A Counter-rotating Differential

Back in November of 2011 we posted an article about our first Spirit II Mars Rover. Among other things, that rover included a counter-rotating differential (or universal joint) integrated into its rocker-bogie suspension system. If the front wheel on the left side of the rover went upward over a large obstacle, then the differential would push the three wheels on the right side downward, and vice versa. Its purpose was to stabilize the rover when traversing rough terrain. We made that first counter-rotating differential from bevel gears, aluminum rods, aluminum plates, and screws. Its shafts were connected to each side of the rocker-bogie suspension system. The function was simple: If either of the shafts rotated, then it caused the other shaft to rotate in the opposite direction. Our original version worked quite well, but it was also large, bulky, and a bit difficult to keep running smoothly.

Recently, our friends at Actobotics / ServoCity sent us a couple of their new bevel gears to try out. It’s cool how a few new parts can fire up your imagination and get you thinking. We decided to see if we could create a new counter-rotating differential for a future Mars Rover project. We were very pleased with the results. The new differential is much smaller than the old version, so it will fit nicely inside the box of a Mars Rover, and its operation is smooth and robust. Camille took all the photos of the new differential.

CLOSE-UP OF THE NEW COUNTER-ROTATING DIFFERENTIAL — THE NEW COUNTER-ROTATING DIFFERENTIAL

12 Comments

Susan Jankowski on July 12, 2014 at 4:04 am

How exciting! The new differential looks beautiful! Good job in the making and the fabulous photo’s.

chris on July 12, 2014 at 2:03 pm

Very nice. It looks much more robust than the original. I’ve always wanted to build a robot on the 6 wheel rocker boogie chassis with a pitch averaging differential. In theory, it should be a very stable platform for sensors and have high mobility. Also, I wonder if you could drive one of the output shafts of the differential to actively control chassis pitch? Just a thought. Can’t wait to see how you integrate this new design.

Camille on July 12, 2014 at 9:25 pm

Chris: Because the counter-rotating differential is screwed to the main electronics box, it does indeed provide the pitch-averaging feature. When you say “actively control chassis pitch” are you thinking about using an encoder / rotary sensor on one of the output shafts so that you could sense the tilt in the main electronics box? We could even put a motor on that shaft, along with the encoder/sensor and/or an IMU, such that the tilt motor always keeps the main box level. I’m not sure, but I think the real NASA rover probably works that way.

arnab kumar barman on October 14, 2014 at 8:30 am

I am a 1st year student of engineering and I dream of building rovers.Please help me in making.

Chris V on October 26, 2014 at 9:19 pm

I just saw your reply. Yes, my thought was to connect a motor to one of the two output shafts. The motor could be used to correct pitch, or lock the differential to limit pitch. But it could be connected to an encoder to read pitch as well. I’m not sure how useful this would be and it might even cancel out the desired averaging function of the differential. However it might be useful in controlling the chasis when climbing over tall objects or uneven terrain. I look forward to see how you integrate this differential gear in your projects. You guys do great stuff!

Camille on October 29, 2014 at 3:42 am

Chris V: Thanks for your reply. Yes, I think you’re on the right track with the motor idea. I believe the actual NASA Mars Rover has a motor on the central differential similar to what you’re describing.

Markus Kerola on December 18, 2014 at 11:50 am

So happy to have bounced to your site.
I was looking information on how the 6 wheeled mars rover suspensions were designed and saw a picture of your suspension system, so I opened up the site and read your story. How cool and inspiring ! Had to subscribe 🙂
Wish you all the best

Camille on December 18, 2014 at 1:46 pm

Thank you, Markus! 🙂

John T. on February 2, 2015 at 8:25 am

Just a curious question. Can you estimate the maximum load that the new gearbox can hold?

Camille on February 2, 2015 at 12:27 pm

Sorry, I didn’t calculate max load on this.

Frank on February 8, 2015 at 5:26 pm

Beautiful! I am going to try and replicate this differential on my rover project! What diameter axles are you using?

Camille on February 17, 2015 at 8:51 pm

Frank: The main axle that you see is 1/2″ diameter. The shafts that are 90-degrees to the main axle are 1/4″ diameter.

Susan Jankowski on July 12, 2014 at 4:04 am

How exciting! The new differential looks beautiful! Good job in the making and the fabulous photo’s.
chris on July 12, 2014 at 2:03 pm

Very nice. It looks much more robust than the original. I’ve always wanted to build a robot on the 6 wheel rocker boogie chassis with a pitch averaging differential. In theory, it should be a very stable platform for sensors and have high mobility. Also, I wonder if you could drive one of the output shafts of the differential to actively control chassis pitch? Just a thought. Can’t wait to see how you integrate this new design.
Camille on July 12, 2014 at 9:25 pm

Chris: Because the counter-rotating differential is screwed to the main electronics box, it does indeed provide the pitch-averaging feature. When you say “actively control chassis pitch” are you thinking about using an encoder / rotary sensor on one of the output shafts so that you could sense the tilt in the main electronics box? We could even put a motor on that shaft, along with the encoder/sensor and/or an IMU, such that the tilt motor always keeps the main box level. I’m not sure, but I think the real NASA rover probably works that way.
arnab kumar barman on October 14, 2014 at 8:30 am

I am a 1st year student of engineering and I dream of building rovers.Please help me in making.
Chris V on October 26, 2014 at 9:19 pm

I just saw your reply. Yes, my thought was to connect a motor to one of the two output shafts. The motor could be used to correct pitch, or lock the differential to limit pitch. But it could be connected to an encoder to read pitch as well. I’m not sure how useful this would be and it might even cancel out the desired averaging function of the differential. However it might be useful in controlling the chasis when climbing over tall objects or uneven terrain. I look forward to see how you integrate this differential gear in your projects. You guys do great stuff!
Camille on October 29, 2014 at 3:42 am

Chris V: Thanks for your reply. Yes, I think you’re on the right track with the motor idea. I believe the actual NASA Mars Rover has a motor on the central differential similar to what you’re describing.
Markus Kerola on December 18, 2014 at 11:50 am

So happy to have bounced to your site.
I was looking information on how the 6 wheeled mars rover suspensions were designed and saw a picture of your suspension system, so I opened up the site and read your story. How cool and inspiring ! Had to subscribe 🙂
Wish you all the best
Camille on December 18, 2014 at 1:46 pm

Thank you, Markus! 🙂
John T. on February 2, 2015 at 8:25 am

Just a curious question. Can you estimate the maximum load that the new gearbox can hold?
Camille on February 2, 2015 at 12:27 pm

Sorry, I didn’t calculate max load on this.
Frank on February 8, 2015 at 5:26 pm

Beautiful! I am going to try and replicate this differential on my rover project! What diameter axles are you using?
Camille on February 17, 2015 at 8:51 pm

Frank: The main axle that you see is 1/2″ diameter. The shafts that are 90-degrees to the main axle are 1/4″ diameter.