I rarely post in here as it's too close to the day job and work, but maybe this will help?
I am also very anti FWD cars, no one would design a performance car as front wheel drive unless other criteria were dominant, such as occupant space, cost or other factors, they are plain wrong....
Bear in mind without an effective limited slip differential weight transfer from stiff bars can cause serious corner exit traction issues.
It’s possible to select anti-roll bars by calculation and get pretty close to right on the first try. That does take a bit more engineering than selecting springs, but you don’t need an expensive computer program or five years of university maths. You do need approximate values for sprung mass c.g. location, roll centre heights, unsprung mass weights, and anticipated lateral acceleration. The lateral acceleration value can be based on what similar vehicles do on similar tyres, or you can just use 1g, which will give you your roll gradient in degrees per g and your load transfer values in pounds or newtons or percent per g.
First, estimate how much total load transfer is going to occur. This is simply total car mass times lateral acceleration, times overall c.g. height, divided by track.
Decide how you want to apportion this, front to rear. If the car has 50% rear weight and equal tyre sizes all around, and roughly comparable camber control front and rear, try to get close to equal load transfer at both ends. If the car is nose-heavy or tail-heavy but also has unequal tyre sizes such that the rubber is apportioned the same way the weight is, try for equal percentile load transfer at each end. When the car is nose-heavy or tail-heavy, and the front and rear tyres are equal size, as with most front-drive cars and older rear-engine cars, it is necessary to have a disproportionately large share of the lateral load transfer at the light end. How severe this should be is usually a guess, but we try to make it an intelligent guess.
Next, figure out what load transfers and roll angle you’d have with no anti-roll bars.
You have to estimate the unsprung load transfer. That’s c.g. height of effective unsprung mass for lateral force at each end of the car, times weight of this, times lateral acceleration in g’s, divided by track.
You have to estimate geometric load transfer. That’s roll centre height, times sprung mass weight at that end of the car, times lateral acceleration in g’s, divided by track.
Since we’re assuming steady state, we can ignore frictional anti-roll or pro-roll (anti-de-roll) moments.
We add unsprung and geometric load transfers, and subtract that total from the total load transfer. The resulting difference is the elastic load transfer. Assuming equal track at both ends, this will be apportioned in proportion to the elastic angular roll stiffness at each end of the car. We are going to solve for required values for these, and then choose anti-roll bars accordingly. The roll displacement will be equal to the sprung mass elastic roll moment (total overturning moment, minus the total of unsprung and geometric moments) divided by the combined front and rear angular elastic roll stiffness. So we calculate what angular elastic roll stiffness we get from just the springs, and then, without changing the springs, we add elastic roll stiffness with the anti-roll bars to give us total load transfers at each end that come close to the target values we set, and also a roll angle we consider acceptable.
Angular elastic roll stiffness at each end of the car is the angular rate per inch of wheel travel, in pounds inches per inch per wheel or Nmm / mm / wheel, is the sum of the wheel rates at the two wheels in roll from all elastic devices (usually meaning springs and anti-roll bars), times the square of half the track. The angular rate in lb in /deg or Nmm/deg is that divided by the number of degrees that an inch per wheel of displacement equates to, which is 180/π or 57.3 divided by half the track. For example, for a 57.3” track, a wheel rate of 500 lb/in gives us 1000*57.3/2 or 28,650 pounds inches per inch per wheel; same for 400 lb/in on one side and 600 on the other. An inch per wheel is 57.3/(57.3/2) = two degrees, so the rate in pounds inches per degree is half as much, or 14,325 lb in / deg.
We want check to see how much or our elastic roll resistance we’re getting from the bars, compared to the springs. As a rule of thumb, for cars without a lot of downforce, on the road or on race tracks, we don’t want to get more than about half of our total elastic roll resistance from the bars.
Finally, we need to make sure we aren’t predicting 50% load transfer at either end unless we are okay with lifting a wheel. More than 50% is physically impossible. If 50% transfers, that means 100% is on the outside wheel. In some cases, we may have this condition at the light end of the car and still have more lateral acceleration capability at the light end than at the heavy end. That’s why we see nose-heavy front-drive cars understeering despite carrying a rear wheel, or vintage Porsches and other rear engined cars oversteering despite carrying a front tyre aloft.
