[altaic]

I've read about a few shops' suspension modifications on their race cars, as well as some others' who seem to have a method to their madness. For good reason, most attention seems to be given to the new rear setup. This is sort of stream of thought, so I apologize in advance if it's hard to follow, or if there are errors. This is meant to stimulate thought on how to get the most out of the suspension on our cars.

So, first off, rear camber: the rear camber changes a lot as the suspension (de)compresses. Stock rear camber is negative, so with neutral load the tire patch is inboard (not great, because the tire patch and thus grip is reduced). Then, when accelerating, the rear is loaded more, shifting the tire patch even more inboard (bad). However, the rear is unloaded when braking, reducing the amount of negative camber, shifting the tire patch toward the center (good).

Now in a turn, the outside tire is loaded more, increasing the negative camber and shifting the tire patch even more inboard. However, body roll counteracts it... For a given degree of body roll, you need an equal degree of camber on the outside tire, and the opposite degree on the inside tire. Careful attention must be paid to suspension height WRT the dynamic camber curve. Similarly, careful attention must be paid to anything that relocates the rear control arm mount (such as cam-type bushings).

I note that my car, stock, seems to do better going into a turn hot than leaving a turn hot. Accelerating out, I lose grip in the rear (oversteer), as well as grip in the front because the front is unloaded and has power going to it (understeer). I realize the situation with the rear bump stops increasing the spring rate and unsettling of rear suspension when hitting a bump mid turn, but that's a different phenomenon. This stock setup could be worse-- at least it's somewhat balanced.

Another thought: Body roll is a Good Thing. It disconnects the bulk mass of the car from the tires' contact patch, giving you lateral movement for "free." I think of body roll in terms of riding a motorcycle in a slalom, where separating your body from the bike and letting the bike roll side to side underneath you is way easier/faster than rigidly connecting yourself to the bike. It's analogous to the idea of sprung vs. unsprung weight applied to turning.

Anyway, we're sitting here thinking about how to maximize rear grip when so many people ***** about understeer on these cars. Planting the rear better will accentuate understeer. Linking the two sides of the rear suspension reduces grip on the rear because the front of the car is rolling and causes the rear inside tire to lift. Picture a completely linked rear suspension and a completely free front suspension... On an autocross circuit, the car will roll more than the rear will allow, so it'll lift the rear inside tire. Or rather, the outside rear suspension pulls up the inside rear suspension. This is an okay way of tuning the front/rear lateral grip ratio since it is independent of the acceleration/deceleration longitudinal grip changes due to dynamic camber discussed above.

An aside: It seems to me (IOW, I could be mistaken) that sway bars should be used in the minimal size to set up this ratio, and then one should move on to choosing spring and damper rates. IOW, I only run a rear bar to bias toward front grip, or only run a front bar to bias toward rear grip, but not both. If I want less "slop" body roll (only really the case if I'm hitting the bump stops on hard turns), I get stiffer springs and more aggressive dampers. I also mess with damping a bit on the bits linked by a sway bar, since that's essentially a spring, and spring rates add up. Alternatively, I'm currently mulling over the idea of ditching the rear sway bar and running stiffer rear springs and more aggressive rear dampers. Throw helper springs in the mix and I might have better rear bump grip in turns. Hmm..

So anyway, to set up the acceleration/deceleration rear grip, we want to mess with the rear camber. TiC, by way of Whiteline's rear camber bushings, decreased the amount of negative camber in the rear. Actually, IIRC, they went with some positive camber. This works because the tire patch is closer to the center with neutral and positive load, which means more grip when accelerating. But, there's a downside-- too much positive camber when braking, the tire patch gets shifted outboard. So, it's essentially trading braking grip/stability for acceleration grip/stability. I think that's a fair compromise, because the rear is more important for acceleration and the front for deceleration. Plus accelerating out of a corner is fun and feels fast. I'm not so sure that it's actually faster, though.

Finally, the Whiteline anti-lift bushings for the front control arm add more front braking dive and acceleration lift. I'm uncertain, however it seems like this would also add more rear braking lift and acceleration dive, compounding the aforementioned undesirable acceleration/deceleration rear dynamic camber changes. OTOH, it seems like it increases front grip, which would mean I could run a smaller rear sway bar, and have more grip all around. I think I'd have to run stiffer rear springs, though, and dampers to match.

Thoughts?

[Turninconcepts.com]

My thoughts - you've got a very good idea of what's going on. Keep going.

[caveguru]

Comments in red:

So, first off, rear camber: the rear camber changes a lot as the suspension (de)compresses -negative rear camber increases (becomes more negative) as rear suspension compresses, opposite for decompressing. Stock rear camber is negative, so with neutral load the tire patch is inboard (not great, because the tire patch and thus grip is reduced). -one reason we run negative camber on cars is because the tire sidewall tends to roll over under cornering, reducing the effective negative camber. Also, there is a some flex in the suspension pivots and the frame itself, also causing a loss of negative camber under cornering. Some of the initial negative camber is eaten up by these effects, which is part of why using more negative camber typically increases cornering traction rather than reducing it, as you have alluded to. Then, when accelerating, the rear is loaded more, shifting the tire patch even more inboard (bad) -while this might sound bad, remember this is 4wd. unless you are drag racing or have gobs of power, its rare that you will have insufficient traction to accelerate at full throttle when going in a straight line. However, the rear is unloaded when braking, reducing the amount of negative camber, shifting the tire patch toward the center (good). -great for stopping, but unless you're turning, this is mostly beside the point.

Now in a turn, the outside tire is loaded more, increasing the negative camber and shifting the tire patch even more inboard. However, body roll counteracts it... For a given degree of body roll, you need an equal degree of camber on the outside tire, and the opposite degree on the inside tire. -body roll has nothing to do with camber other than to the extent that it represents a change in the suspension geometry. Body roll does not "counteract" compression of the rear suspension because it is just that- compression of the outside rear suspension and extension of the inside suspension. The shifting of weight to the outside of the car does increase traction on the outside tires though. Careful attention must be paid to suspension height WRT the dynamic camber curve. Similarly, careful attention must be paid to anything that relocates the rear control arm mount (such as cam-type bushings). -true

I note that my car, stock, seems to do better going into a turn hot than leaving a turn hot. Accelerating out, I lose grip in the rear (oversteer), as well as grip in the front because the front is unloaded and has power going to it (understeer). I realize the situation with the rear bump stops increasing the spring rate and unsettling of rear suspension when hitting a bump mid turn, but that's a different phenomenon. This stock setup could be worse-- at least it's somewhat balanced. -so accelerating out of corners, you lose grip at both the front and back of the car? I guess you're saying your car is neutral on power exit? At low speeds and tight corners, I would have to disagree- it is easy to power slide this car in green mode. At moderate speeds, yes, I suppose the car is fairly balanced, although I would say stock it quickly transitions to understeer as the speed of the corner increases.

Another thought: Body roll is a Good Thing. It disconnects the bulk mass of the car from the tires' contact patch, giving you lateral movement for "free." I think of body roll in terms of riding a motorcycle in a slalom, where separating your body from the bike and letting the bike roll side to side underneath you is way easier/faster than rigidly connecting yourself to the bike. It's analogous to the idea of sprung vs. unsprung weight applied to turning. -the mass of the car still needs to get around the corner. body roll just makes the mass travel at a different radius and apex at a different point in the turn. This could be an entire useful or harmful effect that I would not make generalizations about. I believe the reason you and your bike move opposite to each other is to keep the sum of the force vectors from the bikes mass and yours the same as the gravitational vector created by the sum of your speed, the radius of the corner and the direction to the center of the earth, while at the same time maximizing the bike's tire's contact patch with the road by minimizing bike roll-over. In other words, you lean in to keep from flipping over to the outside of the corner, but you don't lean the bike in because you want the broadest contact patch possible from the tire. Body roll on many cars is also a "bad thing" because the suspension does not gain camber or enough camber on compression to keep the tires flat to the road under hard cornering. This is the advantage of the new rear suspension and the reason why companies like BMW and Porsche can maintain excellent traction with relatively low spring rates by using more complicated multi-link suspension geometries.

Anyway, we're sitting here thinking about how to maximize rear grip when so many people ***** about understeer on these cars. Planting the rear better will accentuate understeer. Linking the two sides of the rear suspension reduces grip on the rear because the front of the car is rolling and causes the rear inside tire to lift. Picture a completely linked rear suspension and a completely free front suspension... On an autocross circuit, the car will roll more than the rear will allow, so it'll lift the rear inside tire. Or rather, the outside rear suspension pulls up the inside rear suspension. This is an okay way of tuning the front/rear lateral grip ratio since it is independent of the acceleration/deceleration longitudinal grip changes due to dynamic camber discussed above. -However, it must be kept in mind that it also reduces total grip, especially under uneven surfaces, which is why little or no swaybars are run in rally cars. Increased spring rates is still a better way to go for purpose-tuned racecars because it allows maximum grip for each individual tire. Swaybars could be thought of as more of a "shortcut" that allows lower spring rates to be maintained for the same roll resistance.

An aside: It seems to me (IOW, I could be mistaken) that sway bars should be used in the minimal size to set up this ratio, and then one should move on to choosing spring and damper rates. IOW, I only run a rear bar to bias toward front grip, or only run a front bar to bias toward rear grip, but not both. If I want less "slop" body roll (only really the case if I'm hitting the bump stops on hard turns), I get stiffer springs and more aggressive dampers. I also mess with damping a bit on the bits linked by a sway bar, since that's essentially a spring, and spring rates add up. Alternatively, I'm currently mulling over the idea of ditching the rear sway bar and running stiffer rear springs and more aggressive rear dampers. Throw helper springs in the mix and I might have better rear bump grip in turns. Hmm.. ...sounds good- try it out and see what happens

So anyway, to set up the acceleration/deceleration rear grip, we want to mess with the rear camber. TiC, by way of Whiteline's rear camber bushings, decreased the amount of negative camber in the rear. Actually, IIRC, they went with some positive camber. -no, there is too much negative for the bushings to take it all out. This works because the tire patch is closer to the center with neutral and positive load, which means more grip when accelerating. But, there's a downside-- too much positive camber when braking, the tire patch gets shifted outboard. So, it's essentially trading braking grip/stability for acceleration grip/stability. I think that's a fair compromise, because the rear is more important for acceleration and the front for deceleration. Plus accelerating out of a corner is fun and feels fast. I'm not so sure that it's actually faster, though.

Finally, the Whiteline anti-lift bushings for the front control arm add more front braking dive and acceleration lift. I'm uncertain, however it seems like this would also add more rear braking lift and acceleration dive, compounding the aforementioned undesirable acceleration/deceleration rear dynamic camber changes. OTOH, it seems like it increases front grip, which would mean I could run a smaller rear sway bar, and have more grip all around. I think I'd have to run stiffer rear springs, though, and dampers to match.
-seat of the pants it seems like the ALK has only increased front grip marginally on the '08, possibly because there is already so much caster in the new front suspension (~6 degress). Now the roll center kit (KCA313) is making a much bigger difference with my lowered suspension that I'm still exploring having just installed it last week.

there are much longer and more in depth discussions on subaru suspension geometry elsewhere here on the net that, while they may not specifically discuss the '08 rear, they do address many of the concepts touched on here. It may be helpful to reference from there here where they provide insight, although at the moment I'm too lazy to pull any up.

Frank

[altaic]

Thanks for the comments, guys. I think we're generally in agreement, caveguru. There are a couple points that I should clarify, though.

-so accelerating out of corners, you lose grip at both the front and back of the car? I guess you're saying your car is neutral on power exit? At low speeds and tight corners, I would have to disagree- it is easy to power slide this car in green mode. At moderate speeds, yes, I suppose the car is fairly balanced, although I would say stock it quickly transitions to understeer as the speed of the corner increases.

I apologize, should have said that those results were not quite stock-- IIRC, I was running roughly -2° camber in the front, and 265/45R18 tires. Ideally, it had pretty neutral handling. Realistically, it was sensitive to throttle adjustment and less predictable than I would like.

Regarding body roll and my motorcycle analogy, I wasn't really referring to leaning through a long corner (although it works there too). Rather, I'm talking about flicking the bike into lean without the person leaning, simply using his own natural multi-link suspension to cantilever. However, he can just as well use this to his advantage in a long turn-- he leans the bike very quickly, cantilevering, and may more slowly get into his tight, leaned position. I use this technique with my supermoto bike on sm/kart tracks and on public roads to avoid potholes, suicidal squirrels, and so forth (and occasional jackassery).

I don't think there's such a thing as too much body roll, so long as your suspension can control it. If the guy riding is too weak to slow the leaning of the bike in a controlled, damped fashion, he might low-side. If the car's springs aren't stiff enough to keep from slamming into the bump stops, or if they're under-damped, the car might spin. I guess my point is that body roll is not something that should be minimized, which some people do because they think it feels "tighter" and whatnot. Rather, it should be maximized within the abilities of the suspension.

Also, I want to encourage people to introduce thoughts and ideas not mentioned. Maybe a different way to control body roll? Maybe a different approach to reducing understeer and make handling more predictable? I saw one company rig up a steering linkage to strut tops to dynamically change the front camber. Seemed to work pretty damn well, too. Maybe someone should rig up a 4-wheel steering linkage

I wonder if I should rename this thread, "suspension: stock classification goes out the window"

[caveguru]

Originally Posted by altaic

I wonder if I should rename this thread, "suspension: stock classification goes out the window"

-since I'm far from stock and trying to forget what stock felt like asap, that works for me!

On body roll, I am concerned. Depending on the camber curve of the suspension design, the car could easily be losing camber or not gaining enough as the car rolls over, seriously hurting contact patch. Another disadvantage of body roll that I hinted at before is that it represents additional body movement relative to the trajectory of the car. In other words, you turn the wheel, things happen to the suspension, the car turns. Body roll adds a little bit to a lot of delay in the middle there as the car's weight is shifting. It also makes the car harder to control because the weight transfer from the roll can still be happening long after you turn the wheel. This requires further correction from the driver to account for. I suppose you could just be an incredible driver and continuously predict and account for this delay, but its still a disadvantage because you're working harder to incorporate this into all your control inputs. I can't think of any advantages of body roll, although I appreciate that some methods of reducing body roll come with their own disadvantages. Audi does have a shock damper equalization tube trick that connects the dampers diagonally across the car somehow. Supposedly it works really well to reduce body roll with minimal decrease in traction. Since it is basically some plumbing the size of brake lines it should be possible for someone like KW to come out with an aftermarket product that adds this functionality. Not really sure what the advantages/disadvantages of it are for racing applications though. Audi uses it on their higher-end sedans iirc.

Frank

[altaic]

Originally Posted by caveguru

On body roll, I am concerned. Depending on the camber curve of the suspension design, the car could easily be losing camber or not gaining enough as the car rolls over, seriously hurting contact patch. Another disadvantage of body roll that I hinted at before is that it represents additional body movement relative to the trajectory of the car. In other words, you turn the wheel, things happen to the suspension, the car turns. Body roll adds a little bit to a lot of delay in the middle there as the car's weight is shifting. It also makes the car harder to control because the weight transfer from the roll can still be happening long after you turn the wheel. This requires further correction from the driver to account for. I suppose you could just be an incredible driver and continuously predict and account for this delay, but its still a disadvantage because you're working harder to incorporate this into all your control inputs. I can't think of any advantages of body roll, although I appreciate that some methods of reducing body roll come with their own disadvantages. Audi does have a shock damper equalization tube trick that connects the dampers diagonally across the car somehow. Supposedly it works really well to reduce body roll with minimal decrease in traction. Since it is basically some plumbing the size of brake lines it should be possible for someone like KW to come out with an aftermarket product that adds this functionality. Not really sure what the advantages/disadvantages of it are for racing applications though. Audi uses it on their higher-end sedans iirc.

Frank

I see your point about incorrect camber changes, and delayed body reaction.

Ideally, the suspension geometry should be set up to adjust camber correctly with body roll. I think Subaru was trying to do that with the rear, but had to make some compromises along the way. It's difficult to properly deal with body roll as well as lift/dive when it comes to dynamic camber and toe changes.

Anyhow, delayed body reaction due to body roll does pose a problem for the driver, but shouldn't be too demanding considering he has to have a feel for suspension loading anyway. He has to set up his lines well before a turn, perhaps many turns before. I'm willing to chalk it up to personal preference, but I suspect that increased body roll, or more generally delayed body reaction, can be very good in the right hands.

[Clarque]

thoughts:

I usually have a pretty good handle/idea on my suspension, but I must say that after reading the above posts...I feel like the biggest noob in history and have decided to completely abandon the pursuit of knowledge in the technical area of suspension. Thus I will also be removing all suspension components from my car and replacing them with bundles and bundles of cut springs from factory civics that are welded together/stretched to act an 3mm sway bars.

jk...on the real though...I feel like I just got a pro lesson in subaru suspension geometry/the performance of.

Great stuff guys, keep it up.

-Ryan