pic: Inverted CIM 2-Speed Gearbox



Dogshifter gearbox with the CIMs flipped so that the front face faces the inside of the drive chassis.

Specs:
Weight: 7.87lbs
High gear ratio: 5.29:1
High speed: 17.51fps
Low gear ratio: 8.47:1
Low speed: 10.94fps

Speeds are if the wheels are 4in diameter, although the treads may add to make 4.2in diameter. In addition, the current draw per motor seems to be significantly high (102amps and 65amps), which concerns me.

I don’t plan to change this; I would rather make a new one, and try a different piston mounting as well as different ratios, which I personally found difficult to find because the way the CIMs are mounted limits what gears can go in unless there’s an idler gear or some other way to increase the distances between the CIMs…

Well, if anyone wants the CAD, then I will gladly send the .STEP file

Looks good. You mounting of the idlers is interesting.
That current draw seems a little low for 17.5fps. What CoF are you using to calculate that?
What is your weight at?

CoF is 1.3
Would it be higher if I’m using the rubber treads that go on the performance wheels?
And the weight is under the specs: 7.87lbs.

I’d lower your low gear quite a bit. For 2 CIM gearboxes, you wanna pull around 40 amps each pushing in low gear, so you don’t trip the 40A breakers very easily. Between voltage drop and the breaker safety margin, you can afford to draw a bit more theoretical current, but no real need to push it. You’re not driving in low gear to go fast. Around 6 FPS (81% speed loss constant) is roughly where you want to be, not 10.

This is dependent on your design objectives. 2 speed shifting gearboxes are desirable because they allow for design criteria which are often conflicting with a single speed gearbox. There are kind of 2 “schools of thought” as far as the design criteria each gear is intended for.

School A:
-Low gear is intended for pushing matches, and in some cases fine movement (e.g. bridge balancing in 2012). You want this gear to be traction limited such that the wheel slip condition (which is when torque the output from the gearbox overcomes the static friction of the wheels on the floor) occurs around or below 40 A per motor. With 6 CIMs, you might need to begin to consider the Main Breaker 120 A limit, but be sure to check out the spec sheet and understand how long your breaker will last at max current for the drive train.
-High gear is what you use at all times when you’re doing anything other than pushing an opponent or doing fine motion which is less effective at a higher speed. The aim should be to minimize the travel time for a given sprint distance that aligns with your style of game play. It’s nice if this is traction limited, but don’t lose too much sleep over the actual current draw numbers at your traction limit since you should never hit those with good, practiced driving.

School B:
-Low gear is intended for completing game objectives with a short sprint distance, and again some fine movement tasks. The gearing should be chosen to optimize that sprint distance. You want this gear to be traction limited such that the wheel slip condition (which is when torque the output from the gearbox overcomes the static friction of the wheels on the floor) occurs at some value which will allow you to be in a pushing match for some amount of time. You will also use this gear when in pushing matches.
-High gear is is intended for completing game objectives with a farther sprint distance. The aim should again be to minimize the travel time for a given sprint distance that aligns with your style of game play. Similarly to School A, it’s nice if this is traction limited, but don’t lose too much sleep over the actual current draw numbers at your traction limit since you should never hit those with good, practiced driving.

The merit to School A is that it’s often easier on the drivers to have a simple dichotomy of which gear to use when. When you’re mindset is “Pushing match = low gear, Every other situation = high gear”, it’s hard to go wrong. It is also good peace of mind to never worry about losing power during a pushing match, especially in years like 2014 (well…almost never :rolleyes:). Chris is spot on with his 6 ft/s suggestion if your design objectives align with School A. Depending on your efficiency and your CoF, 5-7 ft/s is a general range that gets you down to 40 A per motor at your traction limit for a full weight robot with battery and bumpers.

If my memory serves me correctly, one example of a School B design is the Killer Bees’ robot in 2013. Their robot was a floor pick up machine and would often times pick up discs from the floor during teleop if the opportunity presented itself. However, there weren’t always discs on the floor. In that case they would drive to the opposite end of the field to get discs from the feeder station. Playing the floor pickup role was a short sprint distance objective, and playing the feeder station role was a longer distance objective. They couldn’t necessarily know going into each match what role they would play, and sometimes it would change throughout the match, so having a separate drive train gear ratio for each style of play was an elegant solution. It seemed to work well, they were World Finalists after all.

In general, drive train gearing is a trade-off between how long it takes you to go from point A to point B and how much current you’re pushing through your breakers. When going above about 15 ft/s, you start to reduce your pushing force and initial acceleration in high gear at the expense of additional top speed (assuming a full weight robot plus battery and bumpers).

I thought 33 2013’s robot used shifitng for improved acceleration with their autoshift code.

I’m not crazy about the tiny idler gear - seems like it’d wear very quickly and as a result become a detriment to efficiency as the season went on. Is that the largest gear you can get for that initial stage?

More and more I find myself learning from the masters in School B for certain games. 2011/12 would more have a School A thought - the field was congested or blocked and objectives needed slow/meticulous movements. 2013/14 required much less precision in control (so long as the intake made up for margin of error) so School B would have dominated.

As a result

  • School A needs the best high-traction treads available to maximise performance in the situations the design is made for.
  • School B needs tread that WILL slip under enough torque so as to not trip the breakers. Colsons fit that spec nicely, imo.

I’ve found colsons tend to be rather grippy if you’re aiming for slippage under high traction. Something like sky wheels or the white KOP wheels would work better IMO.

I’m not crazy about the tiny idler gear - seems like it’d wear very quickly and as a result become a detriment to efficiency as the season went on. Is that the largest gear you can get for that initial stage?

So the main purpose of the idler gear is to get the gear space in between the CIMs because the way the gearbox is designed. If we do end up doing idler gears, then we could just replace them with new ones if they ever wear down. On the other hand, the big gear is a 72T gear, so I might be able to replace it with an 84T and remove the idlers. However, I remember on another gearbox I was working with another person, we had trouble with the gear sizes with an 84T.

I plan on making another gearbox of a similar design, and perhaps take a look at Andymark dog gears, which come in smaller diameters, so hopefully that will eliminate the need for dog gears.

I’ve found colsons tend to be rather grippy if you’re aiming for slippage under high traction. Something like sky wheels or the white KOP wheels would work better IMO.

I was thinking this gearbox to be one of those that goes decently fast on high gear, and slow for pushing matches on low (I guess this would be School A?). With that being said, I was considering the nitrile rubber treads that go on Andymark performance wheels.

Hopefully this is the right way to go if I’m going for a maneuverable and defensive bot?

Switching to belts might be a good solution here. Plus, you’d get a nice efficiency boost.

You know which wheels aren’t too grippy and work really well for slippage? Lunacy wheels :smiley:

If it weren’t for the emoji, I’d have to report you for trolling. My right eye is twitching, even though I never heard of FRC until April or May of 2011.

You may be right. You likely had more interaction with them that year than I did. The information from my post was what I remember from a conversation I had with one of their members at Championship when I asked how they arrived at their drive train gear ratios (they seemed odd to me, and I wanted to learn more).

I would not be surprised at all to hear that they incorporated autoshifting in 2013, given that they developed a 4-speed autoshifting drive train and code as far back as 2004. However, the logic for gear ratio selection for autoshifting would still probably be consistent with the thought of having good sprint distance for both floor pick up and cycling. Maybe someone from Killer Bees could fill us in with more details?

Interesting…I would have classified 2014 into School A given the prevalence of defense and the fact that just about every team would be playing D at one point or another. That is to say, sustaining a pushing match for 15+ seconds would be more valuable than shaving a few tenths of a second off of a 10 foot sprint. But like we were saying before, depends on which criteria you’re aiming to meet.

Actually, I’m not entirely sure what the CoF of a Lunacy wheel is on carpet. I’m pretty sure it’s more than on that pesky Glassliner FRP (AKA regolith), though.

For those that don’t get the reference… 2009 (Lunacy) was played on a field of Glassliner FRP with about a foot of carpet on all sides next to the rail. If you want to know what that looks/feels like, there’s probably something similar in your nearest school/park restroom (as an anti-graffiti/easy-clean sort of measure). All robots were required to use certain wheels for their floor-contacting propulsion–the CoF between said wheels and the floor was something just under 1 as I recall, while your typical nitrile wheels are 1.something-or-other. Unlimited quantity… but that was the ONLY type allowed! Low-traction game, low-speed, low-friction…And then there were the trailers, but I’ll end there.

I was far under 1 iirc. The regular tread wheels get around 1.

Just under 1 is the CoF of the 2008 Kit wheels, if memory serves, which were a bit slicker than other years. The CoF of lunacy wheels on FRP was reportedly 0.06, though it was probably 2 or 3 times that on a worn field.

You could eliminate those idlers by making your own gear sets. A quick search in the white papers using my name or “2005 716” will show a similar gearbox that we have been using off and on since 2005. The 12 tooth intermediate low gear has never been a problem for wear. This gearbox can be made on manual machines. One change is the that we use the stock CIM gear instead of the fancy tapered thread version in the plans.

As a student machinist for my team, I would recommend against making gears if you can avoid it at all. Indexing and cutting 2-3 times per tooth takes a very long time, not to mention all the work of taking a vise off the table for an indexing head or rotary table and centering everything. If you have the resources, go for it, but personally I can think of few situations where you would really want to do that (a 26t or 16t gear being those few situations :P). I’ve designed swerve drives that *can * be manually machined, but not ones that you would really *want *to.
Buying spur gear stock or something similar would be good for making many custom gears.

I also do not recommend machining your own gears unless you have the time. I have been using Martin spur gears and modifying the hubs as needed. They come in almost every gear size and are available on line and from McMaster.

I dunno, we run about 3.5 fps in games where we’re doing pushing and it works pretty well for us. I’ve yet to find a robot (or two at once) our 2014 machine couldn’t push. :rolleyes: