Feb 4, 2006 update to the Mazda STS2 Miata project car

Anti-roll bars

Today we’ll look at anti-roll bars (aka sway bars). The primary function of anti-roll bars is to combat the body roll produced as the suspension reacts to cornering forces. Body roll in and of itself is not a bad thing, but too much of it allows the suspension to venture into areas where the camber situation is grim or where the suspension bottoms out on one corner. Body roll can also exacerbate a roll steer geometry problem. Large anti-roll bars can prevent these things from happening, keeping our tire contact patches flat on the ground, and pointed mostly straight ahead.

There is an age-old debate about how much roll stiffness to provide via sway bars and how much via the springs. Sway bars have a huge advantage in that they can provide much larger amounts of roll stiffness than typically usable spring rates. To get the equivalent roll resistance of a typical sway bar, you’d need springs that were so stiff that you’d be losing tire grip on anything but a super-smooth surface. So we use a combination of both, but rely primarily on the anti-roll bars.

From our CSP setup we’ve found that the 1.125” hollow front bar from Racing Beat provides the necessary control to keep our chassis roll angle at about 2 degrees. Since the camber gain curve on the Miata’s double a-arm suspension provides a favorable return of about 1 degree gained for every 2 degrees of roll, this means we’ll only need about 1 to 1.5 degrees of static negative camber to keep our outside tires vertical in a turn. This also keeps the inside tires flatter than they might be with more static camber. This is tremendously important at the rear of the car as we fight to put the power down on corner exit.

Its important to note that with the reduced cornering forces of STS2’s street tires vs CSP’s wide, fat and sticky Hoosiers, we’ll be generating less body roll. But that will be counteracted by the fact that the relatively soft street tires typically need more camber for optimal performance than the super-stiff R-compounds. Once again, this is just a starting point and we’ll need to revisit and tune once the car comes together fully.

While we generate more of our roll resistance at the front of the car, we need to balance that in part with a rear bar. In addition to providing necessary roll resistance, a rear bar also helps us to balance the effects of weight transfer from front to rear. In short, we can alter the steady state balance of the car quite easily via the rear sway bar. In CSP, we ran in three different configurations: 1) no bar, 2) stock 12mm bar, and 3) Racing Beat 5/8” bar. Option 1 was used on very open courses on slippery surfaces to combat the natural oversteer tendencies of the car in those situations. Option 2 was used when either the course was tighter or the surface was much gripper. This was a very stable combination that allowed for early power application on corner exit and was our typical setup. Option 3 was rarely used, but was ideal for tight courses on grippy concrete.

While sizing the bars is important, mounting them for optimal performance is critical. Particular attention must be paid to removing any tendency for bind. This is especially important when urethane frame mount bushings are used. These typically are sized too tight. In almost every off-the-shelf aftermarket application I’ve ever seen, these required some modification to facilitate smooth movement. Lubing the inside of the bushing will provide some temporary relief, but under pressure, the lube will just be forced out over a very short period of time. Once that happens, the bar will act a lot stiffer than it is or it will be erratic. It will also start to fatigue the frame attaching hardware or the frame itself. It is common to see either the saddle bracket or frame extension on a Miata cracking from improper front bar installation.

To fix this problem, there are two easy solutions. One is to grind away some of the material from the flat outside part of the D_shaped bushing. Alternatively, you can put a washer or two under the saddle mount to shim it up. I prefer the former method. Ultimately, you are looking to be able to smoothly move the bar arms up and down when the end links are disconnected.

End links are also critical to proper sway bar performance. Stock rubber bushings provide a delay or offset in the action of the bar in exchange for less road vibration. Urethane link bushings are a better alternative, as they reduce this offset effect, but the best performance is provided by spherical rod ends. These remove all compliance so that the bar is always in action. 1999 and newer Miatas come from the factory with a similar setup and it can be emulated on the earlier models as seen in the pictures. At the front of the car we have merely screwed together a pair of male/female rod ends to provide the desired length. At the rear, we need a longer end, so we have screwed a pair of male ends into a coupler purchased at a local hardware store.

Rod ends vary tremendously in quality and you do get what you pay for. Teflon linings will keep road dirt out of the joint, which in turn will reduce wear. Cheap rod ends will typically start their tell-tale clicking sounds after just a few months of use. Of course, if you are willing to replace them frequently, go for it. We have good bearings at the front of the car and cheap ones at the back. It just worked out that way.

In addition to removing compliance, end links can also provide needed adjustability. This is important for two reasons. The first is to remove any pre-load from the bar when the car is in a neutral state (at rest, with driver). The second purpose is to compensate for lowering of the car and larger diameter bars. In the front of the Miata, use of standard length end links will typically place the large aftermarket sway bar in contact with the upper control arm. In addition to nasty handling effects, this puts tremendous stress on those frame mounts again. You should remove the spring, connect one end of the sway bar, and move the suspension through its entire travel to verify no contact. You also want to be sure that the rod ends do not move into a state of misalignment where they bind against their mounts.

Next up is a report on some tire testing we did recently.

Join Free Join our community to easily find more project updates.
View comments on the GRM forums
Sponsored by

GRM Ad Dept

Our Preferred Partners