pic: 2 Speed Switching Module

A back view of the module, it has 2 speed andymark shifting components inside of a switching wheel module and chain links the two wheels. Parts are currently being ordered and designs made for a prototype by end of oct.

I was just taking a look at this, and I had 3 questions:

  1. Why such a large gear coming off of the CIM motor
  2. What are the gear ratios and wheel sizes
  3. What is the total weight
    Looks great though, and it would be nice to see pics of the finished prototype.

The large gear is inaccurate I had to throw it on there to represent how the power is transferred but there will be another gear in between the cim and gear shaft. That way we can gear it down to a desirable ratio instead of being geared up like it is in the render. The gear ratios are traction 1:48, 1:20. Omni 1:20, 1:8.3 . Those ratios can change slightly but most of the components are from andymark supershifters. The entire modules weigh 8 lbs a piece and their will be 4 of them. So they will be similar weight to a swerve module which we used this year.

You know most teams run the omni’s at high gear speed and the tractions at low gear to accomplish effectively having 2 gears each optimized for the task at hand right?

148 drove the omni’s at a high speed off the cim, then a chain reduction to the traction wheel for “low gear”.

With the current setup it pretty much just looks like you’re needing 4 extra pistons for not much in added features.

So basically do you need 2 speeds when in each mode? or will 1 speed optimized for each mode work.

The gearing ratios are in limbo because it was calculated to speeds for the specific gears. The tractions drive at 2ft/s with 60 ft/lbs of torque and the omnis drive at a max 10 ft/sec. For the prototype the team is only putting 2 modules on a frame and putting two tractions in back and with this configuration you could actually run it and have an effective tank drive with some additional torque benefits. The gearing allows for an even bigger ratio between traction and omni.

Do the math and see what level of torque your traction wheel is actually capable of transmitting to the ground, at 2ft/s you’ll just make yourself move really slow and not have any more pushing power than someone who geared themselves right at traction limited.

Perhaps your correct but that’s the reason for the prototyping and that kind of gearing ratio can be changed with the gearing from the cim which would not be terribly difficult to accomplish. The main goal is to see if the module can even withstand the stresses of housing the shifting mechanisms and continue to work properly and reliablely.

5 ft/sec is a common geared speed - there is a reason for this. You should at least layout both sets of gear ratios to be sure they are easy to switch.

You can prototype a low gear with a normal traction base. Just gear it down until the wheels slip in place when pushing against a wall.

Trust me, you will be a LOT happier if your prototype goes 5 FPS in low gear than not.

There is no “perhaps” in this case. You are limited to a certain amount of traction and thus possible force you can transmit to the ground. With that gearing you will not be able to transmit all that force so then you will just be exceptionally slow. Have a mentor with a physics background or at least some physics knowledge explain how to find the maximum force you can apply based on friction and normal force. It will greatly improve future designs.

Also like some people have already said, use the shifting of the modules to accomplish the gear switching. You have no need for 2 speeds on each wheel and it greatly complicates things and makes the entire module heavier.

One final comment is it looks like the gear on the traction wheel is a higher DP than the rest of the gears which is counter intuitive to typically accepted practices. The highest forces will be seen by that gear so typically you want that gear to have the lowest DP. Most teams run 20DP, .375 fce width gears in the final stage of a transmission.

Yes I understand there is a limit for the amount of torque that can be transmitted through the wheels. I have had a year of physics and am in AP physics now. The reason I used perhaps, is the fact of weight distribution and the surface area of wheels and material used all play into the how much of that torque can be distributed. Perhaps a configuration can be made that 3.5 ft/sec still transfers torque effectively through the wheels. That is the reason for the prototyping, in addition if we can successfully create the modules with gearing inside of them; it should be much easier task to create modules without switching speeds. This is merely a prototyped design that is being constructed. There is no belief that it will work flawlessly or be a great improvement it’s more of skill building project for preseason. The gear ratios have been made smaller in consideration for what everyone on here has been saying, the low speed has been moved up to nearer 4ft/sec.

I am also confused on which (higher dp) gear you are referring to? There are no gears attached to the traction wheel. That is driven by chain and sprocket from the omni. HOWEVER you are correct that the omni has a higher dp gear (which is what I believe you were referring to), the reason for this is very simple. It is very hard to get 5.5" diameter gears that have the proper bore size and thickness. The reason for such huge gears, Mentors have strongly discouraged the use of chain to connect the switching gear shaft to the omni wheel (their reasoning still escapes me). The design was modified to accommodate this request.

The first stage reduction (between the CIM motor and shifter cluster shaft) looks like it’s geared faster. Don’t do this in a gearbox if your intention is for the output to be slower. Gearing faster in one stage only to gear slower later is just wasted weight, space, and lowered efficiency. Instead, do like AndyMark does with a 12t pinion on the CIM motor and a 40t gear on the shifter cluster shaft.

As for the higher DP gears on the shifter output, if you could fabricate a gearbox to accurate enough tolerances to properly mesh the 20DP AndyMark shifter gears, you can easily use 25p roller chain or timing belt instead of the higher DP gears.

Didn’t even realize the gearing up and then back down. Additionally, how is the CIM getting attached in the module and is there going to be any type of cross bracing?

sorry but its posted above in this thread, that cim gear is for representation of where the connection is, and is NOT an accurate representation of the actual gearing size connecting those two shafts. Due to time constraints and hw i was unable to get that fixed yet. yes but our mentors will not approve the plan if it has roller chain. We suggested that but I guess previous bad experinces lead them to shy away from using chain. I think chain would be easier/cheaper/ and actually lighter than a 5.5" x .375" gear even with speedholing.

Yes there is a top sheetmetal piece that hold the sides apart and then bracing is brought up to support the cim perpendicular to its output as well as parallel to its output. Due to changes in the design from what I posted just a few weeks ago, theres been inadequete time to flesh the mounting specifics for the motor.

Refusing to use roller chain is silly… It will most likely be cheaper and lighter than gears, and being pre-season this is the perfect time to learn how to use this drive mechanism that is a staple component of FRC.

I think what Sean was saying in one way or another, is that you don’t need to prototype to get a good idea of what to expect in this particular scenario. Prototyping is certainly one way to do it, but the problem with prototypes is that they get expensive quickly. If you go on to study engineering, and eventually become an engineer, you will quickly learn that saving resources (time, money, machines, etc.) is always something to consider.

Going back to your original problem… A 135 lb robot (120 lb + battery, etc) can only exert a given force on the ground through whatever number of contact points it will have. If you have a good estimate on the coefficient of friction between the wheel and the carpet, you can easily estimate what the ideal gearing should be; no prototyping required.


That’s exactly what I was saying. The whole purpose of engineering and math and theory is so you don’t have to experiment to determine if basic systems are going to work or work well. If you want to throw away money on a less than optimized prototype then no one will stop you but if you take the literally 3 minutes required to run basic calculations to determine your traction limited wheel speed you will save precious time in the long run and end up with a much better performing module.

The calculations were done today with some help from two physics majors, so hopefully they were correct.
CIM normal load = 4320 rpm
Torque = 64 in-oz’s
Ff= (c.o.f.) * Fnorm
T= r * Ff
Ratio between high and low gears is approximately 1:2.5
These numbers were calculated using the max weight of a full robot , 150 lbs.
4” Traction Wheel (C.o.f. = 1.25 – 1.5)
**Low- **1:30 gear ratio, **6000 in-oz’s **torque, 2.5 ft/sec, 144 rpm ( before slipping occurs at 6056 oz-inches)
**High- ** 1:12 gear ratio, 2500 in-oz’s torque, 6.3ft/sec¸ 360 rpm

6” Omni Wheel (C.o.f. =1)
**Low- **1:18 gear ratio, **4000 in-oz’s **torque, 6.0ft/sec, 230 rpm
**High- **1:7.4 gear ratio, **1800 in-oz’s **torque, 15.25ft/sec, 580 rpm
(forward/backward direction only)

Slipping occurred at 10.7nM and at 12.7Nm depending on the tread material, so these projected ratios have the torque coming in lower than that value. If the math proves to be grossly off the only true loss is in the traction wheel. So the sprocket ratio from Omni to traction would simply be changed and that should solve any issues with stalling or being too slow and not gaining torque. Any corrections of the math are welcomed, better to fix it now before the pieces are made/ordered.

No need for physics majors.

I’ve got to refer to this guy. Just type in gear ratios, it spits out torque, speed, and current draw numbers. It’ll also indicate where cut offs for traction limiting is when you play around with the numbers.


You did your numbers with CIM normal load value, this means that the CIM can do this loading all day. Realistically for an FRC application, you should use at minimum the max power numbers (171 oz-in, 2655 rpm, 67.9 amps). While there are other things constraining it (speed controllers, breakers, wiring, efficiencies, etc.), this will give you a better value. When I say FRC application, I mean short pushing matches, two minute matches, etc.

Let me just say, we used 4" x 2" roughtop wheels last year at a 10.75:1 ratio and they were still spinning up the wheels.

Do yourself a favor, use JVN’s calculator.

I got to ask, how are you doing your center to center distances on gears?

Edit: Forgot to say, while I wrote about the max power numbers up there, the CIMs are capable of putting out 343.4 oz-in at stall current.

Thanks for that link. I got all the stats from probably the same data sheet you did for the CIM’s, the normal load was picked because it would prove to some more conservative members of our team it could maintain that torque. The gears are standard andymark shifting gears for the most part and in addition their charachteristics were placed with the gear portion of design accelerator on Autodesk Inventor. On top of hand calculations to ensure proper meshing. The modules are being lasercut to ensure the precision.