roll centres?

[bianchi1]

how doe's the roll centres affect handleing? a panhard rod has a decent roll centre, but a mumford can be set below ground, how would this affect handleing? thanks

[Alleggerita]

The lower the roll center, the less resistance to roll. Also, within moderation, the lower the roll center on the rear, the better the traction on the rear. The roll axis is important though, that is the relative roll centers of front and rear axle which will have a signifcant vehicle dynamics and oversteer and understeer. Generally, on Alfa 105's you are trying to keep the front from going below ground (using a variety of wishbone & upright modifcations), as it decreases roll resistance and thus requires very stiff front springs and you want to try to lower the higth rear roll center to improve traction and decrease understeer. From my understanding, below ground is not desirable.

[Echo Leader]

Quote:

Originally Posted by Alleggerita From my understanding, below ground is not desirable.

[Alleggerita]

I am aware of the WRX issue - the lowered and slammed cars just plain do not work - and any well set-up one always looks slightly high off the ground. Never played enough w/ a 116 Alfa to start moving roll centers, but I do know that people who race them often do consider raising the front roll center when the suspension has been lowered. The rear is less of an issue than the 105 series, as the roll center is much lower w/ the Watt's link and the rear end has lots of stick as is.

[Echo Leader]

At this point my pal hasn't messed around with the ride height at all, so it still has the stock "slightly" jacked up look to it. His main efforts thus far have been to stiffen the chassis as much as possible, while remaining in the STU SCCA AutoX catagory. He's also increased front and rear swaybar thicknesses. The car is fantastically fun, but its height makes it feel a little top-heavy.

Is the only factor in the roll center the geometry of the suspension? And does combating the effects of lowering a car like the WRX require one to mess with the chassis mount points for the suspension arms?

[Alleggerita]

One of the issues on an Impreza that is the same as on the Alfas is the camber control curve, that isn't enough camber gain, as well as the lack of roll resistance on the front when the roll centre goes sub-terranean. On Imprezas, people usually use huge sway bars to combat this issue whereas on Alfas stiff springs are more commonly utilized.

This is kind of off-topic for Alfas but here is a thread on Imprezas http://www.suspensionparts.info/showthread.php?t=294

[beatle_bayly]

I believe it's a simple matter to change the roll-centre of a transaxle car by welding a suitable stud to the top (or maybe the bottom) of the dedion tube and relocating the centre pivot of the watts link.

[60sRacer]

The issue of roll center on our alfas is not just about how the rear axle is set up. What you need to understand is how the front and rear work together.

All of the 105 chassis have an unequal length front suspension that puts the roll center of that end of the car below the road surface. The stock rear setup has the roll center about at the T-bar trunion, some 27 inches above the front roll center. This steeply angled roll axis from front to rear is the root of some of the car's quirks. To make this work well, the front springs are very stiff and the rears very soft, so that the car's handling is dominated by the front geometry. A side effect is that this keeps the chasis - especially on the spiders - from twisting too much.

One implication of this is that 'sporty' suspension changes - which usually involve stiffer and lower springs - add load to the chassis, hence the spiders (which are not as stiff a body structure as the coupes) do so well with the addition of a chassis stiffener.

Alfa made this design so that the cars could handle well and still not experience any severe oversteer conditions - very undesirable in street cars. (recall the Corvair experience, a car that could oversteer dangerously - and ignore the fact that the Porche bathtub of the same era had exactly the same handling). For racing alfas, they made some changes: Knuckle risers to improve the front camber angles on cornering, and a sliding block replacement for the rear T-bar to lower he rear roll center. The latter helps keep the rear inside wheel on the ground as the car loads in a hard turn. LSD was also added to minimize power loss as the inside wheel unloaded and tried to spin.

The front camber change helped keep the tire from tilting outward - lifting the inside edge - on hard cornering. A by product of this was that the cars added quite a bit of "bump steer" - a condition where the steering angles change as the wheel rises on compression. Made cornering pretty squirly, but the added traction of keeping he outside wheel in better contact was a net plus as long as you were a good enough driver to dance on that very sharp edge. It also makes the car wander a bit on rough straight sections.

Modern versions of the sliding block are panhard rods, Watts links and the very rare Mumford linkage. They all change the rear geometry to lower the rear roll center to the center of the diff case, a reduction of about 9 inches. This has a small impact on the car's overall roll axis as the front roll center is unchanged and about a foot below the road surface. You still get the rear twisting hard and trying to lift the inside tire, just less so. It's a noticable improvement to a racer or XC driver, but we're talking gains of tenths of seconds in a 40 second run on a XC course, not revolutionary handling changes.

It's pretty hard to make revolutionary changes to alfa handling - it comes from the factory nearly as good as a solid axle car can be! Not like the rice rockets that use low-production-cost parts and can change dramatically with upgrades.

Alfaholics has made some highly modified front arms that finally dial out the bump steer effect, a dramaic improvement in the drivability of these cars at the racing edge. Not a big deal in street cars; also they are made with rose joints that are definitely not suitable for a street car (unsealed bearings that require frequent inspection for wear).

Enough rambling I guess.

Robert

[genericwood]

Robert, just to add some more rambling. It is not that difficult to fix the bump steer problem on lowered 105/115 cars. Many of us just drill the steering taper out to accept a 14mm bolt, and replace the ball joint with a 14mm rod end. It can then be spaced as needed to virtually eliminate bump steer. I measure the static roll center on the front of my street GTV which is stock other than International sport springs. It sits about 2 inches below the ground. Serious lowering without upright changes could definitely put it a foot below! My race car roll center sits 2 inches above the ground. With Beck modified uprights, the lower a-arm is flat. I would like to the car further, however. If I do it without further upright changes, the roll center will go below ground. I'm trying to decide which way to go right now. With the rear roll center about 8 inches above ground (Panhard bar), do I want the front 2 inches above, or 2 inches below ground? Left and right movement of the roll center and camber curves will probably be the deciding factors.

Erik

[conedriver]

Murray,

It depends upon the pivot material. Alfetta, Milano & GTV6 all had Watts links with rubber bushing pivots, and didn't have the rattle you mention. My '76 Spider has a Panhard bar, and noise, but this is probably due to worn heim joints. Any of the axle locating schemes (Panhard, Watts or Mumford) could have either rubber bushed or Heim style pivots.

Quote:

Originally Posted by msiert I almost bought a Panhard Rod once but I think this modification is only worth it when the car is a track car only. From what I hear the Panhard Rod or Watts Link have a rolling rattle to them.......not something you want listen too while taking your wife or girl friend out to dinner.

[genericwood]

Murray, lot's of street cars have panhard rods or watts links. They can be designed to be quiet. I haven't raced without one, but Russ Neely says the Panhard rod was worth more than a second a lap on his IT car. Once you hit the track, you will want to begin tuning the chassis with the same level of study that you have put into the driveline.

Erik

[Alleggerita]

Quote:

Originally Posted by msiert Does yours rattle? How do you make them quite?

[60sRacer]

Murray,

In general, the suspension bits used by racers are noisy. The common part is a steel ball joint; this provides very precise location with very free motion, all highly desirable for a 10/10ths suspension.

But metal joints transmit every pebble in the road as noise. Perhaps you can get a ride in George's car to see what it's like!

The other problem with these fittings is that they commonly are unsealed bearings, and collect a lot of dirt. Hence they wear like the dickins. Also, they are subject to corosion and rust if driven in wet conditions. Racers commonly inspect their bearings every race session for wear.

Nylon or delrin will quiet this somewhat, but it still will not be like a street car. Hence rubber bushings are often used. Modern cars use more precise joints in the suspension as sub frames, then mount the subframe on rubber bushings to isolate the noise.

Robert

[60sRacer]

Erick,

I was aware of the solution for bump steer. Another approach I have seen involves heating and bending the steering arms, but just spacing them works too.

Sounds like you really have your geometry sorted out; must be a fun car to push to the limits. Have you found a way to get a good front roll center - by that I meen anything above the road surface - and still get a good camber curve? The GTA's knuckle risers did the later well, but made the other problems (roll center and bump steer) harder to deal with.

My guess (WAG) on your question is that tire contact is more important than roll, so camber first, but the mix of fast and tight turns on the course might sway the balance. It'd be a neat test program; I'd probably wear out two sets of tires just testing.... ;D

Robert

[genericwood]

Robert, I really DON'T have the geometry figured out! At least not yet. If my latest measurements are accurate. A big if. The roll center on the front of my car is currently at 0.9 inches. In dive it goes below ground, and in combinations of dive and roll can move sideways a lot. I want to drop the car in the front an inch or so. If I do it by spacing the spring pans, the roll center goes to about an inch below ground. But the camber curve actually improves as does left/right movement of the roll center. That is what I will probably do in the short term. But I am still comtemplating upright changes that would allow me to lower ride height about an inch but raise the roll center to about 3 inches above the ground. This would also improve the camber curve and left/right roll center movement, but would be a lot more work. Theoretically, the higher roll center would allow me to run the front end a lot softer and still control roll.

What I would really love to figure out is how to improve the camber curve in roll, without causing lots of negative camber in dive under braking! Guess there are always tradeoffs!

Erik