Alternative Diff to Quaife: Westgarage Engineering

[JR @ CFS]

Hej all,

I have been in contact with Douglas at westgarage engineering about the differentials he has designed as an alternative to the Quaife ATB. I told him all about Formula Student/SAE so, he has forwarded me a brochure for formula student teams in .pdf format which can be found on his website. Here is the link;

http://www.westgarage.co.uk/downloads/brochure1.pdf

I advised him to join up to fsae.com incase you had any technical questions.

Regards,

Jonnie

[billywight]

Why would you want to run a viscous coupling diff on your racecar???

[JR @ CFS]

Not my area of expertise. If you want to put forward reasons why you wouldn't do it and what you would propose instead, maybe I can learn something http://fsae.com/groupee_common/emoticons/icon_wink.gif

[Douglas Anderson]

Hi, The viscous coupling in the Ford diff has the effect of a limited slip differential.
Here (http://www.westgarage.co.uk/downloads/viscous.jpg)

[murpia]

Viscous diffs are speed sensitive, so they generate a torque bias proportional to the difference in rear wheel speeds. Drive torque is not a factor.

Geared torque biasing diffs (Quaife ATB / Torsen) have a fixed torque bias, somewhat independent of the difference in rear wheel speeds or drive torque.

Plate & ramp diffs have a torque bias proportional to drive torque and somewhat independent of the difference in rear wheel speeds. They can be setup with preload, in which case they have an extra fixed torque bias level. Also, they can have different settings for drive and over-run (engine braking).

So, which to choose depends on which characteristics you want. However, high inside wheel slip ratios are generally undesirable in a racecar and of the above 3 concepts the viscous is probably worst in this respect. Comparing the other 2 depends a lot on how they're set up.

I don't really know how to describe how cam & pawl diffs 'work', nor how to work out what characteristics you'll get for any given setup...

Regards, Ian

[ben]

Originally posted by murpia:
Viscous diffs are speed sensitive, so they generate a torque bias proportional to the difference in rear wheel speeds. Drive torque is not a factor.

Geared torque biasing diffs (Quaife ATB / Torsen) have a fixed torque bias, somewhat independent of the difference in rear wheel speeds or drive torque.

Plate & ramp diffs have a torque bias proportional to drive torque and somewhat independent of the difference in rear wheel speeds. They can be setup with preload, in which case they have an extra fixed torque bias level. Also, they can have different settings for drive and over-run (engine braking).

So, which to choose depends on which characteristics you want. However, high inside wheel slip ratios are generally undesirable in a racecar and of the above 3 concepts the viscous is probably worst in this respect. Comparing the other 2 depends a lot on how they're set up.

I don't really know how to describe how cam & pawl diffs 'work', nor how to work out what characteristics you'll get for any given setup...

Regards, Ian

A lot of LMP teams seem to run visco-mechanical diffs which combine a VC and a ramp and plate LSD. As I understand it the VC prevents single wheel overspeed if you lift a wheel. Normally to prevent single wheel overspeed you need massive preload in a standard plate diff, which can lead to corner entry understeer.

visco-mechs were developed by Ricardo in BTCC in the early 90s where they helped minimise disturbences caused by curb strikes, but Xtrac trumpeted their version a little louder and most people think they came up with it.

If I was doing an FS car today I'd go with the plate and ramp option simple because you can tune the bias ratio between drive and brake whereas with a Quaife or Torsen you're stuck with what you get.

Ben

[murpia]

Originally posted by ben:
A lot of LMP teams seem to run visco-mechanical diffs which combine a VC and a ramp and plate LSD. As I understand it the VC prevents single wheel overspeed if you lift a wheel. Normally to prevent single wheel overspeed you need massive preload in a standard plate diff, which can lead to corner entry understeer.
Also, Le Mans teams like the idea of being able to get back to the pits if a driveshaft lets go. They used to run spools for just that reason, despite the understeer issue. A viscous coupling can exhibit a phenomena known as 'humping' if it spins for long enough and gets hot, this is when the fluid kind of goes solid. That way you can keep driving on one driveshaft for a bit:

http://en.wikipedia.org/wiki/Viscous_coupling_unit

I don't see how this feature has much relevance for FSAE.

Regards, Ian

[billywight]

Also, Le Mans teams like the idea of being able to get back to the pits if a driveshaft lets go. They used to run spools for just that reason, despite the understeer issue.

That sounds like a rediculous compromise. Did they also run 2 steering wheels in case one broke? Are you sure that was their reason for the spool?

[murpia]

http://www.sportscarpros.com/across-the-border-features...h-jaguar/default.htm (http://www.sportscarpros.com/across-the-border-features-from-guests/features/what-it-takes-tony-dowe-on-winning-at-le-mans-with-jaguar/default.htm)
My evidence is anecdotal, and I don't really think it was a great idea personally, but there have been plenty more worse ideas in the history of motor racing...

Regards, Ian

[billywight]

The gearbox was built
with a limited slip differential. This was quite different from the UK cars because
the thinking was that with a "spool"� fitted if a drive shaft failed you could get back to
the pits! Well, unbeknown to the guys in England, we had Kenny Hill of Metalore
(they now make most of the F1 world's hubs/drive shafts and axles) make us
some super strong F1 type drive shafts.

Interesting. Seems like the guy had the right idea though. The performance advantage of the limited slip should far outweigh the weight penalty of manufacturing axles that last the entire race...