w116 Roll Center Correction (Page 1 of 2)

I've been on the same page, I'm on my 5th suspension design book...still learning to swim, but learning a lot along the way...

I had the exact same question while setting up the rear suspension on our MR swapped Golf...where exactly do I put the roll center?

The answer I finally found was that roll center mitigation is a very complicated topic, and moving it at all effects so many things that it's almost impossible to improve handling by simply moving the roll center.  Lowering the car might have raised the roll center higher above ground, but it also reduced the roll couple by moving it closer to the center of gravity.  Roll axis is the perceived effect of all the other components of the suspension, not a design point.

To make sure your roll center is reasonable, I'd verify that all lower control arms are level with the ground or slightly declined toward the wheel.  If they car is lowered enough so that they are inclined towards the wheel, the geometry is all jacked up and should be corrected before doing anything else. 

To control body roll, you must increase roll stiffness.  This is most easily done with a larger diameter (or multiple) swaybars.

Balljoints, tie rod ends, or other "at the wheel" corrections are probably primarily used to correct geometry after lowering, improve camber gain/loss, or fix bump steer issues.  If you're serious about this, I'd make yourself some simple alignment tools (toe gauge, camber gauge, bump steer gauge) and see where your suspension falls short...then you can start playing around with adjustments.

That thing is even tippier than the Sputnik Fury. I could dig up the contact info for the Dirty Little Freaks if you want.

Not that I can figure. The front suspension is quite different than anything else I've worked on. A-arms are so much easier to modify.

Sadly our car is 4.5 hours away but here's a google photo.

Here's the suspension without the spindle

Spindle and brakes in place

From underneath showing the camber eccentric. There might be some room for improvement from this area. I'd love to squeeze 2°+ of dynamic camber out of this thing.

From my diagrams it looks like a taller top balljoint would help roll center.

I think I could maybe move the camber eccentric out/up on the crossmember giving me more camber.

In the first picture is the bar running from the spindle area to the left the anti sway bar?  If so you need look at taking that out and finding one off a full size truck to replace it with, and probably a better way to attach it to the lower control arm.  You could also look at just adding a second anti sway bar to the existing setup pretty easily.  That would stiffen up the roll rate without affecting the actual spring rate to much, making it ride the same and corner better.  The one shown would be about the right size for a compact car nowadays.

In a turn most of the weight comes off your inside wheels, but there is still a couple of hundred pounds of weight on each one.  If you can get the inside wheels to compress with the outside wheels then the inside will stay lower and you corner flatter

Think of it like balancing on one leg (90% of your weight) and using your other leg just for balance (10% of your weight).  if you start to pick up the leg with less load you will start falling towards it to compensate the lack of support.  IE falling towards the inside of the turn.

You'll need to upgrade the front and the back to a similar degree otherwise you could end up just picking up a wheel up.  If both inside wheels get in the air you have a different problem.

Changing the roll center is great and all, but if you can fix most of it with a simple addition like a sway bar it will save having to reproduce the suspension every time you need a new part.

For the camber problem, relocating the lower body mount holes to move the mounting points out would be a step in the right direction, depending on how much fabrication you can do to the car where it is.

Just my 2 cents...

The back is difficult to adjust on these cars because I've got a semi-trailing-arm setup. Your only changes really are lowering springs which introduce camber and a bit of toe.

I like the concept of adding a second sway bar up front. I'll have to look into that. I'd still like to figure out a way to shorten the roll couple. I'm going to keep digging on this suspension type.

I've started reading up on the "Big Bar, Soft Spring" set up. Seems to work for big, rwd cars.

Very true. From my reading lowering the car lowers the CG some distance D but lowers the RC D*n where n>1 increasing the length of the moment arm.

However if I can raise the roll center by the same amount that I lower the car then I will end up with a shorter moment arm located closer to the ground. This should give less body roll and would stop overloading the outside tire.

Tires are very interesting. We're dealing with F=m*N.  We get more grip(F) by increasing the normal force(N) with weight transfer but as we load the tire we are reducing the coefficient of friction(m) from that tire. As we load the tire N increases quicker than m decreases but when we overload the tire the amount of grip quickly drops off. I'm not sure what the specific properties are of my Direzzas are and I have no way to find out how much normal force is put onto the outside tire while turning.

I'm not disagreeing with what you said, Nathan, and perhaps my statements did not come across clearly.  You are correct, load transfer causes body roll.  My point was that adding a sway bar won't reduce load transfer, it will only reduce body roll.  I think, correct me if I'm wrong.  So you have increased loading on the outside tire, but a more level attitude of the car body.

I better approach would be to reduce the roll couple, so that the amount of weight transfer is les, thus inducing less roll and requiring less bar.

I also agree- tire loading is a weird thing, not at all linear, and as Buzz pointed out, past a certain point grip drrops off very quickly as the tire gets overloaded.  I was merely postulating that the tire load ratings might be a good place to start with regard to calculating what amount of load transfer you want on those outside tires.  If the roll couple is too small, and not enough weight transfer happens, the car can skid off the track.  Obviously in this case he's a LONG way from worrying about this! 

Roll axis also affects oversteer/ understeer.  In our first Lemons car, the Volvo 122, we lowered it a bunch simply by cutting the springs.  This resulted in a car that was very over-steery.  We "corrected" that with a massive front bar, stiff front springs, no rear bar and pretty weak rear springs.  At the time, we didn't really know what we were doing, we just did stuff till the car handled OK.  In retrospect, I would have preferred to go after the front roll center, and that might have "fixed" the handling issue without needing such extreme bar and spring setups.  We also wore out tires much faster than we should have.  Luckily, it was a light car.  On a heavy beast like our current LTD or the Benz, you don't have this safety factor.

I'm reaching the limits of my understanding but I think I still have a little to offer, lol. 

Mostly correct.  Anti-roll bars also affect the RATE of weight transfer, especially between axles...this is why the ratio between front and rear bars can be used to adjust over/understeer characteristics. 

Total Weight Transfer is the sum of three components :

Unsprung Weight Transfer - due to the component of lateral force applied by the weight of the wheels, uprights, brakes etc.  For live axle, includes total axle assembly weight.  We take the axle height as a close approximation to the centre of gravity, (CG), for the non suspended mass. 

And two components of Sprung Weight Transfer:
Geometric Weight Transfer - due to the component of lateral force, applied directly at the Roll Centre (RC).  Geometric WT is reacted directly through the suspension linkages, and does not induce body roll.
Elastic Weight Transfer - due to the component of lateral force, applied at the Suspended Mass CG, and does induce body roll.  This force is reacted in the springs, anti-roll bars and shocks, and is the only one of the three components of total weight transfer that does induce body roll.

It is clear that low roll centre give little geometric weigh transfer and most of the weight transfer goes through the springs (elastic wt), and is therefore delayed by the time it takes for the vehicle to take a set.  Conversely, with high roll centre most of the weight transfer precedes the body roll, leaving a smaller amount of weight transfer to go through the springs.

The location of roll centre heights and the affect on geometric weight transfer vs elastic weight transfer is of importance in the set up of the car.  Geometric weight transfer is a major influencer for cars of high front weight percentage and/or for FWD.  Also for RWD with live rear axle.  Also for current open wheelers with high downforce and little suspension movement.

From "How to make your car handle" - Fred Puhn, P.43:

"In the old days, it was common to use roll-center height as an adjustment to tune the chassis.  The theory is, by raising the RCH at one end, you get more weight transfer at that end.  There are two problems with this and this method is no longer used. 
First, raising the roll center does increase weight transfer due to RCH, but it decreases the weight transfer due to body roll.  This introduces complications into the chassis tuning process that are hard to figure out.  Whenever you make two changes at once with a chassis adjustment you are in for trouble, and if your changes act in opposite directions, you have made a bad situation even worse!
The second, and by far the worst problem, is the almost impossible task of changing RCH without radically changing the suspension geometry.  People used to move the upper A-arm of a double wishbone suspension, thinking it would alter the RCH, but they were changing the camber characteristics and bump steer of the suspension at the same time.  These multiple changes make adjusting the roll center a very risky and confusing guessing game with independent suspension.  To me, it seems so much easier to adjustable anti-roll bars.  These are not only easier to adjust, but the anti-roll bars give you an infinitely fine adjustment. 

Load ratings are not the tire data you need, and tire companies don't really publish slip angle or coeffecient of friction data vs tire load data.  If you're sighting in a gun, I'd prefer to take a shot, see how much you're off and adjust the scope, rather than calculating muzzle velocity, distance from target, curvature of earth, aerodynamic effects, and then trying to curve the barrel.

WHICH, brings us back full circle to:

and, most elegantly stated,

One other thing to think about when talking about friction between the tire and the track, is that with lean angles measured in radians the outside tire is likely going into a positive camber situation and being loaded in a way they were not designed for, and rolling onto the outside wall under the weight. 

Your outside tires in the picture from Judge Phil look like they have negative camber, but at that angle how much is actually touching the ground and doing any good? 

The more contact patch you can keep on the ground the more work the tires can do to turn the car. 

With the size of your car, getting the front tires into a good spot with a couple of degrees of negative static camber should be more important than screwing around with the camber on the back.

This!  and if you don't like it, it's easier to put back to stockish