"Limited slip diffs have put food on my table for 17 years..."

[Paul Valente]

Preloaded Helicals

This is because the Torsen type diffs open up if one wheel looses its resistance. This is because (as I am sure you know) the helical planetary gears require some torque loading to provide enough thrust action to generate the required friction to transfer drive to the other wheel.
The Torsen in the Supra opened up regularly relegating it to open diff status. And once you have one wheel spinning, you have to get completely out of the throttle before it will lock up.
I assume this is the reason that TRD offered a standard clutch type diff for the Supra. I switched to that unit and was much happier.

Right, you are governed by the worse wheel so if you get zero torque to the worse wheel (like if it comes off the ground) you get zero times the bias ratio to the other wheel (e.g. 0 x 2.0)...which is still zero. Once the diff breaks-away and you get a spin-out, it is hard to stop it and start over again (momentum, mu static > mu dynamic, etc.)

Plate-type diffs are more responsive to pre-loads and can support a lot of torque just due to the pre-load spring. There are pre-loaded helicals (like the Quaife as you know) but they are less effective though the pre-load does dampen lash. You can see the disc spring-pack in the middle of the assmebly shown in the inset picture. Pushing two steel side gears against a steel case has much less effect than pushing against a clutch pack:

As far as Quaife the company goes; Tractech was selling Dana 30 sized helicals (about the size of a front wheel drive diff only a little bigger), made in Warren, Michigan, at a union shop (Teamsters) for about $400 retail. Quaifs were about $1200 at the time. Not saying that the Quaife is a bad diff or anything, Steve. Just that they are expensive. Maybe we could start our own company to compete with them? :hmm:

 

[Steve Hoelscher]

As far as Quaife the company goes; Tractech was selling Dana 30 sized helicals (about the size of a front wheel drive diff only a little bigger), made in Warren, Michigan, at a union shop (Teamsters) for about $400 retail. Quaifs were about $1200 at the time. Not saying that the Quaife is a bad diff or anything, Steve. Just that they are expensive. Maybe we could start our own company to compete with them? :hmm:

Wow... $400 retail is cheap. I assume they are cranking out lots of those at that price. I know Quaife is really proud of their stuff. None of it is cheap. Perhaps their production runs are too small... :hmm:

That would be fun to start our own business. Although competing against the ebay chinese junk isn't my idea of fun. If you think that $1200 for a diff is high, they get $15,000 for a sequential shift transaxle. I think there is room to compete there. And if Quaife can sell diffs for $1200 we can probably do something there too.

 

[Steve Hoelscher]

6 speed X?

Unfortunately, there isn't room in the X transmission for a 6th gear. Even a case extension (which was used to make the 5 speed) would hang the gear out into the wheel well. Not a good idea.

Probably the easiest way to get a six speed X is to adapt a K20 transmission from an Acura RSX Type S. It fits and could work with a bellhousing adapter.

 

[Paul Valente]

All depends on where you are coming from

At Tractech we made probably 200-ish Dana 30 TrueTracs a year. It's a lot to make in your garage but compared to OEM volumes it's nothing. At AAM we've sold something like 200,000 Silverado/Sierra rear axles this year just through June and we are having a bad year.
So "a lot" is realative.

 

[Rupunzell]

Paul, Thanks For Posting

the info on Helical diffs. I have often wondered how they work.. now I have a much better idea of how they actually work. After digesting the paper on Torsen diffs, It seems the helical gears live a pretty tough life in these diffs and the housing for these gears can be a grenade waiting to go off if the part are not properly designed due to the forces involved.

Anyway, I did get me thinking about traction control systems. Having never seen or been around any modern electronic traction control system or electronic diff. It appears a system that looks at wheel speed and actual axle torque transmitted to each wheel as part of a control loop with a "device" that can linearly proportion torque each wheel based on available traction at each wheel should work well. It seems the system must have quick response with little hysteresis to work well in a race car and if the loop goes out of range, it must behave in a predictable manner.

As with any control loop there is response of the individual components involved, closed and open loop response, and what happens when the loop goes out of range. Such is the nature of feed back loops, but there many things feed back can fix quite well when applied properly.

I guess the difficulty of making a mechanical power splitter that has wide an linear range is just not simple.

 

[Rupunzell]

Some one on X web or ? should try this..It sounds like a very reasonable way to get a 6 speed into an X if that is the goal. There are added advantages of aftermarket LSDs, choices of final drive and close ratio gears and the cable shifter can be a plus too.

As part of the TR6 project, I'm going to splice the front bell housing section of the TR6 transmission to a Toyota W58 5 speed. The TR6 transmission is too shallow to make an adapter plate, but maybe an adapter plate for the RSX transaxle to FIAT engine can be made or it will be a cut & splice.

Unfortunately, there isn't room in the X transmission for a 6th gear. Even a case extension (which was used to make the 5 speed) would hang the gear out into the wheel well. Not a good idea.

Probably the easiest way to get a six speed X is to adapt a K20 transmission from an Acura RSX Type S. It fits and could work with a bellhousing adapter.

 

[Paul Valente]

Controls

Anyway, I did get me thinking about traction control systems. Having never seen or been around any modern electronic traction control system or electronic diff. It appears a system that looks at wheel speed and actual axle torque transmitted to each wheel as part of a control loop with a "device" that can linearly proportion torque each wheel based on available traction at each wheel should work well.

There is more to vehicle-level controls than meets the eye. At first, I thought it was going to be trivial too and that I'd only need to look at a few parameters but when you really start to get into it, it is much more involved. Escpecially if you want to influence vehicle dynamics with the device as well improve traction.

Just one small example: your ECU picks up a delta of 10 rpm across the diff with equal torque on both sides. Have you lost traction at the faster wheel? Maybe, but you may also have a flat tire, or have a mini-spare on, or be in a turn, or be spinning both tires, or just have a bad wheel speed sensor.

Say you determine you are spinning one wheel, now what? Don't lock 100% because the last thing you want is to spin both wheels (complete loss of stability). So how much do you lock? 10%? 20%?

 

[Rupunzell]

All so true Paul.. The amount of details involved in trying to make a real improvement rather than "just a change" is complex and not simple.

This is why when folks complain about some aspect of a design really need to understand all the "Stuff" and WORK that went into the design before being critical of it. The number of tiny details that can come back and bit are easy to not discover during the development process and why so much testing is required before any product begins production and sale_able without problems.

There is no glamor in sweating the details for any design, but this is what makes all the difference in the final results. ya got to hand it to all the folks who work so hard out of the lime light trying to make it all work problem free and meet market and design and economic constraints..

Bernice

There is more to vehicle-level controls than meets the eye. At first, I thought it was going to be trivial too and that I'd only need to look at a few parameters but when you really start to get into it, it is much more involved. Escpecially if you want to influence vehicle dynamics with the device as well improve traction.

Just one small example: your ECU picks up a delta of 10 rpm across the diff with equal torque on both sides. Have you lost traction at the faster wheel? Maybe, but you may also have a flat tire, or have a mini-spare on, or be in a turn, or be spinning both tires, or just have a bad wheel speed sensor.

Say you determine you are spinning one wheel, now what? Don't lock 100% because the last thing you want is to spin both wheels (complete loss of stability). So how much do you lock? 10%? 20%?

 

[dcioccarelli]

6 speed

If adding an extra gear requires extending the length of the gearbox (as was done for the 5 speed) I'm wondering how the earlier Colotti 5 speed conversion kits managed to fit 5 gears in the existing 4 speed case?

Dom.