2399 Inspired Octocanum Module -- Looking For Feedback

Imgur Album

This design, inspired by team 2399’s octocanum pod, is what my team is thinking about using for next year. We plan to mount it on the stock Andymark frame by widening the “wheel wells” and having it pivot around the mecanum wheel. There are two different gear ratios for the 4" mecanum and 4" colson. The colson gear ratio gives an adjusted 5.85 FPS and the adjusted mecanum ratio gives ~20FPS. We plan to add an encoder onto the colson wheel and implement motion profiles and velocity PID control for auto. What do y’all think about this design? We’re slightly worried about the sturdiness of the 4" vex mecanum. Anyone had any experience with this wheel?

Any feedback is greatly appreciated.

Note:This design isn’t completely finished, I still need to add a place to actuate the pod, which we plan to do pneumatically.

Interesting design. I like how you play with the gear ratios on the colson and mecanum. Using the WCP lightened gears could help you reduce the gear weight on this.
However much I’d like to see a mecanum drive careen across the field at 20fps, I don’t think it would work well at all. The colson speed is perfect for traction-limited pushing and defense though. Consider switching to something more manageable for mecanum, like 17fps free speed (14 adjusted). That’s about the fastest I’ve ever seen a mecanum drive go.

Another thing to note is that you’re using a CIM + miniCIM per pod. Although that’s very doable, given the speed constraints of a mecanum and the low-speed colson, having a power slightly surpassing a 6-CIM drive won’t help much. A 4-CIM drive will do the same job just as well in this case.

Definitely add encoders for auton purposes, and later on for general speed control. Given that it’s only 2 gear stages to the wheel, even a Munchskull CIMcoder (not the AM one) could work fine for fairly good accuracy.

Random side note: if you already have your ratios set, it can be better to make your colson axle the fixed one. That way there’s a slight suspension effect of the mecanums to ensure they all touch the floor and provide traction.

First off, I really like your idea of rotating around the colson as that effectively gives us pneumatic suspension for the mecanums. I also like the idea of using the lightened gears as this could get quite heavy.

Do you think this is the fastest you’ve seen a mecanum drivetrain go because most people dont use 2 cims for their mecanums or because it wont work at 18-20FPS? I feel like even if we geared for 20FPS and that turned out to be to fast, we could limit the throttle in our code because of all the torque and acceleration headroom we have with the two cims.

One of the design constraints of this design was that it needed to be powerful and able to push. Do you think with the 2 cims (1 full, 1 mini) per pod we could push or at least defend againt 6 cim bots? I’m willing to have the extra weight of the mini cim so as to ensure that we can push and defend against other 6 or even 8 cim bots.

What do you think?

Hey! Nice work!

I followed up on your PM about this; my feedback breaks down into three parts, as follows:

Actual concerns

  • You will absolutely want to put another bearing on that idler axle, and it seems like you left a bearing hole underneath the CIMs for that exact purpose.
  • Actuating is probably going to be trickier than it seems right now. I spent about equal time designing the module as figuring out how to make it actually swap wheels, since there’s a pair of torque equations involved, and you’re constantly trading module size for piston bore size. I sent you a spreadsheet to help with this
  • I’d recommend tossing the encoder you’re planning onto the inside of your chassis, rather than onto an actuating part of the module. Less moving around of wires.
  • 20fps is probably a touch fast for the high gear end of this; I’d need to do some math but you’re pretty close to the “can’t actually reach top speed” line, given the low effective COF on the mecanums and size of FRC fields.

Quibbles

  • You’ve got a couple of interior corners that are still, well corners. These are effectively un-machinable. Throw a fillet on them and your machinists/CNC operators will thank you.
  • I’m not sure there’s enough meat left in those rings around the CIMs to actually support them if they should get pushed into them, rather than just permanently deforming the rings.
  • Don’t forget to retain the CIM pinions. Grooveless push retaining rings are standard here.

Dang, that looks good

  • Your plates look generally appropriately lightened, while still leaving a good amount of material near stress points.
  • Fasteners! Super important in this type of design; if you miss one, there’s suddenly no gear clearance.
  • Just about everything else. This looks super clean, looking forward to seeing the final CAD, or manufactured version if you intend to pursue that.

Are you saying robot powered in this way will have the same pushing power are a 4 cim drive?

All other things being equal, yes, if both systems have the same traction and are traction limited they should have the same pushing power, however, the 4 CIM drive will pull more current per 40A breaker than the CIM+MiniCIM drive will, but the CIM+MiniCIM drive will also use more current overall. If neither system is traction limited, then the CIM+MiniCIM should have the advantage because it will take more force to stall than the 4 CIM drive will.

EDIT:
One other thought on this drive, you might want to consider using WCP Pocketed Gears to save weight, especially since you’re using such large gears. I’ll also second wmarshall11 comment about adding a second bearing to that idler axle.

On that note, what are peoples experiences with using those clamping Hex Collars on gearboxes/drives like this? Personally I’d be terrified to use them since in my experience they are almost constantly sliding out of place. Thoughts?

What would you recommend using instead?

Personally, I would drill and tap a 1/4-20 hole in the end of the shaft and use a screw+washer to hold the gears/wheels in place. Alternatively, you could still use the clamping collars but I would drill and tap another hole or two in them (pointing towards the center of the shaft) so that you could tighten a set screw through the collar directly into the shaft to prevent it from moving.

Generally these kinds of gearboxes have an additional plate on the outside of the gears to allow for another bearing to support the shafts and improve rigidity, that would be the ideal solution here but would also increase the size and weight of the module. Spur gears are very sensitive to any variation in angle or center distance between gears, even a ±0.002" variation in center distance or a few degrees of flex can be enough to cause serious damage to the gearing after a few uses. Gearboxes that have cantilevered gears like this usually get around this problem by mounting the gears as close to the bearings as possible (since the shorter lever arm reduces the chance that the forces will be sufficient to bend the shaft), but this design has two rows of cantilevered gears which may be more susceptible to flexing. Now that said, the arrangement of the gears contains the center idlers pretty effectively, but keep in mind that any additional pressure caused by the outer gears will likely transfer across to where your pinion gears interface, potentially causing the teeth of the pinion and idler to be closer and could cause the teeth to pinch and possibly even break on each other (speaking from first-hand experience).

18sampson updated the CADs and uploaded some updated renders. He told me to post this response because his CD account still has the lengthy new-account moderation period after each post.

Done

We moved fixed axle to the Colson. I was thinking we can just turn down the end of some hex shaft to 1/4" and mount the encoder on the inside of the frame.

There was a mistake when calculating gear ratios (pointed out by wmarshall11), so while we were fixing the issue we swapped the gears for the pocketed WCP ones.

Here are the updated renders

Note: These renders are still a little out of date. I think by now 18sampson has removed the excess idler shaft going to the back plate (in the renders I linked earlier in this post) and is planning on adding standoffs with button head screws which attach underneath the idler gear (and go to the back plate).

There was also a response on CD-media which advised removing the CIM supports as they provide the potential for completely ripping apart your pinion gears. Do you guys recommend this? Is there a better way to support CIMs and correctly mesh them?

Is this now 2 CIMs per module? Because if it is, you won’t be able to have 4 modules (with 2017 rules).

Also, I would support the whole face of the CIM. If you don’t, you have a much higher chance of the CIMs bending away from the center gear, causing axial misalignment in the gears (which they are very sensitive too).

No, sorry. It’s supposed to be one mini-CIM and a CIM.

Any chance you could get a shot from the “front” looking down the drive base? I’m trying to see how everything is spaced with each other.

Given the supports are part of the gearbox I don’t see how this would be an issue as the motors should move relative to the rest of the gearbox. The only reason this might cause issues would be if the assembly flexes too much. If you wanted to play it safe, you could try widening the supports, leaving a gap between the CIMs and the support ring, then fill in the space with some kind of compression foam or rubber, that way if there is flex it won’t apply too much force to the motors but will still support them.

I would be hesitant to remove the supports completely because in the case of this drive, the motors will be actuating with the gearbox (which is fairly uncommon). The repeated shock-loads will not be healthy for the mounting holes or the pinion gears if not supported properly.

Sure… here’s an a360 link he sent me earlier.

https://myhub.autodesk360.com/ue29e70b1/shares/public/SH7f1edQT22b515c761ee25c24e81860abc0

I like the compression foam/rubber idea a lot. It might also possibly help with motor vibrations…?

Thanks that helps. Looking at this I feel like there should be some way to reduce the amount of material either in the new outer aluminum brace or in the center one but I can’t quite put my finger on how to do it. If not for the motors, you would only need the two outer plates and could remove the center one entirely, saving weight and potentially shrinking the package a bit. Perhaps there is some other way to securely attach the CIMs without connecting the mounting plate to the lower shafts? :confused:

Alternatively, instead of modifying the center plate, you could modify the new outer plate so that it only interfaces with the 2 lower screws of the motors (removing any material above that point) and make the pockets bigger. Keep in mind that this is really just a stiffener so it shouldn’t be taking any major loads.

Also, any idea how much this weighs currently?

Possibly, though I don’t think the effect would be significant (and motor vibration isn’t generally a major concern with CIMs anyways). Also keep in mind that you would want to limit the use of foam/rubber since overuse will affect the thermal properties of the motors (more heat buildup).

With about 0.56 lbs estimate for the weight of all of the plates, we estimate it comes in around 7 lbs per module. I do agree with your ideas about removing some material in either the outer or middle plates… I can’t put my finger on it either. I’ll see what 18sampson thinks about it

We were thinking some kind of TPU liner to the inside of the motor supports. This technique is commonly used for reducing gyro noise in quadcopters, so it might be worth a shot to see if adding these TPU liners makes the gearbox run more smoothly (as well as supporting the CIMs)

Might help with encoders if you decide to add those on each module since those can be susceptible to vibration-induced noise (though most of your vibration is probably going to come from the Mechanum wheel). Certainly worth a try in any case.

At .56lbs of aluminum it might not even be worth changing, the 7lb module is already about 0.5lbs lighter than if you were just to use the same motors and wheels and use standard VP 2CIM double reduction gearbox to drive it. You end up saving even more in the long run since you don’t need to build a separate standalone module just for the wheels.

Here’s our lexan prototype for anyone that’s interested: http://imgur.com/a/2mQFX

Supporting the CIMs using just the face mounts should be sufficient. If it works for COTS gearboxes, chances are it will work for you. Just make sure to give the CIMs more clearance around its body, since they’re not guaranteed to be exactly 2.5". They are however guaranteed to be smaller than 2.536 according to spec sheet.

Is there a bolt somewhere on the bottom side of the module? If not, right now the only parts stopping the bottom part of the plates from flexing are the live axles themselves. And the plates will flex when the mecanum wheel puts a sideways load on it. Put some solid support down there.

Another reason to add bolts+standoffs to the bottom side of the module is that right now, you have to take the gearbox apart to replace the CIMs. That isn’t bad, just inconvenient- you can’t assemble the gearbox without the motors.

The difference is, COTS gearboxes are not experiencing repeated shock loads. I’d be more comfortable adding the support at the back of the CIMs just to be safe. Better safe than a shredded gearbox. As for the tolerance… I was thinking we could print some TPU liners for the inside of the plate so that the CIM is supported but not completely “locked”