Application specific suspension.....

[SN65]

Bob, who are you asking? Or is the comment rhetorical?

jp

Hi JP,

No one in particular. it was more of a rhetorical comment. That is, unless someone is in disageement.

[parsonsj]

OK. I don't disagree, but I must (sorry ... it's this engineering background thing again) ask for some measure of quantity. In other words, how much? You've cut torsional deflection to 25% (or so) of the factory value. In and of itself, it is meaningless.

Any idea how it will affect spring rates? (less body movement means more suspension movement for a given input load) Shock valving? (shocks have to handle more direct motion)

Any calculation on a resultant improvement on lateral acceleration?

Don't get me wrong, I'd make those same improvements in a heartbeat ... I'm just wondering about the measurable effects on a car's performance.

I remember somebody once telling me that their sub-frame connectors made the car "feel a foot shorter". That's great, but what does that really mean?

Thoughts?

jp

[JMarsa]

I'll jump in...I'd like to be able to quanitfy the results of my "engineering" changes but like many, unsure exactly how to do that with the tools I have at home.

For example, I've seen the "tortional" test of attaching a lever to a chassis (width-wise) and measuring the results with a dial indicator from a fixed base.

Should that test start with a stock subframe?
Should it then be attatched to the car?
If so, what body mounts ?

If there is an improvment what does it prove?

What I'm getting at is what's already been brought up here, what standards used and what do the results really say?

--JMarsa

[David Pozzi]

A stock 73 Camaro I tested in running condition had (from memory 4000 ft lbs/deg). We added the Herb Adams style braces from firewall to upper A arm bolt on each side, and upped torsional rigidity to 6000. It was an improvement that was felt by the driver and the feeling was described as quicker response to steering inputs.

I've heard 10,000 ft lbs/deg mentioned as a good goal for a tube chassis race car. I also read a blurb about the new Aston Martin model having 12,000 ft lb/deg stiffness. I'd like to know what a new Z06 is, the only figures I've seen are in percent of improvement over the past models.
David

[SN65]

OK....

For working in engineering for almost 30 years now, my basic math skills suck the big one. :-(

I am applying 750 lbs of preasure 2.5 feet off axis. (750 lbs up on the drivers side and 750 lbs down on the passengers side).

That comes to 1,875 foot lbs???

I am getting .3 degrees of deflection.

That same setup netted me 2.5 degrees of deflection before I started. If all this is right I would end up with 750 foot lbs to move the unmodified chassis 1 degree, and that just does not sound right.

So, I don't think I can assume that increasing the force to 5625 foot lbs will get me 1 degree of deflection.

With that said, I think that I am going to have to increase the presure to whatever it takes to get 1 degree of deflection and see what the reading is at that point. I have noticed that the preasure required to twist the chassis is not a "linear" curve. It takes more and more pressure to move the chassis less and less.

Does this sound correct or am I off somewhere????

[High Plains Mopars]

OK. I don't disagree, but I must (sorry ... it's this engineering background thing again) ask for some measure of quantity. In other words, how much? You've cut torsional deflection to 25% (or so) of the factory value. In and of itself, it is meaningless.

Any idea how it will affect spring rates? (less body movement means more suspension movement for a given input load) Shock valving? (shocks have to handle more direct motion)

Any calculation on a resultant improvement on lateral acceleration?

Don't get me wrong, I'd make those same improvements in a heartbeat ... I'm just wondering about the measurable effects on a car's performance.

I remember somebody once telling me that their sub-frame connectors made the car "feel a foot shorter". That's great, but what does that really mean?

Thoughts?

jp

So what do you expect to see?

A 25% increase in torsional rigidity resulted in a 10% decrease in needed spring and shock rates?

or

A 25% increase in torsional rigidity netted a 1/2 second faster slalom time and .25 increase in lateral acceleration?

I think we would all agree that a more rigid chassis allows better control of the suspension components. Are you looking for what does this additional control benefit you in skid pad testing or some other measure?

[parsonsj]

So what do you expect to see?

A 25% increase in torsional rigidity resulted in a 10% decrease in needed spring and shock rates?

or

A 25% increase in torsional rigidity netted a 1/2 second faster slalom time and .25 increase in lateral acceleration?

I think we would all agree that a more rigid chassis allows better control of the suspension components. Are you looking for what does this additional control benefit you in skid pad testing or some other measure?

Any of the above, though the 2nd would be better since it is directly tied to performance. We have all been taught that a stiffer body is a better body, but that's the end of the discussion (as far as I know anyway). I'm wondering if there is anyone with direct knowledge of how making the body stiffer makes the car faster with quantifiable results free from confirmation bias ...

As I said, I'm a believer ... I'm just wondering if how much of my belief is mythology and how much is quantifiable.

jp

[SN65]

So what do you expect to see?

A 25% increase in torsional rigidity resulted in a 10% decrease in needed spring and shock rates?

or

A 25% increase in torsional rigidity netted a 1/2 second faster slalom time and .25 increase in lateral acceleration?

I think we would all agree that a more rigid chassis allows better control of the suspension components. Are you looking for what does this additional control benefit you in skid pad testing or some other measure?

I expect to see an improvment in three areas.

1... Ride quality
2... Overall responsivness
3... Better accelleration all around (straight line, lateral and brakeing)

The only way to quantify any of this would be to make all the mechanical changes without any structural changes what-so-ever. Run the car and then make all the chassis modifications.

Unfortunatly, this is not the proper order of things when you build a car. The chassis is always first.

[parsonsj]

this is not the proper order of things when you build a car. The chassis is always first.

In our world, sure. We don't have the resources to do before/after/before testing. But if we're buying chassis/body stiffening products, shouldn't the manufacturers of such products provide such data? Or should we just buy the mythology? Stiffen thy body, and thou shalt be blessed.

jp

[High Plains Mopars]

In our world, sure. We don't have the resources to do before/after/before testing. But if we're buying chassis/body stiffening products, shouldn't the manufacturers of such products provide such data? Or should we just buy the mythology? Stiffen thy body, and thou shalt be blessed.

jp

IMO, most manufacturers don't have the resources to perform such tests nor do their customer base demanded it, so why bother with it since it is expensive and time consuming. XV is the first aftermarket rganization I've heard of conducting some pretty exhaustive testing...which, ironically, a lot of mopar guys I've seen have then questioned the value of. Go figure.

The way I see it, there is a lot of "associative acceptance" going on with this. I.E. the designer is an engineer, their other product line is race specific so they should know what they are doing, feedback on their products is that they work great. Any or all of these three attitudes prevail when contemplating the purchase.

In addition to that is then going to be either A)cost or B) looks. For better or worse, I see the majority of pro-touring cars as looks first, performance second. It is an appearance driven hobby as there is no universal rule book of what consitutes pro-touring or that regulates what constitutes goodness. Now I'm not saying that it isn't performance based, just, like I said earlier in this thread, so long as the result is better than stock, it is accepted as good with little to no proof behind it. Since the vast majority of PT cars are never judged against a stop watch, the majority of owners don't use quantified results as a primary decision driver.

Is that bad? Well, in this case, not really. They do get improved performance. Is it application specific? Well, since the application is improvement in looks and performance and not necessarily against a stop watch within the confines of the rule book, then yes.

[parsonsj]

so long as the result is better than stock, it is accepted as good with little to no proof behind it.

I agree. I would even assert there are a lot of things in the aftermarket that results in performance worse than stock that is accepted as good with little or no proof.

This has been an interesting thread ...

jp

[wiedemab]

My take on chassis stiffening has been that it allows you to be in control of the items that have more predictable behavior - ie: springs, shocks, sway bars, and the basic geometry of the suspension.

In the example of 2.5deg of deflection, if you assume that it is rotating about the centerline of the car and the suspension pickup points are approx. 20" from the centerline the vertical deflection for each side is .87" (total of 1.75"). I am very far from an expert, so someone else chime in, but if this chassis flex wasn't included in the original calculations when designing the suspension geometry - I would think that this much deflection could cause some undesired results.

If you lessen this to .3deg for the same torque applied you end up with .1" per side (.2" total).

Could you model the suspension with the pickup points moved for deflection and see how much it affects the design characteristics?????

OK - Hopefully that made some sense, but again - I'm just thinking outloud and that's scary for me!

I agree that this in an interesting topic. I'm sure I'll learn something!

Later,

Brandon

[David Pozzi]

OK....

For working in engineering for almost 30 years now, my basic math skills suck the big one. :-(

I am applying 750 lbs of preasure 2.5 feet off axis. (750 lbs up on the drivers side and 750 lbs down on the passengers side).

That comes to 1,875 foot lbs???

I am getting .3 degrees of deflection.

That same setup netted me 2.5 degrees of deflection before I started. If all this is right I would end up with 750 foot lbs to move the unmodified chassis 1 degree, and that just does not sound right.

So, I don't think I can assume that increasing the force to 5625 foot lbs will get me 1 degree of deflection.

With that said, I think that I am going to have to increase the presure to whatever it takes to get 1 degree of deflection and see what the reading is at that point. I have noticed that the preasure required to twist the chassis is not a "linear" curve. It takes more and more pressure to move the chassis less and less.

Does this sound correct or am I off somewhere????

Only count the 750 lbs once, 750 X 2.5 = 1875 foot lbs applied. Chassis twists .3 deg, 1875/.3 = 6,250 foot lbs per degree. Pretty close to the Camaro I measured after the brace install. I'd say this is in the "good" range for an old muscle car.

I'm assuming the end being twisted is supported in the center and not constrained. I use either a small block of wood on a floor jack, or a piece of angle iron open end down on the jack to act as a pivot.

Measure from where you apply weight, to the chassis centerline in feet, multiply weight applied by lever length to get foot lbs applied. Divde that by degrees or tenths of degrees the chassis twisted = ft lbs/degree.

The stiffness may not be linear but applying the same test torque will at least give you some consistency. I don't think you really need to twist it 1 full degree, but enough to get some resolution in your readings would be good.

My original level only had .1 deg increments, not enough when my first test yielded .3 deg.

I bought a new digital level with .01 deg resolution, it's ten times more accurate. A dial indicator can be used to measure twist very accurately but you need to convert inches of movement into degrees.
David

[Mean 69]

My take on chassis stiffening has been that it allows you to be in control of the items that have more predictable behavior - ie: springs, shocks, sway bars, and the basic geometry of the suspension.

Exactly. The real story behind any suspension system is how well it keeps the tires in contact with the road (track) under all of the circumstances it faces: braking, linear acceleration, vertical bumps, turning forces, and all combinations. It's a heck of a thing to model and solve for, the goal of the suspension is to maximize the tire's performance. Because the chassis is such a complex structure to model dynamically, and once built, there's no tuning to it, the goal is to eliminate, or at least, greatly reduce, it's influence on the suspension characteristics. In other words, if the chassis is flexing, it is effectively PART of the overall suspension system.

I'd imagine that there are examples where a car with a noodle chassis structure has been characterized really well for a given circuit, or whatever the case. It's all about real world tuning, imagine a car that some guy has been running for years and has it setup perfectly. Take that same car, in that same trim with no other changes, and increase the chassis rigidity, and I'd bet that in many cases the car would be slower. Of course, once the new state is sorted, the car would probably overall be faster if the increase was significant. Further, changes in the tune of the car (spring rates, etc) would probably (will) be a lot easier to characterize and feel with a stiffer structure because there is less coupling between the two.

How much is enough? Like everything, it depends. I am certain that you will get to a point of diminishing returns on a street driven car when working exhaustively to increase rigidity. At some point, the other aspects of the performance package will become the weak link, driver, suspension, tires, etc. Clearly all cars are different, what's sufficient for one won't be for others.

I have a lot more opinions regarding other points expressed here, but not enough time to write at this moment.

Mark

[Damn True]

The question that often occurs to me is the following:

According to the guy I bought it from ^that guy^ the rear suspension I have selected (and damn its gonna be soooo slick) will work well with the chassis as I intend to build it.

According to the guy I am buying my front suspension from those products too will work well with the chassis as I intend to build it.

But.....will they work together? I don't know enough about it to evaluate it, but things like roll couple and such are sure to be a factor. But to what degree? And, would one have to be named Andretti to notice the difference?

[David Pozzi]

True,
Our 73 Camaro has a front weight bias of aprox 56% but the front roll couple is 74% give or take.
As you can see, the front suspension handles more cornering load than the percent of weight it carries. This tells me the rear is trying to twist the front subframe when cornering. If the front to rear weight bias is not close to 50/50, then I'd expect the torsional load the chassis "sees" to be greater.

A flexable chassis acts like a un-dampened spring, this would result in a time lag to steering input, and disconnected feeling to the steering. Lots of people say they feel subframe connectors after install, so a stiffer front sub connection should be just as noticeable. I don't think Mario is the only one who would feel it.

David

[Norm Peterson]

Very recently posted in response to a similar topic elsewhere by a guy in the automotive business (OE level, I believe),

The shock absorber is the most sensitive to the torsional stiffness. The shock absorber 'sees' the point impedance at which it is mounted. This is measured in units of mm/s/N, typically, and is a function of frequency.

For handling purposes we are probably interested in the frequency range 0-10 Hz.

The distribution of stifnnesses, and masses, and spring-mounted masses, in the vehicle will alter the shape of the impedance graph, but may not have much effect below 10 Hz.

Having units of time involved in addition to length, force, and mass suggests to me that transient behavior is affected much more than ultimate grip.


Norm

[SN65]

And it all brings us back to the basic chassis / unibody. The foundation upon which we build.

Really, all things being equal, the major difference between one pro-touring car and another is the chassis. How much it weighs. Where the weight is distributed. How it is constructed. Etc...

Everyone uses the finest mechanical components they can find, or at least they believe so. In fact, most in the hobby will spend tens of thousands of dollars on high performance components. And then again most will never give thier chassis a second thought, other than the manditory sub-frame connectors.

Hell, even Dynacorn has added some of the convertable structural items to the 67 stang chassis they build. They have even altered material thicknesses and composition. They understand the application that their chassis are are going to be used for and have taken steps to make sure that they are up to the challenge. I would have gone a few steps farther, but then again I am a glue, nails, screws kind of guy. :-)

[parsonsj]

Having units of time involved in addition to length, force, and mass suggests to me that transient behavior is affected much more than ultimate grip.

That makes sense to me. Steering input (turn-in, accel out) would see the most benefit in terms of "feel" to the driver. I doubt very much you can measure any difference on a skidpad ... and the bigger tracks would show less difference since transient behavior is a lower percentage of the total lap time. Perhaps it is measurable on a tight autoX ... but the best way to measure difference due to chassis flex might be slalom testing. Though as Bob has said, it isn't something the average enthusiast can do since it will require before/after/before testing on a single vehicle, and a statistically significant number of cars if testing a vehicle type.

jp

[Damn True]

True,
Our 73 Camaro has a front weight bias of aprox 56% but the front roll couple is 74% give or take.
As you can see, the front suspension handles more cornering load than the percent of weight it carries. This tells me the rear is trying to twist the front subframe when cornering. If the front to rear weight bias is not close to 50/50, then I'd expect the torsional load the chassis "sees" to be greater.

A flexable chassis acts like a un-dampened spring, this would result in a time lag to steering input, and disconnected feeling to the steering. Lots of people say they feel subframe connectors after install, so a stiffer front sub connection should be just as noticeable. I don't think Mario is the only one who would feel it.

David

So David,

Am I correct in assuming from the above that since my car will be lighter in the front by virtue of an aluminum block and a bunch of fiberglass that my weight distribution is likely to be closer to 50/50 and therefore I will be needing a relatively stiffer chassis?

Good thing I am planning an 8pt (maybe 10) cage.