Nothing says "triumph of enthusiasm over physics" like a FWD race car.
A FWD has all of physics working against it. You are trying to put power down through the front wheels, but power transfers weight off the front wheels to the rear wheels - so the harder you accelerate, the more weight goes rearward, the less grip you have on the driven wheels. Eventually, this reaches equilibrium at a grip/acceleration level far less than should be possible for the power/weight of the vehicle.
That being said though, there are a couple of advantages to FWD:
- They can be smaller and lighter than the average production car. FWD packages the drivetrain very compactly (think about the original Mini) and that lets you build a car to much smaller and lighter physical dimensions. This means that FWDs, sized appropriately, can fit into smaller gaps and can generally avoid slowing down very much. A car like the Jim Harnish CSP CRX may not make much power or be able to put it down (even if it had it) but it can negotiate slaloms and similar course elements at much higher speeds than larger, heavier cars.
- FWD cars tend to compete against other FWD cars (meaning a level playing field) and when FWDs are mixed with other driveline configurations, the rules tend to give the FWD cars weight and size breaks compared to RWD/AWD.
- I think that FWDs are physically easier to drive. No matter what, an FWD car is never going to experience power-on oversteer. This gives the FWD driver a powerful tool, in that he can manage longitudinal weight transfer with his throttle foot - more throttle means more rear grip, without ever crossing the threshold into power-on oversteer that occurs with a RWD car. (Try watching a novice/hack C5 Corvette driver; see how often they spin on corner exit). It is possible to set up a FWD car with a lot more corner-entry oversteer than with other driveline types, because catching the car is a simple matter of mashing the throttle. Done correctly, you get a car that rotates on lift throttle, but plants the rear firmly once throttle is applied. This can make for slaloms/offsets that follow a steer-lift-rotate-plant driving rhythm that, especially when coupled to small and light physical dimensions, can be easy to drive stupid fast.
So then, the keys to setting up a FWD are going to revolve around the facts that you can't put power down, and when you do, it is going to understeer. I'm making the assumption here that your class rules allow you to make reasonable power levels. If your rules limit power production, the same advice applies, but to a much lower degree.
The very first step (outside of proper tire selection) is thus going to be putting a GOOD limited-slip front diff in the car. If you're going to be forced to drive the car with the wrong end, you might as well get both of the wheels on that end working for you. Otherwise, you're going to be plagued with inside-front wheelspin which is not going to help you put down power and will magnify your power-on understeer problem. Furthermore, there are some neat things that happen when you can change the thrust vector of the car with the steering wheel; meaning that the LSD is going to help you with throttle-on understeer by channeling some of the thrust from the driven wheels into cornering force. This is commonly reported by drivers as the car "sucking the nose in" under power, and it's a good thing.
The next is going to be balancing the car. You want the car biased much more heavily towards entry oversteer than you would with a RWD/AWD; partially because entry is the only chance you get to rotate the car, and partially because a lot of rear lateral weight transfer is going to (thanks to diagonal weight transfer) put more weight on the inside front, and get more grip out of that inside front tire. You're trading rear grip for front grip, but given that you're trying to do everything with the fronts anyway, that's also a good thing.
It's probably worth mentioning that you want as much front weight bias as you can get away with. This will hurt braking, but you aren't going to be braking all that much and you need every ounce of weight on those front wheels. Note, however, that there is an upper limit to how much front weight you can stand; the FWD drag racers were up in the high 80% area at one point, and some very odd things started to happen when the CG got forward of the front contact patch..... So lots of forward weight bias, but not too much.
Now things get interesting.
Given that we want corner-entry oversteer, and given that we want a lot of rear lateral weight transfer (to help plant the inside front) that means it's pretty much a given that we're going to lift the inside rear off the ground on corner entry. That's not a bad thing, in of itself, but it does mean that once we've got that wheel up in the air, the rear has given all it can in the way of total roll resistance, and if our total roll angle is still too high to keep the fronts in their Happy Place, we're going to have to add front spring.
Things get really interesting when we have a front McStrut suspension to deal with. McStruts have really crappy camber compensation in roll; the suspension does not gain negative camber when compressed like a double A arm does. This means that as the car rolls, it is trying to pry the outside front increasingly positive in camber, to the detriment of front grip.
The usual hot ticket for setting up a front McStrut is to admit that one happy tire usually makes more grip than a pair of pissed off tires, so you slap a lot of front bar on the car, get 100% front lateral weight transfer (lifting the inside front off the ground in the process) and then set the static camber such that the outside front tire is in its Happy Place at full roll - look at any fast M3 or CamaroStang.
But with the fronts being the driven wheels, we want LESS front lateral weight transfer, not more. And that means that there is going to be a tradeoff somewhere, especially when it comes to static camber angles. More will help out the outside front at full lateral G, but it will also hurt power-down - and we've been battling to get all the power-down we can get.
And to add further complexity, if our front diff is a Torsen, like a Quaife, it absolutely cannot tolerate having one of the wheels it connects hoisted off the ground. As soon as that happens, it goes open and you lose front drive. So if you have both front McStruts and a Torsen front diff, there is a limit to how much front weight transfer you can get away with before the diff stops working. That shouldn't be an issue because we're trying really hard to keep the front bar soft - but a high CG and too much roll angle without moving to stupid springs might not give us much choice.
This means that there is going to be a compromise in here, and it is probably course-dependent. On a course where you can maximize momentum-maintenance, and especially if your competition is mostly wide, high-power AWD/RWD cars that cannot fit into the same spots you can, you want maximum lateral grip and you crank up front camber to get it. On a pointy-shooty course with a lot of tight corners that you cannot maintain momentum through, you back off the camber in order to get more power down.
Probably the best solution is to find a small, light car with a double-A-arm front suspension; I think there are some Civics that might fit this mold....
Anyway, in practical terms I think we're looking at a front natural frequency of around 2Hz with a rear frequency around 2.5Hz - at least as a start point. Were I setting up a FWD, I'd start there, and then try front frequencies of 1.8Hz and 2.2Hz and see what happened. Based around what I learned there, I'd set the rear bar and/or spring until I had the inside rear off the ground at entry, tune front spring to either get the roll angle I wanted (or found the point where power-down started to suffer) and then start playing with front static camber. This means a lot of iteration around front natural frequency and front camber angle trying to find the happy medium between peak cornering force and peak acceleration - on several surfaces, and on several course layouts, such that there's a "concrete open setup" and a "asphalt tight setup" etc etc.
I'd also seriously consider a traction control system like the RaceLogic system, especially on a high-HP, non-turbocharged car.