Even with all the mentions of future ECU tuning plans in our previous installments, we still received a flood of complaints about Project Evo VIII MR and why we chose to save costs and use a temporary piggy-back fuel computer when the suspension cost almost $5000.
Thus far, our Evo has been through some mild power tuning and significant chassis work, all done at Road Race Engineering and Road Race Chassis in Santa Fe Springs, California. There's a reason why it seems like an RRE exclusive car: RRE has a complete tuning program and is familiar with what works and what doesn't with the hardware that's already on. So guess who we took it back to for its final big power installation. Why shouldn't we? No one would take their car to one body shop for body work and then to another for paint.
Now here's the disclaimer: we liked the Evo the way it was (aside from temporary AFC tuning) and knew what we were getting into when we asked for big power involving a Garrett GT3076R turbo, Cosworth CNC head and lumpy cams. This is purely an experiment on how much useable and reliable power the stock bottom end would put up with on 91-octane pump gas. Take the term 'useable' with a grain of salt.
RRE Evo VIII GT turbo kit
We start with the heart of the newfound power: RRE's functional but no frills GT turbo kit. RRE assumes you already have all the basic bolt-on plumbing for intake and exhaust, so only includes the turbo hardware itself, hence the term `turbo kit'. Along with a selection of four different types of turbos, our kit comes with a cast, smooth bending T3 turbo manifold, all the plumbing, TiAl 44mm wastegate, trick turbine outlet housing and wastegate plumbing. More impressive is all the thoroughly thought-out additional hardware, including manifold studs, gaskets, proper nuts and washers, V-band clamps for the wastegate, coolant lines, flexible hard oil return pipe, and even an in-line oil filter for the turbo.
To complete the kit, you can pick from an E-Pinions Fashion ball bearing 50-trim turbo, Street Power/Track Whore ball bearing GT3076R, Track Whore Deluxe custom ball bearing GT35R or Drag Queen/Race Gas Slut ball bearing full GT35R. The intercooler and intake plumbing might have to be modified, depending on the turbo. We went with the GT3076 Street Power/Track Whore turbo, which, in theory, can flow up to 525 wheel-hp. How much horsepower can be squeezed out is limited by the clutch, bottom end and how much fuel we can keep pouring into the cylinders.
Since the car had to go back to Mitsubishi, the original plan was to cut some corners and just strap on the turbo kit, temporarily tune the AFC and throw the car on the dyno before adding the Cosworth CNC head and doing a final tuned ECU reflash. Don't ask how we know, but this is a pointless exercise with the stock intercooler. Like the intake and exhaust, everything the air has to flow past to get to the intake valve and beyond the exhaust valve is a flow restriction. Each restriction is a little gate and acts as a limiting factor, some more, some less. Obviously the easiest, and sometimes most significant restrictions to get rid of are the intake and exhaust. Some people will also change the cat and the downpipe, which essentially opens up the exhaust side. But all these parts are at the ends of the engine breathing apparatus. The intake charge still needs to go through a turbo, intercooler and all its plumbing, the throttle body, intake manifold and then into the head, all of which still play a part in pinching off air going into the cylinder. In this case, the biggest restrictions are the turbo, head and intercooler, all of which we will deal with.
Regardless of how big a turbo we strap on, it still has to blow through the stock intercooler, which could only cope with 360 wheel-hp worth of air--according to the tests we're denying we ran at Injen's all-wheel-drive Dynojet. So it was off to RRE to get Injen's 23 3/4 x 11 3/4 x 3 3/8-inch FMIC and intercooler pipe kit installed. In order to clear the hardware, we also had to change the AEM intake out for an Injen polished cast aluminum short ram intake system, which also comes with a mandrel bent upper intercooler pipe. Although smaller in appearance, Injen rates its intercooler for up to 700hp. The added flow and cooling ability comes from a much thicker core and larger core passages. The Injen intake was modified with a 2.5- to four-inch silicon hose expansion to fit over the larger compressor inlet on the new turbo.
The monstrous Garrett GT3076R we decided to use comes with a 56-trim compressor wheel and 0.60 A/R. The turbine side uses a 60mm 84-trim wheel that we chose to mate with a 0.63 A/R turbine housing. In addition to the usual ball bearing, watercooled bells and whistles, the large compressor housing also includes a ported shroud that effectively pushes the surge limits of the turbo to lower flow rates, preventing premature turbo damage.
To better reflect the output most tuners should see with this kit on a non-MIVEC Evo, we had RRE temporarily install a set of aggressive HKS 272 duration intake and exhaust cams, retarded two degrees on the intake and exhaust side via adjustable AEM cam gears. The retarded cam timing will push the power curve further up in the rev range and allow the high-lift, longer-duration HKS cams to take advantage of the added volumetric efficiency of higher rpm, since a pump gas dyno whore was what we were after. Theoretically, this further delays the turbo's spool-up response in exchange for top-end peak power.
RRE calls this the Track Whore turbo for a reason. In our experience, it lags on the street and hits like a freight train, which makes it decent for WOT ass-hauling on big tracks, but a little hard to modulate on tighter courses. But that's what we asked for in order to make big power. With just a splash of 100-octane thrown in to prevent the car from blowing up while being tuned with the AFC, this set-up put down 437 wheel-hp at 7000rpm and 367lb-ft torque at 5500rpm. Keep in mind that fuel cut comes at 7750rpm and the MR has a close ratio six-speed 'box. Unfortunately, we weren't able to log the boost on this run--too busy hanging on to look down at the boost gauge. It swept easily past the last mark at 1.5bar.
Cosworth CNC head and M2 cams
Part Two of stupid power comes in the form of a Cosworth CNC ported cylinder head with 1mm oversized stainless-steel intake and Inconel exhaust valves. The Cosworth head starts as a brand-new Mitsubishi casting (MIVEC or not, depending on which car you have) which is CNC machined to enlarge and smooth out the ports, de-shrouded in the combustion chamber area and treated to a three- and four-angle valve job for the exhaust and intake respectively.To go with the valve job, the larger valves are also back-cut on the intake side for better flow transitions and lapped for proper sealing. The head also comes with stiffer double wound springs rated for up to 10,500rpm, titanium retainers, and hardened steel platforms for these springs to sit on. Cosworth claims a maximum flow increase of 26 percent on the intake side and 18 percent on the exhaust side at half an inch of lift.
Making the most of these flow capabilities means we were again going to step up to cams more aggressive than the 10.8mm-lift HKS versions. Cosworth makes two cams, ground from new chill cast billets. The M1 cam, probably intended more for the street, has a maximum lift of 10.2mm with a open-to-close duration of 264 degrees for both the intake and exhaust, while the big brother 272-degree M2 cam, with 11mm of lift, was designed for exactly what we were after: wide open power and a very lumpy idle. Depending on the tuner and tuning tool, Cosworth claims that stock engines with this cam can be tuned to idle like stock. We left the cam gears retarded by two degrees and went back to the dyno to see where this hardware got us. Maxxed out on boost, liquored up on high octane, we pulled a Hail Mary run (something RRE is well known for) of 480 wheel-hp, just to see how much the stock bottom end and Cosworth head could cope with. This was just enough power to make the stock clutch slip even before peak torque. However impressive these numbers are, we weren't inclined to repeat them again with just the AFC in control.
Finally, Proper Tuning
We were down to four days before the car had to be picked up by Mitsubishi in stock form without any holes in the block. Our intent was to settle the issue and get the car properly setup for street gas with an ECU reflash and ROM tune by Vishnu Tuning. Before that though, we needed a clutch with enough clamping force to put up with the subsequent abuse of repeated dyno tuning. The only shop around that could meet our schedule was Tuning Technologies in Colton, California. With a reputation of a one-day turnaround on a painfully tedious Evo clutch install, we immediately sent the car over. Even the ACT Heavy Duty clutch kit with street disc was on the shelves waiting for us.
Alfred, Grand Pooh-Bah at Tuning Technologies (TT) made good on his word, and the next morning we arrived to see Project Evo already strapped down on TT's Dynojet all-wheel-drive dyno. The glazed and anemic stock clutch sat in the doorway as proof. There to greet us was yet another Grand Pooh-Bah, Shiv Pathak of Vishnu, here to personally tune our potential time bomb.
The final hurdle was to street-tune Project Evo for strictly pump gas and make the car as bulletproof as possible on our hardware. Shiv and Alfred started by disconnecting the manual boost controller and putting the stock electronic boost solenoid back in the helm. Set for around 23psi and regulated by the factory ECU and knock sensor, Project Evo set out on a series of 29 tuning pulls on the dyno. The end result was 385 wheel-hp wheel horsepower and a lot more useful low-end torque between 2500 and 4500rpm (which you can't see with our dyno chart scaled this way). Why so much less power? Gas obviously has the most to do with it. Even just a couple octane points will make a night and day difference on fire-breathing turbo cars. The other part is peak boost and the boost curve profile, since the manual boost controller held boost constant all the way to redline. The electronic controller was tuned by Vishnu to taper off much like the stock program, to ease the load on the engine. Even though the bottom end could take more, it's just not smart on a daily driver.
Project Evo's last day was spent at the track. Knowing the limitations of the weaker MR six-speed, we set out to get one last respectable run on all this power. With the boost lag and fragile tranny, we couldn't just dump the clutch at 6000rpm, which is what the car requires for a fast ET. We did manage to finally squeeze out a 12.8-second run trapping at 100mph. Not impressive, but no slouch compared to the 14 seconds on the stock car. This time is roughly what we found in our last Evo Tuner shootout.
Coming to terms
Project Evo VIII MR is done. Unlike our other project cars that live as long as we're willing to come up with bad ideas to throw at them, the Evo has been returned to Mitsubishi after a gracious loan of two years and will probably end up in the hands of some undeserving owner. We still think of Project Evo fondly now it's gone.
This was by far our all-time favorite project car throughout the years. But our fondness is of the Evo in its near stock form, which provided endless hours of giggling--in traffic or on the track. Even in bone stock form, it went out and destroyed everything else in our project car garage in the rain during the last Project Car battle. The car is just that brilliant. And as good as the parts are that we put in, we've made some compromises in ride, noise, and lag.
If we had to do this all over again, I would have just replaced the pads with something less aggressive than we had used, stuck with the stock turbo and a quieter exhaust, and used a suspension with more emphasis on street driving. We go to the track, but that doesn't mean we avoid sitting in traffic along with everyone else.
Our next Mitsubishi project won't be an Evo, we'll hopefully be getting our hands on the Ralliart Lancer, a mean and affordable cousin of the Evo X and the Lancer GTS.
It was a miracle Project S2000 made it from San Diego to Los Angeles the day we bought it. The stock rear S-03s were so worn on the inside that there was no longer any rubber visible, the remaining steel belts actually reflecting headlights of the car behind.
Like any enthusiast would, we took the worn tires as the perfect excuse to splurge on some new wheels and rubber. Mackin Industries had just the rollers for the job: a set of black Volk Racing TE-37s. These days, people like to run 18-inch wheels or bigger on relatively small cars like the S2000. But since ours is going to the track, not MTV Cribs, a one-inch diameter hike to 17 inches is perfect.
Off the shelf, it's tough to beat a good ol' fashioned TE-37 when it comes to light weight. The 17x7.5-inch front wheels weigh 15.6 pounds a pop, and each of the nine-inch wide rears weighs 17.1 pounds. Combined with the higher profile afforded by 17-inch rubber, the forged wheels should be pretty much indestructible on the road. And while we're not exactly sticklers for aesthetics, they look pretty darn good, too.
We decided to wrap our new metal in a set of BF Goodrich KDs measuring 225/45R17 up front, and 255/40R17 in the back. This front-to-rear ratio should keep a similar balance to stock, while adding a small safety margin to guard against the snap oversteer that plagues 2000 model year S2000s. We stayed relatively conservative with the tire compound for now, in an attempt to keep the car streetable and match grip with engine output. In the future, we'll install a host of power enhancers that will necessitate the increased traction offered by R-compound tires. When that time comes, we'll swap out for BF Goodrich's new (to the US) R1.
It's tough to improve on the stock S2000 suspension. The car comes with external reservoir shocks that, combined with well-chosen spring rates, make it pretty capable in the twisty stuff. But since we were about to add a couple of pounds of unsprung weight to each corner, we needed something with quick reaction time and good rebound damping. Something a little more geared for track use.
The answer came in the form of KW's Variant III coilovers. Built for the weekend racer, they offer independently adjustable rebound and compression, and we can pick our own spring rates before they get here. Given the wide range of adjustability and KW's reputation in racing, they should help fling the S2000 around The Streets of Willow, our local race circuit. And thanks to the stainless steel construction and the preset high-speed damping curve, they might actually be bearable during daily use.
We decided to stick with the recommended spring rates, which meant 515lb/in for both front and rear. The spring rates being equal may cause a pogo-stick effect at certain speeds in a car with a 50/50 weight distribution, but we're willing to take the chance in exchange for retaining the neutral handling of the stock suspension.
Due to the limited travel of the rear suspension and the arc in which it travels, it's important not to go too far when lowering an S2000. KW allows for an adjustment range of 1.2 inches in the front, and an inch out back. That doesn't refer to how much lower the car gets to the pavement, it just means we can raise or lower the ride height over a spread of 1.2 inches or one inch, respectively.
You can make the car higher or lower than recommended by a small degree if you want, but KW suggests that the shock would then be out of optimal operating range. Based on the results we got when working within the recommended range, we see no reason to set the coilovers otherwise.
M-Workz in Gardena, California, is the end result of a lifetime in cars for a guy named Steve Mitchell. He's raced and built cars professionally for decades and he's spent the past 18 years at Nissan, engineering every imaginable aspect of its cars. So we figured he could handle the task of installing our suspension and corner-balancing the car.
The Wheel Supply in Industry, California, handled the task of mating Volk to BFG. We installed the finished product and took a quick measurement of the wheel well gap-we didn't want to stray too far from stock. There was a 2.75-inch gap up front and a full three inches in the rear. Mitchell got to work dismantling the stock stuff and installing the coilovers.
Installation couldn't have gone smoother. Mitchell likes to whistle while he works-incredibly irritating for any journalist who might be taking pictures, but at least it's a sign that he likes what he does. Springs mounted on shocks, we were ready to mount the suspension. The front was a bigger pain in the ass than the rear, as the upper A-arm had to be removed in order to slide the spring/shock assembly in (and slide the old one out). The rear shock and spring on the S2000 sit behind the moving geometry, so barring digging through the trunk to unbolt the strut towers, installation and removal was much more straightforward.
Suspension mounted and height loosely guesstimated, it was time to move on to corner balancing. "There's a critical step many tuners miss when corner-balancing a car," blurts Mitchell. "It's called leveling the floor. It's common sense, really. If your car is sitting at even a slight angle, weight is going to be shifted down the slope, toward the lower end of the car."
Armed with his laser leveler, Mitchell aligned the four tables that would support both the scales and the car. His have additional run-off area (roll-off pads), so the car can be adjusted and rolled on and off in a snap. We still weren't ready to start. Mitchell laid a towel down in the driver's seat before loading my exact weight in sandbags (that number to remain confidential).
When the S2000 first hit the scales, we recorded 2974 pounds-with a full tank of gas and the sandbag fat-ass in the driver's seat, which means we have an unnaturally light S2000 on our hands. According to the US Department of Energy, a gallon of gas weighs anywhere from 5.8 to 6.5 pounds. Let's call it 6.15 pounds. The tank's 13.2 gallons, so figure on about 81 pounds in gas alone. Tack on the sandbags, which, I'll regretfully admit, weigh 200 pounds, and we have a total weight of 2693 pounds. Like I said, a light S2000. Oh, and yeah, I'm on a diet, so you can all just stop writing the e-mails now.
What we're concerned with, though, is weight distribution. As it sat, the car showed a cross weight distribution of 48.8 percent from right front to left rear (and 51.2 percent from LF to RR). Not bad. When we were done, we had managed to equalize the cross weights for near 50/50 distribution. We adjusted the front spring perches to 60mm and the rear to 50mm, closer to the low side than the high side of the adjustment range, but nowhere near over-lowered. The wheel well gap had shortened by 1.5 inches up front and an inch out back, for a total gap of 1.25 and two inches respectively.
Before we sent the S2000 back out into the wild, we wanted to figure out what had caused the inside of the stock S03s to wear so rapidly. Unsurprisingly, the toe was way off, measuring a remarkable 9.5mm toe-in on each side, which is about 3.5mm out of max spec. And that's a ton. Often, the toe is adjusted in an attempt to compensate for the car's natural tendency for sudden oversteer, when the rear suspension compresses. After double-checking the front toe at zero millimeters, we took the toe back out to max stock specs (6mm a side). Because put simply, over-toeing in the car is a crappy band-aid for the real problem, which is toe-out under compression.
The first test of the revitalized S2000 was on the rugged roads that surround M-Workz. We left the rebound and compression at their factory recommended settings. We were utterly shocked to find that the ride was, if anything, smoother than stock, absorbing imperfections in the road without overreacting to potholes or large dips. At the same time, there was no perceptible body roll. It'd take a day at the track to really dial the suspension in, because for street use, it was more than capable.
And so we took the car to the Streets of Willow for a few shakedown laps. Right off the bat, the car was brilliant-even Mitchell was impressed with its ability to turn in and carry speed through technical sections. After messing with the coilovers for the better part of four hours, we ended up hovering right around the factory settings. A further quarter-turn of rebound up front was all that was needed to perfect the set-up for Streets' bumpy surface.
Does the car oversteer? Yep. But we're not going to dull its responses by toeing in the rear. We'll tame the rambunctious tail with a bump steer kit, but that's for another installment.
Before putting the S2000 away wet, it was off to our proving grounds in Fontana, Cailfornia. It was positively adjustable through the figure-eight test course, slight lifts of the throttle rotating the car around to find a perfect line in and out of the long corners. When the dust had settled, the S2000 net a 25.9-second average. To put this in perspective. A stock 2006 WRX runs the figure eight in 27 seconds and a Porsche 911 (997) GT3 will do it in 24. On the skidpad, the car managed an impressive 0.9g.
As hard as we try, we can't find the need for any further suspension modifications (short of a bump-steer kit). Thicker anti-roll bars would do nothing to improve the car's perfect balance. Amazingly, we seem to have conquered the suspension situation in one installment. Watch us attempt to tune some individual throttle bodies from Fuji Racing next time around.
Even with all the mentions of future ECU tuning plans in our previous installments, we still received a flood of complaints about Project Evo VIII MR and why we chose to save costs and use a temporary piggy-back fuel computer when the suspension cost almost $5000.
Thus far, our Evo has been through some mild power tuning and significant chassis work, all done at Road Race Engineering and Road Race Chassis in Santa Fe Springs, California. There's a reason why it seems like an RRE exclusive car: RRE has a complete tuning program and is familiar with what works and what doesn't with the hardware that's already on. So guess who we took it back to for its final big power installation. Why shouldn't we? No one would take their car to one body shop for body work and then to another for paint.
Now here's the disclaimer: we liked the Evo the way it was (aside from temporary AFC tuning) and knew what we were getting into when we asked for big power involving a Garrett GT3076R turbo, Cosworth CNC head and lumpy cams. This is purely an experiment on how much useable and reliable power the stock bottom end would put up with on 91-octane pump gas. Take the term 'useable' with a grain of salt.
RRE Evo VIII GT turbo kit
We start with the heart of the newfound power: RRE's functional but no frills GT turbo kit. RRE assumes you already have all the basic bolt-on plumbing for intake and exhaust, so only includes the turbo hardware itself, hence the term `turbo kit'. Along with a selection of four different types of turbos, our kit comes with a cast, smooth bending T3 turbo manifold, all the plumbing, TiAl 44mm wastegate, trick turbine outlet housing and wastegate plumbing. More impressive is all the thoroughly thought-out additional hardware, including manifold studs, gaskets, proper nuts and washers, V-band clamps for the wastegate, coolant lines, flexible hard oil return pipe, and even an in-line oil filter for the turbo.
To complete the kit, you can pick from an E-Pinions Fashion ball bearing 50-trim turbo, Street Power/Track Whore ball bearing GT3076R, Track Whore Deluxe custom ball bearing GT35R or Drag Queen/Race Gas Slut ball bearing full GT35R. The intercooler and intake plumbing might have to be modified, depending on the turbo. We went with the GT3076 Street Power/Track Whore turbo, which, in theory, can flow up to 525 wheel-hp. How much horsepower can be squeezed out is limited by the clutch, bottom end and how much fuel we can keep pouring into the cylinders.
Since the car had to go back to Mitsubishi, the original plan was to cut some corners and just strap on the turbo kit, temporarily tune the AFC and throw the car on the dyno before adding the Cosworth CNC head and doing a final tuned ECU reflash. Don't ask how we know, but this is a pointless exercise with the stock intercooler. Like the intake and exhaust, everything the air has to flow past to get to the intake valve and beyond the exhaust valve is a flow restriction. Each restriction is a little gate and acts as a limiting factor, some more, some less. Obviously the easiest, and sometimes most significant restrictions to get rid of are the intake and exhaust. Some people will also change the cat and the downpipe, which essentially opens up the exhaust side. But all these parts are at the ends of the engine breathing apparatus. The intake charge still needs to go through a turbo, intercooler and all its plumbing, the throttle body, intake manifold and then into the head, all of which still play a part in pinching off air going into the cylinder. In this case, the biggest restrictions are the turbo, head and intercooler, all of which we will deal with.
Regardless of how big a turbo we strap on, it still has to blow through the stock intercooler, which could only cope with 360 wheel-hp worth of air--according to the tests we're denying we ran at Injen's all-wheel-drive Dynojet. So it was off to RRE to get Injen's 23 3/4 x 11 3/4 x 3 3/8-inch FMIC and intercooler pipe kit installed. In order to clear the hardware, we also had to change the AEM intake out for an Injen polished cast aluminum short ram intake system, which also comes with a mandrel bent upper intercooler pipe. Although smaller in appearance, Injen rates its intercooler for up to 700hp. The added flow and cooling ability comes from a much thicker core and larger core passages. The Injen intake was modified with a 2.5- to four-inch silicon hose expansion to fit over the larger compressor inlet on the new turbo.
The monstrous Garrett GT3076R we decided to use comes with a 56-trim compressor wheel and 0.60 A/R. The turbine side uses a 60mm 84-trim wheel that we chose to mate with a 0.63 A/R turbine housing. In addition to the usual ball bearing, watercooled bells and whistles, the large compressor housing also includes a ported shroud that effectively pushes the surge limits of the turbo to lower flow rates, preventing premature turbo damage.
To better reflect the output most tuners should see with this kit on a non-MIVEC Evo, we had RRE temporarily install a set of aggressive HKS 272 duration intake and exhaust cams, retarded two degrees on the intake and exhaust side via adjustable AEM cam gears. The retarded cam timing will push the power curve further up in the rev range and allow the high-lift, longer-duration HKS cams to take advantage of the added volumetric efficiency of higher rpm, since a pump gas dyno whore was what we were after. Theoretically, this further delays the turbo's spool-up response in exchange for top-end peak power.
RRE calls this the Track Whore turbo for a reason. In our experience, it lags on the street and hits like a freight train, which makes it decent for WOT ass-hauling on big tracks, but a little hard to modulate on tighter courses. But that's what we asked for in order to make big power. With just a splash of 100-octane thrown in to prevent the car from blowing up while being tuned with the AFC, this set-up put down 437 wheel-hp at 7000rpm and 367lb-ft torque at 5500rpm. Keep in mind that fuel cut comes at 7750rpm and the MR has a close ratio six-speed 'box. Unfortunately, we weren't able to log the boost on this run--too busy hanging on to look down at the boost gauge. It swept easily past the last mark at 1.5bar.
Cosworth CNC head and M2 cams
Part Two of stupid power comes in the form of a Cosworth CNC ported cylinder head with 1mm oversized stainless-steel intake and Inconel exhaust valves. The Cosworth head starts as a brand-new Mitsubishi casting (MIVEC or not, depending on which car you have) which is CNC machined to enlarge and smooth out the ports, de-shrouded in the combustion chamber area and treated to a three- and four-angle valve job for the exhaust and intake respectively.To go with the valve job, the larger valves are also back-cut on the intake side for better flow transitions and lapped for proper sealing. The head also comes with stiffer double wound springs rated for up to 10,500rpm, titanium retainers, and hardened steel platforms for these springs to sit on. Cosworth claims a maximum flow increase of 26 percent on the intake side and 18 percent on the exhaust side at half an inch of lift.
Making the most of these flow capabilities means we were again going to step up to cams more aggressive than the 10.8mm-lift HKS versions. Cosworth makes two cams, ground from new chill cast billets. The M1 cam, probably intended more for the street, has a maximum lift of 10.2mm with a open-to-close duration of 264 degrees for both the intake and exhaust, while the big brother 272-degree M2 cam, with 11mm of lift, was designed for exactly what we were after: wide open power and a very lumpy idle. Depending on the tuner and tuning tool, Cosworth claims that stock engines with this cam can be tuned to idle like stock. We left the cam gears retarded by two degrees and went back to the dyno to see where this hardware got us. Maxxed out on boost, liquored up on high octane, we pulled a Hail Mary run (something RRE is well known for) of 480 wheel-hp, just to see how much the stock bottom end and Cosworth head could cope with. This was just enough power to make the stock clutch slip even before peak torque. However impressive these numbers are, we weren't inclined to repeat them again with just the AFC in control.
Finally, Proper Tuning
We were down to four days before the car had to be picked up by Mitsubishi in stock form without any holes in the block. Our intent was to settle the issue and get the car properly setup for street gas with an ECU reflash and ROM tune by Vishnu Tuning. Before that though, we needed a clutch with enough clamping force to put up with the subsequent abuse of repeated dyno tuning. The only shop around that could meet our schedule was Tuning Technologies in Colton, California. With a reputation of a one-day turnaround on a painfully tedious Evo clutch install, we immediately sent the car over. Even the ACT Heavy Duty clutch kit with street disc was on the shelves waiting for us.
Alfred, Grand Pooh-Bah at Tuning Technologies (TT) made good on his word, and the next morning we arrived to see Project Evo already strapped down on TT's Dynojet all-wheel-drive dyno. The glazed and anemic stock clutch sat in the doorway as proof. There to greet us was yet another Grand Pooh-Bah, Shiv Pathak of Vishnu, here to personally tune our potential time bomb.
The final hurdle was to street-tune Project Evo for strictly pump gas and make the car as bulletproof as possible on our hardware. Shiv and Alfred started by disconnecting the manual boost controller and putting the stock electronic boost solenoid back in the helm. Set for around 23psi and regulated by the factory ECU and knock sensor, Project Evo set out on a series of 29 tuning pulls on the dyno. The end result was 385 wheel-hp wheel horsepower and a lot more useful low-end torque between 2500 and 4500rpm (which you can't see with our dyno chart scaled this way). Why so much less power? Gas obviously has the most to do with it. Even just a couple octane points will make a night and day difference on fire-breathing turbo cars. The other part is peak boost and the boost curve profile, since the manual boost controller held boost constant all the way to redline. The electronic controller was tuned by Vishnu to taper off much like the stock program, to ease the load on the engine. Even though the bottom end could take more, it's just not smart on a daily driver.
Project Evo's last day was spent at the track. Knowing the limitations of the weaker MR six-speed, we set out to get one last respectable run on all this power. With the boost lag and fragile tranny, we couldn't just dump the clutch at 6000rpm, which is what the car requires for a fast ET. We did manage to finally squeeze out a 12.8-second run trapping at 100mph. Not impressive, but no slouch compared to the 14 seconds on the stock car. This time is roughly what we found in our last Evo Tuner shootout.
Coming to terms
Project Evo VIII MR is done. Unlike our other project cars that live as long as we're willing to come up with bad ideas to throw at them, the Evo has been returned to Mitsubishi after a gracious loan of two years and will probably end up in the hands of some undeserving owner. We still think of Project Evo fondly now it's gone.
This was by far our all-time favorite project car throughout the years. But our fondness is of the Evo in its near stock form, which provided endless hours of giggling--in traffic or on the track. Even in bone stock form, it went out and destroyed everything else in our project car garage in the rain during the last Project Car battle. The car is just that brilliant. And as good as the parts are that we put in, we've made some compromises in ride, noise, and lag.
If we had to do this all over again, I would have just replaced the pads with something less aggressive than we had used, stuck with the stock turbo and a quieter exhaust, and used a suspension with more emphasis on street driving. We go to the track, but that doesn't mean we avoid sitting in traffic along with everyone else.
Our next Mitsubishi project won't be an Evo, we'll hopefully be getting our hands on the Ralliart Lancer, a mean and affordable cousin of the Evo X and the Lancer GTS.
It was a miracle Project S2000 made it from San Diego to Los Angeles the day we bought it. The stock rear S-03s were so worn on the inside that there was no longer any rubber visible, the remaining steel belts actually reflecting headlights of the car behind.
Like any enthusiast would, we took the worn tires as the perfect excuse to splurge on some new wheels and rubber. Mackin Industries had just the rollers for the job: a set of black Volk Racing TE-37s. These days, people like to run 18-inch wheels or bigger on relatively small cars like the S2000. But since ours is going to the track, not MTV Cribs, a one-inch diameter hike to 17 inches is perfect.
Off the shelf, it's tough to beat a good ol' fashioned TE-37 when it comes to light weight. The 17x7.5-inch front wheels weigh 15.6 pounds a pop, and each of the nine-inch wide rears weighs 17.1 pounds. Combined with the higher profile afforded by 17-inch rubber, the forged wheels should be pretty much indestructible on the road. And while we're not exactly sticklers for aesthetics, they look pretty darn good, too.
We decided to wrap our new metal in a set of BF Goodrich KDs measuring 225/45R17 up front, and 255/40R17 in the back. This front-to-rear ratio should keep a similar balance to stock, while adding a small safety margin to guard against the snap oversteer that plagues 2000 model year S2000s. We stayed relatively conservative with the tire compound for now, in an attempt to keep the car streetable and match grip with engine output. In the future, we'll install a host of power enhancers that will necessitate the increased traction offered by R-compound tires. When that time comes, we'll swap out for BF Goodrich's new (to the US) R1.
It's tough to improve on the stock S2000 suspension. The car comes with external reservoir shocks that, combined with well-chosen spring rates, make it pretty capable in the twisty stuff. But since we were about to add a couple of pounds of unsprung weight to each corner, we needed something with quick reaction time and good rebound damping. Something a little more geared for track use.
The answer came in the form of KW's Variant III coilovers. Built for the weekend racer, they offer independently adjustable rebound and compression, and we can pick our own spring rates before they get here. Given the wide range of adjustability and KW's reputation in racing, they should help fling the S2000 around The Streets of Willow, our local race circuit. And thanks to the stainless steel construction and the preset high-speed damping curve, they might actually be bearable during daily use.
We decided to stick with the recommended spring rates, which meant 515lb/in for both front and rear. The spring rates being equal may cause a pogo-stick effect at certain speeds in a car with a 50/50 weight distribution, but we're willing to take the chance in exchange for retaining the neutral handling of the stock suspension.
Due to the limited travel of the rear suspension and the arc in which it travels, it's important not to go too far when lowering an S2000. KW allows for an adjustment range of 1.2 inches in the front, and an inch out back. That doesn't refer to how much lower the car gets to the pavement, it just means we can raise or lower the ride height over a spread of 1.2 inches or one inch, respectively.
You can make the car higher or lower than recommended by a small degree if you want, but KW suggests that the shock would then be out of optimal operating range. Based on the results we got when working within the recommended range, we see no reason to set the coilovers otherwise.
M-Workz in Gardena, California, is the end result of a lifetime in cars for a guy named Steve Mitchell. He's raced and built cars professionally for decades and he's spent the past 18 years at Nissan, engineering every imaginable aspect of its cars. So we figured he could handle the task of installing our suspension and corner-balancing the car.
The Wheel Supply in Industry, California, handled the task of mating Volk to BFG. We installed the finished product and took a quick measurement of the wheel well gap-we didn't want to stray too far from stock. There was a 2.75-inch gap up front and a full three inches in the rear. Mitchell got to work dismantling the stock stuff and installing the coilovers.
Installation couldn't have gone smoother. Mitchell likes to whistle while he works-incredibly irritating for any journalist who might be taking pictures, but at least it's a sign that he likes what he does. Springs mounted on shocks, we were ready to mount the suspension. The front was a bigger pain in the ass than the rear, as the upper A-arm had to be removed in order to slide the spring/shock assembly in (and slide the old one out). The rear shock and spring on the S2000 sit behind the moving geometry, so barring digging through the trunk to unbolt the strut towers, installation and removal was much more straightforward.
Suspension mounted and height loosely guesstimated, it was time to move on to corner balancing. "There's a critical step many tuners miss when corner-balancing a car," blurts Mitchell. "It's called leveling the floor. It's common sense, really. If your car is sitting at even a slight angle, weight is going to be shifted down the slope, toward the lower end of the car."
Armed with his laser leveler, Mitchell aligned the four tables that would support both the scales and the car. His have additional run-off area (roll-off pads), so the car can be adjusted and rolled on and off in a snap. We still weren't ready to start. Mitchell laid a towel down in the driver's seat before loading my exact weight in sandbags (that number to remain confidential).
When the S2000 first hit the scales, we recorded 2974 pounds-with a full tank of gas and the sandbag fat-ass in the driver's seat, which means we have an unnaturally light S2000 on our hands. According to the US Department of Energy, a gallon of gas weighs anywhere from 5.8 to 6.5 pounds. Let's call it 6.15 pounds. The tank's 13.2 gallons, so figure on about 81 pounds in gas alone. Tack on the sandbags, which, I'll regretfully admit, weigh 200 pounds, and we have a total weight of 2693 pounds. Like I said, a light S2000. Oh, and yeah, I'm on a diet, so you can all just stop writing the e-mails now.
What we're concerned with, though, is weight distribution. As it sat, the car showed a cross weight distribution of 48.8 percent from right front to left rear (and 51.2 percent from LF to RR). Not bad. When we were done, we had managed to equalize the cross weights for near 50/50 distribution. We adjusted the front spring perches to 60mm and the rear to 50mm, closer to the low side than the high side of the adjustment range, but nowhere near over-lowered. The wheel well gap had shortened by 1.5 inches up front and an inch out back, for a total gap of 1.25 and two inches respectively.
Before we sent the S2000 back out into the wild, we wanted to figure out what had caused the inside of the stock S03s to wear so rapidly. Unsurprisingly, the toe was way off, measuring a remarkable 9.5mm toe-in on each side, which is about 3.5mm out of max spec. And that's a ton. Often, the toe is adjusted in an attempt to compensate for the car's natural tendency for sudden oversteer, when the rear suspension compresses. After double-checking the front toe at zero millimeters, we took the toe back out to max stock specs (6mm a side). Because put simply, over-toeing in the car is a crappy band-aid for the real problem, which is toe-out under compression.
The first test of the revitalized S2000 was on the rugged roads that surround M-Workz. We left the rebound and compression at their factory recommended settings. We were utterly shocked to find that the ride was, if anything, smoother than stock, absorbing imperfections in the road without overreacting to potholes or large dips. At the same time, there was no perceptible body roll. It'd take a day at the track to really dial the suspension in, because for street use, it was more than capable.
And so we took the car to the Streets of Willow for a few shakedown laps. Right off the bat, the car was brilliant-even Mitchell was impressed with its ability to turn in and carry speed through technical sections. After messing with the coilovers for the better part of four hours, we ended up hovering right around the factory settings. A further quarter-turn of rebound up front was all that was needed to perfect the set-up for Streets' bumpy surface.
Does the car oversteer? Yep. But we're not going to dull its responses by toeing in the rear. We'll tame the rambunctious tail with a bump steer kit, but that's for another installment.
Before putting the S2000 away wet, it was off to our proving grounds in Fontana, Cailfornia. It was positively adjustable through the figure-eight test course, slight lifts of the throttle rotating the car around to find a perfect line in and out of the long corners. When the dust had settled, the S2000 net a 25.9-second average. To put this in perspective. A stock 2006 WRX runs the figure eight in 27 seconds and a Porsche 911 (997) GT3 will do it in 24. On the skidpad, the car managed an impressive 0.9g.
As hard as we try, we can't find the need for any further suspension modifications (short of a bump-steer kit). Thicker anti-roll bars would do nothing to improve the car's perfect balance. Amazingly, we seem to have conquered the suspension situation in one installment. Watch us attempt to tune some individual throttle bodies from Fuji Racing next time around.
Even with all the mentions of future ECU tuning plans in our previous installments, we still received a flood of complaints about Project Evo VIII MR and why we chose to save costs and use a temporary piggy-back fuel computer when the suspension cost almost $5000.
Thus far, our Evo has been through some mild power tuning and significant chassis work, all done at Road Race Engineering and Road Race Chassis in Santa Fe Springs, California. There's a reason why it seems like an RRE exclusive car: RRE has a complete tuning program and is familiar with what works and what doesn't with the hardware that's already on. So guess who we took it back to for its final big power installation. Why shouldn't we? No one would take their car to one body shop for body work and then to another for paint.
Now here's the disclaimer: we liked the Evo the way it was (aside from temporary AFC tuning) and knew what we were getting into when we asked for big power involving a Garrett GT3076R turbo, Cosworth CNC head and lumpy cams. This is purely an experiment on how much useable and reliable power the stock bottom end would put up with on 91-octane pump gas. Take the term 'useable' with a grain of salt.
RRE Evo VIII GT turbo kit
We start with the heart of the newfound power: RRE's functional but no frills GT turbo kit. RRE assumes you already have all the basic bolt-on plumbing for intake and exhaust, so only includes the turbo hardware itself, hence the term `turbo kit'. Along with a selection of four different types of turbos, our kit comes with a cast, smooth bending T3 turbo manifold, all the plumbing, TiAl 44mm wastegate, trick turbine outlet housing and wastegate plumbing. More impressive is all the thoroughly thought-out additional hardware, including manifold studs, gaskets, proper nuts and washers, V-band clamps for the wastegate, coolant lines, flexible hard oil return pipe, and even an in-line oil filter for the turbo.
To complete the kit, you can pick from an E-Pinions Fashion ball bearing 50-trim turbo, Street Power/Track Whore ball bearing GT3076R, Track Whore Deluxe custom ball bearing GT35R or Drag Queen/Race Gas Slut ball bearing full GT35R. The intercooler and intake plumbing might have to be modified, depending on the turbo. We went with the GT3076 Street Power/Track Whore turbo, which, in theory, can flow up to 525 wheel-hp. How much horsepower can be squeezed out is limited by the clutch, bottom end and how much fuel we can keep pouring into the cylinders.
Since the car had to go back to Mitsubishi, the original plan was to cut some corners and just strap on the turbo kit, temporarily tune the AFC and throw the car on the dyno before adding the Cosworth CNC head and doing a final tuned ECU reflash. Don't ask how we know, but this is a pointless exercise with the stock intercooler. Like the intake and exhaust, everything the air has to flow past to get to the intake valve and beyond the exhaust valve is a flow restriction. Each restriction is a little gate and acts as a limiting factor, some more, some less. Obviously the easiest, and sometimes most significant restrictions to get rid of are the intake and exhaust. Some people will also change the cat and the downpipe, which essentially opens up the exhaust side. But all these parts are at the ends of the engine breathing apparatus. The intake charge still needs to go through a turbo, intercooler and all its plumbing, the throttle body, intake manifold and then into the head, all of which still play a part in pinching off air going into the cylinder. In this case, the biggest restrictions are the turbo, head and intercooler, all of which we will deal with.
Regardless of how big a turbo we strap on, it still has to blow through the stock intercooler, which could only cope with 360 wheel-hp worth of air--according to the tests we're denying we ran at Injen's all-wheel-drive Dynojet. So it was off to RRE to get Injen's 23 3/4 x 11 3/4 x 3 3/8-inch FMIC and intercooler pipe kit installed. In order to clear the hardware, we also had to change the AEM intake out for an Injen polished cast aluminum short ram intake system, which also comes with a mandrel bent upper intercooler pipe. Although smaller in appearance, Injen rates its intercooler for up to 700hp. The added flow and cooling ability comes from a much thicker core and larger core passages. The Injen intake was modified with a 2.5- to four-inch silicon hose expansion to fit over the larger compressor inlet on the new turbo.
The monstrous Garrett GT3076R we decided to use comes with a 56-trim compressor wheel and 0.60 A/R. The turbine side uses a 60mm 84-trim wheel that we chose to mate with a 0.63 A/R turbine housing. In addition to the usual ball bearing, watercooled bells and whistles, the large compressor housing also includes a ported shroud that effectively pushes the surge limits of the turbo to lower flow rates, preventing premature turbo damage.
To better reflect the output most tuners should see with this kit on a non-MIVEC Evo, we had RRE temporarily install a set of aggressive HKS 272 duration intake and exhaust cams, retarded two degrees on the intake and exhaust side via adjustable AEM cam gears. The retarded cam timing will push the power curve further up in the rev range and allow the high-lift, longer-duration HKS cams to take advantage of the added volumetric efficiency of higher rpm, since a pump gas dyno whore was what we were after. Theoretically, this further delays the turbo's spool-up response in exchange for top-end peak power.
RRE calls this the Track Whore turbo for a reason. In our experience, it lags on the street and hits like a freight train, which makes it decent for WOT ass-hauling on big tracks, but a little hard to modulate on tighter courses. But that's what we asked for in order to make big power. With just a splash of 100-octane thrown in to prevent the car from blowing up while being tuned with the AFC, this set-up put down 437 wheel-hp at 7000rpm and 367lb-ft torque at 5500rpm. Keep in mind that fuel cut comes at 7750rpm and the MR has a close ratio six-speed 'box. Unfortunately, we weren't able to log the boost on this run--too busy hanging on to look down at the boost gauge. It swept easily past the last mark at 1.5bar.
Cosworth CNC head and M2 cams
Part Two of stupid power comes in the form of a Cosworth CNC ported cylinder head with 1mm oversized stainless-steel intake and Inconel exhaust valves. The Cosworth head starts as a brand-new Mitsubishi casting (MIVEC or not, depending on which car you have) which is CNC machined to enlarge and smooth out the ports, de-shrouded in the combustion chamber area and treated to a three- and four-angle valve job for the exhaust and intake respectively.To go with the valve job, the larger valves are also back-cut on the intake side for better flow transitions and lapped for proper sealing. The head also comes with stiffer double wound springs rated for up to 10,500rpm, titanium retainers, and hardened steel platforms for these springs to sit on. Cosworth claims a maximum flow increase of 26 percent on the intake side and 18 percent on the exhaust side at half an inch of lift.
Making the most of these flow capabilities means we were again going to step up to cams more aggressive than the 10.8mm-lift HKS versions. Cosworth makes two cams, ground from new chill cast billets. The M1 cam, probably intended more for the street, has a maximum lift of 10.2mm with a open-to-close duration of 264 degrees for both the intake and exhaust, while the big brother 272-degree M2 cam, with 11mm of lift, was designed for exactly what we were after: wide open power and a very lumpy idle. Depending on the tuner and tuning tool, Cosworth claims that stock engines with this cam can be tuned to idle like stock. We left the cam gears retarded by two degrees and went back to the dyno to see where this hardware got us. Maxxed out on boost, liquored up on high octane, we pulled a Hail Mary run (something RRE is well known for) of 480 wheel-hp, just to see how much the stock bottom end and Cosworth head could cope with. This was just enough power to make the stock clutch slip even before peak torque. However impressive these numbers are, we weren't inclined to repeat them again with just the AFC in control.
Finally, Proper Tuning
We were down to four days before the car had to be picked up by Mitsubishi in stock form without any holes in the block. Our intent was to settle the issue and get the car properly setup for street gas with an ECU reflash and ROM tune by Vishnu Tuning. Before that though, we needed a clutch with enough clamping force to put up with the subsequent abuse of repeated dyno tuning. The only shop around that could meet our schedule was Tuning Technologies in Colton, California. With a reputation of a one-day turnaround on a painfully tedious Evo clutch install, we immediately sent the car over. Even the ACT Heavy Duty clutch kit with street disc was on the shelves waiting for us.
Alfred, Grand Pooh-Bah at Tuning Technologies (TT) made good on his word, and the next morning we arrived to see Project Evo already strapped down on TT's Dynojet all-wheel-drive dyno. The glazed and anemic stock clutch sat in the doorway as proof. There to greet us was yet another Grand Pooh-Bah, Shiv Pathak of Vishnu, here to personally tune our potential time bomb.
The final hurdle was to street-tune Project Evo for strictly pump gas and make the car as bulletproof as possible on our hardware. Shiv and Alfred started by disconnecting the manual boost controller and putting the stock electronic boost solenoid back in the helm. Set for around 23psi and regulated by the factory ECU and knock sensor, Project Evo set out on a series of 29 tuning pulls on the dyno. The end result was 385 wheel-hp wheel horsepower and a lot more useful low-end torque between 2500 and 4500rpm (which you can't see with our dyno chart scaled this way). Why so much less power? Gas obviously has the most to do with it. Even just a couple octane points will make a night and day difference on fire-breathing turbo cars. The other part is peak boost and the boost curve profile, since the manual boost controller held boost constant all the way to redline. The electronic controller was tuned by Vishnu to taper off much like the stock program, to ease the load on the engine. Even though the bottom end could take more, it's just not smart on a daily driver.
Project Evo's last day was spent at the track. Knowing the limitations of the weaker MR six-speed, we set out to get one last respectable run on all this power. With the boost lag and fragile tranny, we couldn't just dump the clutch at 6000rpm, which is what the car requires for a fast ET. We did manage to finally squeeze out a 12.8-second run trapping at 100mph. Not impressive, but no slouch compared to the 14 seconds on the stock car. This time is roughly what we found in our last Evo Tuner shootout.
Coming to terms
Project Evo VIII MR is done. Unlike our other project cars that live as long as we're willing to come up with bad ideas to throw at them, the Evo has been returned to Mitsubishi after a gracious loan of two years and will probably end up in the hands of some undeserving owner. We still think of Project Evo fondly now it's gone.
This was by far our all-time favorite project car throughout the years. But our fondness is of the Evo in its near stock form, which provided endless hours of giggling--in traffic or on the track. Even in bone stock form, it went out and destroyed everything else in our project car garage in the rain during the last Project Car battle. The car is just that brilliant. And as good as the parts are that we put in, we've made some compromises in ride, noise, and lag.
If we had to do this all over again, I would have just replaced the pads with something less aggressive than we had used, stuck with the stock turbo and a quieter exhaust, and used a suspension with more emphasis on street driving. We go to the track, but that doesn't mean we avoid sitting in traffic along with everyone else.
Our next Mitsubishi project won't be an Evo, we'll hopefully be getting our hands on the Ralliart Lancer, a mean and affordable cousin of the Evo X and the Lancer GTS.
It was a miracle Project S2000 made it from San Diego to Los Angeles the day we bought it. The stock rear S-03s were so worn on the inside that there was no longer any rubber visible, the remaining steel belts actually reflecting headlights of the car behind.
Like any enthusiast would, we took the worn tires as the perfect excuse to splurge on some new wheels and rubber. Mackin Industries had just the rollers for the job: a set of black Volk Racing TE-37s. These days, people like to run 18-inch wheels or bigger on relatively small cars like the S2000. But since ours is going to the track, not MTV Cribs, a one-inch diameter hike to 17 inches is perfect.
Off the shelf, it's tough to beat a good ol' fashioned TE-37 when it comes to light weight. The 17x7.5-inch front wheels weigh 15.6 pounds a pop, and each of the nine-inch wide rears weighs 17.1 pounds. Combined with the higher profile afforded by 17-inch rubber, the forged wheels should be pretty much indestructible on the road. And while we're not exactly sticklers for aesthetics, they look pretty darn good, too.
We decided to wrap our new metal in a set of BF Goodrich KDs measuring 225/45R17 up front, and 255/40R17 in the back. This front-to-rear ratio should keep a similar balance to stock, while adding a small safety margin to guard against the snap oversteer that plagues 2000 model year S2000s. We stayed relatively conservative with the tire compound for now, in an attempt to keep the car streetable and match grip with engine output. In the future, we'll install a host of power enhancers that will necessitate the increased traction offered by R-compound tires. When that time comes, we'll swap out for BF Goodrich's new (to the US) R1.
It's tough to improve on the stock S2000 suspension. The car comes with external reservoir shocks that, combined with well-chosen spring rates, make it pretty capable in the twisty stuff. But since we were about to add a couple of pounds of unsprung weight to each corner, we needed something with quick reaction time and good rebound damping. Something a little more geared for track use.
The answer came in the form of KW's Variant III coilovers. Built for the weekend racer, they offer independently adjustable rebound and compression, and we can pick our own spring rates before they get here. Given the wide range of adjustability and KW's reputation in racing, they should help fling the S2000 around The Streets of Willow, our local race circuit. And thanks to the stainless steel construction and the preset high-speed damping curve, they might actually be bearable during daily use.
We decided to stick with the recommended spring rates, which meant 515lb/in for both front and rear. The spring rates being equal may cause a pogo-stick effect at certain speeds in a car with a 50/50 weight distribution, but we're willing to take the chance in exchange for retaining the neutral handling of the stock suspension.
Due to the limited travel of the rear suspension and the arc in which it travels, it's important not to go too far when lowering an S2000. KW allows for an adjustment range of 1.2 inches in the front, and an inch out back. That doesn't refer to how much lower the car gets to the pavement, it just means we can raise or lower the ride height over a spread of 1.2 inches or one inch, respectively.
You can make the car higher or lower than recommended by a small degree if you want, but KW suggests that the shock would then be out of optimal operating range. Based on the results we got when working within the recommended range, we see no reason to set the coilovers otherwise.
M-Workz in Gardena, California, is the end result of a lifetime in cars for a guy named Steve Mitchell. He's raced and built cars professionally for decades and he's spent the past 18 years at Nissan, engineering every imaginable aspect of its cars. So we figured he could handle the task of installing our suspension and corner-balancing the car.
The Wheel Supply in Industry, California, handled the task of mating Volk to BFG. We installed the finished product and took a quick measurement of the wheel well gap-we didn't want to stray too far from stock. There was a 2.75-inch gap up front and a full three inches in the rear. Mitchell got to work dismantling the stock stuff and installing the coilovers.
Installation couldn't have gone smoother. Mitchell likes to whistle while he works-incredibly irritating for any journalist who might be taking pictures, but at least it's a sign that he likes what he does. Springs mounted on shocks, we were ready to mount the suspension. The front was a bigger pain in the ass than the rear, as the upper A-arm had to be removed in order to slide the spring/shock assembly in (and slide the old one out). The rear shock and spring on the S2000 sit behind the moving geometry, so barring digging through the trunk to unbolt the strut towers, installation and removal was much more straightforward.
Suspension mounted and height loosely guesstimated, it was time to move on to corner balancing. "There's a critical step many tuners miss when corner-balancing a car," blurts Mitchell. "It's called leveling the floor. It's common sense, really. If your car is sitting at even a slight angle, weight is going to be shifted down the slope, toward the lower end of the car."
Armed with his laser leveler, Mitchell aligned the four tables that would support both the scales and the car. His have additional run-off area (roll-off pads), so the car can be adjusted and rolled on and off in a snap. We still weren't ready to start. Mitchell laid a towel down in the driver's seat before loading my exact weight in sandbags (that number to remain confidential).
When the S2000 first hit the scales, we recorded 2974 pounds-with a full tank of gas and the sandbag fat-ass in the driver's seat, which means we have an unnaturally light S2000 on our hands. According to the US Department of Energy, a gallon of gas weighs anywhere from 5.8 to 6.5 pounds. Let's call it 6.15 pounds. The tank's 13.2 gallons, so figure on about 81 pounds in gas alone. Tack on the sandbags, which, I'll regretfully admit, weigh 200 pounds, and we have a total weight of 2693 pounds. Like I said, a light S2000. Oh, and yeah, I'm on a diet, so you can all just stop writing the e-mails now.
What we're concerned with, though, is weight distribution. As it sat, the car showed a cross weight distribution of 48.8 percent from right front to left rear (and 51.2 percent from LF to RR). Not bad. When we were done, we had managed to equalize the cross weights for near 50/50 distribution. We adjusted the front spring perches to 60mm and the rear to 50mm, closer to the low side than the high side of the adjustment range, but nowhere near over-lowered. The wheel well gap had shortened by 1.5 inches up front and an inch out back, for a total gap of 1.25 and two inches respectively.
Before we sent the S2000 back out into the wild, we wanted to figure out what had caused the inside of the stock S03s to wear so rapidly. Unsurprisingly, the toe was way off, measuring a remarkable 9.5mm toe-in on each side, which is about 3.5mm out of max spec. And that's a ton. Often, the toe is adjusted in an attempt to compensate for the car's natural tendency for sudden oversteer, when the rear suspension compresses. After double-checking the front toe at zero millimeters, we took the toe back out to max stock specs (6mm a side). Because put simply, over-toeing in the car is a crappy band-aid for the real problem, which is toe-out under compression.
The first test of the revitalized S2000 was on the rugged roads that surround M-Workz. We left the rebound and compression at their factory recommended settings. We were utterly shocked to find that the ride was, if anything, smoother than stock, absorbing imperfections in the road without overreacting to potholes or large dips. At the same time, there was no perceptible body roll. It'd take a day at the track to really dial the suspension in, because for street use, it was more than capable.
And so we took the car to the Streets of Willow for a few shakedown laps. Right off the bat, the car was brilliant-even Mitchell was impressed with its ability to turn in and carry speed through technical sections. After messing with the coilovers for the better part of four hours, we ended up hovering right around the factory settings. A further quarter-turn of rebound up front was all that was needed to perfect the set-up for Streets' bumpy surface.
Does the car oversteer? Yep. But we're not going to dull its responses by toeing in the rear. We'll tame the rambunctious tail with a bump steer kit, but that's for another installment.
Before putting the S2000 away wet, it was off to our proving grounds in Fontana, Cailfornia. It was positively adjustable through the figure-eight test course, slight lifts of the throttle rotating the car around to find a perfect line in and out of the long corners. When the dust had settled, the S2000 net a 25.9-second average. To put this in perspective. A stock 2006 WRX runs the figure eight in 27 seconds and a Porsche 911 (997) GT3 will do it in 24. On the skidpad, the car managed an impressive 0.9g.
As hard as we try, we can't find the need for any further suspension modifications (short of a bump-steer kit). Thicker anti-roll bars would do nothing to improve the car's perfect balance. Amazingly, we seem to have conquered the suspension situation in one installment. Watch us attempt to tune some individual throttle bodies from Fuji Racing next time around.
Even with all the mentions of future ECU tuning plans in our previous installments, we still received a flood of complaints about Project Evo VIII MR and why we chose to save costs and use a temporary piggy-back fuel computer when the suspension cost almost $5000.
Thus far, our Evo has been through some mild power tuning and significant chassis work, all done at Road Race Engineering and Road Race Chassis in Santa Fe Springs, California. There's a reason why it seems like an RRE exclusive car: RRE has a complete tuning program and is familiar with what works and what doesn't with the hardware that's already on. So guess who we took it back to for its final big power installation. Why shouldn't we? No one would take their car to one body shop for body work and then to another for paint.
Now here's the disclaimer: we liked the Evo the way it was (aside from temporary AFC tuning) and knew what we were getting into when we asked for big power involving a Garrett GT3076R turbo, Cosworth CNC head and lumpy cams. This is purely an experiment on how much useable and reliable power the stock bottom end would put up with on 91-octane pump gas. Take the term 'useable' with a grain of salt.
RRE Evo VIII GT turbo kit
We start with the heart of the newfound power: RRE's functional but no frills GT turbo kit. RRE assumes you already have all the basic bolt-on plumbing for intake and exhaust, so only includes the turbo hardware itself, hence the term `turbo kit'. Along with a selection of four different types of turbos, our kit comes with a cast, smooth bending T3 turbo manifold, all the plumbing, TiAl 44mm wastegate, trick turbine outlet housing and wastegate plumbing. More impressive is all the thoroughly thought-out additional hardware, including manifold studs, gaskets, proper nuts and washers, V-band clamps for the wastegate, coolant lines, flexible hard oil return pipe, and even an in-line oil filter for the turbo.
To complete the kit, you can pick from an E-Pinions Fashion ball bearing 50-trim turbo, Street Power/Track Whore ball bearing GT3076R, Track Whore Deluxe custom ball bearing GT35R or Drag Queen/Race Gas Slut ball bearing full GT35R. The intercooler and intake plumbing might have to be modified, depending on the turbo. We went with the GT3076 Street Power/Track Whore turbo, which, in theory, can flow up to 525 wheel-hp. How much horsepower can be squeezed out is limited by the clutch, bottom end and how much fuel we can keep pouring into the cylinders.
Since the car had to go back to Mitsubishi, the original plan was to cut some corners and just strap on the turbo kit, temporarily tune the AFC and throw the car on the dyno before adding the Cosworth CNC head and doing a final tuned ECU reflash. Don't ask how we know, but this is a pointless exercise with the stock intercooler. Like the intake and exhaust, everything the air has to flow past to get to the intake valve and beyond the exhaust valve is a flow restriction. Each restriction is a little gate and acts as a limiting factor, some more, some less. Obviously the easiest, and sometimes most significant restrictions to get rid of are the intake and exhaust. Some people will also change the cat and the downpipe, which essentially opens up the exhaust side. But all these parts are at the ends of the engine breathing apparatus. The intake charge still needs to go through a turbo, intercooler and all its plumbing, the throttle body, intake manifold and then into the head, all of which still play a part in pinching off air going into the cylinder. In this case, the biggest restrictions are the turbo, head and intercooler, all of which we will deal with.
Regardless of how big a turbo we strap on, it still has to blow through the stock intercooler, which could only cope with 360 wheel-hp worth of air--according to the tests we're denying we ran at Injen's all-wheel-drive Dynojet. So it was off to RRE to get Injen's 23 3/4 x 11 3/4 x 3 3/8-inch FMIC and intercooler pipe kit installed. In order to clear the hardware, we also had to change the AEM intake out for an Injen polished cast aluminum short ram intake system, which also comes with a mandrel bent upper intercooler pipe. Although smaller in appearance, Injen rates its intercooler for up to 700hp. The added flow and cooling ability comes from a much thicker core and larger core passages. The Injen intake was modified with a 2.5- to four-inch silicon hose expansion to fit over the larger compressor inlet on the new turbo.
The monstrous Garrett GT3076R we decided to use comes with a 56-trim compressor wheel and 0.60 A/R. The turbine side uses a 60mm 84-trim wheel that we chose to mate with a 0.63 A/R turbine housing. In addition to the usual ball bearing, watercooled bells and whistles, the large compressor housing also includes a ported shroud that effectively pushes the surge limits of the turbo to lower flow rates, preventing premature turbo damage.
To better reflect the output most tuners should see with this kit on a non-MIVEC Evo, we had RRE temporarily install a set of aggressive HKS 272 duration intake and exhaust cams, retarded two degrees on the intake and exhaust side via adjustable AEM cam gears. The retarded cam timing will push the power curve further up in the rev range and allow the high-lift, longer-duration HKS cams to take advantage of the added volumetric efficiency of higher rpm, since a pump gas dyno whore was what we were after. Theoretically, this further delays the turbo's spool-up response in exchange for top-end peak power.
RRE calls this the Track Whore turbo for a reason. In our experience, it lags on the street and hits like a freight train, which makes it decent for WOT ass-hauling on big tracks, but a little hard to modulate on tighter courses. But that's what we asked for in order to make big power. With just a splash of 100-octane thrown in to prevent the car from blowing up while being tuned with the AFC, this set-up put down 437 wheel-hp at 7000rpm and 367lb-ft torque at 5500rpm. Keep in mind that fuel cut comes at 7750rpm and the MR has a close ratio six-speed 'box. Unfortunately, we weren't able to log the boost on this run--too busy hanging on to look down at the boost gauge. It swept easily past the last mark at 1.5bar.
Cosworth CNC head and M2 cams
Part Two of stupid power comes in the form of a Cosworth CNC ported cylinder head with 1mm oversized stainless-steel intake and Inconel exhaust valves. The Cosworth head starts as a brand-new Mitsubishi casting (MIVEC or not, depending on which car you have) which is CNC machined to enlarge and smooth out the ports, de-shrouded in the combustion chamber area and treated to a three- and four-angle valve job for the exhaust and intake respectively.To go with the valve job, the larger valves are also back-cut on the intake side for better flow transitions and lapped for proper sealing. The head also comes with stiffer double wound springs rated for up to 10,500rpm, titanium retainers, and hardened steel platforms for these springs to sit on. Cosworth claims a maximum flow increase of 26 percent on the intake side and 18 percent on the exhaust side at half an inch of lift.
Making the most of these flow capabilities means we were again going to step up to cams more aggressive than the 10.8mm-lift HKS versions. Cosworth makes two cams, ground from new chill cast billets. The M1 cam, probably intended more for the street, has a maximum lift of 10.2mm with a open-to-close duration of 264 degrees for both the intake and exhaust, while the big brother 272-degree M2 cam, with 11mm of lift, was designed for exactly what we were after: wide open power and a very lumpy idle. Depending on the tuner and tuning tool, Cosworth claims that stock engines with this cam can be tuned to idle like stock. We left the cam gears retarded by two degrees and went back to the dyno to see where this hardware got us. Maxxed out on boost, liquored up on high octane, we pulled a Hail Mary run (something RRE is well known for) of 480 wheel-hp, just to see how much the stock bottom end and Cosworth head could cope with. This was just enough power to make the stock clutch slip even before peak torque. However impressive these numbers are, we weren't inclined to repeat them again with just the AFC in control.
Finally, Proper Tuning
We were down to four days before the car had to be picked up by Mitsubishi in stock form without any holes in the block. Our intent was to settle the issue and get the car properly setup for street gas with an ECU reflash and ROM tune by Vishnu Tuning. Before that though, we needed a clutch with enough clamping force to put up with the subsequent abuse of repeated dyno tuning. The only shop around that could meet our schedule was Tuning Technologies in Colton, California. With a reputation of a one-day turnaround on a painfully tedious Evo clutch install, we immediately sent the car over. Even the ACT Heavy Duty clutch kit with street disc was on the shelves waiting for us.
Alfred, Grand Pooh-Bah at Tuning Technologies (TT) made good on his word, and the next morning we arrived to see Project Evo already strapped down on TT's Dynojet all-wheel-drive dyno. The glazed and anemic stock clutch sat in the doorway as proof. There to greet us was yet another Grand Pooh-Bah, Shiv Pathak of Vishnu, here to personally tune our potential time bomb.
The final hurdle was to street-tune Project Evo for strictly pump gas and make the car as bulletproof as possible on our hardware. Shiv and Alfred started by disconnecting the manual boost controller and putting the stock electronic boost solenoid back in the helm. Set for around 23psi and regulated by the factory ECU and knock sensor, Project Evo set out on a series of 29 tuning pulls on the dyno. The end result was 385 wheel-hp wheel horsepower and a lot more useful low-end torque between 2500 and 4500rpm (which you can't see with our dyno chart scaled this way). Why so much less power? Gas obviously has the most to do with it. Even just a couple octane points will make a night and day difference on fire-breathing turbo cars. The other part is peak boost and the boost curve profile, since the manual boost controller held boost constant all the way to redline. The electronic controller was tuned by Vishnu to taper off much like the stock program, to ease the load on the engine. Even though the bottom end could take more, it's just not smart on a daily driver.
Project Evo's last day was spent at the track. Knowing the limitations of the weaker MR six-speed, we set out to get one last respectable run on all this power. With the boost lag and fragile tranny, we couldn't just dump the clutch at 6000rpm, which is what the car requires for a fast ET. We did manage to finally squeeze out a 12.8-second run trapping at 100mph. Not impressive, but no slouch compared to the 14 seconds on the stock car. This time is roughly what we found in our last Evo Tuner shootout.
Coming to terms
Project Evo VIII MR is done. Unlike our other project cars that live as long as we're willing to come up with bad ideas to throw at them, the Evo has been returned to Mitsubishi after a gracious loan of two years and will probably end up in the hands of some undeserving owner. We still think of Project Evo fondly now it's gone.
This was by far our all-time favorite project car throughout the years. But our fondness is of the Evo in its near stock form, which provided endless hours of giggling--in traffic or on the track. Even in bone stock form, it went out and destroyed everything else in our project car garage in the rain during the last Project Car battle. The car is just that brilliant. And as good as the parts are that we put in, we've made some compromises in ride, noise, and lag.
If we had to do this all over again, I would have just replaced the pads with something less aggressive than we had used, stuck with the stock turbo and a quieter exhaust, and used a suspension with more emphasis on street driving. We go to the track, but that doesn't mean we avoid sitting in traffic along with everyone else.
Our next Mitsubishi project won't be an Evo, we'll hopefully be getting our hands on the Ralliart Lancer, a mean and affordable cousin of the Evo X and the Lancer GTS.
Even with all the mentions of future ECU tuning plans in our previous installments, we still received a flood of complaints about Project Evo VIII MR and why we chose to save costs and use a temporary piggy-back fuel computer when the suspension cost almost $5000.
Thus far, our Evo has been through some mild power tuning and significant chassis work, all done at Road Race Engineering and Road Race Chassis in Santa Fe Springs, California. There's a reason why it seems like an RRE exclusive car: RRE has a complete tuning program and is familiar with what works and what doesn't with the hardware that's already on. So guess who we took it back to for its final big power installation. Why shouldn't we? No one would take their car to one body shop for body work and then to another for paint.
Now here's the disclaimer: we liked the Evo the way it was (aside from temporary AFC tuning) and knew what we were getting into when we asked for big power involving a Garrett GT3076R turbo, Cosworth CNC head and lumpy cams. This is purely an experiment on how much useable and reliable power the stock bottom end would put up with on 91-octane pump gas. Take the term 'useable' with a grain of salt.
RRE Evo VIII GT turbo kit
We start with the heart of the newfound power: RRE's functional but no frills GT turbo kit. RRE assumes you already have all the basic bolt-on plumbing for intake and exhaust, so only includes the turbo hardware itself, hence the term `turbo kit'. Along with a selection of four different types of turbos, our kit comes with a cast, smooth bending T3 turbo manifold, all the plumbing, TiAl 44mm wastegate, trick turbine outlet housing and wastegate plumbing. More impressive is all the thoroughly thought-out additional hardware, including manifold studs, gaskets, proper nuts and washers, V-band clamps for the wastegate, coolant lines, flexible hard oil return pipe, and even an in-line oil filter for the turbo.
To complete the kit, you can pick from an E-Pinions Fashion ball bearing 50-trim turbo, Street Power/Track Whore ball bearing GT3076R, Track Whore Deluxe custom ball bearing GT35R or Drag Queen/Race Gas Slut ball bearing full GT35R. The intercooler and intake plumbing might have to be modified, depending on the turbo. We went with the GT3076 Street Power/Track Whore turbo, which, in theory, can flow up to 525 wheel-hp. How much horsepower can be squeezed out is limited by the clutch, bottom end and how much fuel we can keep pouring into the cylinders.
Since the car had to go back to Mitsubishi, the original plan was to cut some corners and just strap on the turbo kit, temporarily tune the AFC and throw the car on the dyno before adding the Cosworth CNC head and doing a final tuned ECU reflash. Don't ask how we know, but this is a pointless exercise with the stock intercooler. Like the intake and exhaust, everything the air has to flow past to get to the intake valve and beyond the exhaust valve is a flow restriction. Each restriction is a little gate and acts as a limiting factor, some more, some less. Obviously the easiest, and sometimes most significant restrictions to get rid of are the intake and exhaust. Some people will also change the cat and the downpipe, which essentially opens up the exhaust side. But all these parts are at the ends of the engine breathing apparatus. The intake charge still needs to go through a turbo, intercooler and all its plumbing, the throttle body, intake manifold and then into the head, all of which still play a part in pinching off air going into the cylinder. In this case, the biggest restrictions are the turbo, head and intercooler, all of which we will deal with.
Regardless of how big a turbo we strap on, it still has to blow through the stock intercooler, which could only cope with 360 wheel-hp worth of air--according to the tests we're denying we ran at Injen's all-wheel-drive Dynojet. So it was off to RRE to get Injen's 23 3/4 x 11 3/4 x 3 3/8-inch FMIC and intercooler pipe kit installed. In order to clear the hardware, we also had to change the AEM intake out for an Injen polished cast aluminum short ram intake system, which also comes with a mandrel bent upper intercooler pipe. Although smaller in appearance, Injen rates its intercooler for up to 700hp. The added flow and cooling ability comes from a much thicker core and larger core passages. The Injen intake was modified with a 2.5- to four-inch silicon hose expansion to fit over the larger compressor inlet on the new turbo.
The monstrous Garrett GT3076R we decided to use comes with a 56-trim compressor wheel and 0.60 A/R. The turbine side uses a 60mm 84-trim wheel that we chose to mate with a 0.63 A/R turbine housing. In addition to the usual ball bearing, watercooled bells and whistles, the large compressor housing also includes a ported shroud that effectively pushes the surge limits of the turbo to lower flow rates, preventing premature turbo damage.
To better reflect the output most tuners should see with this kit on a non-MIVEC Evo, we had RRE temporarily install a set of aggressive HKS 272 duration intake and exhaust cams, retarded two degrees on the intake and exhaust side via adjustable AEM cam gears. The retarded cam timing will push the power curve further up in the rev range and allow the high-lift, longer-duration HKS cams to take advantage of the added volumetric efficiency of higher rpm, since a pump gas dyno whore was what we were after. Theoretically, this further delays the turbo's spool-up response in exchange for top-end peak power.
RRE calls this the Track Whore turbo for a reason. In our experience, it lags on the street and hits like a freight train, which makes it decent for WOT ass-hauling on big tracks, but a little hard to modulate on tighter courses. But that's what we asked for in order to make big power. With just a splash of 100-octane thrown in to prevent the car from blowing up while being tuned with the AFC, this set-up put down 437 wheel-hp at 7000rpm and 367lb-ft torque at 5500rpm. Keep in mind that fuel cut comes at 7750rpm and the MR has a close ratio six-speed 'box. Unfortunately, we weren't able to log the boost on this run--too busy hanging on to look down at the boost gauge. It swept easily past the last mark at 1.5bar.
Cosworth CNC head and M2 cams
Part Two of stupid power comes in the form of a Cosworth CNC ported cylinder head with 1mm oversized stainless-steel intake and Inconel exhaust valves. The Cosworth head starts as a brand-new Mitsubishi casting (MIVEC or not, depending on which car you have) which is CNC machined to enlarge and smooth out the ports, de-shrouded in the combustion chamber area and treated to a three- and four-angle valve job for the exhaust and intake respectively.To go with the valve job, the larger valves are also back-cut on the intake side for better flow transitions and lapped for proper sealing. The head also comes with stiffer double wound springs rated for up to 10,500rpm, titanium retainers, and hardened steel platforms for these springs to sit on. Cosworth claims a maximum flow increase of 26 percent on the intake side and 18 percent on the exhaust side at half an inch of lift.
Making the most of these flow capabilities means we were again going to step up to cams more aggressive than the 10.8mm-lift HKS versions. Cosworth makes two cams, ground from new chill cast billets. The M1 cam, probably intended more for the street, has a maximum lift of 10.2mm with a open-to-close duration of 264 degrees for both the intake and exhaust, while the big brother 272-degree M2 cam, with 11mm of lift, was designed for exactly what we were after: wide open power and a very lumpy idle. Depending on the tuner and tuning tool, Cosworth claims that stock engines with this cam can be tuned to idle like stock. We left the cam gears retarded by two degrees and went back to the dyno to see where this hardware got us. Maxxed out on boost, liquored up on high octane, we pulled a Hail Mary run (something RRE is well known for) of 480 wheel-hp, just to see how much the stock bottom end and Cosworth head could cope with. This was just enough power to make the stock clutch slip even before peak torque. However impressive these numbers are, we weren't inclined to repeat them again with just the AFC in control.
Finally, Proper Tuning
We were down to four days before the car had to be picked up by Mitsubishi in stock form without any holes in the block. Our intent was to settle the issue and get the car properly setup for street gas with an ECU reflash and ROM tune by Vishnu Tuning. Before that though, we needed a clutch with enough clamping force to put up with the subsequent abuse of repeated dyno tuning. The only shop around that could meet our schedule was Tuning Technologies in Colton, California. With a reputation of a one-day turnaround on a painfully tedious Evo clutch install, we immediately sent the car over. Even the ACT Heavy Duty clutch kit with street disc was on the shelves waiting for us.
Alfred, Grand Pooh-Bah at Tuning Technologies (TT) made good on his word, and the next morning we arrived to see Project Evo already strapped down on TT's Dynojet all-wheel-drive dyno. The glazed and anemic stock clutch sat in the doorway as proof. There to greet us was yet another Grand Pooh-Bah, Shiv Pathak of Vishnu, here to personally tune our potential time bomb.
The final hurdle was to street-tune Project Evo for strictly pump gas and make the car as bulletproof as possible on our hardware. Shiv and Alfred started by disconnecting the manual boost controller and putting the stock electronic boost solenoid back in the helm. Set for around 23psi and regulated by the factory ECU and knock sensor, Project Evo set out on a series of 29 tuning pulls on the dyno. The end result was 385 wheel-hp wheel horsepower and a lot more useful low-end torque between 2500 and 4500rpm (which you can't see with our dyno chart scaled this way). Why so much less power? Gas obviously has the most to do with it. Even just a couple octane points will make a night and day difference on fire-breathing turbo cars. The other part is peak boost and the boost curve profile, since the manual boost controller held boost constant all the way to redline. The electronic controller was tuned by Vishnu to taper off much like the stock program, to ease the load on the engine. Even though the bottom end could take more, it's just not smart on a daily driver.
Project Evo's last day was spent at the track. Knowing the limitations of the weaker MR six-speed, we set out to get one last respectable run on all this power. With the boost lag and fragile tranny, we couldn't just dump the clutch at 6000rpm, which is what the car requires for a fast ET. We did manage to finally squeeze out a 12.8-second run trapping at 100mph. Not impressive, but no slouch compared to the 14 seconds on the stock car. This time is roughly what we found in our last Evo Tuner shootout.
Coming to terms
Project Evo VIII MR is done. Unlike our other project cars that live as long as we're willing to come up with bad ideas to throw at them, the Evo has been returned to Mitsubishi after a gracious loan of two years and will probably end up in the hands of some undeserving owner. We still think of Project Evo fondly now it's gone.
This was by far our all-time favorite project car throughout the years. But our fondness is of the Evo in its near stock form, which provided endless hours of giggling--in traffic or on the track. Even in bone stock form, it went out and destroyed everything else in our project car garage in the rain during the last Project Car battle. The car is just that brilliant. And as good as the parts are that we put in, we've made some compromises in ride, noise, and lag.
If we had to do this all over again, I would have just replaced the pads with something less aggressive than we had used, stuck with the stock turbo and a quieter exhaust, and used a suspension with more emphasis on street driving. We go to the track, but that doesn't mean we avoid sitting in traffic along with everyone else.
Our next Mitsubishi project won't be an Evo, we'll hopefully be getting our hands on the Ralliart Lancer, a mean and affordable cousin of the Evo X and the Lancer GTS.
