pic: Drivetrain Concept

CD-Media: Photos
1

dad14e42a4cc2ed8e4c4b0d6bd1905e3_l.jpg
dad14e42a4cc2ed8e4c4b0d6bd1905e3_l.jpg2560×1440 338 KB


This is a drivetrain that I’ve been working on for the past few weeks. It is a slide drive with drop down traction wheels, and is designed to be extremely compact. Almost the entire drive is under four inches above the ground. The side omni wheels are geared for 20 fps, the center omni wheel is geared for 13 fps, and the traction wheels are geared for 7.5 fps. The center omni wheel is pushed down by a set of springs to maintain constant contact with the ground. It weighs 43 lbs according to Inventor.

To cut down the height, I connected the pistons for actuating the wheel modules to a linkage and ran them underneath the drive axle for the wheel module. It’s difficult to see in the picture since I have not added chains yet, but a chain would go from each of the gears in the side gearbox to the drive axles to the wheel modules. This makes it easy to adjust the speed of the wheels in the wheel modules to optimize the drive for any game by changing sprocket sizes and adding chain tensioners.

2

Look like a really sweet idea. Only question is will it strafe straight?

3

Looks good. I would be partial to a regular 6WD with the cims in that configuration (for space), but omni is cool too.

How is this going together? Screws? ARe you sure that you can assemble it right?
What’s the material?

43lbs is pretty heavy. There’s probably a problem somewhere.

4

It has a lot of motor weight 6cim/2miny cims… its a lot of weight in motors.
Even more then a swerve would be.

5

Your Motor weight would be there anyway.

Otherwise, looks great. Just a couple questions…

  1. Are those modules made of lexan? Kinda hard to tell on the render… If so, what have you done to prevent flexing?
  2. Is the slot in the box large enough to fit both gears in?
6

Strafing should be straight, and it should be possible to do anything that a swerve can do as well.

Assembly would either be bolts or rivets, but I am thinking mostly rivets. The frame is made by 1" x 2" aluminum box tubing, and 1" x2" u channel.

For weight, roughly half is in motors, and I could drop a set of CIMs and change the reduction pretty easily to reduce weight. Also, I’m planning on adding some kind of lightening cutout to all of the frame members, which should cut down the weight some as well.

7

Pre 2014, I would say the frame members are way over sized and using thinner metal and a little redesign on the frame would have been recommended for weight. After watching the frame failures in 2014, I would not say it is over built. Still there are other methods to absorb the impact forces with less weight.

8

Keep in mind that there are 3 additional CIM motors in this setup that you normally won’t see on standard butterfly/slide drives. To drop some weight we can remove two CIMs and one mini CIM without sacrificing much as a majority of them are run with 4 CIMs. 2013 and 2014 have been the only years we’ve seen 6 CIMs and 4 mini CIMs in the kop so robots have been getting heavier just in motor weight alone. Our past year’s drive was 6wd six CIM shifters came in around 40lbs so this base gives us more for a few more pounds. Remove the center slide module and you can remove more weight and it becomes a standard butterfly.

This is also a lot easier to build, program, and drive compared to a full on swerve drive so its a more feasible project for us to try out with room to make it lighter.

Nice work Eric!

9

I’m not gonna lie this drive looks beautiful.
I’d love to see it in action- is this a planned off-season project?

10

A few things that don’t pass the “that doesn’t feel right in my gut” test –

  1. Change the channel to rectangular tubing everywhere. Channel is much less rigid in torsion than rectangular tube.

  2. The strafing omniwheel is not well-supported in the frame. You’re removing a lot of material from the 2x1 cross-members for chain clearance and leaving very little wall where the U-shaped cutouts are. The weight of the robot, when it’s sitting on that strafing wheel, is supported almost wholly across the four thin gussets you’re using to couple those 2x1 cross-members to the rest of the frame.

2a. That all presumes, of course, that you’re actuating the strafing wheel downward. It’s unclear if that’s the case. Otherwise, I’d be concerned that the normal force acting on that wheel alone will not be sufficient to move the robot sideways.

  1. Gusset the joints more to give you some additional support against wracking to whole frame into a parallelogram.

  2. It’s hard to be sure from this view, but it looks like you’re mounting the CIMs against the outer flange and race of the bearings. That gives me the willies.

11

How does the center drop down wheel work? Is it one wheel or two wheels?

12

This is impressive. I really like it. I’m assuming you want suggestions so here are a few:

  1. I would change the framing to 2 x 1 x 1/16 tubing (someone pointed out to me that this is lighter as well as stronger than pocketed 1/8th material).
  2. The structure for the middle module could be made simpler. I’m not sure which part is spring loaded so I may be missing something. You seem to have several pieces of flat bar (3/16?) acting as stiffeners across several of the frame members in the middle. This could be redesigned to use sheet metal or standoffs to save weight and space.
  3. Aluminum gears? I’m assuming that’s what you’re using but just checking.
  4. I would take your high gear speed down to 16 fps. Just my preference after running a 20 fps drivetrain this past year. We made the change to 16 fps and really liked it. That was with 4 cims though so you could save weight by taking 2 out or leave the ratio the same and using 6.
13

Wow, thats a ton of CIM motors. I really like the layout of the side modules and the gearbox in the frame.

A few comments/questions:
How do you plan on retaining the bearings and locating the module axially?

The shaft that drives the module is going to be really difficult to do the way you have it. I’m assuming there’s 3 bearings supporting it, two on the inner frame part, and one on the outside. If you had only two, I’d worry about the shaft bending. It’ll be tough to get those three holes aligned with each other, especially if you’re welding it. If one of those holes is in the wrong spot, you’ll “overlocate” the shaft and it’ll bind up. You do have the advantage of using hex bearings, which usually have a pretty sloppy fit.

If your frame bends at all, your center wheel won’t touch the ground. The slots that have been cut out to let the chain by are weakening your center module. These wheels are usually articulated so that you have enough force on the wheel so that it won’t slip. You also want to avoid the opposite, where the robot rocks side to side and one half of the drivetrain is off the ground.

2 x 1 with .063" wall box tubing is a pain to find and is weaker than I’d be comfortable with for a frame. We used .125" thick on our robot, in roughly the same configuration as show above but with a bellypan and we still managed to get bent.

If you want to save weight, you could get .100" thick tubing from vex pro. If you’re really desperate, you could probably put some lightening holes in the sides.

You said you had the ability to adjust the drive ratio with the sprockets on the wheels. I’d wait until you know what distance you want the robot to travel quickly, then gear to get the best acceleration for that distance. That being said, 20 fps is on the fast side. If you do go to four CIMs, I’d take gear to go slower, but that’s my opinion.

14

This is interesting, I like using the pistons on the side of the module to allow the CIMs to be mounted inside the rail and not waste the space to the sides.

Have you thought about supporting the outer side frame rails somewhere along their length? I am not sure how bumpers would mount to this but I would be very worried about those channels bending in and contacting the CIMs pretty quickly.

15

Very cool, I haven’t looked at it with enough detail to comment on it’s structure, other than the fact that you should really not use channel or 1/8th inch aluminum in something like this. 43 pounds seems pretty steep, but once the tubing was switched I think it would be a bit more reasonable. If I were to build it I would put 2 full CIMs on the center omni wheel and use 4 CIMs and 2 mini CIMs on the outside, but I think this depends on the year, and in some years your configuration might be better.

Any plans on building it?

16

The center wheel uses a rod running vertically through the side cross pieces of the frame and the 1/4" aluminum plates. Four springs would attach to a tabs on the top gusset of the cross piece at each corner and the bottom attachment plate on the center wheel module. The plates should flex a little and allow the wheel to maintain constant contact with the ground, and because of the springs, there would be little force pushing up on the cross pieces to bend them, but it would probably be better to change at least one set of the cross piece gussets to 1/4".

For the outside U channel pieces, I was planning on using a structural bumper mounting system consisting of long pieces of box tubing bolted to the frame along its entire length, which would be attached to the plywood of the bumper. This would essentially turn the outer frame pieces into pieces of box tubing without adding the extra weight. For the inner frame pieces, I will switch them to the .1" Vex tubing or 1/16" tubing. How much would increasing the length of each leg of the T shaped gussets on the frame increase the rigidity? It seems like that would improve the rigidity by a fair amount, but I have not made a frame using gussets before, so I have no experience with that.

I was planning on adding some kind of spacer between the two side frame pieces, but I want to be able to easily remove the outside piece to more easily take off CIMs and wheel modules, so I haven’t come up with a good way of doing that yet that wouldn’t require sheet metal or welding, which we can’t do.

To make the wheel module shafts easier to line up, I was planning on latheing down the ends of the shaft to slightly under 1/2" and rounding the edges, then using round bearings. That would cost less and be easier to line up.

To change the maximum speed of the drive, I would just change the ratio using a different reduction in the sprockets. Since we will hopefully try building this drive before the start of next season, I wanted to make it easy to adjust so that we could try different confiurations and possibly adjust during the build/ competition season if we needed to. I set the speed at 20 fps because it seems like it would never be reasonable to gear faster than that.

The CIMs do not touch any of the bearings (that’s one of the reasons the gears are the size they are), but if I were to use smaller gears, I would lathe down the end of the gearbox shaft to 1/4" or 3/16" so that the bearings could be smaller (that’s what I did on the center wheel).

Hopefully, we will actually make this drive, so a lot of the design is based on our machining capabilities. We have access to CNC/ manual milling machines at our sponsor’s shop, but in house, the only metal working stuff we have is a drill press, a small lathe, and a horizontal band saw.

17

Not in my team’s current design, we have a Cim and a MiniCim powering the wheel, and a bag powering the steering. the whole thing is 4cims 4 minicims, and 4 bags…

18

I would reach out to teams who have built an a drive with perpendicular wheel(s) to hear their perspective. At least from watching robots with them in the past, they didn’t seem to use the sideways drive much to make it worthwhile.

19

Just be aware that this configuration would be illegal under the 2013 and 2014 rules. You’re allowed to use up to 4 total of either mini-CIMs, BAGs, or a combination of the two. You weren’t allowed 4 of each.

20

How many teams actually use a H drive? I know that 624 and 148 do, but how many others are there? From watching match videos of these teams, it looks like they primarily use their center wheels for getting out of defense, but do not use them continuously.

Also, has anyone ever tried using a CIM as part of a spacer? If it was possible to mount something to the outside frame piece that touched the back of the CIMs, it would act as several large spacers and make it much more difficult to bend the frame in.