Weismann Differential

[Warpspeed]

Marcus, I had heard rumors that the Weismann diff worked something like a detroit locker with damping, but was not prepared to say so, because I did not know for sure and did not wish to cloud the issue further.

Carroll Smith gave a brief synopsis of the various available diff type, and his opinions of their various virtues and failings.

He called the detroit locker (in his own unminced words) "an abortion", and he rated the Weismann locker very highly.
And that has clouded the issue for me ever since.

Most of us realise that the differential will have a profound effect on vehicle dynamics and handling, and different differential types are each more suited to totally different applications, and sometimes driver preference.

Anyone that has driven a true locker will know the sudden violent change in vehicle response when it switches instantly from one condition to the other.

Which brings us back to the main problem.
Just because a Formula One driver, and maybe the most technically knowledgeable one ever, says a Weismann is in his opinion the best solution for Formula One, does not make it best, or even suitable for every other class of racing or even a road car.

I can understand a hot shot Ferrari test driver getting all excited about shaving .001 seconds off his lap times.
And I can also understand Ferrari marketing not liking sudden violent transitions between understeer or oversteer as the diff suddenly switches to the opposite mode for an inexperienced and startled driver.

Crashed Ferraris and dead drivers do not help sales figures, so older and wiser heads may have preferred a more benign engineering solution.
Nothing wrong a locker diff, just probably not the best solution for a Ferrari or the type of overconfident drivers that sometimes drive them.

[Marshall Grice]

Originally posted by Markus:
And because of those cluthes I still claim the differential has only 2 mechanical torque-split states: 50%-50% and 0%-100% (+- preload)**. And you need rotational speed difference to change to the latter state, so in engineering terms it doesn't sense torque but the instant reaction of losing it.

So here's where I see the "Magic": the differential locking-unlocking action can be achieved with very high frequency, essentially creating "virtual" torque-split states (think PWM) in the range of 0%/100% to 100%/0% (+- preload). During a manouver the differential changes between states fast enough so that the tires "see" it as a stepless change in torque-split.

This type of "virtual" control with 2 mechanical states is used in many branches of engineering but it's the first time I ran to it in differentials. And I have to admit that I like it.

** There's actually 3 mechanical states:
Both outputs coupled to input.
"Left" output only coupled to input.
"Right" output only coupled to input.

I disagree with it (at least the version shown in the patent) only having discreet locked and unlocked states. you have the rollers acting as wedges in between the driving and driven elements. as you apply more torque the friction between the rolling elements and the driving/driven elements increases in proportion to the input torque until they lock. Think of how a clutch type diff cross shaft applies increased clamp load to the pressure rings in proportion to the input torque. only in this case instead of the wedging force being in the the axial direction it is in a radial direction.

[Marshall Grice]

also, i don't see this unit having a discreet 50/50 torque split state. there is no geared interconnection between any of the parts. if you lose the clamp load you lose the connection between the driving and driven elements.

key to notice it's called a "locker" not a differential.

[Z]

At last, an interesting thread about mechanical stuff! (And not an Indian in sight, yahoo!!! )

I agree with a lot of what Chris says about commercialising good ideas. My guess is that most of the "magic" in the WL is in the form of detailed dimensional tolerances, heat treatments and hardnesses, preload spring-rates, and even the type of oil used, or not used (ie. no moly?).

I also agree with Tony that just because it works great in F1, that doesn't necessarily mean it is good for FSAE.

In particular, is it really a good thing to send most of the power to the wheel with most of the traction?
Especially when the car has a very high rear-track-to-wheelbase ratio (~4/5 in FSAE, cf. ~1/2 in F1)?
And the car is driven by amateurs?

To explain, here is a brief summary of pros and cons of some "diffs", with more emphasis on the cons.

1. SPOOL.
Pro. - Both wheels always driven, so you can get home if you lose a rear wheel (which is why all Le Mans Porsches had spools).
Con. - When equally loaded, both wheels try to make the car go straight (= bad on corners, especially tight ones).
When unequally loaded, or when on "split-mu" surfaces, the wheels give different thrusts, and thus try to steer the car.

2. OPEN DIFF.
Pro. - Both wheels always give equal thrust, so any thrust vector is always down centreline of car. The diff never tries to steer the car.
Con. - Both wheels have thrust of weakest wheel, so if one wheel in air (or on oil/wet clay/ice), then NO thrust.

3. FULL TORQUE VECTORING (eg. 2 x IVTs, or 2 x electric motors).
Pro. - Complete control of torque AND speed at each wheel.
Con. - If control system is visited by glitch, then both wheels might receive full torque, but in opposite directions (= very entertaining, and this has happened)!

Of the above, the open-diff has the most benign disadvantages. It doesn't mess with the car's handling, and it "fails" very safely (= no thrust). Its main disadvantage of low thrust due to spinning inner-wheel in a corner (or on split-mu) can be largely fixed with more rear percentage. And while cost-wise the spool is best, the open-diff is a close second.

All the other diff types seem, IMO, to be expensive ways of compensating (and only partially so) for a lack of rear percentage.

Z

[murpia]

Reading the patent, and generally thinking about the device, the 'magic' is clearly the means by which the rollers get their initial wedging motion. Once wedged, there is a positive feedback effect to keep them engaged.

Without the initial wedging, the half-shaft discs could just remain stationary with the outer housing rotating and the rollers spinning relative to the half shaft discs.

Differential action (relative motion of the two half shaft discs) such that it is, must come from a small amount of roller skidding on the half shaft discs. I do wonder how much actual true differential action ever comes from this skidding and how much from relative tyre slip (as per a spool). I doubt a Weismann equipped vehicle would work well in tight turns without the tyres operating at high slip ratios.

I'd be tempted to try and model this, if I had the time...

Regards, Ian

[Chris Weismann]

The "magic" is in how it senses torque. It is in how it locks up, not how it disengages.

Markus, Sorry but you shot yourself down. You are still not "seeing" it virtually in your head. But don't take offense, you aren't the first and you won't be the last. You need to remain open minded. Seeing you are in huntington beach, I invite you to come to our shop in costa mesa to play with one on the bench. And if after some schooling you are still not convinced, I suggest you take up another profession.

Also as I stated before it is not on/off, ie not 100% on, and then "violently" off. If it was, it would be diablolical in the rain. Instead it is smooth and very drivable.. and it the rain it is at it's best.

Gilles Villeneuve was not a test driver. At Watkins Glenn in 1979, he was around 19 seconds a lap faster than the rest of the field in changing conditions. He was the only one running with our Locker. Those that were present were astounded at his performance. He had help.

Jack Brabham went on to win the 1966 World Championship in F1 the first year he ran it. His teammate Denny Hulme won it the next year. Those Brabham cars were down on power to the BRMs and the Ferraris. But they had help. Also my dad made a trick auto trans for those cars to give them the advantage on the standing starts, but that is another topic.

All Can Am championships between 1966 and 1983 were won with our Locker.

All 4 GT40 Le Mans wins were won with the Weismann Locker. Carroll Smith was the team manager for Ford in those days. My dad designed the MKII and got no credit to this day for doing that work. He also designed the manual and auto transmissions for those cars.

Anyway... if you are traction limited, the locker will out perform any other diff. So if your SAE cars are traction limited for whatever reason, then it can help.

With the advent of wings, ie ground effects, the wheel slip conditions began to go away, and therefore the distinct advantage of the Locker also went away... the cars were not so traction limited.

In !981 (I think) Danny Ongais crashed heavily between 3 and 4 at Indy. The car was completely destroyed and Danny is lucky to have both his legs today as he was hanging outside the front of the car as his car ground to a halt. A very successful, if not the most successful chief mechanic at Indy George Bignotti, was asked his opinion as to why the car crashed so suddenly. He said something like "Well since is has a Weismann transmission, it must have a Weismann Locker, and if it broke an axle, the Locker will put you in the wall."

The next day there was a picture of Danny laying in the car against the wall, the transmission was broken wide open and in it you can clearly see... a spool. But the damage was done.

My dad decided to stop making the Locker because he was tired of fighting close minded people. When Ken Miles was testing in 1967 at Riverside in turn 9 in a GT40 MKIII, the "j" car, and was killed, the locker was blamed for the crash. It was later found out to be because of a broken upright from a poor brake mount design.

Why was Locker so quickly blamed? Because it was the trick device that very few people understood properly.

Lesson here is don't assume... assumptions are the mother of all f*** ups.

Any diff can make you crash if an axle breaks... but depending on which axle breaks, you could spin into the infield or into a wall. Danny's crash was not because of a broken axle. For consideration of the man that designed that car, I will not expose the failure as he has chosen to hide it all these years. I know what happened because I helped him design his next car and we spent a lot of time on the area of failure to ensure it didn't happen again. But I digress...

Back to the Locker. It is not a spague device for Pete's sake. (pun intended, my dad's name was Pete!) It is a roller clutch that locks in both directions and freewheels in both directions.

Mechanically speaking, a sprague has two cylindrical races, no cams! The srague's cage takes a ton of load from the pawls, the Lockers cage manly keeps the rollers phased and straight in relationship to the cams, therefore it sees no load. The spague is 100% on or 100% off!

A roller clutch isn't.

It is only similar because the Locker locks when the relative movement of the inner and outer races engage it. Period. So eliminate the srague from your minds.

Ok, one thing that you are all missing... the rollers roll up the cam as more torque is applied... that is the key to understanding the torque sensing. The Locker is still fundamentally a differential.

So imagine this... on a road course in a high horsepower car with no wings, going flat out into say a 150 mph turn without lifting, and then continue though the turn still without lifting onto a straight. Then come back to me and explain how you think that the Weismann Locker can't be torque sensing.

And one last thing about our dear friend Carroll Smith. I suggest to all of you that you buy all his books and learn from them all you can, especially if you intend to get into racing. He wrote these books to teach you "how to win." His style of writing is funny and at the same time on point. If you assume you know it all and don't need to, then good luck when you have to compete against a student that has understood his teachings.

I remember Carroll coming by our shop after my dad's death in 1997 to see how my brother Patrick and I were holding up. I told him that I was so disappointed for not asking him some questions about things I was still unclear on. He put his arm around me and said, "Don't worry son, you asked enough questions". Made me feel better, but I still regret not asking more.

And that is my finally lesson for today, learn all you can from old people... they have the experience that you don't. I'm not that old, I'm 46, but I was 7 days old when I was in Jack Brabham's helmet bag on the starting grid at Riverside. I got a massive head start.

[Chris Weismann]

Originally posted by Z:

2. OPEN DIFF.
The diff never tries to steer the car.

Z

Sorry Z.

My grandfather was driving in the snow in a Crosley (sp?)(similar to a bug I think) when his open diff caused him to lose control and crash when one of the driven wheels hit an ice patch.

While stuck in the snowbank, he pondered about how the open diff caused him to crash and came up with the locker concept. Then my dad got involved and they made a locker for the car. Then they set about pulling crashed cars out of snow backs.

My dad went to GM with the intention to put the locker into pickup trucks. Unlaidened pickups in the snow have the least amount of tractive effort.

GM wasn't interested because it wasn't race proven. My dad then met Roger Ward and the rest is racing history.

[Warpspeed]

I have to agree with Z here.

The exactly equal torque fed to both rear wheels under any and all conditions that a fully open diff creates, cannot steer the car.

It is hardly surprising that hitting an ice patch caused a car to totally lose traction and then go completely out of control.
You cannot seriously blame having an open diff as being the cause of the resulting accident.

[Jay Lawrence]

Chris,

It does not matter how much experience you or your dad had, you cannot convince people based on some stories from the past, and nor should you be able to. If you are interested in fostering drivetrain development in this series (which would be a start to wider acceptance of your diff) you need to talk to us like engineers (numbers, graphs, useful engineering information). Just because a gun driver went round a corner at 150mph doesn't mean your diff is torque sensing.

Tony, not so sure I agree with that. Well, I agree in principle, but in practice I find the open diff on my RWD road car to be quite unsettling when pushed. I can't quite match the equal torque theory with the practice. Perhaps when the speed differential between outside/inside is very high (due to an ice patch or whatever) and then you regain some inside traction, the inertia of the wheel/axle assembly causes some unbalance in the torque distribution. Maybe I'm talking rubbish too.

[Markus]

Jay, that wraps it quite well.

Originally posted by Chris Weismann:
You are still not "seeing" it virtually in your head.

...

Seeing you are in huntington beach, I invite you to come to our shop in costa mesa to play with one on the bench.

I guess this is truly a blind trying to lead a blind then...