2005 Mitsubishi Evolution(s) and Dynos

Before heading to Willow Springs Raceway's Streets circuit for our track comparison of Mitsubishi's trio of Evolution models (as featured in the March 2005 issue of Motor Trend magazine, on newsstands beginning February 1, 2005), our test crew made a special detour to Torrance, California. A respite by the sea before getting down to work, perhaps? Hardly. The destination was Dan Harman's AWD dynamometer, where, one-by-one, our Evos - the standard VIII, bare-bones RS, and technology-laden MR - would be strapped to Harmon's horsepower lie detector and forced to scream their performance truth (though hopefully not spilling their guts in the process). Ultimately, we learned more than we expected.

Click on illustration above to view the dyno results for all three cars.

Our original goal was to certify that residing beneath each car's hood were engines that could be considered truly even-steven in output (what you'd expect, given their identical specs). But we wanted to be sure that any differences in handling or acceleration at the track were owed entirely to the variations we knew about going in - springs, shocks, wheels, gearing, and aerodynamics - not an undisclosed horsepower hiccup that would leave us later scratching our heads.

According to the spec sheets, these 2.0-liter, turbocharged and intercooled four-cylinder engines deliver 286 lb-ft of torque at 3500 rpm and 276 bhp at 6500 rpm before the revs top out at 7000. Engine-to-engine variations are to be expected, but how significant are they? Consider the nearby graph.

Each car was granted three runs, and here we've plotted the best two of these, noting the peak power of the better one. Nevertheless, all our curve's apexes fall significantly short of the manufacturer's claim - a surprise? An expose in the making? Not at all given that our dyno is accumulating data at the tires' contact patches, well downstream of the engine itself. Indeed it's interesting to note that the power consumed betwixt the flywheel and rubber is on average about 17 percent of the total, less of a bite than we would have expected. But it's maybe more interesting to note how variable each car's runs were, with significant divergences appearing at the upper revs as the electronics struggle to deliver maximum power while keeping the head from being tossed across the room. Also notice that in the case of Evo VIII, the power peak appeared 1450 rpm short of the manufacturer's claim, a consequence of power's broad plateau and the curve's random fluctuations.

So how consistent were they? The maximum difference in peak power at the wheels was 8 bhp (3.5 percent of the average horsepower) which would magnify to maybe 9.5 bhp at the flywheel. Not bad, but it's unsettling to think that at the Evo behind yours at the next autocross might have a handy 9.5 more ponies. But life's never fair.