Swerve Module 1CIM 1Speed 3.25in vex pro wheel

[jman4747]

1 CIM per module 1 Speed. 1st stage: 12t pinion to 40t 2nd stage: 32t input to 54t attached to wheel.

AM 9015 mounted to 100:1 Vex Pro planetary and 42:24t pulleys (HTD 5mm pitch 9mm wide belt) for pivot.

Module weight < 9.2lbs per. Robot weight supported on Needle roller thrust bearing 1-1/2in ID (McMaster Carr #5909K410).

Vex Pro 3.25in wheel. Vex Pro 3/8in hex id 15t bevel gears supported by steel thrust ball bearings 3/8 shaft (McMaster Carr #6655K150).

Using encoders from Automation Direct (TRD-MXxxxxAD series) for wheel speed and pivot angle.

[asid61]

9.2lbs is a ton for a module like that. These have to be ways to cull the weight. A single speed single cim module should not be more than 8lbs IMO.

However, your module looks like it's hewn from solid rock. So it's never going to break, on the plus side. Definitely a good choice if you want to be careful.

I would flip the cim and use a belt to transfer power to the first stage of transmission and put the a pulley on the drive encoder to engage with the belt. I would also put the turning encoder on a pulley. This will allow the uppermost plate to only have to hold a single bearing.

The side plates on the caster box could be shortened more. The ends don't need to be curved like that either; a long chamfer will remove more weight and be easier to machine.

You can replace the versaplanetary with banebots to cull some pounds but that comes down to personal preference and experience.

Considering the size of the caster box, you could replace the gears with belts or chain, although you would need to broach a pulley from SDP-SI. Worth it for the extra gear reduction you would get, allowing you to use small pulleys on the CIM instead of gears, and then flip the cim.

Skip the 24:48 stage off the versaplanetary ad see if it can be made into a 1:1. Right now it will turn rather slowly, especially for an AM9015.

You can go with a 1/4"-20 bolt for a dead axle instead without hardly changing your design. The gear already floats anyway.

Looking at it again, I'm not sure your weight is correct. Re-material parts and make sure things are aluminum in place of steel, etc.

EDIT: You are geared for around 8.19 fps, it seems. If it's single speed you could consider bumping it up to around 14-15fps, the reason being 8.19fps is way too slow for offense. Swerves can lock themselves in position simply by making the wheels into an X pattern, or even just normal to the opposing robot, playing a very hard defense. Arguably a fast swerve could block better this way by being able to move into their escape path easier. Your call, because 8.19fps does make you into a really good pusher.

[jman4747]

Thank you for the feedback I will definitely change out the wheel shaft for something smaller most likely 1/4in bolt. Also will trim down the bottom of the pivot module. I believe the weight is overestimated for some of the bearings. I always just set those to "Alloy Steel" and leave em but some are solid models thus no air between the parts.

I'll likely rearrange the pivot motor/gearbox to engage the pivot shaft via gears to eliminate the belt (one failed point) altogether. I also think that change will reduce the weight.

Note that I based the pivot motor choice and gearing off of 1640's Swerve. I chose the AM9015 with this total reduction because it mimicked the torque output of a 540 geared 256:1 with a higher speed. I chose the Vex planetary because the cots hex shaft option means I can use the good selection of cots vex pro gears and pulleys. I didn't want to have to modify and custom make as much as possible. I also have and can get free more AM9015 motors through the kit, first choice, and the Andy mark voucher than any other comparable motor choice. At this rate I won't need to pay for any AM9015's, even for replacements.

Not sure about the drive speed. Total reduction is 5.29:1 to a 3.25in wheel. I used the JVN calculator and got 14ft/s free and 11ft/s loaded. I posted a picture.

[asid61]

Whoops, I forgot to change the wheel diameter in the calculator. 14fps is plenty. You could still go faster, but you don't really need to then.
Good choice on the versaplanetary if those ar eyour reasons. You would indded have to get a keyway into a regular pulley if you went banebots. What kind of manufacturing resources does your team have?
The AM9015 is a relatively high power motor. If you want the lower speed, go for it, but buy an extra set of pulleys and belts to test out if you can.
Again, see if you can flip the cim. If will greatly lower your COG.

[jman4747]

Rev 2

Total weight per module is less than 8.4lbs.

Now with gears (24t:42t) connecting pivot motor/gearbox to pivot module. Pivot motor/gearbox is mounted next to the CIM and upside down. This all saves weight, part count, and space inside frame perimeter. Edit: this gets the motor leads away from the floor and removes the belt as a failure point.

Lower polycarbonate mount plate is pocketed for weight.

Pivot side plates are chamfered at the bottom and smaller on the sides.

Wheel uses 1/4in bolt as dead axial.

CIM motor is connected to a AMT103-V encoder in a fashion inspired by that posted by Twins Inc. of 2605 earlier in November (http://www.chiefdelphi.com/media/photos/40928). This will save weight and space, as well as cost in our situation.

Module is now also 100% symmetrical (no left or right versions).

Thank you to asid for the suggetions.

[asid61]

Originally Posted by jman4747

Rev 2

Total weight per module is less than 8.4lbs.

Now with gears (24t:42t) connecting pivot motor/gearbox to pivot module. Pivot motor/gearbox is mounted next to the CIM and upside down. This all saves weight, part count, and space inside frame perimeter. Edit: this gets the motor leads away from the floor and removes the belt as a failure point.

Lower polycarbonate mount plate is pocketed for weight.

Pivot side plates are chamfered at the bottom and smaller on the sides.

Wheel uses 1/4in bolt as dead axial.

CIM motor is connected to a AMT103-V encoder in a fashion inspired by that posted by Twins Inc. of 2605 earlier in November (http://www.chiefdelphi.com/media/photos/40928). This will save weight and space, as well as cost in our situation.

Module is now also 100% symmetrical (no left or right versions).

Thank you to asid for the suggetions.

8.4lbs is much better. I love that encoder mount BTW, as it's really good for a lot of gearbox scenarios.
I don't like how the center of mass moved up, but with the mounting you are doing it's not like it will affect your perfermance much.

[nathannfm]

Originally Posted by jman4747

I believe the weight is overestimated for some of the bearings. I always just set those to "Alloy Steel" and leave em but some are solid models thus no air between the parts.

VexPro has the exact weights for all the common FRC bearings. I'm not sure what CAD program you are using, but you should be able to set weights independent of material.

For rotation speed I agree with asid61, too slow, you are going at about 1/2 the speed of MOEs modules (see my post history) and that works out to about 1.2 rotations/sec. I have heard some teams gear them as fast as 4 rps.

Also, if you wouldn't mind posting a cross section of the coaxial shaft, I'm always interested in seeing how people connect all the parts in this area.

[asid61]

Originally Posted by nathannfm

VexPro has the exact weights for all the common FRC bearings. I'm not sure what CAD program you are using, but you should be able to set weights independent of material.

For rotation speed I agree with asid61, too slow, you are going at about 1/2 the speed of MOEs modules (see my post history) and that works out to about 1.2 rotations/sec. I have heard some teams gear them as fast as 4 rps.

Also, if you wouldn't mind posting a cross section of the coaxial shaft, I'm always interested in seeing how people connect all the parts in this area.

Eh, I said the wrong speed. He's moving at 14fps, not 8. 14fps lets them push the 16fps 4 cim runner, which I think is a pretty popular combo (maximum distance/time for some people). Rps is not so much a measure of speed, because wheel diameter is more important; I ahve designed swerve modules with 5", 4.3", and 2" wheels before.

[jman4747]

The drive shaft Steel is 3/8 hex at the input bevel gear then 3/8 round through the pivot top. It includes a grove for a 3/8 retaining ring that keeps it from falling. just above that it is 8mm with 2mm keyway (same as CIM) to fit on to the Vex Pro 8mm-1/2hex adapter. Needle roller bearings sit in the top and bottom of the pivot top to support the drive shaft. The pivot top sits under a needle roller thrust bearing and has a 1in diameter extrusion that fits into a 1in id ball bearing. This part of the assembly was based off of 1640's design. The pivot top machined to a "D" profile to match a modified gear with the same profile inside. Gear is held down by a 1in id self locking retaining ring.

[asid61]

Originally Posted by jman4747

The drive shaft Steel is 3/8 hex at the input bevel gear then 3/8 round through the pivot top. It includes a grove for a 3/8 retaining ring that keeps it from falling. just above that it is 8mm with 2mm keyway (same as CIM) to fit on to the Vex Pro 8mm-1/2hex adapter. Needle roller bearings sit in the top and bottom of the pivot top to support the drive shaft. The pivot top sits under a needle roller thrust bearing and has a 1in diameter extrusion that fits into a 1in id ball bearing. This part of the assembly was based off of 1640's design. The pivot top machined to a "D" profile to match a modified gear with the same profile inside. Gear is held down by a 1in id self locking retaining ring.

The coaxial shaft can be 7075 easily. 3.25" wheel, and you are gearing down with the gears next to the wheel, so steel is not required there.

Why not try 1/2" hex on top, and 3/8" hex for the bevel gear instead of the keyed thing you have now?

EDIT: Looking at this design, you may be better off buying one of 221's new coaxial modules, as they take 3.25" wheels.

[jman4747]

Originally Posted by asid61

Why not try 1/2" hex on top, and 3/8" hex for the bevel gear instead of the keyed thing you have now?

I can't insert anything larger than 3/8 up through the pivot top because of the bearings. A half hex peg through a 3/8 round hole or even a 3/8 hex peg through a 3/8 round hole. I would need to join another shaft at the top or do something like this or use a gear with a 3/8 keyed id at the top.

Edit: As for buying, believe it or not it would be cheaper not to for us at this moment.

[asid61]

Originally Posted by jman4747

I can't insert anything larger than 3/8 up through the pivot top because of the bearings. A half hex peg through a 3/8 round hole or even a 3/8 hex peg through a 3/8 round hole. I would need to join another shaft at the top or do something like this or use a gear with a 3/8 keyed id at the top.

Edit: As for buying, believe it or not it would be cheaper not to for us at this moment.

And the bearings can't increase in size I guess. This is still a really good design. Just if you you're making this in-house, pay extra careful attantion to how the coaxial shaft is machined.

[nathannfm]

With rps I was referring to rotation of the pod by the planitary gearbox, not speed of the wheel.

And thanks, that cross section helps. So the pod top and D shaft are machined out of 1 solid, reasonably thick block? Interesting.

[jman4747]

Originally Posted by nathannfm

With rps I was referring to rotation of the pod by the planitary gearbox, not speed of the wheel.

How would I calculate the load on the pivot motor?

[asid61]

Originally Posted by jman4747

How would I calculate the load on the pivot motor?

http://www.chiefdelphi.com/forums/sh...d.php?t=129105
Should be in there. It's not too hard.