pic: Four Speed Three CIM Gearbox

2002 meant your robot could be several hundred pounds due to how heavy the movable goals were. 71 could drag over 900 pounds that year!

503 found them quite useful, so much so that we had two separate PTOs. 4 CIMs to power our can grabbers and 2 for the elevator. The drivetrain could use 2, 4, or all 6 CIMS.

I mean to be fair, there was no reason to have 6 CIMs in the drivetrain this year either. I know of a team on Einstein that was running two motors. I heard it was two Mini’s at one point, but I don’t recall.

Not one reason? Hint: I just gave you one.

Don’t think of it as a 6-CIM drivetrain, think of it as a 6-CIM powertrain, powering three very different motorized mechanisms in an extremely effective and efficient manner. Many ways to skin a cat.

That’s very interesting. I’ve only thought of PTOs as using all of the motors at once for one function or another. It hadn’t yet crossed my mind that you could be running multiple functions simultaneously by splitting up the motors like that and being able to control these functions independently.

Although what advantages would it serve to have power coming from the same place if you could just use more motors for a different mechanism? The only reason I can think of is because there is a limit on the amount of CIMs you can use and it’s hard to get that kind of power from other motors, so for climbing or stacking you’d want to be able to tap into those motors. But for example there wouldn’t be much of an advantage from using a PTO to also power an intake system that could easily be powered by a smaller motor. It would actually be kind of wasteful.

The ultimate FRC PTO would have six CIMs each running at a bit under 20A, each generating about 150W. A slice of this massive 900W (that’s about 1.2 HP) would then be available to whatever functions the robot needed to perform, whether it was drive, lift, pickup, place, throw, can grabber, or whatever screwy function the GDC required that year. If tapping the “correct” amount of energy off of a rotating shaft and applying it to a task was easy, this would definitely be the way to go. And OBTW, if you weren’t using all 900W, perhaps you could spin up a flywheel, which would allow you to draw a few hundred extra watts later when you need to “go to eleven” for a bit. Theoretically, this sort of PTO sounds like heaven for a game like Aerial Assist which involved defense, driving, shooting, and (at endgame) climbing. The devil’s in the details, as always!

And, for the record, my prediction: The first FIRST team to execute this sort of PTO will join 71, 118, 254, and 1114 as one of the top five teams of all time (unless, of course, it’s one of those four who do so).

First of all, you obviously can’t make a list of the top FRC teams of all time without including 67.

Second of all, the ultimate PTO would use every motor available and just shift however many motors necessary to any function.
(However you’d have to have some crazy programming scheme or something to keep from browning out like crazy)

I also think the first team to do this won’t be one of the greats; I think the first team to try it won’t get picked at their event. That’s insanely complicated, even for a team like 254 or 118. I would love to see someone do it successfully though. it would just be a mechanically sick robot to check out. :smiley:

The reason for stopping at 6 CIMS is that they are capable of efficiently turning every amp you can get through the main breaker into mechanical energy. Once you do that, any additional motors are meaningless.

And I did mean the first team to successfully execute it, not the first to attempt it.

in 2014, a 3 CIM gearbox helped my team be the most dominant team in the PNW for the first 3 weeks (ish). no one could catch up to us, and our rams on defense rekt teams with loose ball carry. we didnt have a 4 speed though… it was just all fast, all the time

I think you went quite overboard on the weight reduction on the outer plate. A pocketed 1/8" plate with no flanges, especially with the reduction being supported off of the end of the plate past the standoffs like that, will have some issues with flex and rigidity. With gearbox designs posted to CD in general there’s often an overemphasis on lightening the plates as much as one possibly can, when it’s really not that substantial of a weight savings over a more conservative lightening pattern or even unlightened plate.

The CIMs aren’t very efficient actually, there are more efficient motors in the kit by a good margin (lower power though).

It’d be nice to get a high power, high efficiency motor on FRC legal list.

Also, it’s simpler for teams just to add more motors and gearing independently than it is to centralize PTO it all. It just doesn’t make sense for FRC.

We are one of those teams.
Due to weight constraints, we went from a robot that had 6 motors similar to that of 2014 to a robot with only 1 CIM on each side. We started with a 2 CIM 1 Banebot 775-18 on each side, then got rid of the Banebot for several regionals, then finally 1 CIM on each side at CMPS. We did this in order to make weight as we added 1, then 2 ramps and also to add our upper claw to hold the trash can in place while stacking.
This game was perfect for having just 1 motor on each drive, considering all we did eventually was make stacks from the feeder station.

Thanks for the feedback. Would you suggest that I cut back on the lightening holes, add some flanges, and/or thicken the plate?

All of the motors that are significantly more efficient than the CIM have their peak efficiency at a much higher speed (11-17k rpm vs 4500 rpm). This means that another level of gearing will be required to reduce these high-speed inputs down to a mechanically useful speed, offsetting most of the increase in motor efficiency. And you’d need more motor controllers, too. At least you could use a single encoder to know the speed of your power train.

No argument. Everything that is currently an independent motor would have to be replaced with a clutch, CVT, or similar mechanism, which brings us back to Kevin’s original suggestion to use the shifters on a PTO rather than to make a 4 speed single function transmission.

I often find that at least thinking about (and sometimes even building) the extreme cases leads to ideas that prove useful later. IIRC, the OP tossed this up as a bit of a “blue sky” thread, and I’ve been treating it as such. Checking back, OP included “probably weighs a thousand pounds” and “I do not actually intend to build this thing.” As a result of this thread, I now know to keep an eye out for high power jobs that don’t require simultaneous wheel drive. 2013 and 2015 both provided this sort of challenge (pyramid climb and can burglar respectively), so it can’t be that rare. It wasn’t in my mental toolbox six months ago.

Another thing that makes me really like the PTO concept is the addition of a flywheel. It is possible to tap more power from a PTO with a flywheel for a short period than from motors, especially given the new motor brownout situation. I like having a defined brownout over an undefined one, but it will probably be several years before the FRC community learns to take the best advantage of the new opportunity/challenge. At which time we’ll switch to yet another control system ;-P>

Thinking a bit more, 2014 (Aerial Assault) also provided a PTO possibility with the ball thrower. We had six CIMs on our 2014 robot, two for a “kicker” and four for drive. (We used pneumatics and an AM gearmotor for the ball pickup.) We nearly always threw the ball while stationary, so sharing all six CIMs between drive and a more elegant and powerful launcher might have been an improvement on both fronts, especially as most of our “accelerator” designs would have thrown the ball from higher off the carpet than the kicker did.

Why did they do that? was it just like, one of the kids said in a meeting, as a joke, “you know what would be cool?” and everyone just kinda went along with it.

also on the subject of small motors on drivetrains, I know that 2471 was running their Swerve on a mini-CIM for each wheel by DCMP in order to have more weight for a better can claw & some can burglars, with no discernible difference in performance from when the had a CIM per wheel.

Nope. 2007 was PHYSICAL. Not only was there ramp climbing (up all random sizes and slopes of ramp), but there were 5 other robots on the field, all trying to score or block scoring in the same general area. This “area”, the Rack, was a structure that could be slid and spun around, took up a fairly large chunk of real estate in the center of an otherwise flat field…

… and needed some degree of precision to place scoring objects (inner tubes) onto its scoring pegs.

All I’ll say as far as it being physical is: Aerial Assist was rough. Now extend and beef up the appendages another couple of feet or so. And “meet in the middle” autonomously…at speed…or under driver control…while trying to score…or block scoring…

118 had a pretty decent swerve the previous year, but for this one they added a pair of F-P motors to a 4-CIM, ran all the wheels off the same gearbox and the same turning motor, and had a turret on top to allow for scoring in any orientation they felt like being in.

Basically, lots of push and lots of maneuverability were judged “pretty important” that year. (Though… IIRC, Einstein was a bunch of 6WD skid-steers that year, with maybe a couple of non-tank drives around.)

I’m still wondering what was up with that central gearbox that was later distributed. It appeared to be an essentially 1:1 gearbox which brought four CIMs together onto a single shaft, which was then redistributed around on four separate swerve shafts. It seems to me that it would have been easier and more efficient to put a CIM on each swerve shaft, but to include a sprocket which tied the four shafts to the same speed to allow redistribution of energy as needed (e.g. as loading increased on two wheels due to acceleration away from them).

They combined it all into a single shifting gearbox (reusing a Dewalt drill planetary). They also added 2 FP motors, so distributing it would’ve been harder.

Traction. By combining the motors onto one shaft, in addition to allowing for a single dewalt shifter, it also allowed any one wheel to have all the motors power available. Thus, even if a wheel lost traction, the full pushing power would still make it to the ground through the others.

Another good example is 148 in 2008… they did 2 FP plus 4 Cim, going to three swerve pods, which were steered with a van door motor. Since they had no “front” (it was an equilateral nonagon lap runner), they had no need to “steer”. As a result, they could both run laps well AND provide very good defense (and this was one of many reasons why their CMP alliance were champions).