Does Roll-Center height have much effect?

[911st]

Sorry, spell check got me. I do think that was very interesting info?

Not un-interesting.

[petevb]

Originally Posted by Flieger

However, a sufficiently stiff anti-sway bar can cover up bad roll center placement and achieve the desired lateral load transfer rate. If suspension/chassis/wheel/tire geometry necessitate a certain suspension design with bad roll center placement, anti-sway bars can help.

Yes, you can almost always use roll bars to balance out mid corner. Changing the roll center actually has a few other interesting attributes, however, which often make it preferable.

First, as the car rolls and roll center changes height you can move from oversteer to neutral (for example). So this is one of the few knobs you can turn that will allow you to adjust mid corner and corner entry balance separately.

Second, simply adding extra spring rate (in the form of swaybars or primary springs, doesn't matter) hurts overall grip all else being equal, as soft is sticky. So there are reasons why you'd prefer to get the same control of the body motion with softer springs. Which is of course why anti-squat, anti-dive, etc suspension is a good thing in the first place...

To say it gets complex is a mild understatement.

[911st]

Very interesting to me.

Can you give any indication how sensitive a lowered 911 front is to roll center?

For example, if I raise the spindles 35mm will that make a significant difference?

Is it like stiffening the front?

Is it as significant as change between say 22 or 23mm front torsion bars (apx 17% diff)?

Currently my front A arms are up sloping toward the wheels.

[jluetjen]

Originally Posted by 911st

Would not stiffer springs do the same thing as stiffer sways?

As John has in his sig line above:

"Any suspension -- no matter how poorly designed -- can be made to work reasonably well if you just stop it from moving." -- Colin Chapman

Yes -- and no. Stiffer springs will make the car handle worse over a bumpy track due to the sudden loading/unloading of the suspension against the bumps. A softer suspended car will be more forgiving and easier to handle. A car with stiff springs while cornering on a buming track will have all 4 wheels affected by bumps, but a softly sprung car with stiff sway bars will essentially have a softer spring rate on the inside wheels since the sway bar will subtract spring force on those wheels -- so they will be more compliant and hold the road better (all other things being optimal).

The other thing to keep in mind is that most folks who are racing or doing track events are using wider, lower profile tires. These tires are less forgiving of camber issues then let's say 70 series tires. So even a little bit of positive camber with wide, 50 series tires will most likely remove a substantial portion of the tire's contact patch from the pavement.

Keep in mind that Colin's musing in my signature was somewhat tongue-in-cheek.

[petevb]

Originally Posted by 911st

For example, if I raise the spindles 35mm will that make a significant difference?

If you have a lowered car and everything else right, yes you can feel it. Same parts on a car that's dialed in correctly will be much easier to drive.

Originally Posted by 911st

Currently my front A arms are up sloping toward the wheels.

Well well know not good. You just can't put numbers on this stuff, though- it's down to feel. A tuned in driver can feel the difference fairly easily. Unfortunately getting it tuned in in the first place is tricky.

Originally Posted by jluetjen

A softer suspended car will be more forgiving and easier to handle.

And just to emphasize, ultimately the softer car will be stickier. What it won't do is transition as quickly, run as low or cope with aerodynamic downforce as well.

Originally Posted by jluetjen

Keep in mind that Colin's musing in my signature was somewhat tongue-in-cheek.

And with the chassis flex of older cars nearly impossible to achieve in any case...

[Bill Verburg]

Roll center is the result of the geometry of the struts and lower A arm.

raising the spindle height raises the roll center and reduces the roll couple(again, 0 roll couple means 0 roll)

a separate issue that needs to simultaneously be adressed is bump steer, do use a bump steer kit

you also need to do the f/r suspension at the same time

[jluetjen]

Bill, you make an interesting point about zero roll couple equals zero roll. While reducing role may be a good thing -- why don't all race cars have zero roll couples???? Because doing this makes the car very "wooden" and it lacks feel. To put it differently, the driver will have a hard time feeling what the car is doing. So some roll is desirable -- how much depends on the sensitivity of the driver. Novices will most likely respond better to relatively larger roll couples while Michael Schumacher would quite likely be very fast with a very small roll couple.

[Bill Verburg]

Originally Posted by jluetjen

Bill, you make an interesting point about zero roll couple equals zero roll. While reducing role may be a good thing -- why don't all race cars have zero roll couples???? Because doing this makes the car very "wooden" and it lacks feel. To put it differently, the driver will have a hard time feeling what the car is doing. So some roll is desirable -- how much depends on the sensitivity of the driver. Novices will most likely respond better to relatively larger roll couples while Michael Schumacher would quite likely be very fast with a very small roll couple.

Gonna have to disagree, the roll center can never be raised to the height of the cg because of the geometry of the components, the best you can do is keep it as high as you can while lowering the cg as much as you can. This will minimize the length of the roll couple. I haven't seen 911 mapped out but on 964/993 the rollcenter goes under the pavement unless the RS suspension pieces are used.

steering feel comes primarily from caster and secondarily from toe, the more caster the more self centering the wheel will be, the less toe(to 0) the faster the response to steering inputs

a non assisted 911 wheel feels alive because there is little rubber, lots of caster, little toe and what bump steer exists is transmitted directly into the steering wheel. Lots of people dislike the later assisted steering because of on center numbness, the truth of the matter is that the steering is actually much faster but also has more rubber in the actuation path, reduce or remove the rubber and max caster/min toe brings back a lot of the liveliness, but not all because there is still the power steering hydraulics damping the system, the latest electronic boost steering is much better in this regard but can never be as good as the original 911 steering where there is very little isolating the tire from the steering.

the whole point of stiffer springs is to reduce suspension reaction times, softer springs to a point will generate more grip but it will also take longer to transition and be more disorienting to the driver.

[911st]

Originally Posted by Bill Verburg

raising the spindle height raises the roll center and reduces the roll couple(again, 0 roll couple means 0 roll)

you also need to do the f/r suspension at the same time

I understand the front of a SC/Carrera can be done by raising the spindle, what can we do with the rear?

Outside of not lowering the rear much (keeps the COG high), just do not see what options we have.

I can see that the inner rear arm point could be moved up or down some. Is there somthing there?

Thx.

[911st]

Is rake a way to get to less anti dive?

[jluetjen]

Originally Posted by Bill Verburg

Gonna have to disagree, the roll center can never be raised to the height of the cg because of the geometry of the components, the best you can do is keep it as high as you can while lowering the cg as much as you can. This will minimize the length of the roll couple. I haven't seen 911 mapped out but on 964/993 the rollcenter goes under the pavement unless the RS suspension pieces are used.

True in the case of the 911. I was speaking in general terms. I've got some suspension analysis software if anyone has the measurements. I'd be happy to make a few models. I could picture a car which has had the spindles raised as far as is possible, and then having the T-bars indexed so that A-arms are angled up on the inside having a very small roll-couple. In addition to the "wooden" feel, the there are two other problems with this:

1) Roll centers are not static. As was mentioned earlier, they can move around (in 2 dimensions) quite a bit depending on where the car is in its roll and bump travel. Having a roll couple that suddenly doubles or triples in size can cause an awfully weird handling car, especially if the roll-couple were to suddenly reverse so that the Cg was below the roll center!

2) Suspensions with high roll centers are prone to jacking. The classic example is the rear suspension of a early Triumph Spitfire

Basically, because the roll center is relatively high in relation to the contact patch, there is a tendency for the chassis to be jacked up and over the contact patch. Not a big deal on 50's sports cars with non-sticky 50 series tires, but this can be a very big deal with the high loads induced by today's super-sticky rubber. If the roll center goes over the Cg, the chassis will actually "jack down" and start to roll to the inside of the turn.

So the secret is that...
1) you want your roll couples to be consistent front-to-rear. In situations where the front roll-couple gets longer, you want the rear to get proportionally longer too. It's possible to accommodate changes in the roll couple relationship by using a stiffer roll-bar or springs, but these will then increase the risks of "unintended consequences". This is where the art of suspension design and set-up comes into play.
2) You want your roll centers to be high enough so as to not have terribly large roll couples, thus containing the roll to a reasonable amount (in conjunction with the rest of the suspension pieces), but not so high that it induces severe jacking forces.
3) You don't want your roll centers so high that the roll couples ever reverse (where the roll center is above the Cg).

[petevb]

Originally Posted by Bill Verburg

Gonna have to disagree, the roll center can never be raised to the height of the cg because of the geometry of the components, the best you can do is keep it as high as you can while lowering the cg as much as you can.

I'll disagree- or at the very least I think you need to add a "practically" to the above statement.
We know that it's possible (normal) for the roll center to be above the ground front and rear.
We also know that if we add ballast at the bottom of the car it will move the CG down.
Add enough ballast and the CG will move below the roll center.

While I'll stipulate that the above is slightly outrageous, I do think jluetjen makes a valid point that zero roll couple is possible.

[911st]

On a 911 is the front roll center easer to adjust or move up and down. Further dose it move up and down significantly with normal travel.

And, is the rears by design more fixed and less prone of changing as dramatically with lowering or travel?

If so, what dose that mean to a torsion bar 911?

[petevb]

Originally Posted by 911st

On a 911 is the front roll center easer to adjust or move up and down. Further dose it move up and down significantly with normal travel.

And, is the rears by design more fixed and less prone of changing as dramatically with lowering or travel?

How To Make Your Car Handle Page 37 and 38.

Lowering the front, making the A arm angle up towards the wheel, lowers the roll center. Raising the spindle height moves it back.

In the rear changing the pivot axis (such as by mounting the shorter turbo semi-trailing arms) raises the rear roll center.

Google is your friend.

[911st]

Thanks.

Did you add "(such as by mounting the shorter turbo semi-trailing arms)"?

Shortening dose change the pivot axis and helps the camber curve significantly (2.7 RS & RSR, made car nervous).

Moving the inner up changes the pivot axis. (RSR plus moving the inner up 10mm = 930, not nervous any more).

Moving both axis up changes the pivot axis. (935)


I am thinking moving the inner down, lowers the roll center and steepens the camber curve. Moving it up raises the roll center and flattens the camber curve. However, I do not know if this is true, nor now how significant the change in roll center.

And then there is anti-squat.

[Kevin Stewart]

considering most roll centers should start 2-3 inches off the ground i find it hard to believe you are getting the center of gravity below that, with people going soft springs heavy bars and controlling roll with rebound it makes the center of gravity more important, (the go cart feel), but many people are starting to soften the bar a little in favor of rebounds that have went off the chart. Kevin

[petevb]

Originally Posted by 911st

Did you add "(such as by mounting the shorter turbo semi-trailing arms)"?

?? Search for the links I suggested and you'll understand why this is true.

Originally Posted by 911st

Shortening dose change the pivot axis and helps the camber curve significantly (2.7 RS & RSR, made car nervous).

Moving the inner up changes the pivot axis. (RSR plus moving the inner up 10mm = 930, not nervous any more).

Moving both axis up changes the pivot axis. (935)

Generally correct. Remember with your comments about "nervous" that the tire construction 30 years ago was completely different than today. So different, in fact, that if it was ideal then it's guaranteed to be wrong now. And of course each solution matches best with a particular spring rate and ride height.

You're not going to get "right" answers for your particular situation here. If you want those you'll need to figure this stuff out, which will involve trial and error if you're serious. Take this info, better yet read some books and at least you'll figure out what the different knobs do. There is no "right answer". If you're happy being close that info has already been quoted.

Originally Posted by 911st

And then there is anti-squat.

That's the easy part. Move the trailing arm pickups up relative to the spindle and you increase anti-squat. How much you want again depends on power, ride height, etc.

[jluetjen]

Originally Posted by petevb

We know that it's possible (normal) for the roll center to be above the ground front and rear.

Actually that is not true. It's quite common for the front roll center to be at ground level or lower in many cars. I don't think that this is the case with the 911, but with cars with double A-arm suspensions it can be more common. There's nothing really magical about the roll center versus the contact patch as long as there isn't a huge difference in the height between the two.

[jluetjen]

Originally Posted by petevb

In the rear changing the pivot axis (such as by mounting the shorter turbo semi-trailing arms) raises the rear roll center.

I disagree (to a degree). With a semi-trailing arm suspension, the roll center is defined by the angle generated by both mounts. Raising the inner mount will raise the roll center, even if you don't change to the shorter Turbo trailing arms. The Turbo Trailing arms were designed to do this AND increase the camber gain curve by virtue of the shorter swing arm. But the two affects can and sometimes are separated by using an adjustable trailing arm mount as shown in this picture.



Depending on how they are installed, they can be used with or without Turbo trailing arms. In the photograph they were used with Turbo trailing arms.

[burgermeister]

I don't understand the various sentiments about front vs rear roll center heights affecting front vs rear roll stiffness. I think the two are separate issues.

The front & rear roll center heights affect the roll axis, which affects the overall roll couple (distance from the CG), and thus the overall amount of roll.

But, assuming the chassis is torsionally stiff compared to the suspension roll rate, the chassis will roll as a unit. When it rolls, the front & rear suspension roll stiffnesses determine the couple generated at the front and at the rear, which will affect the balance of the car.

Once the amount of roll has been established by the roll axis & CG, the roll center height no longer enters any of the equations, and therefore it ought not affect front/rear roll stiffness balance.

Roll centers significantly removed from the plane of the road generate lots of scrub (sideways motion of the wheel resulting from vertical motion of the suspension), which is probably bad for stability and certainly is bad for tire wear.

As has been stated, roll centers significantly removed from the plane of the road also can generate jacking loads (swing axle beetles were also famous for this problem).

From Bastow's suspension book, changing the front view angle of the semi trailing arm largely changes the ride steer (toe vs. vertical deflection) curve, with a small effect on the roll center height. Small changes in the angle cause big changes in the curve. All the various books stating why Porsche did so-and-so are just hypothesizing. There were probably 2 or 3 people that actually knew the reason, and everyone else is guessing.

Here's a swag at a ride steer curve for a +10mm change (up) in inner pickup point.
(green: inboard point is 20mm above outboard pickup)
(yellow: inboard point is 10mm above outboard pickup)


That said, I'll hypothesize some.... The RSR & 930 had wider rear tires. They are thus more sensitive to camber changes. The shorter banana arm (inner pickup moved rearward) yields more camber gain, which means 1) the rear tires end up with more favorable camber in a given corner, or 2) static alignment settings can tolerate less negative camber.