Diffs and "spider bite"

[elaw]

I know this is completely out of left field, but I just had a brainstorm and thought I'd share it for discussion.

I've seen a lot of talk about open vs. torsen vs. locked diffs and "spider bite" on Audi forums, and this morning a little light bulb went off above my head re what's going on.

For simplicity's sake, let's consider a RWD vehicle that's going around an oval racetrack. And consider how each diff reacts to differing speeds on its outputs - specifically that locked and torsen diffs try to make both outputs spin at the same speed, and open diffs don't care.

Going down the straightaway, not much is happening. Both rear wheels, and thus both diff outputs, are turning the same speed and receiving the same amount of torque, and all is well regardless of what kind of diff is installed.

Now the car enters a turn, and suppose it has a torsen differential. Relative to the car's average speed (represented by the input to the diff), the inside wheel is turning more slowly and the outside wheel is turning faster.

The torsen diff does not like this, as it wants both of its outputs to rotate at the same speed. So initially it's going to respond by sending more torque to the inside wheel (which is the one with less traction due to cornering forces causing weight shift to the outside wheels) and less to the outside wheel. In other words, to a degree the diff is trying to make the inside wheel break loose. Of course it's trying to make the outside wheel break loose at the same time, but when a wheel actually breaks loose, it won't be the outside because that one has more traction.

Now because our hypothetical car is an Audi with one million horsepower (one of Hank's 07K engines), the inside wheel does in fact break loose. At that exact moment, it's still turning slower than the outside wheel, the diff is still unhappy about that, and thus is still sending the inside wheel more torque. This situation will not reverse until the inside wheel is spinning faster than the outside wheel. This means the torque shift will be a fairly violent event and will not happen until after the inside wheel has completely broken loose. I think that effect is what some people describe as "spider bite" in torsen-equipped cars.

I would think all the above would apply to a locked diff too, as it shares the torsen's characteristic of always trying to make both outputs spin at the same speed... it just tries harder.

So in my clueless opinion, it seems like right until the moment a tire breaks loose, an open diff is actually the best because it isn't actually *contributing* to the tires breaking loose (in a corner) like a torsen or a locked diff. Of course the moment a tire does lose traction, that all changes.

Re the location of the diff, the above would pretty much directly apply to a front diff also. With a center diff it's a little more complicated, but the basic principles and effects are the same.

[SEStone]

I always interpretted the spider bite as a positive feedback loop. It especially applies when you have two torsens in series between power and the ground, such as when you add the rear diff.

That being said, having the center and rear torsen was the best feeling I ever had in my car. Like torque vectoring on the new cars, it eliminate wheelspin and improved handling and balance. So pissed I broke it.

Sam

[pilihp2]

I always interpretted the spider bite as a positive feedback loop. It especially applies when you have two torsens in series between power and the ground, such as when you add the rear diff.

That being said, having the center and rear torsen was the best feeling I ever had in my car. Like torque vectoring on the new cars, it eliminate wheelspin and improved handling and balance. So pissed I broke it.

Sam

Absolutely love my center and rear torsen. Praying that it holds.

[jretal]

Spider bite had more to do (at least, my interpretation) of when the inside wheel would lift and all the power would go to the free spinning wheel (unfortunately, what torsen will do), and when the wheel will make contact again, it would cause a backlash/shock through the system. This shock would cause the gears w/in the torsen to break - hence spider bite.

THat being said, I have dual torsen on my 4000, and it tracks great! I never had an issue w/ inside wheels breaking free w/ this setup, vs. when I had the lockers (unlocked mind you) and I'd religiously get the inside rear to spin a little on corner exit w/ my lowly 20vt

[A1QSHIP]

I have a rear Torsen installed in my S6. It is a '95.5 so it is equipped with EDL. When I tracked the car I encountered the "spyder byte" when breathing the throttle a little too much in the corner. Letting out of the throttle even more while mid corner really made the car unhappy & resulted in a tank slapper moment. As I am but a lowly novice driver I am learning not to do that. When I was able to get a corner right it was like magic having the Torsen & EDL working together.
Currently working towards having a rear Torsen & EDL in my ES2 project.
Chris

[A1QSHIP]

Just found this on Rennlist. Maybe a better description of what is happening re spider bite.
http://rennlist.com/forums/964-forum/40 ... tro-2.html