BMW E36 M3 - Project Car - European Car Magazine

A numerically higher differential has often been referred to as the poor man's supercharger. In the case of E36 M3s with stock gearing (3.15 for 1995 models and 3.23 for 1996-on and M3 Lightweight models), a popular upgrade is a 3.46 or a 3.64 rear end. The end result is a peppier engine that sees its sweet spot sooner and more often.

These swaps are especially beneficial with aggressively cammed or even supercharged cars that don't make tremendous torque, but at the cost of decreased top speeds, increased shifting and higher fuel consumption. A turbocharged car starts getting traction issues with stock gearing and street tires on anything over 400 lb-ft to the wheels.

Diffsonline is the biggest name when it comes to BMW differentials, and it offers longer gearing with the choice of a 2.93 or 2.79 rear end, with either 40 percent or 60 percent lock-up for increased traction. The 2.93 with 40 percent lock-up seems the best compromise for power levels at 400 to 550 wheel-hp.

To measure improvement, we first baselined the car in the quarter-mile. Boost was lowered to 15 psi for this test-good for 465 wheel-hp-a power level we felt could produce more consistent results, given the conditions and lack of traction. In order to level the playing field and limit driver error, we decided to drop the clutch at 3000 rpm, mash the accelerator, and let traction (or lack thereof) take over. Of course, bogging first gear isn't ideal for a quick elapsed time, but it's easy to repeat and reveals potential gains any average driver would benefit from on the street.

Our average from two passes was 12.45 seconds at 123 mph. Acceleration improved slightly with the help of the 40 percent lock-up throughout second gear. And surprisingly, the taller gearing didn't compromise acceleration at the end of the quarter, as proven between 110 to 125 mph, where both are in fourth gear. With tires, boost level, weather, launch technique and location unchanged, our two-run average improved to 12.32 seconds at 123.5 mph. Thirteen hundredths of a second at the finish line may not sound like much, but considering we're traveling over 180 feet per second, that's an advantage of more than one-and-a-half car lengths.

Since a taller gear loads the turbo harder, it was obvious we'd see a quicker rise in power versus revs. How much sooner would be answered at the dyno, at Imagine Auto's facility. We decided to test the car in third gear. Graph 1 shows a significant improvement at an 18-psi boost level-about a 350-rpm advantage.

What about the improvement in everyday driving or at the track? Before concluding that this is a completely compromise-free upgrade, we needed to see the horsepower versus vehicular speed as well. Why? Imagine you're doing 60 mph in third gear on the highway (or coming out of a turn) and you floor it. Surely you'll see boost sooner in the rev band with the 2.93, but remember at 60 rpm you're doing 250 rpm less. The goal was to see an overlap of graphs, so that at any given speed we're still making at least the same power as before, even though rpm is lower. In this type of scenario, Graph 2 shows a little drawback in going with the 2.93 rear end-to the tune of about eight wheel-hp. In other words, with the 2.93 differential, at 60 mph (4220 rpm in this case), Project M3 would have to make exactly 10 more lb-ft of torque to make the exact same horsepower it did with the 3.15 differential. And there's still the fact that the gear is indeed taller. But don't forget, unless you're on drag radials you're going to see wheelspin more likely at this power level with the stock differential, and you'll be shifting sooner as well.

Therefore, in fourth gear, the car with the 2.93 set-up will probably be ahead. If we throw drag radials or slicks into the mix, it would be the other way around.

For us, the overall improvement is enough to call this upgrade worthwhile. As long as the car relies on street-tire traction, it's faster with the 2.93 rear end. The increased speed versus revs also provides better fuel mileage for cruising and a new top speed, given the 7100 rpm cut-off as controlled by the AEM EMS computer. Of course, with this much power and the heavy-duty Ferrea valvetrain we could easily set the EMS to a 7500-rpm rev limiter and do over 190 mph, but that would be unnecessary, if not lunacy.

This month also sees upgrades in the car's cooling systems for both water and oil. Victory Product Design, known for its BMW high performance and racing products, has a trick oil cooler set-up, which, at $650, will save bundles over the European-spec system and is easier to install. It's also good for an average 20 to 25 degrees F in oil temp reduction at the 200-plus F range, and now it's offered with an even larger heat exchanger.

With everything Project M3's got helping the cooling system-including the Fluidyne radiator, hood vents and all the exhaust coating and wrapping-we knew we couldn't go wrong with a heavy duty water pump. EMP Stewart Components has a stock replacement high performance water pump with a stainless steel (and much heavier duty) impeller. The flow is about 20 percent higher for a given pump speed, and the redesigned impeller housing is supposed to reduce cavitation (where water flow exceeds the impeller's flow capacity and potentially damages the impeller blades) at high rpm. Anyone driving a 1991 to 1995 BMW is probably relying on a plastic impeller water pump and there's no telling when those will fail. Owners of cars that see the occasional track weekend or severe summer heat should look into one of these right away. We're curious to see what the improvements of the oil cooler and high-flow water pump will be for us this summer.

Will the Project M3 series ever end? Actually, the car is finished and has been for some time. The project finale will be coming soon.

Rev Limited Reached In Third Gear

At 465 wheel-hp-a slightly above-average power level for turbocharged M3s today-the differential swap proves numerically beneficial on street tires by basically improving traction, bettering the previous set-up by over 24 feet at the finish line. At higher power levels, the 2.93 rear end becomes more beneficial. Next time, we'll try a rip at 20 psi with this set-up and see what happens.

Special thanks to House of Power for the installation of the 3.64 on Lawson's car. Although HOP normally specializes in Mercedes-Benz go-fast tuning, the shop also installed Jon's cam kit and has helped us tremendously behind the scenes with Project M3.