Source for open differential?

Does anyone have a source for a relatively small, light-weight, open differential for under $100-$200 each (we’ll need two)?

Software will not work; and it absolutely has to be an open mechanical differential. The ones on SDP/SI are too small and too expensive, and most of the cheap Go Kart ones are too large and too heavy. There looked like there was a promising one in an older thread, but the link was dead.

// If there aren’t any readily available, we’ll probably end up having to mill our own differential housing.

What is the application? What kind of loads and speeds will it see?

This is the only thing I know of that may work for you:

The large gear is just under 2.5" in diameter.

Parts are listed here: and Tower carries them.

It’ll be somewhere along the line of a multi-stage gearbox powered by a CIM motor; no final location (stage-wise) was chosen since we’re still looking to see if any good COTS differentials exist (and if they have speed/load limits). It will be used in a pretty harsh (non-drive train) environment, so a good factor of safety will be requirement. Further details are classified until ship date (or the vicinity thereof). :smiley:

If none are available, we are looking at getting some 1/4" or 3/8" wall 6061 square box tubing, and milling the bearing/bushing holes into it, then face mounting one side of the box tubing to a dead axle’d gear or timing belt sprocket, with something that looks just like the stereotypical textbook differential:

Thanks for the link to the HPI Racing model (this was just the kind of thing I was looking for); I just sent them an email requesting additional information.

If you find the HPI Baja car on their site, there’s a link to a PDF of the manual, which includes an exploded view. This differential is probably the most readily-available and lowest cost COTS thing I know of that isn’t the size of something like one out of a riding lawn tractor. I’ve never examined this unit myself, but I would suspect it to be of high quality. It is in the final stage of a 1/5 scale off-road RC car, driven by a ~1-2HP engine, so I would suspect it could stand up to whatever you’re doing if you design it right.

If you build your own, I imagine you wouldn’t be hardening the gears. That’s definitely something to consider, since most differentials (including the HPI I believe) have hardened gears, increasing their durability.

I built a Baja 5b ss and it is a beast! with the modifications i have done to the engine, i am getting almost 4hp out of it and 12k rpm. The diff was simple to build and works very well. They make one with a plastic case and a alloy case, get the alloy case. When i first saw the thread this is the diff i was going to reccomend and then saw someone else already had done so.

I don’t know your exact application, but I would like to caution you if you are planning on using it in your drivetrain.

First of all, bevel gears are very tricky to get right. I know that some have managed to make it work but this is not the kind of thing you want to do lightly (or without a lot of heavy duty engineering support).

Second of all, do you really want a single wheel off the ground to mean that your robot doesn’t move?

Look at all those pick up truck commercials. They all talk about limited slip differentials because it is well understood in pick up truck culture that a standard differential is a real pain if you get one wheel with traction and another in the mud.

I wish you well, but I hope that you find another path…

Joe J.

Not an issue for this application.

I’m having fun trying to imagine what other rotating mechanisms on a robot playing Breakaway might benefit from a differential.

It might be something where there is a brake on each side, to determine which side power goes to. Just a guess…

Let’s say we decided to build our own using either 20t or 25t 20DP carbon steel miter gears from SDP-SI inside a CNC milled aluminum housing. If we put this differential higher up (stage wise) in the gearbox, thus limiting the total torque each gear sees, could we avoid needing to harden the gears? Do all the teams who use coaxial swerve drives harden their gears?

I understand the theory behind hardening, but are there any gearheads out there that have achieved acceptable results at flame hardening things by using a garage oxyacetylene torch and quenching in plain water? Or is this just a disaster waiting to happen?

One of the things about a differential is that you have 3 shafts to manage. Perhaps they are using the gearbox as a way of engaging/disengaging a high ratio gearbox from the output (put on a break on shaft 1, then shafts 2 and 3 have a linkage to each other, let go of shaft 1 and the connection between 2 and 3 is broken).

This could be used to engage a ball kicker cocker or perhaps to switch from driving the wheels to driving a winch.

Lots of nails to hit once you make up your mind that you’ve invented the best hammer since Thor.

Let’s get whacking…

Joe J.

P.S. Of the people I know that put bevel gears in a “pretty harsh environment” on a FIRST robot, I only know one who didn’t live to regret that decision (and I am not 100% convinced that JVN is being completely honest with me :wink: Unless you know what you are doing you are very unlikely to be the 2nd, especially if you consider getting a gear hardened as something of an exotic art… Free advice… worth every penny…

Thanks for the feedback. The differential-based gearbox was at the bottom of our idea list, but after reading this and thinking it over even more, we can probably accomplish the same outcome using stock AndyMark dog shifter parts. It adds a pneumatic cylinder, but reading from these experiences, removes a lot of complexity and headaches.

If that’s the case, may I suggest two one-way bearings, one on each axis, so that when the motor turns one way mechanism (a) is attached and, when turning the other way, mechanism (b) rotates? Of course, that only works if your mechanisms are only meant to rotate in one direction. Ours are! :wink: