More front spring, less front swaybar?

[scottzg]

I haven't thought this through that much, but it seems like this might be a good idea. The car would be less unbalanced over bumps, and steering response might improve a bit. To boot, it would pair nicely with all those impreza spring options available to cheapasses. Thoughts?

[Subtle]

I've run a few lowered cars with stiff springs, shocks, and bars.
Fine for the track, but I can't tolerate the discomfort any more.

Now I find standard spring rates and ride height comfortable and have added Whiteline bars which provide nice control on the left and rights.
The front is 20mm and the rear is set at 20 on an adjustible 20/22/24.

The twgn spring rates are 225 front and 180 rear and are firm without
a load.

The next step will be adjustible shocks to really firm up the suspension for the track. For that I can dial out some oversteer by stiffening the rear bar

[Legacy777]

Even with the whiteline springs, the front whiteline bar really adds to the handling of the car. It stays very flat through the turns....and I really haven't even pushed it yet.

[Imprezive]

Subtle, what springs do you have?

[Subtle]

Still the originals.

Once I figure out which shocks to install I'll use the DMS with the
standard ride height and spring rates.

[BAC5.2]

DMS what? I wouldn't use DMS anything....

ESPICALLY if you try to get them from LaChute Subaru.

[Subtle]

DMS springs--I can order them with the standard height and rate. Most
manufacturers seem only to offer ones that drop the ride height and are
very much stiffer.

Been there

[Imprezive]

I could understand stock ride height, but stock spring rates are just a little too mushy.

Subtle have you considered tender springs inserts?

[Subtle]

Nada to the last one.

The twgn springs are firm enough for me --by the time track time comes I'll have adjustable shocks on and will dial them up--with a fair amount of bar it will suit my needs on the track.

Now all I need to do is lighten the car by 500 pounds, which is impossible
when one has become accustomed to the comforts of air conditioning,
power windows, etc.--all the stuff I used to scorn.

[BAC5.2]

Scott - I have thought this through. I'll play devils advocate, because I know you hate it when I do.

What is the purpose of a sway bar?

It is to transfer load from one side to the other, right?

How does it do this? It acts as a beam to connect the strut assemblys. When one wheel compresses (the outside wheel), it attempts to transfer this motion to the inside wheel, essentially attempting to compress the inside strut at the same time. The thickness of the bar dictates how efficiently this is done. Thicker bar, flexes less, transfers more energy, and compresses the opposite strut with more "gusto". This all results in the reduction of body roll.

Right?

So less sway bar, means less compression of the opposite side, and increased body roll.

Countering this with stiffer springs, now runs into the problem of improper damping.

Add spring, you HAVE to add damping. Even AGX's have their limits, and handling cornering loads of significance can overcome the highest tech adjustable strut assembly.

So what you get, is an under-damped corner. What does under-damping a corner do? You turn your smooth suspension, into a pogo stick.

Now, say you go with coilovers so you can damp your super stiff springs. Too stiff a front spring rate, and you can expect the car to bounce and skip off of bumps. The valving for compression and rebound is so hairline, that the suspension would not be able to compensate small bumps very well. Slow the rebound too much, and you will "pack" the suspension over rapid bumps, to quick to rebound, and you'll bounce over rapid bumps.

Packing is unfortunate, and what happens, is the suspension is unable to rebound quickly enough before you hit the next bump, so you compress the strut further.

Plus, slow rebound puts a LOT of stress on a shock, espically with astronomical spring rates.

So you have this incredible load being handled by the shocks valving, and you still have to deal with body roll loads. It's hard on a suspension that is less than geometrically perfect (like our front suspension).

So, in order for a more compliant ride, you decrease spring rate, giving you less of a razors edge to balance rebound and compression on, and use a sway bar to help balance and reduce body roll. Let the strut and spring deal with bump absorbtion, and trade some of the loads of cornering through the sway bar to the other assembly.

For the rear, stiff springs with capable coilovers, and no rear swaybar can produce incredible results. It's the tits for sure.

[scottzg]

hah, good answer phil.

I think in your description you are thinking i want to remove the bugger entirely-- no thanks, that introduces a whole new set of handling issues besides what you covered. However, a little less bar with a little more spring shouldnt have any problems being damped, esp seeing how a swaybar is essentially an undamped spring anyway.

True about the body roll, but if the change in rate = the work the bar was doing, no change in body roll should take place.

Wouldnt less bar/more spring allow for better camber angle on the inside tire?

[BAC5.2]

Isn't a swaybar technically damped by the opposing damper? It's just a lever arm. The only way it is technically undamped, is that it flexes, and the rebound is not damped directly by the bar. It IS indirectly damped by the suspension though, right?

I don't think sway bars effect camber angles whatsoever. I don't see how they would. When turning, castor effects camber far more than static camber would. Invest the money you'd spend on a lesser bar and stiffer springs, in camber/castor plates, and you'd see far greater results methinks.

Evenly swapping bar for spring, would only benefit you by working the dampers less than if you had a bar. And even then, it would work the opposite damper less and the compressed damper more. Body roll wouldn't really change, but that's not the issue.

Remember, the biggest misconception about suspension is that body roll is ALWAYS bad. Sometimes, body roll is indifferent. It's neither bad nor good. That is all related to roll moment, and center of gravity. If the body rolls, but that doesn't change the roll moment (which means an engineer did his job very well), then body roll will actually be somewhat beneficial assuming alignment specs are aggressive enough. Transferring weight more on the outside tires, without reducing the efectiveness of the rest of the suspension. You'll allow the tires to bite more, and you can run more toe without worrying about the car being twitchy under load of cornering (lots of toe, without body roll can cause the inside tire to jerk the car off line). Pull weight off of the inside tire, and you don't need to worry about it pulling you off line.

Right?

[Subtle]

In the North West, Don Nimi has had a lot of serious track time in slalom as well as SCCA road racing in GT3 class.

Check out pdm-racing.com----go from history to Subaru and an essay on
what type of suspension is suitable for street with some track time in mind.

The specialization of his shop in the Vancouver region as well as his website is suspensions.

Its worth a review

[Legacy777]

The anti-sway bars are undamped. They will get some damping indirectly, but more then likely if you take the indirect damping and the "spring" reaction of the anti-sway bars, you're going to have an under damped system vs critical or over damped.

[BAC5.2]

But the only way the sway bar can flex, is if it is acting on a strut body, correct? So how is it totally undamped, if the "rebound" of the bar is followed by the rebound of a damper. Same goes for the "compression" of the bar, it is followed by the compression of a damper. So if nothing else, the sway bar effectively increases spring rate while cornering, but it's not totally undamped.

If you could flex the bar without compressing or extending a damper, THEN it is undamped. But the bar is indirectly damped.

Now, assuming you have some kind of properly functioning damper, it should be fully capable of handling this. Unless, of course, you run an outrageously big sway bar for the rest of the suspension, THEN you can underdamp the setup.

Sway bars must be matched to the other suspension.

[scottzg]

We all know what we're saying.

[Imprezive]

Body roll = *weight transfer = bad. Right? BAC5.2 would you mind explaining "roll moment"?

I mean weight transfer going around turns, not dive or squat when braking and accelerating.

[BAC5.2]

What? How does body roll = weight? Body roll is a function of suspension design and geometry. It CAN be engineered out, but comprimises must be made for NVH, drivability, durability, and saftey. Don't confuse skid-pad numbers with body roll. Lots of cars roll a LOT, and produce great skid pad numbers (numbers of .9g and up, I would call great compared to averages).

Roll moment is the distance from the center of gravity, and the roll center. The greater the number, the greater the body roll. The Roll moment is like a lever arm.

Sometimes, when lowering a vehicle, you actually increase the roll moment, and INCREASE body roll. That's why lots of lowering springs have significantly stiffer spring rates, to help counter the increase in body roll from increasing the roll moment.

Body roll is barely related to weight. Body roll = faults in design for X number of reasons. There are heavy ass fuckin cars out there that don't roll that much. The Dodge Viper is one of them. Weighs a good bit, and doesn't really roll that much. Same with the Corvette.

Couple good suspension design with good tire choice, excellent drivetrain setup, and proper alignment, and you can pull a G with a 3500 pound car like it was nothing.

Its when you work with designs intended for one purpose, and try to get them to do something else that things must be given up. In most cases, you increase Noise, Vibration, and Harshness, as well as reducing the overall ability of the car itself.

Too Stiff can be bad too, for certain road conditions. It's all relative, and 99% of the time, compromises must be made.

[Imprezive]

I see, roll moment is the roll couple. At least thats what I always thought it to be called.

Just to add on to what you said, the roll moment is usually increased because when you lower your car you lower the roll center along with the center of gravity. The drop of the roll center is a direct product of the instant center's position(which also is lowered when lowering a car). The instant center is the point at which the suspension links rotate, be it a place on the car or a non-existing place.

The problem is that the distance between the center of gravity and roll center increases, if it is lowered to much, like you said, the roll moment or "roll couple" increase. Acting like a lever on the body, forcing it to roll more unless addition steps are taken to cancell out the body roll.

[BAC5.2]

The Roll Center is the line that runs laterally through the center of the car, and is the point that the vehicle rolls "around". The difference between this and the center of gravity, is the roll moment. It's also called the Roll Couple, but Roll Moment sounds cooler to me.

When you lower a car (using lowering springs), you aren't actually changing any geometries, simply moving the control arm/lateral links/etc. to a different part of their arc. The instant center is in the same location on the car, but dropped proportional to the drop of the suspension (again, put it on paper. The outside of an arc doesn't move, but the center of the circle drops. The amount the center of the circle drops is proportional to how much you lower the car, and is fairly easy to calculate. The Roll center drops the same amount as the total lowering (drop the car 2", the roll center drops 2", whereas the instant centers may only drop 1/4 of an inch. I don't feel like doing the math).

What sucks about engineering a cars suspension is taking into account real world situations. When you compress the suspension, you are dropping the roll center and the center of gravity (increasing roll moment, and increasing leverage for roll). So this must be taken into account when designing the car. So as you compress the suspension, you are increasing the leverage to roll the body. Even compressing on one side increases leverage to roll the body. So when you turn, every little bit the body rolls, is even more leverage for the body to roll more. Kind of sucks. Sway bars can help, but they are a bandaid, and are not perfect.

The roll center is one line that the whole car rotates around (imagine a bar that runs from front to back, at exactly the center of "body roll"). The Instant center is 4 individual points that each corner rotates around, the pivot of which being the center of the "circle". Imagine the point of rotation for the front right wheel, if you were to allow the wheel to maintain it's arc through 360 degrees of rotation. That point is the instant center.

I don't see how the roll center is dependent on the instant center.

Body roll still isn't bad. Look at everything on a flat plane. When one corner of the car is compressed (like when going over a bump), this is body roll. You don't really think about it, but it is. Imagine the wheel remained on flat ground. The "bump" would be that corner compressing (and trying to lift the opposite corner). This is exactly the same thing that happens when the body rolls.

Am I completely or partially wrong? I just might be. If I am, I'd like some further explanation.

Far to many people throw a sway bar at a suspension and call it a day. Unfortunately, perfect is a pipe dream for street cars. The perfectly designed smooth track race car could likely get away with no sway bar.

Springs, dampers, swaybars, corner balance, suspension geometries, and a host of other items must be taken into account when setting up a suspension. Of course, for 99% of the driving public, who daily drive and don't compete super competitively, this is ALL overkill and far beyond anything that would be noticeable.

[Imprezive]

Roll center is dependent on the instant centers because you use the instant centers to find your roll center.

If you have a scaled down drawing of your front suspension (I.E. struts, tires, control arms...) to find the instant centers you draw a line through the control arms and line perpendicular to the strut shaft axis.

The point on each side of the car where the strut and control arms intersect are the instant centers. To find the roll center you draw a line from the center of the contact patch of the tire to the corresponding instant center of that corner. Where the lines coming from the tires intersect is the Roll Center.

When you lower the car it lifts the control arms making them more horizontally level, this makes all the lines intersect lower, lowering the instant centers and the roll center.

I quit here because we obviously all have a good knowledge of how this works. especially BAC5.2

[93forestpearl]

I'm going to make some sort of a spacer for the lower ball joint. I want my camber gain to be to MY liking when I lower it. Thats what is nice about the Bill Mitchell software I've got. I can play with all the points and watch all the effects like camber gain, caster gain, ackerman, bumpsteer, toesteer, wheel rate, roll centers, and many more than I want to list. It took me three months to figure out half of this program.

Roll center does't neccessarily move equally to the change in ride height. That is designed into the suspension. Where is goes needs to be controlled and compromised with other aspects to get your desired effect.

People don't realize how important roll center is. I laugh at all these people with riced out shitty hondas one the ground, not realizing they made their car handle much worse. The roll center will start to move at a faster(or slower) rate when you get to the ends of the suspension's travel. Thats the thing. The roll center moves, depending on your rideheight(travel) and your roll. It also affects wieght tanser. For my formula car at school, it should be inside of a tennis ball, because there is only two inches of travel.


Don't forget that the roll center for the front and rear will be different and react differently to changes. The rear should be higher than in the front.

In street cars it really sucks. You make huge compromises due to the nature of them. You have a huge yaw moment and thus lots of weight tranfer and you have to tune for it with lots of stiffness in either direction. Which way is dependant on application. On the track you want little travel so both tires can be flat on the ground. Slick conditions are the opposite. On our formula car, the bulk of the mass is in the middle of the car, so it turns much easier and there is exponentially less wieght transfer.

Lowering can also affect bumpsteer, but it can be tricky trying to move the rack or attachment points on the upright. Changes of 1/2 inch on these things can have a huge effect.


I don't think it was mentioned but swaybars not only affect weight tranfer side to side, but also front to rear. Thats how they actually affect a cars under/oversteer.

More weight tranfer helps in slick conditions like snow, ice and gravel. Stiff cars always suck ass in the snow.


There's really only four major aspects of suspension. Controlling your weight tranfer with springs, swaybars, and wheel travel. Controlling your contact patch with the ground through your geometry. Absorbing impacts to keep that contact patch in exhistance. And last but not least, controlling the steering through all such events.


All very complicated, intertwined, and a huge compromise in every direction.

We've started to scratch the surface on how down and dirty you can take suspension. Why do you think a set of remot-resevior MOTON coilovers will can run you well over $6k?

Oh yea. One little discrepancy is all the numbers and theory mean squat if your suspension components aren't rigid enough to maintain your designed geometry. Then you get into materials, manufacturing and weight then you are back at the beginning. Goddam it! I stop now.

[BAC5.2]

Roll center is dependent on the instant centers because you use the instant centers to find your roll center.

If you have a scaled down drawing of your front suspension (I.E. struts, tires, control arms...) to find the instant centers you draw a line through the control arms and line perpendicular to the strut shaft axis.

The point on each side of the car where the strut and control arms intersect are the instant centers. To find the roll center you draw a line from the center of the contact patch of the tire to the corresponding instant center of that corner. Where the lines coming from the tires intersect is the Roll Center.

When you lower the car it lifts the control arms making them more horizontally level, this makes all the lines intersect lower, lowering the instant centers and the roll center.

I quit here because we obviously all have a good knowledge of how this works. especially BAC5.2

Tire width effects roll center?

THAT, I had no idea about.

I always picture the instant center diagrams from bicycle suspension. I know more about bikes than I do about cars. But in bikes, there are 4 pivot points you need to be worried about (some as few as 1, some as many as 6 and beyond, and lots of different types of pivot point to get in the way).