Richard Holdener
December 1, 2004
Eaton Roots vs. Kenne Bell Twin-Screw (11 psi)
Like the Eaton, the positive-displacement twin-screw blower from Kenne Bell produced immediate boost response and a broad torque curve. Looking at the graph, you might be tempted to think that the Kenne Bell twin-screw is more efficient at higher engine speeds than the Eaton (it is), but less so at lower speeds. The reality is that the twin-screw was so much more efficient, we had to slow down the twin-screw by nearly 2,000 rpm (compared to the Eaton) to keep the boost pressure from exceeding 11 psi at the top. Running a maximum boost pressure of 11.9 psi, the Kenne Bell pumped out 629 hp and 525 lb-ft of torque.

While a boost-to-boost shootout seems only natural, there are problems associated with choosing boost pressure to regulate the playing field. The main problem with using boost pressure is that since the four different forms produce decidedly different boost curves, where (in terms of rpm) do you take a boost reading? An example works well here. In the case of the Eaton supercharger, it provides a bell-shaped boost curve with boost being slightly lower at both extreme ends (high and low rpm) and highest in the middle. Thus the 11-psi max rule would have the Eaton blower spinning to produce a maximum boost pressure of 11 psi at roughly 4,500 rpm, where the pressure would have fallen to just 9 psi at the maximum rpm (for this test) of 6,500 rpm.

By contrast, the Kenne Bell twin-screw blower will make peak boost pressure at 6,500 rpm, not because the Roots blower produces better low-end response, but because the inefficiency of the Roots blower requires that it be spun so much faster (that the twin-screw) to produce the desired 11 psi. That the boost pressure offered by the Eaton falls off at a higher rpm is not a plus on the side of low-speed response but rather a negative on the side of keeping pace with the motor. Spun the same speed, the twin-screw design will offer more power and boost pressure (at any rpm), not to mention a lower charge temperature. To put this into perspective, the Eaton was spun over 2,000 rpm faster than the Kenne Bell blower to achieve 11 psi (at 4,500 rpm). That additional blower speed helped improve the low-speed power production where the airflow supplied by the Roots blower could keep up with the demand of the four-valve motor.

Next up was the Kenne Bell 2.2L twin-screw blower upgrade. The twin-screw design offered superior efficiency compared to the Roots blower.

Obviously the centrifugal supercharger will also produce the desired peak boost level at the peak engine speed. The impressive power per pound of boost offered by the centrifugal supercharger (besting both the Roots and twin-screw designs) comes with a price. Though high-rpm (and peak) power is impressive with the centrifugal design, the low-speed boost response is significantly reduced compared to either of the positive-displacement blowers. Peak to peak, the Eaton is easily out-gunned by the Vortech, but Cobra owners do not live by peak power alone. Centrifugal superchargers (regardless of the nameplate attached to the supercharger), all behave in this manner. The boost curve increases with engine speed. In the case of our four-valve Cobra motor, the Vortech produced just 1.7 psi at 2,500 rpm before reaching a peak of 11.3 psi at 6,500 rpm. All the superior efficiency in the world won't overcome a boost pressure deficiency of nearly 8 psi at 3,000 rpm. It is at these lower engine speeds that the two positive-displacement blowers really shine. Note that neither of them matched the peak numbers offered by the Vortech at 6,500 rpm, but it is hard to argue with all that extra torque, especially on a street car.

The boost curve differential also applied to the twin-turbo kit from HP Performance (or any source for that matter). Unlike the superchargers, where boost pressure is a function of blower size and engine (and therefore blower) speed, the boost pressure supplied by the turbo is a function of turbo sizing, the power output (and attending exhaust energy) of the normally aspirated, and the load placed on the motor. For our purposes, the last variable is important, as the load placed on a stationary engine dyno is not the same as experienced on the street. While dyno testing, we loaded the motor at a predetermined engine speed (in this case, we started at 2,500 rpm) and allowed the motor to rev at a predetermined rate. In the real world, it might be difficult to duplicate this type of steady-state load at 2,500 rpm, unless you put it to the wood in Fifth gear going up a hill (something definitely recommended).

The stock air-to-water intercooler core was transferred from the Eaton to the Kenne Bell.

Despite the difference in the load experienced on the engine dyno, in chassis dyno testing and street driving, the HP kit has proven to be both responsive and powerful. Once loaded, the turbo motor can be essentially programmed (via a manual or electronic) wastegate controller to produce a set boost level. The wastegates open to reduce the buildup of exhaust pressure, thus limiting the boost pressure and power. We purposely chose a manual wastegate controller, which provided us a relatively consistent boost pressure of 11 (or 14) psi once there was sufficient exhaust energy to spool the 57mm turbos.

Ignition timing and air/fuel ratio were two other potential obstacles. Given the difference in boost curves generated by these different forms of forced induction, they will respond to different ignition timing curves. Since the centrifugal supercharger produces less boost pressure down low, it can be run with more ignition timing than the others. Is it fair in a direct comparison to skew the results by allowing one form to run a more aggressive timing curve than the others? In the end, we decided it was not and ran all methods with the same timing curves (23 degrees).

The same goes with air/fuel ratio, as the centrifugal supercharger can be run with a slightly leaner mixture down low. Obviously, octane comes into play here, but we purposely eliminated this variable by running 100-octane race fuel. The idea was not to see how much we could make from each on pump gas, but rather to illustrate the differences in the power curves.