Let's talk about swaybars

[Dean]

This is an obvous attempt to suck our resident racer, Gary into a discussion on swaybars.

I have been researching these for my car and have some questions.

So far, here is what I think I know. As with brakes, I am probably wrong.

If you have the room, hollow gievs you more spring for less weight.

Having multiple arm lengths by way of moltiple holes for end links lets you change the relative stiffness by lenghtening and shortening the lever arm.

I am not sure what adjustable length end links do since they don't appear to change the lever arm length.

There are definitely a holy war on big springs vs. swaybars. I'm betting a well balanced system might be best, but th eopinions are all over the map.

If I understand the arguments. The swaybars are bad argument is that they lift the inside tire unncesarily reducing it's ability to assist in turning due to reduced contact patch/friction.

The swaybars are good camp says they reduce body roll without requiring increased spring rates thus improving tire contact all around.

I really like the whole Hotchkis front/rear/endlinks/upgraded brackets setup Gary is using. I was looking at the Cobb hollow ones until I saw it. I can't tell if all the peices Gary is using are really available to the public though according to thier web site.

[sperry]

The idea of a swaybar is that you can effectively increase spring rate only when the car's leaning. That way you maintain a soft, confortable suspension on the straights, while reducing roll in the turns. I thought it worked pretty well, however, it's not ideal for handling because as Dean said, the reduce the independent nature of our suspensions.

I went with swabars early on because they allowed me to tune out some of the understeer for cheap. A heavier rear bar will reduce rear-end traction, allowing the car to rotate.

I started with a 20-24mm adjustable bar in the rear, and ran it at 22mm. That pretty much made the car neutral on the stock suspension. In an attempt to reduce overall body roll, yet maintain the neutral handling, I got a 20-22mm adjustable front bar, and set it to 22mm. I set the rear bar to 24mm to keep the +2mm rear bias. The car was a bit stiffer in the corners, and remained fairly neutral. However, I had a bunch of fitment issues with my front bar, forcing me to remove it. Once the stock 20mm bar was back up front, I left the rear at 24mm just for kicks. Turns out the car's slight oversteer wasn't too bad, and allowed me to turn better during AutoX at the expense of being a little harder to drive.

I've since installed an updated 20-22mm front bar that's had the fitment issues resolved, and I'm back to the 22f/24r setup, but now I've got the JIC's installed. I'm still not totally sure where the car's handling is at now. I think I'm going to leave the swaybars alone where they are and concentrate on tuning the handling with the coilovers. Perhaps I should go to 21f/22r to leave myself more options once I get the struts where I want 'em.

AFAIK, I'm currently running Gary's old setup: Whiteline 20-22mm front swaybar, Whiteline 20-24mm rear swaybar, KartBoy soild rear endlinks.

BTW: I had a problem w/ the KartBoy links when I was still at stock ride-height. The endlinks are a little shorter than stock, which meant that when the car was up on the jack, the rear swaybar would contact the rear latteral links. I was told the endlinks are shorter by design to keep the bar from hitting the fuel inlet pipe. I was confused until I lowered the car. Now the links are perfect.

Which reminds me: I think the adjustable links can be sued to "preload" a swaybar, and/or adjust the bar for clearance purposes. Potentially you could use them for adjusting a bar's effective stiffness, but I'd much rather have the more descrete steps of several holes since it's easier to know what you're getting.

I'm interested to see what Gary has to say, since I'm sure there's plenty for me to learn about suspension setup.

[AtomicLabMonkey]

Originally Posted by Dean

If you have the room, hollow gievs you more spring for less weight.

Yes, basically use tubular bars wherever possible to achieve the stiffness you're after; you'll end up with less weight than if you used a smaller diameter solid bar to get the same stiffness.

Originally Posted by Dean

Having multiple arm lengths by way of moltiple holes for end links lets you change the relative stiffness by lenghtening and shortening the lever arm.

Yup.

Originally Posted by Dean

I am not sure what adjustable length end links do since they don't appear to change the lever arm length.

The adjustable length lets you ensure there is no preload on the bar, and also lets you adjust the height of the bar/endlink connection point so that the bar's lever arm is level. If the lever arm is not level, it basically affects the motion ratio and you lose some of the bar's effectiveness (wheel rate), since the connection points will be translating fore/aft in addition to vertical movement as the body rolls.

Originally Posted by Dean

There are definitely a holy war on big springs vs. swaybars. I'm betting a well balanced system might be best, but th eopinions are all over the map.

If I understand the arguments. The swaybars are bad argument is that they lift the inside tire unncesarily reducing it's ability to assist in turning due to reduced contact patch/friction.

The swaybars are good camp says they reduce body roll without requiring increased spring rates thus improving tire contact all around.

Reading Carroll Smith's books helps for a good overall understanding of things. Swaybars do have some negative effects, such as reducing the independance of the suspension; but with heavy production cars such as ours, they are almost always necessary to keep body roll within acceptable limits. Basically they do more good than harm. The entire purpose of the swaybar is to limit body roll, since it's a simple torsional spring that only has any resistance in roll. You could easily run a car with no swaybars at all and limit body roll with springs alone, but to do so will typically require springs that are so stiff as to reduce tire compliance with the imperfect surface of the track to an unacceptable level.

The way I'm thinking about things right now is that it will probably be fastest to run the stiffest springs possible (without reducing tire compliance - that's the key here), and then use bars as necessary to reduce body roll to your acceptable level. The tire compliance bit obviously varies from track to track, which is why race teams spend endless hours practicing and tuning and optimizing at as many tracks as they can.

At the end of all that you might ask the fundamental question: why is body roll bad in the first place? It doesn't cause weight transfer, contrary to a lot of popular beliefs; it's a result of lateral centripetal acceleration towards the center of the turn (Overall lateral weight transfer from the inside tires to the outside tires is geometrically fixed by the car's layout; it's C.G. height, it's track width, it's weight, and the amount of lateral acceleration present.). Body roll is primarily bad because it 1) affects tire camber since the tire's camber angle is linked to the roll angle of the chassis through the suspension, and the tire camber angle is critical to how much lateral force it can develop; and 2) because roll is a transient; the body takes a finite amount of time to roll, during which the driver is sort of waiting for things to happen. The less roll angle we get, the shorter amount of time it's going to take. There are also some other aspects of the process that can be discussed at length and in detail, such as the different components of weight transfer from sprung mass vs. unsprung mass, and through the roll centers, etc. etc... but I don't think it's critical to this particular discussion.

[Kevin M]

Before this discussion gets too far ahead of itself, like parts of the brake discussion, let's keep something in mind for all responses- almost all of us here drive our cars first, and race them second.

That factor is the point of view I look at sway bars from- Austin mentioned using sway bars to run the highest spring rates possible without affecting tire compliance. I say, sway bars are for using the highest spring rates we're willing to put up withon a daily basis, and then reducing as much body roll as we can without reverting to twin live axle suspension. My spring rates are something like 225# front, 185# rear, with 19mm front and 20mm rear sways (that would be similar to 20mm front and 21mm rear in a heavier WRX chassis). I'm also switching to a Whiteline 20-24mm RSB before race season starts, but I'll start with it at 22. Might go to 24 if I get brave. I couldn't ask for a better stup for real-world corner carving. The ride is firm, but never harsh or jarring, and the faster you go in it, the more solid and connected to the road the car feels. No surprise considering it was developed by Prodrive, then spec'd to some of the best manufacturers in the biz, Eibach and Bilstein.

[AtomicLabMonkey]

Originally Posted by BAN SUVS

Before this discussion gets too far ahead of itself, like parts of the brake discussion, let's keep something in mind for all responses- almost all of us here drive our cars first, and race them second.

I should have put that disclaimer on my post; I was talking strictly about dedicated racecars. If you're talking about a daily driver, well, you'll probably be very dissatisfied with the ride quality from stiff spring & shock rates well before you actually run into tire compliance problems, which leads you to look into stiffer swaybars as a compromise.

[sperry]

One other thing about body roll that I learned at Atwater:

Roll has its own momentum (for lack of being able to describe it better). Say I'm in a fast lefthand sweeper that has a quick transition to a fast righthand sweeper. With lots of roll there's not only the time it takes for the car to transition, but the transition also seems to require more traction. Tossing the weight of the car from leaning hard right, to hard left means that when it gets to the left, the tires will have to resist sliding even more because the car has more rotational enertia than if the transition was quicker.

...someone back me up on this.

[AtomicLabMonkey]

With soft springs and bars that allow a good amount of roll, the roll inertia of the body can have a noticeable effect as you swing it hard from a turn in one direction to the other. This is a reason that SUV's typically have rollover problems; if you're going straight and crank it into a hard turn in one direction, the SUV will probably be fine... but if you then immediately crank the wheel back over for a hard turn in the other direction, the sprung mass builds up so much inertia as it's going through that large roll angle that it overcomes the fully compressed/twisted resistance of the springs and bars and just carries itself on over until the C.G. is outside the track width and it overturns. I don't know if that's a completely technically correct explanation, but it should be close.

[Kevin M]

Originally Posted by AtomicLabMonkey

With soft springs and bars that allow a good amount of roll, the roll inertia of the body can have a noticeable effect as you swing it hard from a turn in one direction to the other. This is a reason that SUV's typically have rollover problems; if you're going straight and crank it into a hard turn in one direction, the SUV will probably be fine... but if you then immediately crank the wheel back over for a hard turn in the other direction, the sprung mass builds up so much inertia as it's going through that large roll angle that it overcomes the fully compressed/twisted resistance of the springs and bars and just carries itself on over until the C.G. is outside the track width and it overturns. I don't know if that's a completely technically correct explanation, but it should be close.

Here's a non-technical one: SUVs suck.

[JoelK]

Originally Posted by BAN SUVS

Here's a non-technical one: SUVs suck.

LOL

[Dean]

Originally Posted by AtomicLabMonkey

With soft springs and bars that allow a good amount of roll, the roll inertia of the body can have a noticeable effect as you swing it hard from a turn in one direction to the other. This is a reason that SUV's typically have rollover problems; if you're going straight and crank it into a hard turn in one direction, the SUV will probably be fine... but if you then immediately crank the wheel back over for a hard turn in the other direction, the sprung mass builds up so much inertia as it's going through that large roll angle that it overcomes the fully compressed/twisted resistance of the springs and bars and just carries itself on over until the C.G. is outside the track width and it overturns. I don't know if that's a completely technically correct explanation, but it should be close.

For the most part, I don't beleive the CG moves much until you run out of suspension travel, or one set of wheels lifts off the ground, and it is the roll that cause one of those two events.

But what is roll? I beleive roll is the CG applying it's force/momentum through the leverage of the susension all the way to the tire contact causing the body mass to rotate. Now you have rotating mass, in addition to the momentum of the CG when you run out of suspension travel.

Does that sound right?

I'm not sure a stiffer suspension would be much better because the bad driver input that caused the roll would instead probably cause a spin, that same driver would overcorrect and as the car spun in the opposite direction, undoutedly something like a Kia, Daewoo, or curb would act as a pivot point at it would roll over anyway.

[AtomicLabMonkey]

Originally Posted by Dean

But what is roll? I beleive roll is the CG applying it's force/momentum through the leverage of the susension all the way to the tire contact causing the body mass to rotate. Now you have rotating mass, in addition to the momentum of the CG when you run out of suspension travel.

Your definition was pretty close. The sprung mass (body/chassis) rolls simply because the instantaneous axis that it rotates about (the vehicle roll axis, defined by drawing a line between the front and rear roll centers - all of which is defined by the suspension geometry) is typically below the C.G. of the sprung mass. Imagine an upright pendulum with a pivot at the bottom; if you grab that bottom pivot and accelerate it to the side, the pendulum is going to lag behind and lean over to the opposite side. When the vehicle is in a turn, the "centrifugal" force (due to the lateral acceleration) acting on the C.G. of the sprung mass causes a torque about the roll axis, which is typically called the vehicle's Roll Couple.

The amount of body roll present for any amount of lateral acceleration is dictated by the value of the roll couple, and the roll stiffness of the springs and swaybars which resists the couple.

This might lead someone to wonder why we don't just design the suspension such that the roll axis passes right through the C.G. of the sprung mass? That way there would be zero roll. Well, since the C.G. is typically fairly high, probably somewhere between the engine block midpoint and the midpoint of the chassis, there are some unpleasant side effects that result from having high suspension roll centers (and therefore a high roll axis). Jacking is the most noticeable; the higher the roll center of an independant suspension is, the more lifting force is created on the chassis during cornering; that's why many racecars have suspension roll centers that are pretty close to the ground.

[Dean]

Originally Posted by AtomicLabMonkey

The sprung mass (body/chassis) rolls simply because the instantaneous axis that it rotates about (the vehicle roll axis, defined by drawing a line between the front and rear roll centers - all of which is defined by the suspension geometry) is typically below the C.G. of the sprung mass.

OK, so it rolls around the roll center as defined by the suspension geometry, but isn't the roll generated by the leverage beteen the CG and the contact patches working through the suspension? Or is that what you said?

[AtomicLabMonkey]

Originally Posted by Dean

OK, so it rolls around the roll center as defined by the suspension geometry, but isn't the roll generated by the leverage beteen the CG and the contact patches working through the suspension? Or is that what you said?

That's close to what I said; the chassis (sprung mass) rolls during cornering because lateral acceleration generates a force acting on the CG of the chassis, and the "pivot point" of the chassis is is below the CG. The chassis is not actually rolling about the contact patches of the tires, it's rolling about the suspension-defined roll axis (the "pivot point"), which is typically somewhere above ground and below the CG height.

[MattR]

I have mixed emotions about swaybars and their use in tuning.

I've always felt that swaybars alone are a bad idea, but when coupled with a well sorted suspension system, they provide an additional level of tunability and performance.

Case in point...my STI.
Correct me if I'm wrong/
AWD cars will inherently plow...my STI has a good balance, and when the dccd is turned back, it becomes almost tail happy. However, with this comes a lot of body roll. I would love to get rid of some of the roll, but that is what makes the car pivot IMO. If I tighten up the front only, with a sway bar, won't that make the car turn in quickly, then plow like hell? ( I can only change a front bar in A Stock). Is there anything I can do with only a front sway to improve handling? I'd say no, Help.

[AtomicLabMonkey]

Originally Posted by MattR

AWD cars will inherently plow...my STI has a good balance, and when the dccd is turned back, it becomes almost tail happy. However, with this comes a lot of body roll. I would love to get rid of some of the roll, but that is what makes the car pivot IMO. If I tighten up the front only, with a sway bar, won't that make the car turn in quickly, then plow like hell? ( I can only change a front bar in A Stock). Is there anything I can do with only a front sway to improve handling? I'd say no, Help.

It's hard to predict what will happen by changing the front bar only on a factory car; it's going to have different and sometimes conflicting effects. During cornering it's going to increase lateral load transfer at that pair of wheels, which will decrease the overall lateral force they'll produce, which would tend to produce understeer... but the increased roll resistance will probably mean less body roll in the corner which could give you a better dynamic camber angle for the outside tire - which means it would probably produce more lateral force. It will also affect the transients like turn-in response, since things will happen quicker with reduced body roll.

So, it's kind of hard to predict what would happen short of either running through some involved weight transfer & roll stiffness calculations for your car or finding some tribal knowledge from someone who's tried it with another STi and seen what happens. Sorry I can't help more than that.

[sperry]

Originally Posted by AtomicLabMonkey

It's hard to predict what will happen by changing the front bar only on a factory car; it's going to have different and sometimes conflicting effects. During cornering it's going to increase lateral load transfer at that pair of wheels, which will decrease the overall lateral force they'll produce, which would tend to produce understeer... but the increased roll resistance will probably mean less body roll in the corner which could give you a better dynamic camber angle for the outside tire - which means it would probably produce more lateral force. It will also affect the transients like turn-in response, since things will happen quicker with reduced body roll.

So, it's kind of hard to predict what would happen short of either running through some involved weight transfer & roll stiffness calculations for your car or finding some tribal knowledge from someone who's tried it with another STi and seen what happens. Sorry I can't help more than that.

Here's my guess:

You'll get the car to turn in a little better, and potentially have more front end grip from the better camber, but unless you can use that DCCD to hang the back end out, she won't turn.

We could always give it a shot at the Test n Tune... I've got that spare 20-22mm front bar... it might now fit any better than it did on my car, but it'll at least tell you if a stiffer bar helps or hurts. We should swap it on at lunch or somthing and see what the effect is.

[AtomicLabMonkey]

Originally Posted by sperry

We could always give it a shot at the Test n Tune... I've got that spare 20-22mm front bar... it might now fit any better than it did on my car, but it'll at least tell you if a stiffer bar helps or hurts. We should swap it on at lunch or somthing and see what the effect is.

That gets my vote; if you've got a chance to test out a different chassis setup for free, by all means do it.

[MattR]

Originally Posted by AtomicLabMonkey

That gets my vote; if you've got a chance to test out a different chassis setup for free, by all means do it.

Yeah, I think that will be the plan. I can always put it back to stock setting at the track if need-be.

Sweet.

[AtomicLabMonkey]

Well, this discussion wasn't anywhere near as spirited as the brake tech thread was...

[sperry]

Originally Posted by AtomicLabMonkey

Well, this discussion wasn't anywhere near as spirited as the brake tech thread was...

Brakes are complicated and mysterious.... swaybars are a friggen metal bar.

[Kevin M]

Originally Posted by sperry

Brakes are complicated and mysterious.... swaybars are a friggen metal bar.

I think it's because after the brake thread, nobody wants to stick their neck out first on a technical discussion, on pain of having it removed.

[AtomicLabMonkey]

Originally Posted by BAN SUVS

I think it's because after the brake thread, nobody wants to stick their neck out first on a technical discussion, on pain of having it removed.

Wusses. I stuck mine out there.

[AtomicLabMonkey]

I actually have a weight transfer and roll stiffness calculation spreadsheet I've been working on for the past few weeks, I could post it if anyone thinks they'd be interested in using something like that.

[GarySheehan]

(sniff...sniff) Did someone say "spirited?!"

Just kidding.

AtomicLabMonkey,

I thought you were right on in the brake thread and I'll say it again here. Right on!

With the WRX, and I'd assume with any AWD strut car, you've got to find a good balance between reducing chassis roll while maintaining tire compliance.

At one point we were running 1,100 lb/in rear springs to get the car to rotate. It did, but it was so stiff that it pretty much made the tire the only moving part of our suspension.

Softening springs will allow so much body roll that you get significant dynamic camber gain. You can see this in almost any photo of a WRX in a corner on an autox track. So you need to control the roll with stiff swaybars.

The setup that has worked the best for us is 750lb/in front springs, 650lb/in rear springs, 26mm x 3mm tubular front swaybar and 26mm x 6mm rear tubular swaybar. And lots of negative camber.

Gary
Sheehan Motor Racing
www.teamSMR.com

[AtomicLabMonkey]

Thanks Gary!