Aside from roll stiffness, there is also the suspensions ability to absorb bumps at all.

One camps says to start with a solid suspension and then soften it up enough to absorb the the type of bumps and dips that the car will encounter.

In my simple mind, this seams to mean that one way to design a suspension that doesn't rattle the teeth and cause the tires to skip over bumps is to make it softer than ideal, and then help restore roll stiffness it with shocks and sways.

...but thats starting in a different direction than just roll stiffness.

Keep in mind a pikes peak dirt road or wet race tracks like more roll to transfer weight onto a sliding tire to give it grip, whereas at the other extreme, a paved surface with slicks already has enough grip and weight transfer can cause significant camber changes etc causing grip to go away.

-In some case you want to limit roll more than others
-in some cases you want more or less reaction to road surface
-in some cases you want wheels to act very independently
-on the smoothest grippiest race tracks I would guess there are few compromises compared to other bumpier track, roads, or rally courses.

So to actually know what is the right thing to do, I defer to the experts because I find this all mind boggling to sort out...

that is what i have been wanting to try. i look at it as a Sbar lifts the inside wheel as where if you removed the Sbars and increased the Tbars to provied the same amount of body roll, the inside spring would soley be pushing down onthe inside wheel.

Wayner, thanks for pointing out that you might want to have the stiffest torsion bars/springs possible but your teeth would be shaken loose. So the factory uses sway bars to reduce sway and allow smaller and more comfortable torsion bars. It makes sense.
T77911S, there is a very fast local racer who argues for 1200 lbs and higher spring rates coupled with a stock rear 18mm sway bar and a not-much-bigger front aftermarket adjustable... in effect using hardly any sway bar at all relative to the torsion bars. He's done very well with this approach.

I made a sheet like that years ago, although I don't know if that one is mine. My conclusion is that both sway bars and T-bars can increase roll stiffness, but do it in different ways and will have different knock-on effects. In both cases stiffer is not better. You want to have your car set-up so that it uses the suspension travel effectively without bottoming. You use the springs to do this. Springs will also affect squat/dive which sway-Bars don't do. Sway bars are really to tune the fore/aft balance of the car. Sure you can prop the car up using sway bars, but the knock-on affect is that they can tend to over-load a tire and make the car have a "knife-edge" affect which can be difficult to drive.

The whole objective is to get the load on the outside tires to be proportional to the tires' available grip when cornering. In the case of the 911, this often results in the inside-front wheel being off the ground. So be it!

Gordon Murry I believe once stated that he didn't believe in putting sway bars on both ends of a car because it was just extra weight on one end. As a result all of the cars that he designed only had sway bars in the back. While a production car like a 911 is not a formula car, the concept does deserve some thought.

Note that we haven't even discussed the impact of shocks, and the tuning of the shocks which many people suggest has a bigger impact on the car's handling.

When it's all said and done - I think that Colin Chapman summed it up well...

Many of the more competitive 911-based GT race cars that run in the PRC GT2 class here in Northern California run no rear anti-sway bar at all. These car are running wide Goodyear bias-ply slicks. These slicks, with how these cars are setup, like almost no static negative camber which suggests the camber curve built into the suspension geometry is sufficient for optimal performance while cornering. These cars do not exhibit much body roll.

I suspect the primary function of the front anti-sway bar is not control body roll. Instead, I suspect it is used to tune roll couple distribution....to tune understeer/oversteer characteristics.

Scott

The other two things to consider are the roll centers and track. Have you lowered your car? Are you using RS front struts with the raised spindle. This is a big deal. Just putting RS struts on the front will stiffen the front suspension in roll (assuming no change in ride height), so you'll need less front sway bar or more rear bar, or both. Changing the rear roll center by adjusting the inner trailing arm mount (which requires more work than just swapping out struts) can have the opposite affect.

Increasing the track at the front will often (but not always) result in increased understeer, and the opposite is true by increasing the rear track. This is one of the areas where the designers of the 911R were clever. In addition to increasing the rear tire size, they also reduced the rear track which transferred more of the cornering load to the front tires and thus reduced the inherent oversteer in the SWB 911.

Increasing the track on one end gives more grip to that end unless the tires are highly non-linear and are both outside their temperature range. Increasing the track decreases total weight transfer, not the rate of weight transfer but rather the magnitude.

if it helps at all;

I have a stock 71t 2.2l w/15's. I run it at track events often. I have replaced the shocks with bilstein, the bushings (stock), turbo tie rods, etc... Anyway the point is all the stuff is new, stock. I enjoy the ride, however, at the track it suffers from huge body roll. I only dicovered this from seeing pictures friends took while I was hard in the corner. I could feel the roll, but seeing it was an eye opener.

Remember, its stock. I run yokohama avs ES100s- 205/50 15's all 4 corners. I had NO sway bars on the car, as is from the factory.

Tired of the body roll and not wanting a stiff ride around town, I installed a aftermarket set of anti-sway bars, the Tarret Engineering brand. I ordered them, they came hollow, light weight and with adjustable links. I installed them and set them to their softest setting, WOW. Ya, that was the difference on track, WOW.

On road, I did notice a very, slight increase in a stiffer ride, no big deal. So by the seat of my pants, I highly recommend anti-sway bars for the racey driver. I have left everything else stock for now, and it rides great. On track, the pictures show very little body roll now, and I can carry a little more speed through the corners, fun.

hope that helps, keep us posted, thats how we learn, Mike.

When had 22mm front and 30mm rear torsion bars installed on my '78 911SC, the ride got much firmer but it was not harsh. The harshness showed up when had Charlie Bars (anti-sway bars) installed front and rear.

Scott

I believe that the car's roll response will be the same for one wheel bumps wether the roll stiffness is from anti-sway bars or ride springs. The difference is only in two wheel bumps like going over a speed bump.

Your car got harsher because you added the anti-sway bars on top of your already increased ride spring rate.

Like Mickey, this 72 is running avs ES100s- 205/50 15's all 4 corners.

Here is the suspension setup. Nothing fancy:
-stock "S" suspension (sway bars),
-one size larger rear torsion bar
-RSR valved shocks all around

Check out the body roll with 205's at the limit.
Nice neutral balance.
I think it is perfect for street and track. (larger stickier tires might require more but not with these tires)

(The cone is the apex. Cornering harder than it looks in the picture )

http://forums.pelicanparts.com/uploa...1316866442.jpg

Wayner,

The front wheels are straight in that photo......

Scott

...and the back end was moving left!

Maybe the alignment is off, now they are pointing right! :-)

Here it is a couple years later with the same tires (now a bit too old)

The owner is driving it this time, not me. (BTW, this ended well with no drama).

http://forums.pelicanparts.com/uploa...1316876965.jpg

Not likely much.....unless the tires are crap. That corner just isn't a good example. I have seen a lot stiffer 73 S at an autocross exhibit quite a bit more roll than that.

Scott

I guess I just don't get your point.

I wouldn't want to drive a street car that is any stiffer than this one.

P.S.
You don't get the full perspective in that first pic.
It is the middle of a double apex and many people screw it up there as you can see by the skid marks in the foreground.

Note that I was hedging a bit. Increasing the track at one end can cause that end load up faster than the relatively narrower track of the other end. So this can cause the end with wider track to lose traction sooner. But on the other hand as you noted, widening the track can have the affect of lowering the roll center at that end, which can make it softer in roll, causing the end with the wider track to roll easier, and thus have more traction.

So the affects of changing the track will depend on what is going on already with the roll centers, springs and sway bars.

No, No, Yes.

Firstly, the roll center depends mostly on the specific suspension type: struts, dual A-arm, swing axle, trailing arm, etc. Widening the track on a trailing arm suspension would do nothing to the roll center. What it would do is decrease the total lateral load transfer and so decrease the roll. Assuming a rigid chassis (not always a valid assumption), the opposite axle will also have less body roll and so less weight transfer.

It is true that on certain suspensions such as swing axle, widening the track can have a big effect on roll center height. On a strut suspension, not so much.

It is true that widening the track decreases the percentage of anti-roll (the percentage that the jacking force makes of the total force fromt he tire).

The track width does not make weight transfer at a faster rate. Stiffer springs do that.

The affects of the sway bars do depend on the rest of the suspension system, though.

Looking specifically at the suspensions designs on a 911, first the front end which uses a struts suspension.

http://1.bp.blogspot.com/_FoXrzNJ_K0...600/fig-14.gif

All other things being equal, increasing the track width will actually increase the roll center height slightly, which will make the front end act as if it is stiffer in roll than the same suspension with the same settings, but a narrower track width.

Looking at the semi-trailing arm rear suspension, the same is true -- widening the track (all other things being equal) will slightly raise the rear roll center, making the car stiffer in roll.

http://www.bevenyoung.com.au/fig13-1.jpg

So let me correct myself -- widening the track tends to raise the roll center at that end, which will make the car stiffer in roll and thus load up that end of the car faster. This is compounded by the fact that since the track is wider, and the distance from the center of the car greater, that the tire with a wider track will get a greater load transferred to it since it's lever-arm from the roll center is greater.

I always do roll centers at the intersection point which should be centerline static, but when the chassis rolls they change.

I was widening the track with a wheel spacer which I suppose raises the roll center a little bit but the jacking force goes down since the slope of the line decreases.

Anyway, if one isolates the effect of an increased track width, it is to add grip since there is less weight transfer on the tires. It doesn't necessarily do anything for body roll. That depends on the suspension specifics. The roll couple distribution is affected by roll centers, roll stiffness (springs), and the track width. A wide track can counteract for a stiff anti-sway bar and is good to have a wide track on the heavy end for stability.