pic: Planitary Transmission


This is an idea for a transmission that I have been working on. My goal to was get the reduction into the smallest possible space. The total reduction without the ND motor running is 16.82:1. The object of the ND motor is to allow some variation in speed and because it means that I can say it is a six motor drive train. I used the ND motor because it can’t be back driven (I think). It was meant to be a simpler version of the 217 CVT. I would like to use a more powerful motor for the ND motors function but don’t want to add too much complication.

Thanks for any comments.

It looks great.
Can you give us some more details about it?
Sizes, Forces etc…

looks like it could be a great transmission if it is geared to the wheels correctly. love the design, great work man.

It’s true that the ND motor cannot be backdriven, but you may face another problem. The gear inside the ND is plastic and in high-torque applications (such as a drivetrain) it has been known to strip. I don’t know what forces you’re putting on it, or exactly what would be enough to destroy the motor, but be careful. Prolonged use, even below the actual stall torque of the motor, can wear down that gear.

Is there any chance you could explode this assembly or provide some better explanation of what it does?

I’m curious as to how you plan on attaching that back plate to the motors…

Other than that, looks like a solid starting base!

Back plate looks easy, probably just some holes for the super long screws that hold the CIMs together. Dunno if that’s legal or not but that’s how I’d do it. I’d be more concerned about the planetary assembly itself. With all the elements on bearings, I think you’d need to plan for a little more backlash than usual to account for extra play from the bearings.

I like the idea, but I would never put the Denso motors in a drivetrain. Even though the worm inside the motor is metal, the gear is fairly soft plastic and strips way too easily.

If you want to make your tranny anti-backdriveable (sp?), I would recommend that using a wormgear other than the one in the Denso.

Worm gear in a drive-train also means no coasting once power to the motors is cut. Don’t know if that is good or bad but seems like it would put a high load on the worm gear if a high speed run came at the end of a match…

Love the work and wish I had some ability to design drive-train transmissions.

I believe that the denso worm gear would see relatively little torque because it is just modifying one part of the planetary gear train that is normally stationary. Precession and procession modify the gear ratio. Can you show the exploded view of the planetaries?


I have been looking to do the same thing for a while.

Its sort of a constant velocity transmission. But teh Denso motor won’t really allow uch variation.

Consider this. Use to motors, on on the hub and one driving the outer ring. I agree that you will have to put some current thru the outer ring when running just the hub but the payoff is many different ratios

Nice cad work

I know that the ND motor won’t allow much variation but it is some and it does add to the output force so I think i could gear this a little higher if I wanted to.

I would be gearing the output shaft directly to a set of 6 inch wheels since the reduction through the trans is high. Meaning if the robot was pushing an object with 200 pounds (something that I have seen FIRST bots do) it would produce 34 Nm of torque on the output shaft. There is a reduction to the ND motor of 3.67:1 which means that everything at stall there would be 9.2 Nm of torque on the ND motor which is less than the motors stall torque. As said by many people this is not an optimal solution but it is the easiest thing to machine.

A major part of the box was that it was easy to machine and assemble. All the parts can be made on a small CNC mill with limited skill or on a small lathe. These two things make a worm gear hard for a more powerful motor. In any case the output rpm of the ND motor is 84 rpm and with a reduction of 3.67 the change in output speed of the output shaft is ~45 rpm which translates to a speed difference of 1.18 ft/s so it varies from 7.55 ft/s to 8.73 ft/s which isn’t a big difference but…

I will post a pic of the planetary showing the inner workings and a pic of a chassis that it could work with.

Thanks for all the compliments.

Pause for a second and consider what usually keeps some part of a tranny stationary. It’s usually metal. If the wheels are stalled, the sun is under torque. In fact, you should be able to figure that torque out pretty easily. Work your CIM stall torque forward to the axle, and then backward to the sun. Assuming your power takeoff is on the planet holder, the ratio should be (Outer ring/CIM pinion)*(Sun/Inner Ring) * CIM Stall * 2. I have a feeling the first bit is going to be around 1. Possibly greater than one. So you’re looking at twice the CIM stall torque being applied to your ND drive shaft. Plus, even worse, shock loads. I’d actually be very worried about this, really.

Cool idea and I see how this is working now that I’ve looked at the internal views.:cool:

I would be cautious about using the long screws in the CIM motors to mount a back plate. Once you take them out, it can be exceedingly hard to get the screws back in, due to all the magnets inside the motor. I’ve trashed a CIM before doing that, and I know a few other people have done similarly.

Out of curiosity, where are you getting the combined spur/ring gear?

First off I haven’t even said that this was the intended method of ataching the back plate but it is one that I have been thinking. I would be more cautious of doing this more because I might have a hard time finding just the right bolt.

I have taken a CIM apart on many occasions in the past to do many things to them (turning the shaft down to mate a pinion gear during the off season) and while the turning the shaft down destroyed the motor getting the bolts back in was extremely easy. It also happens that on our robot this year we had a big CIM that could only be taken out by taking the motor apart (it wasn’t a bad design, it was my incompitent machining) and that big CIM still works like a dream.

I’m not. I would get a normal gear and a ring gear. I would bore out the normal gear so the ring fits inside and drill holes for pins to transfer the torque. The pieces that hold the ring/spur gear are also pins that could transfer the torque. I would have to try this before actually using it during the season cause it seems a little fishy in terms of actually working.

On a similar note if anybody knows where to get a 20DP, 40 tooth ring gear that would help imensly cause mcmaster doesn’t seem to have ring gears and stock drive products doesn’t have 20DP gears.


Boston Gear doesn’t have them. I am not to sure were else to have you check.


Here are some words of wisdom from someone who has lived through the variable planetary gear train design and implementation cycle.

  1. You do not get ANY added output torque from the ring gear motor … zero, nada, nothing. The ring gear has to react against the input torque of the sun gear and adds nothing to the output torque.

  2. The ring gear motor only changes the speed ratio. Designing it to not need motor torque at 0 RPM is a good idea, but the Denso motor worm gear may not be able to handle the load.

  3. When sizing the gears, make sure when the ring gear is spinning in the opposite direction of the sun gear that the robot still moves forward. If you do not do this then any amount of force can stop you. However, you can pretty much stop any amount of force pushing you.

  4. When designing your gear ratio’s, there are three basic areas that the CCT gives you: Ring gear in reverse, ring gear at 0, and ring gear forward. Design your normal operating conditions at ring gear at 0 speed.

  5. Read the paper that I attached here: http://www.chiefdelphi.com/media/papers/1361


It looks like Alex has designed his system “inside out” from the way you expect. If I’m reading the drawing correctly, the dual CIMs drive the ring gear, and the “variable” part is due to rotating the sun gear.