"Suspension Evolution" or "Did Darwin Screw Up?"

[BillyShope]

Was thinking again of my youth, as old men do, and recalled how all the Chevies and Fords of the time had torque tube suspension. Even the Corvette had it for the first 3 years. But, eventually, 'most everyone went to open drive shaft and leafs. The next evolutionary step was to coils and here is where I think the ape might have stumbled over his tail.

The torque tube was simple, effective, and relatively cheap. How did it get lost in the evolutionary scheme of things? No, it can't be used with an open driveshaft, but there's nothing to prevent it from disgorging the driveshaft and moving a few inches to the right. Why didn't our friend who dreamed up the torque arm realize that he could get rid of those lower links by merely reviving use of the front ball?

http://www.racetec.cc/shope

 

[MikeHalloran]

Plastic Camaros, 4thGen, have a nicely engineered torque arm pretty much as you describe.

I haven't looked under a 5thGen Camaro yet.


Mike Halloran
Pembroke Pines, FL, USA

 

[BrianPetersen]

The current Camaro uses an independent multilink arrangement and a rather ordinary open double-jointed driveshaft from the transmission to the differential (which is attached to the chassis).

The 4th Camaro used a torque arm, slightly different from the old torque-tube arrangement because the axle is allowed to pivot relative to the torque arm in plan view and also the axle is allowed to move fore-and-aft relative to the drive shaft; the front end of the torque arm was in a flexible bushing to allow this.

The old traditional torque tube has some rather severe limitations when viewed from a modern perspective:
- The axle is locked to being perpendicular to the torque tube in every dimension, which rules out being able to tune the geometry for roll understeer or roll oversteer (these are done by pulling one end or the other of the axle fore or aft with suspension travel through design of the linkage angles - with torque tube, no can do - See 4th-gen Camaro for fixing this);
- The torque tube pivot point is locked to being where the end of the gearbox is, as opposed to being able to select how the pivots actuate in order to get anti-squat during acceleration or anti-lift during braking (you will get anti-squat / anti-lift with torque tube, but you get what you get ... unless the torque tube mounting is separated from the gearbox to allow the length to be tuned, see 4th-gen Camaro);
- Chassis and suspension related loadings have to be absorbed by the engine/gearbox mountings and the various vibration or torque related motions of the engine/gearbox can't be separated from the suspension actions. (again, see 4th-gen Camaro)

The 4th-gen Camaro arrangement dealt with some of these side effects by using the open drive shaft, the torque arm off to the side of the driveshaft, and being able to pivot the axle and torque arm slightly in plan view. But, the torque arm is really long, adds a bunch of weight, still needs a trailing arm on either side and a Panhard rod, and in general, doesn't have any advantages that I can see over a 3-link arrangement like the current Mustang uses and has a fair number of disadvantages. Unless you want to go really retro and use leaf springs, but those have their bad side-effects, too.

I'm thinking Darwin didn't screw up in this case. Rear-drive cars just aren't being designed with axles that bounce around across the whole width of the car any more.

 

[BillyShope]

Brian, having some problems with your response.

"The axle is locked to being perpendicular to the torque tube in every dimension, which rules out being able to tune the geometry for roll understeer or roll oversteer...."

For what purpose are you going to use roll steer? Oh, there are times when it comes in handy. I remember the introduction of the torsion bars in the '57 Chrysler line, for instance. The press attributed the improved handling to these new magic bars, but it should have been attributed to the rear leaf redesign, which promoted roll oversteer.

Remember, roll steer does not affect tire loading. All it changes is the steering wheel angle. This is pointed out in one of the problems (included at my urging) in the student workbook which accompanies the Millikens' book.

In other words, if an oval track car pushes, you can add as much roll oversteer as you like, but it will still go into the outside wall nose first. The driver will just be holding the steering wheel in a different position as he crashes.

I don't see this as a fault worthy of comment.

"The torque tube pivot point is locked to being where the end of the gearbox is, as opposed to being able to select how the pivots actuate in order to get anti-squat during acceleration or anti-lift during braking (you will get anti-squat / anti-lift with torque tube, but you get what you get ... unless the torque tube mounting is separated from the gearbox to allow the length to be tuned, see 4th-gen Camaro)"

I obviously failed in explaining the use of a torque tube with an open driveshaft. The forward end of the torque tube would have absolutely no relationship to the end of the transmission tailshaft housing. Since it would be offset to the right of the driveshaft, it could be positioned to give any desired antisquat percentage. You wouldn't want adjustment in a production car, of course, but, during development, an adjustable front bracket should suffice.

Again, this is not a shortcoming of the offset torque tube.

"Chassis and suspension related loadings have to be absorbed by the engine/gearbox mountings and the various vibration or torque related motions of the engine/gearbox can't be separated from the suspension actions."

Again, the offset torque tube would need have no relationship with engine/gearbox loadings.

"But, the torque arm is really long, adds a bunch of weight, still needs a trailing arm on either side and a Panhard rod...."

All true, but, remember, we're talking about "what might have been" in the evolution of the RWD beam axle car. The few production examples remaining will soon be gone. But, when you consider that the additional trailing arms would not be required with the offset torque tube, I find it difficult to understand how something as heavy and complex as the torque arm was ever considered.

I, too, appreciate the simplicity of the 3link. But, when you consider that a torque tube is merely an extension of the axle housing, it is, in fact, a "zerolink" suspension. (Can't count the Panhard, of course.) Can't get much simpler than that.

http://www.racetec.cc/shope

 

[BrianPetersen]

It would seem to me that the 4th-gen Camaro solution is the one you're interested in ... as in, "why did nobody think of this sooner"!

Seems to me that the geometry isn't a whole lot different from the "truck arm" setup that GM used in the half-ton pickups from '63 (I think) to '73 (ish). Main problem with those, as I understand it, was the shortness of the Panhard rod that they used, and I've seen aftermarket solutions for that, it's not inherent in the design. And they take up a lot of space underneath the vehicle. Those trucks had the fuel tank inside the cab. When they wanted to move the fuel tank underneath and out of the cab (good plan!) the place that made sense was the place where the old "truck arms" were. As far as handling ... NASCAR still uses it, for what that's worth.

In cars, this was all long before my time, but I think the evolution process went from transverse leaf springs to longitudinal leafs with the torque tube (the leafs did only the springing and lateral locating) to longitudinal leafs with no torque tube (the leafs did all of the axle locating). No question they lost something in that latter transition and ended up with a new problem, axle wrap, that the torque tube deal didn't have. I don't know how the transverse-leaf setups located the axle in the back. I know the front used a leading arm on each side, and the axle was not torsionally stiff; sort of the predecessor to the modern torsion-beam axles on the back of lots of front-drivers.

 

[BillyShope]

"It would seem to me that the 4th-gen Camaro solution is the one you're interested in...."

Not being argumentative, Brian. Just trying to clarify. A torque ARM does not carry X-axis loads, meaning a couple of other links must be added. A torque TUBE does carry X-axis loads, meaning that only the torque tube and a Panhard are needed to locate the axle assembly.

So, the third and fourth generation Camaro rear suspension is not at all that in which I'm interested.

It just appears to me that a more logical design progression would have been to an offset torque tube and then...perhaps to reduce weight...to a multi-link, with the torque arm skipped completely.

http://www.racetec.cc/shope

 

[BillyShope]

Forgot to comment on the truck arm:

Yes, this is another "zerolink" suspension and is, in effect, exactly that to which I'm referring. It is, unfortunately, symmetrical in plan view, so it loses the driveshaft torque cancellation feature available in the offset torque tube.

I know NASCAR rules require this sort of rear suspension, but I don't know how far they go regarding symmetry. On a course with only left hand turns, this is not as significant a problem as was the one encountered by the designers of the asymmetric 3link used in the beam axle C-Type Jaguars.

http://www.racetec.cc/shope

 

[MikeHalloran]

Yep, I misremembeted. The Camaro's torque arm has a slip joint at the front. It's still strong enough to deal with symmetrical thrust, e.g. with a ball joint at the front. Okay, maybe for a slightly lighter car. What it can't deal with is asymmetrical thrust, e.g. from taking off or braking with one wheel on ice. So make it a tube, and maybe it can take some of that, but it would probably be squirrelly without diagonal braces, or you'd have to make the joint at the punkin stronger to take the lateral moment.

But all this talk of logical design progression?

No logic involved. No single master designer.

Evolution is not really a linear process.
More like a completely random process, with the branches trimmed to net a path that generally lies in a direction that no one knows beforehand.




Mike Halloran
Pembroke Pines, FL, USA

 

[BillyShope]

"No logic involved."

How true. How true.

http://www.racetec.cc/shope

 

[GregLocock]

Are you talking about production cars or racers?

For production cars it is more or less as Brian suggested early on, there is no room for a live axle any more, never mind an even bigger clearance zone around the torque tube.

The driver there is that you need to take the exhaust either over or under the axle, in practice you go over and eat luggage space.

That sacrifice is no longer acceptable.

Having said that the MX5 (Miata) uses the IRS version of a torque arm, and of course Peugeot, who know a thing or two, used it for longer than most.

Rear axle roll steer is an inherent part of your understeer budget, and is certainly not something that 'just' affects steering wheel angles. Perhaps that is not important for some classes of racecar, but high speed yaw stability is of great interest in the operating envelope of some racecars and many roadcars.

Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[BillyShope]

I meant to make it clear that I was talking about ancient production car history and that I certainly was not suggesting that we return to live axles.

And, again, I'm talking torque tube and not torque arm.

"Rear axle roll steer is an inherent part of your understeer budget, and is certainly not something that 'just' affects steering wheel angles."

Well, we certainly know it doesn't affect tire loadings, so I must assume you're referencing the driver portion of the car/driver system. And, yes, I recognize that the driver must be both capable of handling the change in steering input and...perhaps more importantly...be at ease with it.

http://www.racetec.cc/shope

 

[GregLocock]

The objective part is that the rear kinematic oversteer operates at about the same time constant as vehicle roll, subjectively that means it is very slow. The usual observation is that after corner entry you have to back the SWA off, as you get a second bite from the rear coming round. I imagine it might be possible to time this to improve rotation on entry, but with full size sedans the time constant is just too long. Trying to compensate for this by dialling in lots of rear axle compliance steer doesn't really work, as the compliance effect builds up with slip angle, which is a different SDOF system to the roll response which is controlled by lat acc.

The linear range understeer budget entirely controls the high speed yaw stability of the system, even without significant driver input. If you exceed the critical speed of an oversteer car then you will be in the ditch, since the SWA and the yaw response are in antiphase. This is not a subjective problem or a matter of preference.




Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[BillyShope]

Consider, Greg, Chapter 17's Problem 6 in the "Problems, Answers, and Experiments" supplement to the Millikens' RCVD. This problem examines roll steer effects on a skidpad test.

Conclusions include no change in slip angles or final plow/spin characteristics with changes in roll steer. While a skidpad test does not involve transients, it will demonstrate the instability associated with a dynamic effect like critical speed. But, since the slip angles are unaffected by roll steer, no critical speed would ever be encountered. This makes roll steer unique in those items included in an understeer budget.

Transient effects, in corner entry and exit and in overtaking, will...as you indicated...be present IF roll oversteer is sufficiently large. But, as I pointed out with regard to the '57 Chrysler line, a "tad" of roll oversteer can be used as a psychological tool to hide a bad push in a passenger car.

http://www.racetec.cc/shope

 

[GregLocock]

I'll need to think about that but am very leary of using constant radius or swept steer tests to assess stability. If I were to test it I'd use pulse steer inputs on straight line running, at different speeds.

At the same time if you are saying that rear axle roll oversteer is 'free' oversteer (ie you get the advantage of oversteer (undershoot) without the disadvantage(speed sensitive yaw instability)) then that may explain why these rear axle configs that give roll oversteer had such a long life.



Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[BrianPetersen]

Should note that with the torque tube setup or the truck arm setup, you can control to some extent whether there is roll understeer or oversteer by the height of the main pivot relative to the height of the axle centerline and the height of the pivots for whatever you are using to locate the axle laterally.

Roll oversteer can cause stability problems during transients. In the case of the '57 Chrysler, most likely the suspension was so soft and had so much body roll and understeer in general that it probably wasn't an issue. It won't show up in a skidpad test. It WILL show up if you do an "elk test" (the now-infamous European magazine test that caused the first generation Mercedes A-class to roll over) - basically an abrupt double lane change. From the angles of the leaf springs on a good many newer light trucks, I can see that they are trying to get a little bit of roll understeer ... the rear attachment point is higher than the front attachment point.

And Billy, you'd love the suspension that I built for my motorcycle trailer. Fed up with broken leaf springs and stuff in the trailer from getting shaken up from horrendously stiff normal trailer suspension, I built a coil-and-shock suspension for it a few years ago ... and the geometry that I used is the same as torque tube with Panhard rod, although it's constructed more like short truck arms with the front of the arms brought together to have a single pivot point. Reason I did this is that the number of links and pivot points is the absolute minimum. I picked the geometry to give zero roll steer at nominal ride height due to foreseeable stability problems of having either roll understeer or roll oversteer. Works fine. No more broken or loosened tie-down straps, and the trailer is stable on bumpy highways.

 

[BillyShope]

"In the case of the '57 Chrysler, most likely the suspension was so soft and had so much body roll and understeer in general that it probably wasn't an issue."

You got it! That shortened forward portion of the leaf springs just reduced the necessary steer input enough to leave a favorable impression with all the magazine "testers."

"...from horrendously stiff normal trailer suspension...."

Hey, my last utility trailer had no suspension and I helped a friend move his piano without knocking it out of tune. Fortunately, none of those Florida sinkholes opened up under me during the trip.

http://www.racetec.cc/shope

 

[patprimmer]

Trailers with rigid mounted axles respond very favorable large baggy tyres run at low pressure.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers &

http://eng-tips.com/market.cfm

for site rules

 

[GregLocock]

Here's a comparison, quick answer for that car is that rear roll steer is very similar in behaviour to modifying the tire characteristics.



Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[BillyShope]

"Here's a comparison, quick answer for that car is that rear roll steer is very similar in behaviour to modifying the tire characteristics."

Don't follow you here, Greg. As soon as you start talking about changing tire characteristics, you introduce the possibilities of the dynamic "nasties" like critical speed.

A far better comparison is to think of the car in a corner as, instantaneously, having beam axles at both the front and rear. The two beam axles have the same relative locations whether the car has roll steer or not. But, as roll steer is introduced, the plan outline of the remainder of the car is rotated.

http://www.racetec.cc/shope

 

[BrianPetersen]

All true, as long as you have steady state cornering.

But, too much slip angle at the rear during an abrupt transient - whether induced by geometry or compliance or tire side-slip - can start a pendulum effect happening and that can make the rear start chasing the front. Autocross and roadrace drivers can use this to get the car turned faster. If this happens to a "normal" driver, e.g. during the "elk test" quick double lane-change or something similar, they'll often find themselves staring at the ditch, and trees, and then back at the road they came from. Transient response is more important for this sort of thing than outright steady-state grip on the skidpad, although much more difficult to analyze and predict.

For what it's worth, during the 4-wheel-steering fad of the late eighties and into the nineties, the various systems by Honda, GM, Nissan, etc all intentionally steered the rear wheels the same direction as the front (albeit much less so) at small steering angines / higher road speeds with the intention of keeping the rear end of the car in, with the hope of avoiding the pendulum and snap-oversteer effect.