In stock form, our WRX has seen it all. From the mountains of Colorado and Wyoming to the deserts of Arizona and Utah, it's been working as our go-anywhere-at-any-speed street car for almost six months. But as good as it is stock, we're ready for a change.
It's no secret that Subaru's new hot rod has as much or more tuning potential as any car to hit the American market in the last five years. Not since the glory days of the twin turbo Z car, third-generation RX-7 and Toyota Supra have enthusiasts had so much opportunity to make major power for only minor green--which is exactly what we'll do in this installment with our WRX.
Here's how. Beginning with four exhaust systems, which vary greatly in price and construction, we'll work into other more comprehensive mods. We'll make a first attempt at cranking up the boost while retaining the stock turbo by using Vishnu Performance Systems' Stage Zero kit. And, naturally, we'll swap wheels and tires because, well, that's what we do.
After the above-mentioned road trip through Wyoming and Colorado, we returned our WRX to California and immediately headed to A'pexi in Orange, Calif. for some quality time on the company's four-wheel-drive Dynapack chassis dyno. There, we made some shocking discoveries.
During baseline pulls, we noticed our car made 177 hp at the wheels, which is on the high end relative to other WRXs. Most WRXs make between 165 and 175 peak wheel hp on a properly configured Dynapack chassis dyno. We recently learned the Dynapack we used for our WRX comparison test in the October, 2001 issue was not properly configured, so although the results in that test can still be compared with one another, they cannot be compared to any of the numbers in this test.
Additionally, we experienced wild inconsistencies from one run to the next, sometimes varying as much as 10 hp throughout the powerband. Combine the differences in peak power with our inconsistencies on the dyno and we began to theorize that the stock WRX ECU is not only extremely adaptive but also exercises a huge authority range over both fuel and ignition curves.
To prove our theory, we reset the ECU by disconnecting the battery. Electrons remaining in the system were purged through the brake lights by stepping on the brakes, leaving the ECU with no memory. Then we re-tested it. This time, our WRX pulled 167 hp at the wheels. Sure enough, the Subaru ECU had learned a few tricks in the last six months. We gave it several runs to try relearning those tricks, but apparently 130-mph blasts across Wyoming are better learning experiences than dyno pulls. The 10 hp did not reappear.
Luckily, Vishnu Performance has already done most of that work for us with its pre-calibrated piggyback computer called the Unichip, which is part of its Stage Zero mods for the WRX. The Unichip is a piggyback ECU, which allows its tuner to change fuel and ignition maps while running in conjunction with the stock ECU. Unichips from Vishnu are all preprogrammed to work with the other mods in the kit; we'll get to that in a minute.
There are several critical issues to address when running a piggyback ECU such as the Unichip in conjunction with the WRX's stock ECU. The most difficult is that changes made with the piggyback ECU can quickly be cancelled out by a stock ECU with this much authority range. For example, let's pretend a tuner analyzes a dyno graph, and then decides to advance the ignition two degrees at a particular load using the Unichip. The change results in huge gains in power on the dyno and he quickly saves the new map into the Unichip.
This change might be fine on the dyno under the conditions the engine was experiencing during that particular test session on that particular day. However, the next day, when the car hits the street under different conditions, those extra two degrees of advance might bump the engine into its knock threshold, sending the knock sensor to DEFCON 5. That's not good.
With the knock sensor on full song, the stock ECU might retard timing as much as 20 degrees, drastically reducing power and canceling any gains from the added advance. This issue is complicated further by the Unichip's ability to see only two sensor inputs (mass air flow and engine speed), while the stock ECU makes ignition advance and fueling decisions based on input from as many as five or six sensors at a time. At any given time, the stock ECU sees engine speed, mass airflow, manifold absolute pressure, intake temperature and possibly the exhaust gas temperature.
In other words, the logistics involved in piggyback tuning are not only complex but, if done poorly, can hurt more than they help. As with any engine management system, the Unichip is only as good as the person tuning it. The calibrations saved into the Vishnu-tuned Unichip are a result of neurotic street tuning as well as many dyno sessions. Luckily for us, there's plenty of room for increased efficiency in the WRX's stock exhaust plumbing and turbo system, which allows for changes made with the Unichip to be maximized.
In addition to the Unichip, the Vishnu Stage Zero kit comes with several other parts, which at first may seem like an unusual combination. Looking at the big picture is the best way to understand why the kit comes with an up-pipe and not a downpipe or exhaust--the usual first choice for upgrades on a turbo car.
First of all, what the hell is an up-pipe? If you're not constantly surrounding yourself with WRX geeks, then the term is probably new to you. Turbocharging a flat four-cylinder engine such as the WRX's inevitably leads to some creative exhaust plumbing solutions. Because two of the engine's exhaust ports are on the left side of the engine bay and two are on the right side, Subaru had to route the left side exhaust around the front of the engine through a crossover pipe. The crossover pipe then joins the exhaust from the right-side exhaust ports in a complex cast manifold.
After collecting in this manifold, exhaust gasses are sent up to the turbo through the up-pipe. The up-pipe is also the location Subaru chose for the restrictive close-coupled catalytic converter--one of three in the exhaust system. The other two cats are located downstream of the turbo. The catalyst located in the up-pipe is what makes it the first candidate for replacement in the WRX's exhaust system. Placing a catalyst before the turbo increases backpressure and exhaust gas temperatures, not to mention acting like giant plug robbing power from otherwise high-energy exhaust pulses.
Vishnu's up-pipe is made from 321 stainless steel and uses 304 stainless, 0.5-inch CNC-cut flanges. It also fits well. However, getting to the up-pipe on a Subaru isn't the easiest. Plan on a minimum of three hours if you're doing it yourself. We had help, and it took us a day.
Boost control is handled with a simple, low-dollar, ball-valve boost controller made by Dawes Devices (www. dawesdeveices.com), which takes the stock boost controller out of the loop completely. The ball valve interrupts the boost signal from the compressor housing, preventing it from reaching the wastegate. Once the release pressure of the spring inside the ball valve is reached (it's adjustable), the ball valve springs open and begins to feed boost to the wastegate actuator. It's a remarkably simple and effective piece. It also allows full boost at partial throttle--something the stock, ECU-controlled boost controller won't do.
Also included in Stage Zero are NGK BKR7EVX-11 spark plugs which are one heat range colder than stock and should help reduce the chances of detonation at higher-than-stock boost levels. A billet underdrive pulley and the shortened accessory belts are included in the kit as well. Underdrive pulleys work on almost every car by simply slowing the belt-driven accessories so they consume less power. This one goes on easily.
If it weren't for a damaged stud on our turbo manifold, we probably could have installed the entire Stage Zero kit in about four hours with three monkeys turning wrenches. While taking the car apart was a pain, it actually went back together fairly easily and, when finished, made a peak gain of 41 hp and 33 lb-ft of torque at the wheels. This gain is compared with the dyno pulls we made after resetting the ECU--which is the only time we saw consistent dyno numbers in stock form. Plus, resetting the ECU by disconnecting the battery is part of the Unichip installation. Power and torque curves have increased dramatically over stock, beginning at about 3000 rpm and continuing to redline.
In the interest of full disclosure, Vishnu Performance is owned and operated by Shiv Pathak, a regular editorial contributor to this magazine. We relished the opportunity to bash his products in public, but unfortunately for us, his stuff is actually pretty good.
With Stage Zero mods complete, we began comparing exhausts. Because of the inefficiencies in the stock up-pipe, there was unlikely to be much power gained by installing these exhausts on the stock car. We hoped that by installing the Stage Zero kit first, we'd maximize the gains offered by aftermarket exhausts. It should be noted that because the WRX ECU is constantly making changes, we saw variations of 3 to 5 hp from one run to the next. This is less than the differences we saw from one exhaust system to the next, so don't try splitting hairs. The only conclusion we can draw from these tests is that they all work about the same, at least with the engine in this state of tune.
Subaru Performance Tuning exhaust
The Subaru Performance Tuning exhaust is the simplest of the four exhausts we tested. It replaces only the rear-most section of the stock exhaust, including the muffler. Included with the SPT exhaust are two nuts, which can be used with the stock bolts at the flange, and no gasket. We re-used the stock gasket, which isn't the best solution but worked. The SPT exhaust saved about 4 lbs relative to the stock exhaust (STP muffler: 16.5 lbs vs. stock muffler: 20.5 lbs). On the dyno, gains were significant. Seven hp and 3 lb-ft of torque.
Also included with the SPT exhaust is a huge disclaimer, which clearly states that it's never to be used on the street and that all warranties on the car will be voided if it is. In our experience, dealers only sporadically enforce this policy. In fact, some dealers will even install the exhaust at the time of a new purchase and finance it into the deal.
A'pexi GT Spec Exhaust
With a peak gain of 5 hp and weight reduction of about 5 lbs less than stock, A'pexi's GT spec exhaust works as promised. Peak torque remains the same as stock but is improved slightly from 5500 rpm to redline. It also comes with every fastener and gasket necessary for assembly as well as a band-clamp hanger for the muffler and a silencer. In other words, it comes with everything you need. Plus, the silencer effectively quiets things to a more law enforcement-friendly volume, which is good, but it almost completely cancels the unit's power gain, which is bad.
Probably the best constructed exhaust of the bunch, both the tubing and flanges on A'pexi's exhaust are built from 304 stainless steel. The flanges are flame cut and the entire exhaust is polished.
On the dyno, the GT Spec exhaust produced the usual throaty bellow we expect from A'pexi exhausts. This one also generated serious clouds of smoke once up to operating temperature, something that had us seriously concerned. Then we tested the other exhausts and it became obvious there was lubricant remaining inside the exhaust after construction, which is common on mandrel-bent exhausts because the process requires lubricant to keep from damaging the tubing as it's bent.
Supersprint Street Exhaust
Both Supersprint exhausts in this test are constructed from polished stainless steel and are mandrel bent, and both look sexy under the car. The street exhaust is a conventional cat-back system, which weighs 28.2 lbs--about 3 lbs lighter than the stock Subaru b-pipe and muffler. Supersprint also offers what it calls a street downpipe that eliminates the two remaining stock cats, but we're enviro-friendly enthusiasts so we didn't install it. Ironically, since many European race series require catalysts, Supersprint's larger, race downpipe does have a cat.
On the dyno, the street exhaust made a peak gain of 7 hp, while the torque peak remained the same. However, after 5000 rpm, torque remained between 3 and 5 lb-ft higher than stock all the way to redline.
Both Supersprint exhausts also exhibited the same smoky behavior as the A'pexi system. While this isn't a long-term problem, as the lubricants will eventually burn out of the tubing, it's quite unnerving if you're not expecting it.
Additionally, both Supersprint exhausts had a thin layer of dirt and dust on top of the lubricant. This is fairly irrelevant downstream of the turbo, but if you purchase either a header or up-pipe from the company, be sure everything is cleaned thoroughly with solvent before installation. Sending any kind of debris through the turbo will end its life quickly.
Supersprint Race Exhaust
Supersprint's race exhaust made 10 peak hp and weighed 15 lbs less than stock--both significant improvements. The race system replaces the stock exhaust from the turbo back and includes a metallic racing catalytic converter in the downpipe. This means using this system in conjunction with the Vishnu up-pipe leaves our WRX with only one catalytic converter. Worth noting is the Supersprint downpipe design, which uses a flat flange over the wastegate where it bolts to the turbine housing. This design has been known to create boost spikes and inconsistent boost control problems.
The downpipe would benefit greatly from a slightly expanded pipe over the wastegate. This would allow bypassed exhaust gasses to escape with less restriction, enhancing the entire system's performance and reliability. It would also reduce the risk of damage to the engine from overboost. The stock downpipe, by the way, has the same flat flange. We have noticed some minor boost spiking, but haven't had a chance to test another downpipe to see if the flat flange is the cause. We would've also liked to see a little more hardware with the exhaust. It didn't have a gasket between the b-pipe and muffler, for example. We borrowed one from the A'pexi kit.
When assembled, the race exhaust is extremely close to the center carrier bearing housing on the driveshaft, causing it to rub under engine braking. Supersprint's U.S. importer, BMP Design, is aware of the problem and promises to correct it before importing the first batch of exhausts.
Perhaps the Supersprint's best qualities are its deep and relatively quiet note at both idle and full throttle. And its oval tip is just subtle enough to be overlooked in California's exhaust-ticket-friendly environment.
We've put about 2,000 miles on Project WRX since installing the Stage Zero kit and Supersprint race exhaust. Throttle response and part-throttle driveability are dramatically improved, not to mention the serious increase in thrust. Aside from some minor surges that seem to be part-throttle boost spikes, the driveability is flawless. We haven't heard a single ping. The exhuast note of the Subaru's flat four is unique and intoxicating. It has a deep, off-beat rumble that makes it sound much bigger than it is, and also happens to be too deep to cause any of the droning you usually get with open exhausts on four-cylinders.
The improved power does have its downsides, however. The WRX can kill its stock brakes, even when going uphill. We'll address that problem in Part II.
We can also say the WRX is a very durable machine. We've beaten it severely on numerous occasions and are yet to have any power-related parts failures. On a recent 500-mile run through California's deserted hill-country roads, we managed about 10 miles per gallon. Granted this was ten-tenths driving for a full tank of gas while making an additional 51 hp at the wheels. Good times. Stay tuned.
Wheels and Tires
On this project, we're taking a slightly different approach than many of you will with wheels and tires. Our goal is to retain plenty of sidewall height in an effort to keep the stock car's driveability in gravel. The WRX is such a naturally gifted gravel car that it seemed counterproductive to install big wheels and low-profile tires that might make it any less effective. This car won't be a track queen unless, of course, that track is a rallycross course or one of California's many deserted backroads.
This goal narrows our wheel choices considerably. To keep any acceptable sidewall height we must retain 16-inch wheels and tires. And, because we want to start relatively inexpensively, we'll probably try out several sets of wheels before the project is over--perhaps even jumping up to some 17-inchers for occasional tarmac use. For now, however, we'll start with 16s and we'll start cheap.
Discount Tire Direct (www.tires.com) sells Tecnomagnesio Indianapolis 16x7-inch wheels in the required offset (48mm when using a seven-inch wheel) for $110. That's cheap. Unfortunately, they're also fairly heavy at just under 20 lbs. each. But, they come in several finishes including gold, silver and white and it's hard to argue with the price. We took 'em in gold. Our tire choice was also remarkably simple, as the demands were very specific. We needed a tire that was better than the stock Bridgestone RE92s in every environment. But, we also couldn't give up the RE92s usability in almost any weather condition or on any road--gravel or tarmac. Nor could we go larger than 16 inches. In other words, we were willing to make compromises everywhere to have a multi-use tire. Having driven BFGoodrich's g-Force KDWS all weather tire several years ago at its introduction, we knew it to be a capable in most weather conditions. Plus it's one of the only all-season performance tires on the market. The choice was made.
It became immediately obvious after installing the g-Force tire on our WRX that they were a significant improvement over the OE Bridgestones. The biggest difference was in sidewall stiffness as the ride became slightly compromised. However, this translated directly into better turn-in response and an overall better feel for what the car was doing.
Not being big believers in tires touted as "all-season" and "performance," we're truly impressed with the do-all nature of the KDWS. While it's not the tire you want for an all-out mountain-road attack or as a track tire, it more than makes up for those shortcomings in real-world usability. Our KDWS' have gotten us through everything from 8,500-ft. snow-covered mountain passes in Wyoming to loads of sideways-sliding gravel road antics, not to mention one California Rally Series rallycross class win. By the way, they're the tire of choice for the Street Stock class as their shoulder sipes make the most of the class rules. Plus, we found them online for $104 each from The Tire Rack. That's relative to $80 each for the OE replacement. It's certainly worth stepping up.