Team 4096 Grasshopper Drive design

[msigalov]

Team 4096 Ctrl-Z, working in Creo 3.0, decided to design and build a grasshopper drive over the offseason to be ready for a good drivetrain in build season. The goal this time was to minimize costs by using only flat machined parts, no bends on the entire drivetrain. In addition this allows us to hand-make some of the parts or laser-cut them out of wood for prototyping before we get them machined.



The frame is connected by bars of 1010 profile 80/20 and the wheels driven by belts powered from a VexPro 3-CIM ball shifter with 2 CIMs installed. Using Vex omnis on the outside on a driven shaft belted to the gearbox, the omni is belted to the tank. The tank is mounted on a lever module which is controlled by a central Bimba Square-I piston. The lever module is mounted onto a piece of tube stock which is on a shoulder bolt. For the H-drive, we have a VersaPlanetary 16:1 gearbox with a CIM, an omni wheel directly on the shaft. That gearbox is mounted to two bars of 80/20. This configuration allows us to switch an H-drive and a tank drive with ease. Should our air run out, the robot can revert to one of the drive systems for the rest of our match.

We ran into a couple of issues while designing this, such as mounting the gearboxes and making sure that our lever points are sturdy. Since we're sure we missed something, we would love your suggestions on everything related to design and rendering. If you have Creo 3.0 and would like to check out our files, go here: https://goo.gl/2LNJ9O. Download the entire folder, then the complete drivetrain assembly is called assembly_drivetrain.asm.
Thanks!

[GeeTwo]

If the air runs out, which drive is the default?

Also, getting the proper weight load on the strafe wheel of an H drive isn't as easy as putting all of the wheels in the same plane. If making a(other)n H drive, I would put springs or a piston to push the strafe wheel down so that the load on the strafe wheel could be adjusted according to driving needs.

[Greg Woelki]

Originally Posted by GeeTwo

If the air runs out, which drive is the default?

Also, getting the proper weight load on the strafe wheel of an H drive isn't as easy as putting all of the wheels in the same plane. If making a(other)n H drive, I would put springs or a piston to push the strafe wheel down so that the load on the strafe wheel could be adjusted according to driving needs.

While it is certainly a good option, in my team's experience, a suspension is not necessary in order for a H-drive to perform well. We used a statically mounted slide wheel this year. Once the drive train was assembled, we shimmed the wheel down until we were satisfied. Combined with a somewhat floppy chassis, this served us very well over the course of four events; it was very simple and we only had to adjust the shims once due to wearing of the slide wheel.

[Scott Kozutsky]

Generally speaking teams like to put the traction wheels on the outside of the drivetrain. If you get pushed back onto your omni wheels in a pushing match then the traction wheels aren't fully doing their job.

[msigalov]

Originally Posted by Scott Kozutsky

Generally speaking teams like to put the traction wheels on the outside of the drivetrain. If you get pushed back onto your omni wheels in a pushing match then the traction wheels aren't fully doing their job.

But what difference does it make? We can't get pushed back onto omnis by another robot since the piston controls the switch.

[GeeTwo]

Originally Posted by msigalov

But what difference does it make? We can't get pushed back onto omnis by another robot since the piston controls the switch.

It depends on how high you lift the omnis, or more likely push the traction wheels down. If the omnis aren't above the line between the bottom of the traction wheel and the bottom of the rear frame or bumper, then yes, you can be pushed onto the omni. Many (most?) butterfly/grasshopper designs do not articulate this far. With the traction wheels on the corners, you have a longer wheelbase, making it harder for another robot to lift your traction wheels through pushing. You also reduce turning ability while in tank mode, which is probably why most butterfly and grasshopper designs I've seen have the traction wheels near the center and the holonomic wheels on the corners.

[Electronica1]

Wouldn't this qualify as nonadrive, not grasshopper drive? The reason being you are actuating all the traction wheels, rather than just 2 traction wheels.

[PAR_WIG1350]

Originally Posted by Electronica1

Wouldn't this qualify as nonadrive, not grasshopper drive? The reason being you are actuating all the traction wheels, rather than just 2 traction wheels.

It certainly does qualify as nonadrive, but I'm not quailifed to say it isn't a grasshopper drive, especially since the term grasshopper drive seems to have multiple meanings.

[msigalov]

Originally Posted by GeeTwo

It depends on how high you lift the omnis, or more likely push the traction wheels down. If the omnis aren't above the line between the bottom of the traction wheel and the bottom of the rear frame or bumper, then yes, you can be pushed onto the omni. Many (most?) butterfly/grasshopper designs do not articulate this far. With the traction wheels on the corners, you have a longer wheelbase, making it harder for another robot to lift your traction wheels through pushing. You also reduce turning ability while in tank mode, which is probably why most butterfly and grasshopper designs I've seen have the traction wheels near the center and the holonomic wheels on the corners.

The problem I see with having the omnis in the middle is that it requires having a piston on the center wheel to keep it level with the omnis. As it is the current Square-I pistton is $80 and all the load is going onto the thread.

Another idea is to push the tank wheels just far down enough so that the drivetrain has all 9 wheels touching the ground instead of resting on just the 4 tanks. The omnis wouldn't reduce maneuverability compared to 4 tank and it would have the longer wheelbase and stability of having the tanks on the outside.

Personally, I'm more worried about the way I mounted the gearboxes. I'm afraid that the weight will bend inside plate of the drivetrain. That's why there's a plate attached to the 8020 that supports the CIM. Same worry goes for the strafe gearbox and CIM. Another worry is that the 8020 crossbars aren't enough support for the chassis.

This entire project was worked on by 3 students, no mentors have helped yet. We want it as a relatively low budget drivetrain for the off season to prepare for a strong start next build season.

[Abhishek R]

I recommend you check out this document. 624 actually tried to make a simpler version of nonadrive/butterfly a few years back, and we dubbed the result the "Grasshopper" drive.

http://www.chiefdelphi.com/media/papers/2968?

The experiment was pretty successful, achieving all the strengths of the standard nonadrive setup while being relatively simple and light.

On the idea of the static strafe wheel - we ran that setup in the original iteration. It worked generally fine most of the season, but every now and then we would find a rough spot in the field where we couldn't strafe. In the offseason, when we made it actuate up and down pneumatically, we realized the static version had actually been causing a lot of friction in the drive. As soon as it was lifted up off the ground, we saw significant speed/acceleration improvements and increased battery life. Someone at one of our offseason events even asked if we added another CIM to our drive!

[msigalov]

Originally Posted by Abhishek R

I recommend you check out this document. 624 actually tried to make a simpler version of nonadrive/butterfly a few years back, and we dubbed the result the "Grasshopper" drive.

http://www.chiefdelphi.com/media/papers/2968?

The experiment was pretty successful, achieving all the strengths of the standard nonadrive setup while being relatively simple and light.

On the idea of the static strafe wheel - we ran that setup in the original iteration. It worked generally fine most of the season, but every now and then we would find a rough spot in the field where we couldn't strafe. In the offseason, when we made it actuate up and down pneumatically, we realized the static version had actually been causing a lot of friction in the drive. As soon as it was lifted up off the ground, we saw significant speed/acceleration improvements and increased battery life. Someone at one of our offseason events even asked if we added another CIM to our drive!

Your drivetrain was the inspiration for our design, hence the most likely incorrect grasshopper name. We wanted to avoid tensioning, but we will try to figure out how to vertically actuate the strafe wheel. Would it need to be actuated when we are strafing, or whenever it is in h drive?

[Abhishek R]

Originally Posted by msigalov

Your drivetrain was the inspiration for our design, hence the most likely incorrect grasshopper name. We wanted to avoid tensioning, but we will try to figure out how to vertically actuate the strafe wheel. Would it need to be actuated when we are strafing, or whenever it is in h drive?

We only actuated it when we were strafing. I'll try and post a pic if I can. There's a bit of lag between the time the driver hits the button and the robot actually starts strafing. If you can, the 148 method of actuation via the torque of the gearbox itself is probably better. I saw 987 used that system this year too, maybe someone from one of their teams may chime in on how it worked in more detail. (They may have posted elsewhere on CD about it too.) Either way, it doesn't need to be actuated when you're in H drive all the time, only when you're strafing.