Help with differential selection

[bentheswift]

I'm trying to lay the groundwork for our university's next Formula SAE car.

We have a "legacy" differential that we have developed over several years. It is a fairly effective clutch-pack limited slip differential. The latest iteration has the gear carrier integrated with the housing, mounted to the chassis by its bearings. There are two spider gears that spin on a hardened steel shaft. Each of the side gears sits on two clutches that act as the limited-slip mechanism. The torque bias ratio is adjustable from the 1.25ish range to about 3. (We've found that the car is fastest with very little bias.) The differential is driven by a chain connected to the output sprocket of the engine. The entire assembly without axles is less than ten pounds.

We have also used a helical-gear torsen I style differential in the past. It is however too heavy and has too high TBR for our uses.

I've been looking at a couple of other options. The first is using a hydraulic pump to energize the clutches during wheel slip. The pump would be driven by differential wheel velocity thus making the locking mechanism activated by the velocity differential rather than drive torque. This seems to be a good idea, since (ideally) you'd be able to tune the pump to cause little to no bias below a differential velocity (say the wheel speed difference at the smallest radius corner you expect to see), while easily locking the wheels together if that limit is exceeded. The clutch pack diff will bias torque during cornering under power, regardless of whether one wheel is spinning or not, thus applying an undesirable understeer moment to the car.

I've seen some schematics of some GM (Versatrak?) gerotor differentials, and they seem fairly straightforward. It seems like we could package it in a smaller space and possibly even make it lighter. We have about sixteen million oil pumps laying around from broken engines (in fact we build our custom dry sump scavenge pump using two gerotor pumps from the stock oil pump).

The other type that I'm very curious about is the Weismann locker. I can find almost zero information about this differential. There is a brief textual description on a patent application I found on google, but I can't find any schematics or descriptions -- anywhere. Weismann's website looks like it has not been updated since 1995 and it sounds like they don't want to hand out that sort of information. But I know that this differential has serious competition history and I hear that most F1 diffs are derivatives of it.

I would appreciate any information or advice that is available. Thanks,

Ben

 

[murpia]

My opinions:

If you found a low TBR to be effective you certainly don't want any 'locker' type of diff. Modern F1 diffs are not 'lockers' they're conventional diffs with actively controlled clutch packs. Weismann is 70's / 80's 'technology'.

You don't want a speed-sensitive diff such as a viscous coupling. Once you get enough wheel speed differential to activate the diff it's too late for the tyres.

To be honest, your current diff sounds pretty good...

Regards, Ian

 

[bentheswift]

A viscous coupling is more of an OEM-suitable application, since it needs no maintenance and fails over time in the open state.

I was thinking a gerotor pump differential could give us the same benefits as the clutches without the torque bias when the car is cornering without wheelslip.

My main reason in asking about the Weismann was because there was a design judge at competition a couple years ago asking every team "Why aren't you using a Weismann Locker?" Rumor has it he was related to the Weismann family or something. We didn't know how it worked so we couldn't defend our design against it.

Thanks for your input.

 

[MikeHalloran]

The Weismann's patented diff appears to be sort of a two-way roller clutch sort of thing:

http://www.freepatentsonline.com/3283611.pdf

Note that the clutches don't just carry whatever torque is needed to limit relative rotation, they carry _all_ the torque.

Probably requires something like SantoTrac for lubricant, and seriously hard parts. I'd guess that when it does go, it either explodes or won't drive the car, and trying to make a lightweight version will make it much more likely to fail.



Mike Halloran
Pembroke Pines, FL, USA

 

[murpia]

"My main reason in asking about the Weismann was because there was a design judge at competition a couple years ago asking every team "Why aren't you using a Weismann Locker?" Rumor has it he was related to the Weismann family or something. We didn't know how it worked so we couldn't defend our design against it."

Cost would be a good enough reason to reject a Weismann, lack of data would be another... Not every FSAE judge will be completely open minded but there's enough different judges involved that personal bias usually gets moderated out of the scoring.

Regards, Ian

 

[MikeHalloran]

Other defensible answers:
- It's not in production. (I think I read somewhere that Weismann et.al. don't _want_ to make it anymore.)
- Some information is available from the patent, but probably not _all_ the information that's really required to make successful components... assuming that it is even possible to do so. Existence of a patent does not imply that the component is producible. Existence of an _old_ patent, and only very limited production, suggests that it is not.
- Undertaking a development program for a major component has a way of consuming all the money and time budgeted for the assembly that depends on it.


Mike Halloran
Pembroke Pines, FL, USA

 

[bentheswift]

Yes, these are good answers. I'm still curious about how they work, but thank you for the replies.

 

[murpia]

My understanding of how it works is this:

Each halfshaft is attached to a metal disc inside the diff. The diff housing is machined with back-to-back wedge shaped pocket around it's inside diameter. The halfshaft discs run inside this diameter and in each of the pockets a hardened roller sits.

If the rollers wedge into the pockets, torque can be transmitted to the respective halfshaft disc and the diff is 'locked' on that halfshaft. Otherwise the diff is 'free' on that halfshaft.

What's supposed to happen is that the inside halfshaft goes 'free' in a corner to allow for differential wheelspeed and then the whole diff goes 'locked' on the exit on power (possibly sending you straight on into the wall...). What happens on entry I don't know. Also I believe that similar to a cam and pawl diff there is no adjustment available and the characteristics change with wear. All in all not a combination I would favour.

Regards, Ian

 

[Tmoose]

A few of the late, great Carrol Smith's books (Tune/Engineer/Design to win) have discussions of the various LSDs.