Sheet Metal, 8-Wheel Drive Design

[woodenLogicGate]

Hello All,

This year our team is looking into using sheet metal to manufacture a drive train and possibly other components of our robot. We are new to sheet metal design and manufacturing and would like to get your input on our current design.



The general design is an 8-wheel drive with the middle two wheels lowered so that the robot normally drives on the middle 4 wheels. The sheet metal is 0.09" thickness. We're using Vex Pro's 3-stage ball shifters to directly drive two of the wheels on hex-shafts, and the rest are chained and dead-axle.

Here's a side view of the drive train. The middle wheels are lowered 1/16" from the outer wheels



Here is a top view of one of the drive modules. vertical discs that you see here are actually sprockets. Please comment on our chaining/sprocket setup. The piston is a concept for a brake. The white cross members are part of a tensioning system.



We've calculated that our speed in low gear should be about 5.2 ft/s and our speed in high gear will be nearly 12 ft/s (probably a little less). Are these reasonable ratios to give good acceleration at start and good top speed?

Thank you for your feedback

[MichaelBick]

1/16 is probably not enough drop on an 8wd. You actually want more drop on an 8wd than on a 6wd. 12fps seems a bit low for a 2 speed drivetrain(what efficiency is this). Usually you want to gear a bit faster to take advantage of the 2 speeds. Also that lightening pattern looks hard to do. It's a lot better to make the lightening pattern out of round holes because it is easier/faster for the shop.

[n1ckd2012]

Originally Posted by MICHAELABICK

1/16 is probably not enough drop on an 8wd. You actually want more drop on an 8wd than on a 6wd. 12fps seems a bit low for a 2 speed drivetrain(what efficiency is this). Usually you want to gear a bit faster to take advantage of the 2 speeds. Also that lightening pattern looks hard to do. It's a lot better to make the lightening pattern out of round holes because it is easier/faster for the shop.

Vex does provide an alternate gear set for the ball shifter at a slightly faster ratio. Our numbers do not take effiency into account. Still need to run the various ratios through the JVN spreadsheet.

Alternatively we might use this 3 motor gearbox:
http://www.vexrobotics.com/vexpro/ge...reduction.html

More specs:

The wheels are 6x1.5" colons with WCP live and dead hubs.
The tensioners are HPDE rods. You twist them to vary the tension.
Chassis is 27x33 in long configuration.
The braking method is borrowed fr 1477's 2012 robot.

[MichaelBick]

Originally Posted by n1ckd2012

Vex does provide an alternate gear set for the ball shifter at a slightly faster ratio. Our numbers do not take effiency into account. Still need to run the various ratios through the JVN spreadsheet.

Alternatively we might use this 3 motor gearbox:
http://www.vexrobotics.com/vexpro/ge...reduction.html

More specs:

The wheels are 6x1.5" colons with WCP live and dead hubs.
The tensioners are HPDE rods. You twist them to vary the tension.
Chassis is 27x33 in long configuration.
The braking method is borrowed fr 1477's 2012 robot.

Yeah that's really slow for a 2 speed. I would reccomend finding gears from vex that have the same combined tooth count that get's you to around 16fps(100% efficiency) give or take.

[jwfoss]

Excellent mix of COTS parts with a custom chassis.

A few lessons learned from 2168 over that past two years with sheet metal work:
- On the inside rails the C shaped bend configuration makes it difficult to mount superstructure components on the flush surface, we now use an S shaped rail.
- With the ground clearance you have you might want to consider a 4"/5" and adjusting your gearing accordingly
- Gearing wise you are being conservative, those are similar numbers to our 2011 robots initial speeds, we eventually changed our output sprocket to increase our high gear to ~14ft/s
- In sheet metal we like to run all dead axles, this also allows us the ability to change our gearing by only swapping a sprocket.
- I would increase the center drop to 1/8" (that's what our 2012 robot ran with the 8wd) Colson wheels do wear down, so be aware of this as well.
- With lightening patterns talk to your sponsor, if they are using a punch it may be faster to manufacture with holes of standard sizes

Keep up the good work.

[sdcantrell56]

Also that method of braking is OK, and worked decently well with 1477's robot because they used pneumatic tires which had give. Colsons have nowhere near as much give so you will most like not have much braking force at all. Look into a different method for braking, something similar to what 25 has done in years past.

[Madison]

Originally Posted by jwfoss

- On the inside rails the C shaped bend configuration makes it difficult to mount superstructure components on the flush surface, we now use an S shaped rail.

Justin, this is a bit unclear to me. What do you mean by 'flush surface' and how does using an S help?

[apalrd]

I don't really see why 12fps is a bad speed on it's own.

For a half-field game (Where the field is divided), it's a good speed. It would be a good speed for the 2012 game.

You don't gear for top speed. You gear for best game performance. This is usually a sprint-distance of no more than 20ft for a very long distance game. Acceleration matters a whole lot more than top speed in many games, especially short-move games.

[jwfoss]

Originally Posted by Madison

Justin, this is a bit unclear to me. What do you mean by 'flush surface' and how does using an S help?

In the attachment below, I quickly drew up a set of chassis rails (looking at the robot from the front). The left side shows the configuration used by the original poster. In this configuration the height of the rails is determined by the height of the gearbox to avoid having to remove a section of the flange. The right shows the flange flipped the other way. Doing this allows height of the rails to be independent of transmission, and allows the mounting of a tower to the internal surface, while retaining a full flange.

There are many ways to configure the rails, our most recent concept has the outer c rail flipped as well. You can see some of FRC2168s cad development as well as some of our past designs by reading my blog (click my name in my signature).

Please let me know if this needs further discussion.

[Madison]

Originally Posted by jwfoss

In the attachment below, I quickly drew up a set of chassis rails (looking at the robot from the front). The left side shows the configuration used by the original poster. In this configuration the height of the rails is determined by the height of the gearbox to avoid having to remove a section of the flange. The right shows the flange flipped the other way. Doing this allows height of the rails to be independent of transmission, and allows the mounting of a tower to the internal surface, while retaining a full flange.

There are many ways to configure the rails, our most recent concept has the outer c rail flipped as well. You can see some of FRC2168s cad development as well as some of our past designs by reading my blog (click my name in my signature).

Please let me know if this needs further discussion.

Thanks; that is was I suspected you might've meant, but wanted to be sure. I think that I'd likely mount structure to the drive by sitting it atop the flanges rather than by abutting the inside face, so it hadn't occurred to me that the flange might get in the way.

[lynca]

Originally Posted by sdcantrell56

Also that method of braking is OK, and worked decently well with 1477's robot because they used pneumatic tires which had give. Colsons have nowhere near as much give so you will most like not have much braking force at all. Look into a different method for braking, something similar to what 25 has done in years past.

Thanks for the tip, do you have a link or reference design for a brake system from team 25?

[sdcantrell56]

Originally Posted by lynca

Thanks for the tip, do you have a link or reference design for a brake system from team 25?

Search some of their pictures on here. Basically it pushed pins into a hole pattern in a gear to lock the drive

[Madison]

Originally Posted by lynca

Thanks for the tip, do you have a link or reference design for a brake system from team 25?

There's a brief bit about it the Aim High edition of the Behind the Design books if you have that handy. They drove a pin into a gear fairly high up in the gearbox using a servo.

[roystur44]

I like your design and it's simplicity. Don't need a lot of rivet holes in the front and back. Just need the correct amount. Standard 1.5 inch spacing is adequate with the correct size rivet. Since you are cutting the main sheet metal parts you can omit the spacers between the side plates with a series of riveted brackets. I think the brackets are lighter than the round standoffs and provide more surface area to stiffen the frame
You could build in a battery holder to the rear flange. I'm a fan of the staggered round circular lightening pattern. Much easier to manufacture than the diamond pattern. Oh and don't forget the access holes on the side plates to tighten the transmissions to the frame. #25 chain?


I've been trying to get my guys to use the s rail configuration to standardize the mounting of the manipulator but they have been hesitant.
Keep up the good work.

[apalrd]

I attached some pictures from my 'robot archive' of team 25's drivetrain components up close.

I have no idea where these came from. I did not take them, they could be in CD-media, or from somewhere else. I'm not entirely sure any of them are from the same robot.

I hope these illustrate what they are doing a little.