Differential Transmission (Simpler than Planetary!)

So I lied about the being simpler than planetary part, but it’s not that much more complicated when you break apart the components.

So with all the excitement about 4194B’s planetary transmission, I though I’d show you all what I was working on before Worlds (as a side project).

Video:
http://www.youtube.com/watch?v=eGApXE-1ehI

Large Picture:

How it works:
http://img818.imageshack.us/img818/5346/differentialsmall.jpg
High-Torque:
Motor1 and Motor2 are powered so that their outputs* are going in the same direction.
Diff1 is turning at the same speed and in the same direction as Motor1 and Motor2.
Diff2 is not turning because its two inputs are turning in opposite directions (Motor1 is mechanically reversed because of the gearing right above the motor).
Diff1 outputs to the lower input of Diff3 at a ratio of 5:3 (slower).
Diff3 is turning at 1/2 the speed of its lower input because Diff3’s other input (Diff2) is not turning.
Diff3 is turning at 3/10 the speed of Diff2 (3/5 * 1/2).
Diff3 outputs to the wheel at a ratio of 5:3 (slower).
The wheel is turning at 3/5 the speed of Diff3.
The wheel is turning at 9/50 the speed of each motor (3/5 * 3/10).

High-Speed:
Motor1 and Motor2 are powered so that their outputs* are going in opposite directions.
Diff1 is not turning because its two inputs are turning in opposite directions .
Diff2 is turning in the same direction as Motor2 and in the opposite direction as Motor1 (Motor1 is mechanically reversed because of the gearing right above the motor).
Diff2’s inputs are geared from Motor1 and Motor2 at a ratio of 5:8 (faster).
Diff2 is turning at 8/5 the speed of Motor1 and Motor2.
Diff2 outputs to the upper input of Diff3 at a ratio of 1:3 (faster).
Diff3 is turning at 1/2 the speed of its upper input because Diff3’s other input (Diff1) is not turning.
Diff3 is turning at 3/2 the speed of Diff2 (3 * 1/2).
Diff3 outputs to the wheel at a ratio of 5:3 (slower).
The wheel is turning at 3/5 the speed of Diff3.
The wheel is turning at 9/10 the speed of each motor (3/5 * 3/2).

*Because the motors are facing each other, when the outputs are going in the same direction, the motors themselves are being powered with opposite powers. And when the outputs are in opposite directions, the motors are being powered with the same power.

very impressive but can it move a competition robot in high speed mode?

That is quite interesting. I would not say it is simpler than planetary mainly because I do not know what just happened.

It would depend on the gear ratio, but yes. And you would have to support each axle very well so friction is as low as possible.
The ratio I had on it in the video was so that you could noticeably see the speed difference, so the high speed setting is much too high.

WOW :eek:

I agree, that looks amazing but I cannot really tell what is going on, can you post some close-up picture or something to better explain it

That does look extremely cool. How does it perform in terms of friction? Any plans to implement pneumatics for faster switching?

There is unfortunately quite a lot of friction due to the fact that there are so many bevel gears.

This was a quick fun little project that I built at Anaheim, but if I were to actually put it on a robot, here are the changes I would make:

  1. Support every axle so it doesn’t budge at all, but turns free of friction (as free as possible, as least).
  2. Increase the initial gearing and decrease the gearing within the differentials so that everything moves faster, thus preventing the differentials from bending and slipping (more speed = less force).
  3. Use pneumatics to alternate between locking one differential and locking the other.

Keep in mind it would be a lot easier to just make a simple gear shifter, so I wouldn’t use this on an actual competition robot, as much as it might impress the judges.

I added a better explanation of how it works into the original post.

I found this tutorial on differentials quite useful and simply explained.

I’ve actually seen this before, and yes it does do a great job of explaining differentials.
I left out explaining differentials because I couldn’t find a good way to do it without confusing people more…so thanks!

One of the major problems with the differentials is that when they are under high stress, the differentials tend to flex alot thus adding more friction than necessary. Thats why my team made our own differentials that are only slightly wider however much much stronger.

https://lh6.googleusercontent.com/-pFD3WOS3Kjw/T5ohoV-CmjI/AAAAAAAAATs/CVMx7jlRwLo/s816/photo%20%283%29.jpg

I’ve just been thinking about how to implement a differential speed-changer and as a result I’m looking for a way to make nice differentials (like yours).
When I look at that differential, it just looks like the axles are going to hit the cut C-channels. Have you had any issues with the screws loosening and the differentials getting jammed?

As an edit to my first post:
Line 6 of the High-Torque how-to should read:
" Diff3 is turning at 3/10 the speed of Diff1 (3/5 * 1/2) "

I have been thinking about differential gears and multi speed transmissions and I have come up with an idea but it is so simple I can’t see why no one has come up with it. This leads me to think that there is something wrong with the idea.

Anyway my idea is to put 2 motors into the bevel gears as inputs but first gear up our down the motors to make sure that they are different. Then the output is the differential’s housing

For example you can gear one motor up 2:1 for speed. When turning in the same direction the final gear ratio is (2+1)/2 = 1.5 or 3:2 for speed. When turning in the other direction the gear ratio is (2-1)/2 = .5 or 1:2 for torque.

Yes there is something wrong with it :slight_smile:

The resultant torque of that kind of differential setup is only as high as twice the smallest input torque. So if one of your motors is geared 1:2 for speed into one of the inputs, and the other input is a motor geared 1:1:
Running in same direction: Speed is x1.5 and Torque is 1x (the torque of a single motor).
Running in opposite direction: Speed is x.5 and Torque is 1x (the torque of a single motor).
This is just like what Paul Copioli was describing in his CCT.

You want both inputs of the differential to be the same torque and running in the same direction to get any kind of output.
I am currently working on a differential two-speed transmission with (only!) two differentials and requiring the use of pneumatics to switch between speeds (but there are no gears that need to mesh, so it’s better than a traditional gear-shifter).

As far as I know, it is impossible to output a torque larger than one of the inputs of a differential or a planetary gear set (if there are two powered inputs). This is because whatever is fighting the output only has to overcome one of the inputs.

There is a much, much simpler and easier to make this. you simple have one gear rotating the encasing of the bevel gears and then another motor rotating the bevel gears themselves. mix up the encasing gear to change your speeds.

Do you have a Picture/Drawing or Video??

Would you use Multiple Motors with different gears to drive the Encasing?? How do you disengage the the Motors not needed?