SDIF

I have read and re read and re read all the post here that I can find and few books on suspension set up and roll centers to include...

Tune to Win Carol Smith
Engineer to Win Carol Smith
How to make you car handle Fred Prun
Chassis Engineering Herb Adams
Race Car Engineering Paul Valkenburgh
The racing and performance Tire Paul Haney


I still have questions. They are...

What is the effect of raising / lowering the front / rear roll center. I do understand that if the roll center was the same height as the cg then there should be no roll. AND the lower the roll center the more roll. And the higher the roll center the more pronounced the jacking effect...

But I still do not understand how this effects weight transfer, if at all, since roll is caused by weight transfer not the other way around as most folks believe.

I do understand why roll is not good, but need a better understanding of roll center placement.

I do understand how it is calculated although I have never done any measurements as I do not know what I would be hoping to discover.

I can also parrot what these books say but I dont understand it enough to use the knowledge to my benefit.

Can someone help me to the punch line?

(BR)G-Machine

Hmmm good question.

I am thinking that it doesnt affect weight tranfers at all, but lowering the ammount of body roll keeps the suspension in it's ideal position: uncompressed. No suspension is ideal, and the wheel is going to be cambered as it goes through it's travel (obviously the ideal is no camber measured against the road so that the entire contact patch is being used efficiently.) So Im thinking if you had a car that had 0 roll, then you wouldnt have to dial in as much camber in it's resting position. And weve all seen the civics or golfs with the 4th wheel up in air mid turn, obviouslythat isnt ideal, but if the tire has no weight on it then it might as well be up in the air. But I guess having all 4 wheels on the ground is better for breaking or accellerating.

Im just thinking out loud here... dunno if that helps any.

O and how do you calculate the roll center of a live axle suspension? I just assumed it was at the pinion of the axle, could be wrong.

Dewey316

On our cars, its where the panhard bar crosses the axle centerline.

Typicly a lower roll center will promote more 'bite'(you illuded to the jacking effect earlier, think equal oposite ), obviosly you can't go too low, other wise the roll bars, and springs needed to control body roll, get rediculous.

You also want to get the front/rear heights as close as you can, that gives you your roll plane. Since on a PHB rear, the PHB height determines your roll height, you are limited in how low you can get the rear height, that is going to be a limiting factor in where you want to try to set the front roll height, considering with much lowering on our cars, the front roll center goes below ground pretty quick, you can see one of the real drawbacks to our suspension.

That was quite a bit of rambling, I hope some if it made sense. I'll try to dig up a couple of links for you on the subject.

SDIF

Thanks for the input.

So how do we use roll center heights to minimize lateral weight transfter?

As for 3 wheels on the ground, that may be a good thing as with late models or on auto x on corner exist only provided that the two rear tires are on the ground. If a front tire is in the air, then that weight is transfered to the opisite rear tire.

I would believe that for a front tire to be in the air that the sway bar would be very stiff or would have a roll center close to the cg, because the car is not rolling. The sprung weight is lifting the unsprung weight due to transfer. (Jacking effect?)

This would be hard to do on our cars because the front roll center is likely to be below ground and below the cg.

I have read circle track articles where roll center height is adjusted to match the sidewall stiffness of the tire to add more bite.

But I still do not understand the cause and effect.

Please continue to post.

83 Crossfire TA

iTrader: (2)

This is about as good an explanation of this as I’ve seen… the mechanism is simlar to that of tuning handling characteristics with sway bars. When the suspension can roll easier it actually sticks better, but if both ends roll over easily and you can’t control suspension geometry as far as it rolls you’ve got a problem.

Actually, that’s a fairly important deal with the 3rd gens. Since there is little/no camber gain in the modified strut front suspension you really need to control roll to maintain a good suspension geometry.

Acutally, a lot of testing with front engine, RWD cars seems to always show that you want to get the rear roll center significantly higher then the front.

again, the best analogy that I can think of is to think of this like you do sway bar sizes. A lower rear roll center is like going with a thinner rear sway bar, which will result in a stickier rear and understeer. You can compensate for it by stiffening the rear suspension.

[/b][/quote]As for 3 wheels on the ground, that may be a good thing as with late models or on auto x on corner exist only provided that the two rear tires are on the ground. If a front tire is in the air, then that weight is transfered to the opisite rear tire.[/b][/quote]

traction from increased load on a tire does not go up as fast as having another tire on the ground. The only way that less then 4 tires on the ground is a good thing is if you’ve got enough power to lift the front tires accelerating and you don’t need to steer at the time.

it has to roll to lift the opposite tire. Usually if you get enough antisquat to keep the suspension from squatting the front does not lift as easily as it will with everything fairly loose and able to roll…

Dewey316

Define significantly higher (no argument, I am actualy interested in how much diffrence you would consider significant) I am going to assume this is the case, because of the diffrence in the camber curve of the live axle vs modified strut?

There is also a lot of data that shows lowering the rear roll center on our cars, coupled with the appropriate spring, sway-bar, and shock changes. Can produce much faster lap times.

Granted, on a typicaly road-race/auto-x setup for our cars, the front roll center is somewhere right at, or belew ground level, so lowering the rear is still higher, even if we were to lower the rear a very significant amount.

BMmonteSS

I've heard that about 6" difference front to rear works out nicely. Us G-body guys try to raise our front roll center since it's about a foot underground stock, and lower our rear roll center since it's about 18" off the ground. I'm not exactly sure how you calculate your roll center on a strut front end. Any one care to share?

Dewey316

it is the a-arm angle, drawn out to where it intersects the percidiculare from the upper strut mount location (perc. to the strut). You will do that on each side. That will give you two points. Draw a line from each of those points, back to the center of the contact patch of the front tire. where those two lines intersect, is the roll center.

BMmonteSS

Sorry for the questions, I'm a double a-arm guy.

Ok so you just draw a line perpindicular from the strut mount. So what happens if your a-arm angles down ward, wouldn't it never intersect?

Dewey316

The angles will always line up somwhere. Think about negative numbers, or so to say.

same as if you upper a-arm is pointed down, and the lower a-arm is pointed up.

83 Crossfire TA

iTrader: (2)

the numbers I’ve seen with corvette and camaro sized/proportioned cars has showed the fastest lap times with the front roll center in the 3-6” range and the rear 6-8” higher then the front. Since we’re talking about cars in with CGs in the <19” range, we’re talking about a 50% or greater change which I would consider significant.

OK, turnabout… fair play… “a lot?” I don’t know that I’ve seen any 3rd gen specific besides Dean posting “I lowered mine and it feels better.” I know that there is a bit out there, mostly from unbalanced engineering and their tinkering with rear roll center on 4th gens, but I would expect their front roll centers to be higher as well as their CG’s and the feel/act significantly different then 3rd gens.

Hopefully, soon I’ll be in a position to tinker with this, though I hate to say it but once I get the modified rear in the car it’s probably going to see more dragstrip time then autox/road course. If it wasn’t for the fact that it opens another whole can of worms, it would seem that the easy solution for the front would be to use extended stud LCA ball joints, but of course that opens a whole can of worms WRT to getting bump steer right again after that (one thing that is more right then not with our factory setups). What would be very slick is if someone still built properly engineered drop spindles for our cars.

Dewey316

Thanks Mark, that is what I was interested in. Did your data take into account the roll center migration of our modified strut style suspension vs a SLA style? (Obviously I am very curious about this, as other people experiance, and theory is about all I have, until I take the dive and make mine adjustable.)

Fair enough, Actual, UE stuff is on several very fast thirdgens. I did talk with them specificly about this, and diffrences with the roll centers and CG of 4th vs 3rd. He did in fact somewhat agree with you, in that a 3rdgens rear roll center would probably perform better slightly higher than they put them on 4thgens, but still lower than factory. (This was assuming a ride height change in front, which causes the front roll center to drop). You did call me on the user of a lot, and in fact probably got me on it, as lowering it was one of the things considered illegal for CMC, I also beleive it is illegal in AI. I know of a couple ORR guys who run a lowered PHB setup, and have claimed that their car was much better in transistion. I have never heard of one person who has lowered it, and moved it back up to the stock location. Some have moved it up slightly from the extremely low points that UE has used, but still 2"-3" lower than stock.

I would imagine that none of use would settle at the stock location, if we had adjustable setups, and could spend the time to tune it in. Especialy when the cars start to get into the 2"+ lower ride heights out front.

That would be nice. I wouldn't hold your breath though.

SDIF

I have lowered my rear roll center.

I have been very pleased with the results. Springs and bars replaced.

This is my belief...
The reason that it feels better is that the weight transfer is felt sooner with the increased body roll and gives me more feel. The springs and bars effectively stop the roll, but the weight transfer is felt and the car takes a set quicker and gives me more confidence and therefore I can drive it harder.

The above ^ may not be a fact, but is a belief of why I think the lowered rear roll center is an improvement.

I have read everyones posts and they are appreciated, still trying to find the correlation between roll cent height and weight transfer.

I read in one of the above posts about roll centers act as sway bars, BUT I do not agree. Sway bars have no effect on weight transfer. BUT adjust the weight transfer that is already there and distribute the force between the front and rear tires.

Yes lowering the roll center does effect body roll which is caused by weight transfer. Weight transfer is not caused by body roll; body roll is a symptom not the cause.

Lowering the cg or increasing track width will reduce total weight transfer.

I believe that the roll center location can also effect total weight transfer, BUT I am not smart enough to connect the dots.

Please keep posting.

askulte

The way I like to approach this is think of front and rear roll stiffness, as a complete package. Roll stiffness is adjusted by spring and swaybar rates, as well as roll center height. You want your tires to share the load as equally as possible (given equal tire width front and rear). If you crank up the roll stiffness up front, and leave the rear soft, you'll get understeer of course. As mentioned, our cars need a fairly stiff front roll stiffness to keep the camber correct on the front tires, so we're left matching the rear once the front is set up. If you can get the front and rear roll stiffness similar (i.e. sharing the load based on CG location), you'll get more ultimate traction, since one end or the other isn't working harder, and giving up a little traction.

The bulk of the weight transfer happens with respect to your cornering acceleration (g's) and CG location. Roll has a small effect in weight transfer, since the CG height is above the roll centers. What will change this for individual tires is the front/rear roll stiffness percentages. Imagine you set the car up with 2000 lb/in front springs, and 100 lb/in rear springs. The rear effectively does nothing to prevent the body from rolling, so both rear tires get loaded equally, and have their max traction available. Problem is the outside front has the entire weight transfer load on it, so it's getting extremely overworked, and the inside front is basically along for the ride. You've now got a 3 wheeler, where one tire has all the extra load from cornering. One other concept to remember is that tire friction isn't linear with load. As you add load, the friction slowly gets less.

An example, with completely made up numbers:

With zero roll stiffness:
1000 lbs * .9 friction = 900 lbs cornering force available on that one tire, or 1800 lbs for both.

With lots of roll stiffness:
1800 lbs * .7 friction = 1260 on the outer, and 200*1.0 = 200 inner, so a total of 1460. You lost 340 lbs cornering force just by having the outer tire do the bulk of the work.

If you can share the roll stiffness between the front and the rear as equally as will work on your setup, you'll maximize the grip available from the tires.

The missing part of engineering the total cornering force available is that we don't know the friction curves for the tires, and the mfg's definitely don't share that information, outside of their respective racing teams (think F1). We can just make a best guess based on experience and other's data.

DeanE

So pretty soon you can quote yourself with those exact words instead of me when you get some real world experience then.

And when you bring your car out to an autoX course, I would love to be there to school you llike I have so many others- You wouldn't even be close. Just keep shooting your mouth off at me while I keep actually doing things for real and posting proof to help others here with documented facts and pictures.

SDIF, the roll centers change differently with body lean based on the different setups of everyones cars. Suspension travel is the greatest cause of roll center changes in a cornering affect whether you are talking entering, steadystate, or exit they can be changing diffenerently per car and do affect everything. There is no "1 perfect solution" or setting for for everyones car, they are all different and need to be experimented with to get your particular settings optimum.

Based on A-arm angle, softness of suspension, bushing deflection, shock and spring movement control, chassis flex.... the front roll center can move up and out, up and in, down aand out, or down and in while sitting through corner lean thus changing the characteristics of the cars handling- again, every car is different from what was added aftermarket and how it was set.