Project Nissan Silvia: Part II

Elsewhere in this issue you will read about how to choose the right turbo for your car. The explanation uses compressor maps, math, and the need to at least pretend you understand some complex engineering concepts.

Here's another method.

Step 1: Drive a car with a really good turbo match.
Step 2: Build a similar car.

A year and a half ago, we drove a very unlikely Sentra. Dubbed the Disco Potato for its unique combination of psychedelic color-shifting paint and unfortunate Jerry Hirshberg styling, the car was the culmination of the clandestine efforts of a few very driven engineers.

The story starts with Dan Passe who, at the time, was a Nissan PR genius with a penchant for bending rules. He conveniently "lost" the paperwork for a 1.8-liter Sentra which Nissan Design International had modified for the L.A. Auto Show. The car quietly landed in the hands of Nissan engineers Steve Mitchell and Mike Kojima.

Meanwhile, a few miles away at Garrett, turbo engineer Jay Kavanagh wanted to boost his Miata. Having full access to the newest Garrett technology, he concocted a physically small turbo with a ball-bearing center section and internal aerodynamics 20 years more modern than the T3/T4 standard the aftermarket is used to.

A few cubicles from Kavanagh, Rob Cadle, a good friend of Mitchell and Kojima, realized Kavanagh's Miata turbo would be perfect for the SR20DET the Nissan boys were planning for the Sentra. He brewed up a turbo, stuffed it under his shirt, and went out the back door.

Jim Wolf Technology built a very mild SR20 for the Sentra, making it functionally equivalent to a stock Japanese-spec SR20DET. The turbo was installed, and amazing things started happening. The car's power was impressive, 280 hp at the wheels, but not earth-shattering. The driving experience however, was. Throttle response was excellent, turbo lag virtually non-existent, and the tire-shredding power was easily modulated. The turbo spooled up early, making so much torque, that the best quarter-mile time (13.7 at 104.5 mph) was achieved launching in second gear.

Mitchell brought the Disco Potato to the Ultimate Street Car Challenge in 2001, and placed an impressive fourth overall. The rest of the time, the car was stashed away in Nissan USA's service garage and used strategically as an attitude adjustment tool. Whenever Nissan or Garret executives needed an injection of gasoline in their veins, they were offered the keys. They would inevitably come back grinning from ear to ear and breathing heavily, eager to bring horsepower to the masses.

There's no telling how many decisions were nudged in the right direction by the Disco Potato, but it did prompt two that are critical to this story. We drove the car for a week, including a 1,000-mile road trip and track day at Thunderhill Raceway. After outrunning every car at the track, and blasting down the freeway at 140 mph, we knew we must have that engine. Ours, however, had to be in a rear-drive chassis. Thus was born Project Silvia.

Several Garret executives also experienced the Potato. They were so impressed, they decided to produce the turbo, double the engineering staff in the aftermarket department, and start applying this modern Garrett technology to a whole range of aftermarket turbos. Technically, this is a GT28R with a 62 trim compressor (.60 A/R) and a 76 trim turbine (.86 A/R), but you don't need to know that. Ask anyone at Garrett for a Disco Potato turbo, and this is what you'll get.

The Engine Concept
Before turning a wrench on an engine project, you should have a goal. Making 400 hp on the dyno might be good for the ego, but living in a world of 91-octane horse urine, unpredictable traffic and some of the best driving roads in the world, we preferred the immediate, responsive, controllable power delivery of the Disco Potato. Short of the dragstrip, there are few times when 280 hp in a 2,600-pound car will feel slow.

With the Potato as a target, we didn't need to do any engine work. Our stock '91 SR20DET already has turbo-friendly 8.5:1 compression, beefy pistons and rods, piston-cooling oil squirters, and a good-flowing, high-port head. The Potato used a later, low-port head with Jim Wolf C1 cams, but for now at least, we'll stick with stock cams and hope our better head will make up for it. This could be a very simple path to the 300-hp range, with just a turbo, intercooler, injectors and ECU needed to make it work. Of course, reality turned out to be a bit more complicated.

The Turbo
Miraculously, the Potato turbo is a direct, bolt-in replacement for the Silvia's stock T25. In fact, from the outside, the two turbos look nearly identical. The oil feed and both water fittings for the water-cooled center housings are identical to the stock turbo.

There are four complications to bolting in the potato. The oil drain fitting has a slightly different bolt pattern, with the new turbo's bolts 0.25-inch closer together than the old ones. One of the five studs on the turbine exhaust has also moved by a fraction of an inch. Both problems can be solved in a few minutes by filing the holes and gaskets.

The other two complications are on the compressor side. Both the stock intake and compressor discharge hoses attach to cast aluminum elbows that bolt to the stock compressor housing. Those elbows won't bolt to the Potato's turbine housing. You won't want to use the stock plumbing anyway, so the fact that the compressor intake is designed for a 3-inch hose and the discharge for a 2-inch hose will work out just fine. Of course, now you have to make a bunch of plumbing, but look on the bright side; the most common big Silvia turbo, a T3/T4 hybrid, would require a new exhaust manifold and downpipe, too.

Exhaust
Turbos like big exhausts. A turbo takes its power from the pressure drop across the turbine, and whatever pressure drop the turbine wants, the compressor will back things up until it gets it. That means if the turbo is causing 10 psi of backpressure, and the exhaust is creating 4 psi more, you'll end up with 14 psi in the exhaust manifold. Better to have an exhaust that makes 0 psi.

We also like to make every modification a chance to save some weight. The stock exhaust, despite its puny 2-inch pipe, and the ravages of 270,000 rust-producing miles, weighs 25 pounds. JIC's 3-inch titanium Bullet exhaust weighs only 13.

Titanium is hard to bend and challenging to weld, so the Bullet exhaust is designed to minimize both. The exhaust is nearly straight, bending just a few degrees to dodge the rear differential. And while JIC had to weld the resonator and catalytic converter flange, welds are eliminated at the muffler by using a slip-fit instead of a welded-on flange. The exhaust hangers, the most highly stressed welds on the system, are welded to stainless-steel bands that are clamped around the muffler and exhaust pipe. In a beautiful bit of obsessive weight watching, the stainless-steel hangers are hollow, and the ends are bead rolled so they won't slip off.

Upstream of the titanium is the 3-inch GReddy downpipe we installed with the engine ("Hybrid How-To," May 2002), and a 3-inch high-flow catalytic converter from Landspeed racing. On naturally aspirated cars, a good cat usually doesn't cost any power. On a high-output turbo engine like this, we're probably sacrificing a few horsepower, but we like breathing almost as much as driving, so we're running with the cat.

Most Nissans of this era used the same catalytic converter. Our stock 240SX cat, for example, was identical to the stock cat from our 1991 SE-R project car. This means Landspeed's 3-inch cat fits SE-Rs, too. The cat is a fraction of an inch longer than the stock cat--a concession to allow for a larger, higher-flow CarSound core. Most exhaust systems, including ours, have enough flexibility to make this acceptable.

This exhaust combination--despite the cat, presilencer, and muffler--is still loud. With the stock turbo, the sound level is quite tolerable, and the deep tone impressive, but the volume increases substantially with the less restrictive Potato turbo on board and 3,000 miles of heat to settle the muffler packing. Below 3000 rpm, it's quite civilized (good for slipping by John Law) but above that, the interior gets loud and it sets off car alarms.

Cooling
There are plenty of front-mount intercooler kits for the 240SX and Silvia. They come with mounting brackets and prefabricated plumbing, making the front-mount conversion an easy afternoon bolt-on affair. We didn't want to do that. The reason is plumbing. Most of these kits were designed when these cars were still new, and were laid out with the assumption the owner wouldn't want to cut any metal. Avoiding the hacksaw means adding more than 6 feet of unnecessary plumbing. That plumbing dulls throttle response and adds turbo lag, forcing you to wait a fraction of a second while several gallons of extra air get pressurized.

The front-mount intercooler presents another serious compromise: cooling. This summer, making only 155 hp at the wheels on stock boost, the Silvia was already having overheating problems on the track. Putting a big intercooler in front of the radiator and generating 75-percent more heat (Heat, power. Same thing.), the stock cooling system clearly won't work.

After many hours of staring at the engine compartment, a plan emerged. Removing the belt-driven fan would free some power, and removing the shroud would make room to move the radiator back about 5 inches. The gap between the engine and radiator would still be enough for some electric fans, and the gap between the core support and the radiator would be enough for an intercooler. This plan also allowed the intercooler pipes to stay in the engine compartment the whole time, minimizing their length.

This still didn't solve the basic cooling issue, but replacing the stock radiator with a Koyo racing radiator did. The stock radiator core is only a 0.50-inch thick. The Koyo core is just over 2 inches. The all-aluminum construction can also withstand higher pressures, but the unexpected consequence is an extra 8 pounds. The weight should be well worth it, as the Koyo radiator seems to have ridiculous cooling abilities. We installed a functional temperature gauge (actually a Techtom CMX-100 that reads the stock temperature sensor) only after installing the radiator, so there's no direct comparison with the stock radiator, but the Koyo runs shockingly cold. Coolant temperatures are usually in the low 160s, giving us some good headroom for the summer track days to come.

The Koyo radiator has all the proper fittings for installation in the stock location and with the stock shroud. We test fitted it in the standard position before getting all silly with our custom mounting and found the fit perfect. Even the little aluminum posts designed to hold the rubber mounting feet were exactly the right size, allowing the feet to slide on with a little rubber squeak, and never come off again.

The lower radiator hose is straight before entering the radiator, allowing us to simply cut it shorter and re-install it. The upper hose ends up being a very short, very sharp zig-zag, which we accommodated with a Thermal-Flex flexible hose (though not in time for the photos).

For electric fans, we stumbled upon a Flex-a-Lite dual 10-inch fan that's exactly the same size as the Silvia/240SX radiator and is rated at a healthy 1800 cubic feet per minute. In fact, the fans are such a good match for the core we had a hard time finding any extra room on the edges to mount them. We eventually used the last-resort, plastic-strap-through-the-core method provided with the fans. We're not proud of that.

The Silvia ECU has no provisions for turning on the fan, so we're resorting to a temperature switch used by early air-conditioned 240SXs. Mounted in the lower radiator hose, it was used to turn on the supplementary electric fan when the air conditioner was on and the radiator got warm enough. Wired to a relay, it should work for our fans. We're still a junkyard trip away from making that work.

We chose an intercooler less for its cooling abilities than its willingness to let air through to the radiator. A'PEXi's unique trapezoidal core makes some sacrifices in charge air cooling, but the tradeoff is better external airflow through the core. While this may ultimately cost a few horsepower, the ability to keep using the power we have lap after lap without overheating seems far more useful. The core we used is actually listed in the A'PEXi catalog as a Silvia core, but we used different end tanks for our unique mounting arrangement.

Two-and-a-quarter-inch silicone hose will, with some smooth talk and lubrication, barely squeeze onto the end tanks. Since we were using 2.25-inch intercooler pipes, no step hoses were needed to adapt from the pipe size to the end tank size. This allowed us to use 90-degree elbows instead, simplifying the plumbing even more. All the hoses (3-inch couplers at both ends of the intake pipe, 2.25-inch elbows at both ends of the intercooler, a 2 to 2.25 adaptor at the turbo discharge and a 2.25 to 2.75 adaptor at the throttle body) are ThermalFlex silicone custom-made in exactly the same metallic red as the valve cover. Yeah, yeah. Color-matched silicone hoses on a flat-black spray painted car. We know.

We built the pipes themselves by simply cutting and welding steel mandrel bends we bought from Road/Race Engineering. If you ever want your life to be truly complete, learn to weld. Being able to whip up this plumbing in a long afternoon is a very satisfying, life-affirming experience. If you want to skip the affirmation and get straight to the horsepower, Road/Race would charge about $400 to build plumbing like this.

The idle air supply hose should be attached to the intercooler discharge pipe. If it's routed to the intake pipe, boost will leak out the manifold through the idle control valve. Conversely, the crankcase breather should go to the intake pipe (or a ventilated catch tank) lest you pressurize the crankcase. The biggest complication in the pipes themselves was the tight bend needed at the compressor inlet. Normal U-bends weren't sharp enough. We had to use an expensive piece of a stamped-steel, tight-radius donut. Using a silicone elbow may have worked here as well, but late at night, when we discovered the problem, steel was the only option.

The Unexpected Stuff
You don't expect to have to make power steering lines when you change intercoolers, but moving the radiator back caused interference problems with the high-pressure hose running from the pump to a fitting on the T/C rod bracket. The hard line on the other end of this fitting can be bent up and back, to place the fitting on the frame rail, but shortening the hose gets complicated. After another daily affirmation with the welder, we had Earls fittings welded on what were once the ends of the stock hose. We were then able to use high-pressure Earls hose.

The low-pressure hose from the power steering tank to the pump also needed some love from the welder. The intake pipe wanted to go right through this hose, so we bought a length of power-steering hose from the local auto parts store, cut, rotated, and rewelded the intake power steering pump intake, and made room.

On the other side of the engine compartment, the intercooler discharge ran right into the battery. Instead of moving the battery to the trunk, however, we got an Odyssesy PC680 battery, which is amazingly small, totally sealed, and can be mounted in any orientation but upside-down. The battery is so light at 15 pounds that the weight of a long battery cable running from the trunk seems significant, so we decided to mount the battery under the front passenger's knees. Installation was a snap with the simple battery-mounting bracket available from Odyssey, and the battery is so small, the seat goes right over it when slid forward. Plus, the positive battery cable can go through the same grommet as the ECU harness.

In the engine compartment, we attached one of the screw-on terminals from the battery to an insulating hardwood block and ran both the new and original positive battery cables to it. Surprisingly simple.

Tuning
This is the part that makes it all work. Jim Wolf Technology has done this before. It has the Silvia ECU so completely cracked, it can use it to run Hondas, Briggs and Stratton single-cylinder engines, 400-hp Silvias... anything. The Silvia's stock 380 cc/min injectors run out of flow at about 280 hp, about the same time the stock airflow meter runs out of signal. Nismo injectors, 555 cc/min, drop in to move the fuel ceiling to 350 hp.

On the airflow meter side, there are two options. A Ford Cobra mass airflow sensor is good to 350 hp, and can be extended to more than 500 hp with the addition of a circuit that knocks 480 mV from the signal.

A 300ZX twin-turbo sensor is also good to 520 hp out of the box, or perhaps out of the junkyard. Since we were only anticipating 300 hp at the most, the Cobra sensor made the most sense. Using the full range of the sensor ensures maximum resolution, offering better control over idle and light-throttle fueling.

JWT offers a Cobra sensor kit, complete with connector, solder and heat-shrink tubing to splice the Ford connector into the Nissan harness. The ECU, of course, has to know which injectors and flow meter you're using before any of this stuff will work. JWT has programs for various levels of tune on the shelf, but if you're within driving distance of San Diego, Calif., it's worth a trip to have the ECU fine-tuned on its dyno. We did exactly that, and made about 14 more hp than we would have if JWT had just mailed us a program.

The Potato turbo's standard wastegate pressure is 11.5 psi. We used a simple adjustable bleed valve from JWT to push that to 15 psi on the warm air of dyno, which, in favorable winter weather conditions, seems to be 16 or occasionally 17 psi on the street.

Driving It
Holy one-wheel burnout! This is a whole different Silvia. Our goal of matching the Disco Potato was realized with unexpected precision. Overlaying our 274 hp with the Potato's 280, the two cars seldom vary by more than 5 hp. With a 274,000-mile open differential and only 2,600 pounds to push around, that means tire smoke in first, questionable traction in second, and unholy thrust in the rest of the gears.

The turbo is amazingly responsive, making 5 psi of boost as low as 2000 rpm on the freeway, and eagerly making full boost by 3000. Modulating the throttle is as easy as it can be with this much boost. Feather the throttle and the boost gauge will move with your foot. Never has a car needed a limited-slip differential more than this one does now.

We measured intercooler inlet and outlet temperatures under real street conditions. First, we drove the car at full throttle at 4000 rpm, keeping speed steady with the brakes. This was done in both second and fourth gear to see the importance of air speed. Next, a full-throttle blast up an entrance ramp, with temperature measurements at the beginning and end, examined a more real-world situation.

With 274 lb-ft of torque, it turned out to be impossible to duplicate the second gear test. Using enough brake pressure to hold speed steady, the front tires locked and the car just slid, ruining its tires. Now that's torque. The test shows the compressor discharge temperature is similar to the old turbo, but remember, the new one is making more boost and moving much more air volume. Measurements on the dyno also showed the stock intercooler was causing a 2- to 3-psi pressure drop, while the new one, despite vastly more airflow, is between 1 and 2 psi, depending on rpm.

Automotive-style terminals can be screwed onto the Odyssey battery, but since we mounted the battery on the passenger's floor, we used the terminal, mounted to a block of insulating hardwood, to simultaneously connect the stock positive battery cable to the new battery and provide an underhood jump-starting/power takeoff point.	The stock airflow meter (left) is only capable of supporting about 280 flywheel hp before it reaches its maximum voltage. Jim Wolf Technology can substitute either a Ford Mustang Cobra meter (right) for up to 350 hp, or a 300ZX meter for more than 500 hp.	There is no way a stock 240SX fuel pump should be feeding a 270-hp engine. A 300ZX twin turbo pump will fit in the stock location and has plenty of reserve. The pump is surprisingly loud. The 300ZX ran the pump at low-voltage under low-load conditions, but the 240SX has no such circuitry, so it runs full blast all the time.