Suspension Tech Question

[djmcmath]

I'm looking at upgrading my suspension, but I'm an engineer, so I can't just "Buy bigger parts and bolt them onto the car," I have to do some math first. Unfortunately, suspension design is still pretty much a mystery to me, so the exact math to do is kind of fuzzy for me. It seems, ultimately, that there ought to be an equation, of sorts, for picking matching t-bars, shocks, sways, etc. for the weight of the car. I know there's thumb-rules, and there are a lot of data points, but I was hoping for something a bit more scientific than we've done in the past.

First -- "spring rate" -- T-bars always talk about a "spring rate," which makes no sense to me. Springs usually have a constant, some force-distance number that tells how hard you have to push on it to get it to move through a distance. The "rate" part of the thing would imply an entering argument into the sinusoidal aspects of a bouncing car, like a frequency of repetition. That kind of makes sense, but I'm having troubles putting the pieces together.

Second -- Shocks. Bound and rebound are clearly rate limitations on the motion of the springs (t-bars). For engineering purposes, you could determine how fast was "correct," but that wouldn't help you any, because nobody advertises their shocks with rate numbers. ????

Third -- Sways, strut tower braces, roll cages, etc. These seem to just add to the overall stiffness of the vehicle. If balanced fore and aft correctly, it seems that they shouldn't effect over- or understeer, and should have ultimately no inputs into the suspension equation. Am I right?


Dan

[Grady Clay]

Dan,

Good questions.

Yes, there is the whole professional field of automobile suspension dynamics. There are (I’m sure) good computer simulation programs available.

First, yes there is a spring rate. When you measure the rate of a coil spring (force vs. distance) it may be constant pr progressively increasing. If you put that spring on a suspension, the wheel does not see that rate (usually) because of the suspension geometry. A torsion bar spring acts the same way. It has a spring rate also (torque vs. angle) that, through the suspension geometry, translates into a rate at the wheel.
You will find most of us amateur Porsche types simply spec the diameter of the torsion bar rather than the spring rate at the wheel. The larger the diameter - the greater the spring rate.

Second, automobile suspensions are all “over damped.” The question becomes “how much?” Yes when you increase the spring rate you (usually) should increase the shock rate (force vs. velocity.) The shock manufacturers and many good rebuilders have shock dynos and can valve shocks to your specs (within reason) or simply give you a report on their performance.

Third, There are some basic laws of physics that apply to automobiles.
Lighter is better – The overall car and the “unsprung weight” are very important.
Vertical height of the center of gravity – There is a reason that F1 cars go to extreme lengths to get several hundreds pounds underweight and then ballast with a huge steel plate as low as the rules allow.
Stiffness – the chassis is a torsional and bending mass-spring-damper system, so are the tires.
Balance is VERY important – the systems at front and rear must be appropriately balanced. Sway bars can be used to finesse the balance.

2c

Best,
Grady

[Schrup]

Dan, I don't have a scientific answer, what I did was to do a lot of research on this board to figure out what everybody else was using & what they thought of the performance. I first guessed what my car would weigh after I lightened it up (2500lb). Then I determined that I didn't want too harsh of a ride, as many that have used 23/32 torsion bars complain about, nor did I want to go too soft. I went with 22/29, then I had my dampeners custom valved to match the weight of the car & the size of my Tbars. I am very pleased with the results. The car performs well on the track, but doesn't beat me up on the street.

I still think that you have Bilstein Sports, at least on the back because they are yellow, I believe the Boge shocks are black.

Your car is probable a bit heavier than mine so you can probable get away with a bigger rear Tbar 30 or 31 in the rear, but I still wouldn't go any bigger than 22 in the front.

[Wil Ferch]

I have published "equivalent" spring rates for the various diameters of torsion bars...both front and rear. Do a search or look for Tech article archives on Rennlist..it's there too. Look also on Thom Fitzpatrick's web page...he took my math on ride height calculations and made a simple data-entry spread sheet that works very well.

Remember that torsion bars are nothing more than coil springs laid out in a "straight line". The spring rate ( lb/in) increases to the "fourth" ( ! ) power of diameter...pretty dramatic.

Shock dyno info is very useful but even when you get such info ( mostly you can't)...it's published performance would likely be too narrow to be any good. For example, you might get a force/displacement shock dyno curve that is accurate for a specific shock speed or total travel...and these would look different for different speeds and different total displacements. Meaning? you'd need a whole bunch of curves and somehow relate those to the springs.
When all is said and done...you'll likely look for emprical data....compilations from field data ( this group here ! ) of what works and what doesn't. Still...your own perception of what may be "harsh" and "soft" will be different than the next guy...so this is not exact but can give you some guidance.

If it way easy....racing teams would put Cray computers to work and not deal with "dialing-in" cars during qualifying practice. Unfortunately...still necessary.

--Wil

Edit: You may look at what you have right now...relate that to spring rates..and guess what new torsion bars may do for you. Remember too that ( relatively speaking)...the more stiffness you put in front will tend toward understeer, and the more stiffness you put in back will tend toward oversteer. Also, if you check your understeeing/oversteeing characterisitics.....dialing -in "neutral" characterisitic for low speed turns ( where for a given "g" load you have lots of steering wheel angle cranked-in)...will tend the car toward oversteer on high speed turns ( where, at the same "g" loading, you have less steering angle and more desirable tire contact profiles)...so be careful on tuning.

[widebody911]

I took an approach similar to Paul's; I took stock of what others with similar cars were running, and went with that. In all honesty, at that point in time, you could have plopped me into the best prepped 911 in town and I wouldn't have been any faster around the track. I put all the hand-wringing about which shock valving and stuff for a later date when I was a better driver and understood what I was doing.

Besides, if you do everything up front, what is left to do later?

As an aside, I just installed all new Elephant Racing bushings, corner balanced the car and I'm getting it aligned today. I'm going to try a slightly more aggressive setup and see how things go.

[djmcmath]

Thanks for the replies, all. I was hoping some of the smart people would chime in.

I was afraid it would boil down to "Just use empirical data of other people's subjective opinions of their cars, as there's no way to compute what will and won't work." (sigh) Of course, before I really put honest-to-goodness money into the suspension, I need to log some honest-to-goodness track time, as Thom notes.

Part of the problem, of course, is that I'm not sure what's really in my car right now. Paul notes that Yellow is usually Bilstein sports -- but other threads note that yellow is also sometimes used by Koni. And by Boge. I don't know if they're stock, but I do know I don't have receipts for any suspension work from the PO. I know the suspension is a lot more "stiff" than my room-mate's Subbie and my own Honda, but still seems "soft," compared to where I'd like it to be. In fact, it's soft enough that at roughly euro height, the stock tires rub, which means something's too soft. Is it shocks? Is it t-bars? Don't know, not a suspension guru. (sigh)

So it seems like I ought to replace the shocks, perhaps, as those seem like more of a wear item than the t-bars. But if I'm going to replace the shocks, it should be with an eye to the ultimate upgrade sequence, which is a function of the car's final weight, matching t-bars, etc. (sigh) I'll hunt down the tech articles you reference, Wil. That looks like a great starting point. Still, it's a huge guess as to how soft or stiff the suspension will be. I don't know how long the current bars have been in, or how much of their original "spring rate" has been lost, so I can't relate that directly or objectively to anything. Worse, as Wil notes, the size of the t-bar changes the spring rate by the 4th power, so even a single size up or down could make a huge difference. Two sizes out from "optimal," heaven forbid, would be ridiculously bad. (sigh)

Once I understand all this, I'll build the computer program that'll model suspension changes accurately, honest. Then this business won't be nearly so subjective. Anyone own a Cray they could loan to me?

Dan

[anthony]

Third -- Sways, strut tower braces, roll cages, etc. These seem to just add to the overall stiffness of the vehicle. If balanced fore and aft correctly, it seems that they shouldn't effect over- or understeer, and should have ultimately no inputs into the suspension equation. Am I right?

Wrong. A sway bar adds additional spring rate to the wheel just like the torsion bar. Both have a direct effect on understeer/oversteer.

I think the first thing to do is make sure you have good shocks and buy the absolute best tires you can afford.

[Superman]

Dan, you seem like a nice guy so I'll offer my two cents. I'm not sure I trust anyone's knowledge of 911 suspension characteristics as much as that of Tyson Schmidt. He is the young genius who designed and built Black Beauty II's suspension mods, which you may know is a fascinating story of improvements on the already-brilliant German design. Anyway, here is some advice he gave me lately (I hope he does not mind my reprinting it here....I removed all the nasty things he said about all you guys):

"Howdy Jim! I think the best toe specs are 1/8" toe-in in the rear, and 1/16th toe-in in front. Or in degrees, +.40 total rear (+.20 per side) and +.16 in front.(+.08 per side)

For the future, I'd get a Carrera Sport rear bar (21mm factory bolt-in) and a 22mm adjustable through-the-body up front. (Weltmeister or equivelent) I know you like to believe that stiff torsions and weak sways are the answer, but the truth is that this method limits suspension travel too much. I think 21mm front and 28mm rear bars with the above mentioned sways makes a great package. And of course revalved Bilstein shocks and struts. "

Then he went on to say:

"If you can swing it, rear mono-balls for the inner control arm bushings are actually totally liveable on the street. You can't tell they're there, except the improved confidence from less rear toe-change while cornering.

That's really about all you'll ever really need, unless you want to go crazy. Beyond this set-up, it's major diminishing returns."

[Bill Verburg]

The general consensus seems to be "monkey see monkey do". Not a bad way of doing things as you can avoid the worst mistakes.

You can calculate to death but for us amateurs there are so many fudge factors that you don't get an awful lot of useful output.

Generally even the lightest cars like a track only RS l/w will have the stiffest bars(when I jump my 210# frame onto them they hardly budge at all), Bilstein will custom valve shocks for you and provide the#s, just tell them what you have and what the use is and you will be set.

Sways are best when they are stiction free and adjustable. They do contribute to suspension response when ever the paired wheels see different inputs. This is their design goal.

Connection between the chassis and suspension pieces also needs to be rigid and stiction free, monoballs, needle bearings, heim joints etc. all contribute to this cause but would likely cause unacceptable harshness tomost street users, the hardcore will always put up w/ a little discomfort for a smidgen of performance gain.

[djmcmath]

Boy, the replies just keep getting better. Anyone want to page Tyson to actually get his first-hand thoughts? I'd concur, Supe -- Tyson's work on BBII that I first saw in Excellence was an awesome piece of work. If I understood what he did and why, I'd feel a lot happier about this process. Thanks for the inputs.

Roger that business about sways; I stand corrected. Now that I think about it, it does make sense. I hadn't considered Tyson's point, though, that weak torsions and stiff sways are a wiser choice. I'm trying to figure out how to make that make intuitive sense in my own mind, and I'm mostly failing. It seems like the sways/torsions create a square (sort of) frame that resists rotational forces. The simple principle, for stupid guys like me, then, would be that roughly balancing the force applied from the sway and torsion to the "square-ness" of the frame is best, rather than having one or the other extra-super-sized to compensate for a smaller counterpiece???

Bill, I'm not sure I'm catching your full meaning. Is stiction the same as friction, sort of? When a rubber bushing catches on the frame, is that stiction? Seems to me that stiction would result in jerkiness of movement, as the rubber pops from one stable point to the next, which would contribute to lack of confidence, as the suspension changes suddenly, rather than progressively. The most confusing part of your post is the contradiction between rigid and stiction free. If my guessed-at definition of stiction free is right, then the parts should move freely -- which means they aren't rigid. Please understand that I'm not questioning your expertise; I just don't have the background to understand your semantics. Still ???

Aargh, this whole suspension thing is a nightmare for an engineer. Horsepower is easy; it's mathematically determinable, it's stable, it's easy to work with. Suspension work comes back to the vague measure of "driver confidence."

This thread, BTW, looks remarkably similar to the concurrent thread about tire size relation to t-bar size. Had I realized someone else would do the same thread, I wouldn't have cluttered the board with this one. I was also hoping it wouldn't break down to a "use the empirical data" result. (sigh) Oh, well.

Dan

[Bill Verburg]

Stiction is the resistance to intended motion, it can arise from friction, misalignment, improper clearance, lack of proper lubricant etc.

Some parts of the suspension are intended to move(but not freely in all dimensions), one case is trailing arms rotating about their pivot, the standard rubber bushes do allow this motion as well as some side to side motion which is only intended for comfort but detrimental to precise handling. Some racers use needle bearings at this point, they only allow the free rotation w/o any other component of motion, Mono balls work similarly but w/ more stiction/less freedom of motion(but still way better than rubber or p/u). The more free intended motion is the more responsive and controlled the effects of the suspension can be.

The whole purpose of bushes whether rubber or high tech is to allow intended motion as freely as possible and disallow other movement as much as possible within the constraints of other fuctions such as ride comfor,t NVH et al.

[Wil Ferch]

Dan:
Maybe we can simplify things..... if it helps, I'm an engineer, too. Sometimes we over-analyze and don't pay enough attention to 30-40 years of empirical data ( platform age of the torsion bar 911).

Using my car as an example...front torsion bars are 18.8 mm/ rear is 24. My 85 Carrera is considered amongst the heavier in the 911 progression, and common upgrades would be 21/27...or 22/28. Very common, even Bruce Anderson quotes these upgrades.

I've offered the opinion that the stock mid-80's cars come from the factory with a degree of built-in understeer, to counter the inherent oversteer these rear-engined cars have. You can feel the understeer ( my car has limited slip..making things a bit worse in this regard), when you sloppily apply gas in a tight, low speed turn. Much of this can be overcome by "technique"...just "how hard" and "when" you apply the throttle. After that....a bar combo more like 21/28..or 22/29 ( up one size in the rear) dials out a bit of understeer. For a mid-80's car, either of these combo's won't be unbearable..but this is a personal matter and you'd be advised to try out a similarly equipped car. The 4th power influence of bar size may net you a 80-100% increase in spring rate ( don't have the charts in front of me now)...but the majority opinion is that somehow....you don't feel the affect in "seat-of-pants" ...as if it is this much.

So...if you go up....try either 21/28 or 22/29 ( some go up 1 size further *in the rear*...21/29 or 22/30)....that big bob-weight in the back softens the blow somewhat more than you'd expect...but try it first.

Bills' comments are ..as usual...spot on. The rest of the story deals with the imprecise geometry caused by all that rubber collapsing and expanding in the suspension. To get the "compromised" suspension to work as it should...the other point worth considering is non-compliant joints, that pivot *exactly* as intended without distortion. This would be a combination of mono-balls on the top front shocks, poly ( or poly-bronze) bushing for all the suspension pivot points, etc.

A reasonable first step is a combination of Bilstein/Koni shocks, slight increase in torsion bars sizes, and perhaps poly bronze bushings. Then drive...and drive some more. Then we can talk about further upgrades after you have more seat time.

From an engineering standpoint, I disagree that suspensions are all aboyt "feel" and that engines...in contrast...are understandable in technical upgrade terms. As an example, pick up Fred Puhn's "Handling" book....an engineerlike you can truly appreciate the science involved.

---Wil

[fredmeister]

Can I chime in a little?
There are extensive computer programs like ADAMS that will do kinematic/dynamic studies of suspension geometry. They are used by OEM's in design of suspension components.....at least they should be.
Once the basics are decided upon like geometry and spring, damping rates. Further tuning can be done by bushings and sway bars. The whole thing is indeed a compromise because OEM's must sell their cars to the general public which is in a word very diverse and likely to sue if they get in too many crashes in a vehicle. That is why inevitably cars are designed with understeer (911 included)....meaning the front end is likely to slide out in cornering first.....to save people from their own poor driving.
Getting back to your question: You can eliminate the designed in compliance and understeer of the 911 by increasing the spring rates and bushing stiffness and putting in adjustable sway bars to fine tune everything. Leaning towards stiffer springs over bigger sway bars to increase roll resistance will in turn reduce suspension travel at the same time which limits changes in suspension geometry under cornering. Your goal ultimately is to keep the most amount of rubber in contact with the road surface to increase that wheels ability to corner, and as the car rolls in a turn the suspension moves, angles change and the contact patch usually gets smaller than when it was set-up initially at rest when your alignment was done. That is why initially at rest alot of negative camber is dialed in to make things approach zero camber under hard cornering.
It is very complex, to do it ultimately correct you will have to get the basic ball park set-up from popular opinion of others that have found what works, and then experiment on your own car. Then of course publish the results so everyone else can benefit from your knowledge. That is why this site is so cool.
Good luck.
Take baby steps when changing things and change one thing at a time to be able to minimize the variables. Change one thing see what it does versus change many things and then guess at what each particular thing did to the overall effect.

[djmcmath]

Bill, roger all that. The detailed explanation makes good sense to me, thanks a heap!

Will, thanks for the amplification. The empirical data is excellent. As to "suspension feel," I suppose more accurately, suspensions are significantly more complex to quantify than engines. I just placed the Amazon order for Fred Puhn's book, as well as Carroll Smith's 3-book "-to-win" sequence; I'll read up on those, and come back with better questions in a couple weeks.

As to my own personal upgrade path, I'm liable to end up with adjustable Konis over revalving my Bilstein's just because the extra curiosity value seems valuable to me. Additionally, having an adjustable sway bar seems like the same kind of a deal -- while I may not change it much, knowing what effect it would have on handling seems valuable, as I'm the freaky kind of nut-case who likes to try things out on his own. (smirk)

Thanks again, all. I'm going to take some time to do some research, as directed, then get back with ya'll when I'm less uninformed.


Dan

[Randy Webb]

Dan... the system is just a damped harmonic oscillator, so the basic eqns. will be in your soph. physics book (did you save it?). Race Car vEhicle Dynamics is good and pretty comprehensive (tho it ignores all work done outside the US).

There is a good article in Pano on the 996 suspension which will be at least peripherally useful. One interesting finding was that the bump stops were apparently intended as one "phase' of the suspension (near the limit). And, the tires, of course.

All this stuff has been worked out decades ago - but the above will give you that aethetic "from first principles" stuff.

[jluetjen]

Dan;
I won't claim to give you the answer (I rarely do that for anyone else either!). You raised all of the correct issues. If you want a set of answers, as you've noticed most people on this board will be happy to share with you theirs. If you want to understand the system yourself then you most likely need to hit the books first to get a portion of the 100 years of suspension technology and vehicle dynamics under your belt. A couple that I have found very good about explaining the system (of a 911, F1 car or anything else for that matter) are the following by Allan Staniforth:

Race and Rally Car Source Book
Competition Car Suspension

For race car dynamics common sense, all that you need to do is set yourself down at the feet of the (late) master: Carroll Smith:

Prepare to Win
Tune To Win (Specific to your questions)
Engineer to Win
Drive to Win

Now about "upgrading" your suspension -- what particular issues or deficiencies are you trying to fix? Stiffer isn't always better as I'm sure that you're aware, unless it has to do with the chassis (your point three) -- specifically the tub area.

[Randy Webb]

Originally posted by anthony
Wrong. A sway bar adds additional spring rate to the wheel just like the torsion bar. Both have a direct effect on understeer/oversteer.

I took "stiffness of the vehicle" to mean stiffness of the chassis -- so yes that is correct, esp. as it relates to torsional measures.


BTW, a sway bar (aka anti-roll bar) adds to spring rates ONLY when one wheel moves rel. to the other. They reduce the indepence of an independant suspension. If, for example, both wheels hit the same size bump at the same time, the anti-roll bars will have no effect whatsoever.

[Wil Ferch]

I find this sort of dialogue valuable and interesting....

Yet as an engineer....I failed in one important regard...

The original poster wants to upgrade, and admits to some need for better understanding and such....but all of us as repsonders should have asked...

....upgrade to what purpose? What is the "target" or goal or purpose? In retrospect, I think we may have been in a better position to offer advice if we had asked some more pointed questions in this regard. Dual purpose track/road car? Track only car? Canyon blaster? Able to deal with local roads ( smooth vs. bumpy)? Are there many steep inclines to driveway approaches?

Note that some tend to "improve" by maintaining a degree of suppleness ( like RUF)...others stiffen things up to limit suspension travel...and thereby keeping the wheels more upright more of the time, with improved tire contact patch ( at the expense of possibly bounding from peak-to-peak on bumps).

All compromises.....and all based on the "goal"...which we didn't ask.

--Wil

[dd74]

Wil has a point. Another might be what other modifications the poster has in mind, namely in the realm of lightening up his car. A lighter car does not necessarily need heavier torsion bars. I'm in this predicament right now, as I have two sets of rear torsion bars for my car - 28mm and 30mm. I've been advised, because of my car's weight, to go with the 28s and nothing larger, particularly when driving on uneven road surfaces.

[djmcmath]

Wil, you raise an excellent point -- and one that draws out the weakness in my original post. In all honesty, I don't know where I'm going, or what the car will be when I'm "done." As to right now, I know that the current suspension is too soft to allow me to set ride height at something normal, so I've got it at Paris-Dakar height to keep from rubbing. If I'm going to replace bits and pieces, there's no sense in doing it now, then replacing them again next year because they don't match correctly with the part that I installed this fall. That explains the original generic discussion nature of the post -- why pick part XYZ, how to match miscellaneous bits and pieces together, etc. -- like a general suspension discussion, rather than a specific data discussion.

Right now, the car is a Just For Fun car, something I take out on weekends to enjoy. I'd like to start tracking her in the spring, but I'll only have one season (spring/summer of next year) before I have to go back to sea again (Navy) and will run out of time to have fun for another few years. So between here and spring, it seems like it would make sense to replace the old parts with new ones, towards a specific goal -- now I'm just trying to figure out what that goal is.

(shrug) But the car certainly isn't a daily driver, so I have all the time in the world (well, figuratively) to go and read up on suspension so I can figure out what the heck I'm doing. I hate working on stuff that I don't understand. I especially hate just taking "tribal knowledge" or "This part is most expensive, so must be the best" or "This is the way we've always done it" reasoning to make a decision with.

Additionally, I know that some of you more senior types really enjoy the suspension engineering stuff, and could discuss it all day long. If I can persuade some of you smart people to "talk suspension" in a forum that I can overhear (ie right here), I can absorb some knowledge via osmosis, which makes me happy. Questions like "What difference does swing arm length, mounting position, etc." really make?" fire up the geek in me, despite the fact that I have no intention of Tyson Schmidt-esque suspension modifications.

So anyhow, the books are on order. Knowledge will eventually arrive via UPS. When I know more, I'll post more intelligent questions. In the meantime, I'll try to prod you fine gentlemen into posting your brilliance here. Thanks, all,

Dan