lines it's moving in are way too straight to be from open loop motor commands. Of course, I've never built a base like that, so what do I know.

Quite the exacting eye you have. Yes there was an immediate concern about that very issue. The video edit cuts outs the shots of a slight drift in one direction. It is more pronounced in one direction more so than any other so the one motor/transmission that is making all the noise in the video was completely rebuilt after the test.

Now how best to get closed-loop feed back from such a drive. Encoders on the wheels may be one way to ensure that all the wheels are spinning at the desired rates but with the large amount of scrub in such a system we are not certian that solves the problem. What does provide true feed back would be one of the electronic chips. Gyro / 2D accelerometer We are getting those items wired and the programmers will work with those a bit while they also work on the camera system being the feedback system for shooting solutions.

Lots of work to do in just two weeks.

[Andrew Rudolph]

Looks like you guys are wel on your way to having a drivable holonomic system. Good luck with that!

Just a question, have you guys tried doubling up the omniwheels with an offset so that theres always a roller in contact with the ground?

We found it to work better with 2 offset wheels, basically the problem without the offset is that without it, the wheel isnt really round, so it bounces and doesnt slide quite as well.

Good luck to you guys and i hope to see you in houston!

[Andrew Schuetze]

Originally Posted by Andrew Rudolph

Looks like you guys are wel on your way to having a drivable holonomic system. Good luck with that!

Just a question, have you guys tried doubling up the omniwheels with an offset so that theres always a roller in contact with the ground?

We found it to work better with 2 offset wheels, basically the problem without the offset is that without it, the wheel isnt really round, so it bounces and doesnt slide quite as well.

Good luck to you guys and i hope to see you in houston!

Have you tried the new plastic omni-wheels from AM? They've made the rollers wider and with a lower profile so you don't see much bounce with these. I've got another set on order in case we need them or as replacements.

work now is focused on autonomous modes and getting the mechanics of the ball harvesting complete and shooting


Thanks,


APS