BY LARRY WEBSTER, February 2005
The Evo MR and the WRX STi feel very different when evaluated one against the other on the track, so we did a few more handling tests to see if we could gather some data that would shed light on our seat-of-the-pants assessment.
First off, although these cars appear to be quite similar (turbocharged four-cylinder engines, four-door bodies, four-wheel drive, limited-slip front and rear differentials, etc.), there is a major difference in the way the four-wheel-drive systems distribute engine torque.
The Subaru employs an electromagnetic clutch on the center differential. Under normal driving conditions, say you're just cruising down the interstate, the diff clutch is disengaged, and 65 percent of the engine power is routed to the rear wheels, with the remaining 35 to the front (a 35/65 front-to-rear split). The car's engine computer adjusts that center-diff clutch, based on information from the yaw-rate and throttle-position sensors, and can send as much as 50 percent of engine torque to the front wheels. So the Subaru varies the torque split between 50/50 and 35/65. The driver can also manually select the torque split via a center-console switch.
The Mitsubishi, on the other hand, never sends the majority of engine torque to the rear axle. Instead, it can send all the engine torque to the front wheels or 50 percent of it (using an electrohydraulic clutch on the center diff). So the Evo varies the front-to-rear torque split between 100/0 and 50/50. In addition to the yaw-rate and throttle-position sensors, the Evo has a steering-wheel-angle sensor that also provides the computer with information.
Now, we're simplifying things a little because it would take pages to describe exactly when those center-diff clutches operate, but basically, the aim of these systems is to make the car go where the driver wants. For sporty cars like these, four-wheel drive can simply be another tool in the engineer's box that improves handling, and the computer algorithm that controls these center diffs is tuned in much the same way as the car's suspension. Four-wheel drive is another interconnected variable—like shocks, springs, anti-roll bars, and tires—that affects vehicle handling.
Judging by the Subaru's four-wheel-drive system, we initially figured the STi would be better at the racetrack simply because putting more torque to the rear wheels frees the front tires to do their main job, which is turning the car. The Subaru also has better weight distribution (58.2/41.8 versus 60.7/39.3 for the Evo), which should improve its handling.
But despite the STi's power-to-weight advantage, it wasn't the faster car at GingerMan, as Swan reports. The STi's best lap time was 1:39.15 and the Evo's was 1:38.88.
To dig a little deeper, we decided to perform some tests that go beyond our usual procedures, but before we went back to the track, we put both cars on Kumho Ecsta MX tires to remove that variable from our results. And even though these cars have driver-operated switches that can alter the function of the center diffs, we did all our testing in the automatic mode.
First, let's talk about the split-traction-surface test. We used Bosch's proving ground in nearby Flat Rock, Michigan. That facility has lanes of pavement running side by side with tiled surfaces, ideal for measuring varying traction levels. For this test, we put the left-side tires on a tiled lane that simulates driving on packed snow and the right-side tires on dry concrete. Then we accelerated from 5 to 50 mph. We also performed the same test with all four tires on dry concrete.
The Subaru accelerated at the same rate on both the dry section and the split-traction surface, whereas the Mitsubishi was 0.4 second slower on the slippery section than it was on the dry.
That test tells us the Subaru's four-wheel-drive system is a little better at sending torque to the wheels with traction.
Next we ran increasing and decreasing slaloms. Here, the cars were close, but the STi had a significant 2.2-mph-faster run through the decreasing slalom. So far, the Subaru was looking pretty good, and we were starting to wonder if perhaps GingerMan was a track that suited the Evolution but not the STi.
That thinking led us to DaimlerChrysler's 1.7-mile evaluation and handling loop in Chelsea, Michigan. It's a little like GingerMan in that it has a few long, constant-radius turns, but it also has a high-speed sweeper that can be taken at more than 100 mph.
We did two four-lap sessions in each car, and the STi was always quicker. The margin, however, was a scant 0.15 second. We also timed the cars over the tightest and slowest half of the track and found that both went through that section at an identical 52.80 seconds.
The numbers don't tell half the story, and this extra testing only reinforced what we learned during the comparo. The Subaru understeers more than the Evo, and that hurts its corner-entry and midcorner speeds. But the STi is extremely good at accelerating out of turns. And in some situations where stability was the name of the game, as in the decreasing-slalom exercise and on high-speed turns, the Subaru had the advantage.
The Evo felt quite a bit livelier, and drivers could rotate the car easier, which meant getting on the gas sooner than in the Subaru.
The Mitsu's neutral handling, however, was only able to overcome the Subaru's gutsy motor at one track. And we think some subtle chassis tuning on the STi would likely make it quicker everywhere. The tuning may be something as easy as an air-pressure adjustment. But that's tough to say without testing, so maybe we'll have to borrow these cars again and head back to the track.
Mitsubishi Lancer Evolution MR / Subaru Impreza WRX STi
street start, 5-50 mph, dry pavement 5.0 / 4.7
street start, 5-50 mph, split-traction surfaces 5.4 / 4.7
roadholding 300-ft skidpad, g 0.92 / 0.90
increasing slalom*, mph 60.9 / 60.4
decreasing slalom*, mph 57.2 / 59.4
lap time, sec/mph 75.20/81.4 / 75.05/81.5
autocross time, sec 52.8 / 52.8
*This 938-foot slalom consists of 10 gates, each of which is six feet longer and spaced three feet farther apart so that it becomes two different courses when run in two directions. In the increasing direction, the car is accelerating, so there is weight transfer to the rear axle. Conversely, in the decreasing direction, the car is slowing and loading the front-tire contact patch.
