Miata Mailing List: February 1994, Message #116
sponsored by



     


   From: mac.avid.com!Michael_Wissner@avid.com
Subject: Re: How LSD works?
   Date: Tue, 8 Feb 1994 10:00:02 -0500

Thanks for the info, Jim!

> The viscous type LSD simply involves having a fluid coupling between
> the halves of the differential which is filled with a very thick,
> viscous silicone oil. This limits free wheelspin by imposing a
> viscous drag on the spinning wheel. This is a fixed effect, so one
> can't use too much drag without effectively locking the differential.
>
> The Torsen or Positraction differential uses a friction clutch
> somewhat similar to the synchronizers in a manual transmission. This
> is designed so that under normal conditions the differential is
> completely unlocked. Once wheelspin exceeds a nominal amount, this
> clutch activates a mechanism which effectively locks the differential
> until wheelspin again falls below design values. In street
> applications this is typically done by having pressure on the
> clutches increase relative to the degree of wheelspin, attempting to
> counter the spin. Thus, the differential does not fully lock, but
> does provide a significantly higher control of wheelspin than the
> viscous variety.

The viscous type diff looks like a low-cost implementation. Since there is
no way to disengage the viscous coupling at low speed differentials, it
would be acting in opposition to the normal function of the differential
which is to *decouple* the two wheels. Am I right in assuming the Torsen
differential is more efficient?  The viscous type does have the virtue of
simplicity, however.  The clutch mechanism seems more complex: more parts,
more adjustments, more things to wear out, etc. 

But I like the negative feedback loop you describe in the Torson diff; the
faster the wheel spins the more the clutch is applied, slowing the wheel
back down. How is this accomplished? (I'm not familiar with how
synchronizers work in a manual trans).  A simple "governor" type of linkage
wouldn't work because that depends on absolute rate of rotation, and
probably wouldn't apply enough force; what you need is to apply a force
proportional to the relative rate of rotation between the wheels. What's
the secret?

-Mike