Here's another thought: What's so great about progressive springing (especially if you are not running a car with serious downforce like an F1 or Champ car)???

Answer:
Nothing really that I'm aware of since stiffer does not equal faster. Since our beloved 911's in 99.999% of the cases are not developing downforce, ideally I'd like to have a car that does not get stiffer and more skittish at the limit. Others may defer.

Footnote: All of this would be true if all suspension rates were linear, but they are not. Many cars (take the 906/908/917's for example) actually have REGRESSIVE suspension geometery. This was a side affect of trying to package the coil-over springs on the car. Using properly selected PROGRESSIVE springs on the REGRESSIVE suspensions will often closely approximate a linear wheel rate. The result is a car that was predictable at the limit which made them easier to drive. It is not unheard of for many race cars of early 70's (when they started experimenting with progressive wheel rates) to handle nicely at 9/10's and be treacherous at 10/10's. The result was a car that whose handling would change drastically from full tanks to a light full load. Not a recipe for success.

jluetjen:

Not to start a flaming war or anything...
but what in God's name are you talking about ??
Regressive suspensions ?? What ?? Combined with progressive springs to result in a linear function ??

Coil springs ( the springing medium) is either linear ( "X" lbs/ in), or progressive ( "X" lbs/in for the first inch of travel...."Y" lbs/in for the next inch of travel.. and so on). For "progressive", I guess it can get either more stiff as it collapses or less stiff ( ? regressive ?), but in any case ...the springing medium can only be one of these two basic types....linear or progressive in some fashion.

Suspensions are type-cast by the contruction specifics and the resulting *geometry* they offer as they go through their range of travel. There is nothing inherently progressive, regressive, or linear in this regard. You can measure, for instance...the change in camber as the suspension goes through it's full range of motion...and that would determine the camber curve. The Porsche models you mention had no unusual packaging that wasn't done by everyone else at that time. The coil-over damper was placed somewhere at an angle inside either the upper or lower A-arm. As a matter of fact, the earliest of these cars...around the 906/907/910 era...Porsche actually bought Lotus formula car pieces as a template to quickly make the same parts for the new 910 !

As to diabolical handling at 10/10's ...that can be easily explained when springs ( of *any* kind) get close to ( or "at") a full-bind condition. When that happens, the spring rate goes to *infinite*...and that end of the car goes full-stiff. If at the rear..this would result in horrible oversteer..not something you want in a high-speed sweeper.

---Wil Ferch

Ok, i understand what was said about progressive coil springs, and i understand that springs are nothing more than spiral wound torsion bars, but you guys lost me with these progressive torsion bars. How can you make them progressive if you can't change the length under torsion, and wouldn't having multiple diameters actually weeken it since a torsion bar can only be as strong as it's weaker point in the bar?

Strangely, dual stage torsion bars have been available for Toyota 4WD pickups for years. It was basically a two-stage system where a smaller diameter bar twisted inside of a larger bar until it hit a stop and then the larger torsion bar came into play. I think they were made by Sway-A-Way like most of the other torsion bars out there...

The applications on Toyotas was a situation where one wanted a relatively smooth on-road ride but also need the higher spring rate to control the heavy tire and wheel package off road when greater wheel travel came into play.

Mike

OK Wil; Take a deep breath.

Here's the deal, most coil-over suspensions with outboard springing are "regressive".

Why?

Because shocks and springs travel in a linear fashion while the uprights move in arcs. If you take your average outboard coil-over suspension set-up (say from a 906) and measure the change in spring and shock travel versus the change in wheel travel, you will find that it is rarely linear. When the top (chassis end) of the coil-over unit is inboard of the bottom (upright end) of the unit, the suspension geometry will be regressive or falling rate. So for every inch that the wheel moves up, the coil-over unit will move less and less of an inch. For example, the first 1 inch of wheel travel may move the shock 1 inch. The second 1 inch of wheel travel will move the shock .9 of an inch. The third inch of wheel travel will move the shock < (.9 * .9) of an inch and so on. The result is that a linear rate spring in this situation will feel like it is getting softer and softer as the wheel travels.

What to do about this?

1) Use a rising rate spring which is what Porsche did with the 906's etc.

2) Use some sort of an inboard system such as rockers, pull-rods or push-rods where you can play with the geometry to deal with that issue. This is what most modern race cars do.

3) Use active suspension technology to corrective the problem. This is illegal in most racing series nowadays.

Does that help clarify what I was trying to say?

BTW - I think that the 911's T-Bars neatly sidesteps this whole problem by dispening with coil springs! T-Bars will be far more linear then coil-overs. So the question is in the absence of huge corning loads as generated by super wide and sticky tires (which in turn require super stiff springs to keep the car off of the bump stops) -- why bother? The T-Bar's stiffness is certainly adequate.

1fastredsc:
You are correct. And the composite of the previous posts pretty much bear this out and say the same thing.
- a sloping-diameter torsion bar will not get you variable rate. A sliding mechanism that engages at various distances along the length of the bar will. A coil spring that progressively binds ( like Chuck Mooreland's example) will do this too.
---Wil Ferch

jluetjen-

I think there may be some confusion here on the "progressive spring" terminology. It seems as though you are talking about the rate at which the spring compresses, whereas others in this thread are referring to the rate at which the force increases to compress said spring. Of course I could be totally clueless.

Eric;
Actually I'm talking about both. "Progressive" springs are springs with a rising rate as you described, so it takes 100 lbs to compress the spring 1 inch, 300 lbs 2 inches, 900 lbs 3 inches etc. I'm also discussing the suspension rate which is as I described.

When you are talking about a car's suspension, you really can't discuss one without the other since it is really the WHEEL RATE (the combination of both) which affects the car's handling. If you have a linear rate spring in a regressive (falling rate) suspension, the result will be that it may take 100 lbs to move the wheel inch, but only 190 lbs to move it the two inches, 260 lbs to move to move it 3 inches and so on. Now if you put a rising rate spring into this falling rate suspension, and do it right you could wind up with a linear wheel rate which is the most predictable.

So then a progressive spring, for example, would let's say preload 100lbs in the first inch. Then in the next inch as more starts to bind, it might be like 300lbs with the second inch as opposed to 200lbs like a linear spring. Did i nab it on the tail? And i think i understood what you said about a smaller torsion bar in a larger diametered one, but wouldn't that be hella expensive?

There are a number of different kinds of "Progressive" springs:

1) The "ideal" progressive spring would be progressive across it's entire travel. Picture a smooth exponential curve. This sort of spring is very difficult to make since the shape of the spring is critical with either a conical shape or variable distances between the coils or both.

2) The reality is that most "progressive" springs are in fact "Duel rate" springs. So initially they have a "soft" slope until the soft spring, or portion of a spring, binds up at which point there is a bend in the curve and it resumes at the "stiff" rate. Picture a line drawn at a 45 degree angle that then shifts to a 60 degree angle. The trick with these springs is figuring out where to put bend? If you are using "tender" springs, then the bend will occur at very light loads, while hitting a bump stop would be an example of where the spring rates rockets from X lbs/inch to almost infinity within an inch.

3) Bump stops and tender springs when used in series with other springing constitute a "progressive" enhancement to whatever the "spring" is.

Now combine this with the fact that suspension geometry if often not linear, but regressive and you've got an idea of what the sort of problems that a suspension designer has to deal with. (And I haven't even mentioned suspension frequency and shocks yet!) Sure, lots of people throw springs at the problem until they stumble upon a solution. But the clever guys will have the spring rates generally figured out before they are put on the car.

Going back to the my question of why bother? A well handling car will transfer the weight predictably to the tires to maximize traction. Introducing progressive wheel rates at one end of the car or both could quite likely result in a car that suddenly shifts from understeer to oversteer (and back again!) depending on which end of the car transitions the bend in the suspension rate first. Making a poor match of progressive springs with an inately regressive suspensions could result in a wheel rate curve with multiple bends in it!!! At which point you can pretty well throw out the window any hopes of the handling being "predictable" to the driver or crew chief.

Going back to the original question: Why aren't torsion bars progressive? I guess the easiest answer is that their comparatively simple design parameters (length, thickness, material) don't permit it, while the additional parameter of shape for coil springs does. The good news is that the option of mounting of T-Bars concentrically with the suspension pivots could pretty well eliminate the need for progressive springs on a 911.

The funny thing is that this thread has forced me to rethink the suspension design for my 911 race car project. I was planning on using coil springs due to the ease of tuning, but now I'm reconsidering it. First I need to find some time to measure the wheel rates on the car.

Great info John!
-Chris