Under the Hood: DodgeSRT-4

If the Dodge SRT-4 needs anything, it's a limited-slip differential, not more power. For 2004, it's getting both. In addition to a factory-installed Quaife differential, the new SRT-4 is getting a ratings boost from 215 hp to 230. As anyone with a right foot already knows, the "215 hp" SRT-4 was already underrated, making 223 hp at the wheels on our Dynojet. So is the new power real, or just a ratings adjustment? That depends on the weather.

We dyno test with a decent fan and a monkey spraying water on the intercooler. This, we've found, is a reasonable simulation of the 80-mph breeze that should actually be blowing across the intercooler at the top of third gear. DaimlerChrysler tests with a relatively hot intercooler, so it gets less power. The SRT-4's engine management actually adjusts for hot weather by increasing boost until it's making the same power it would in cold weather, so there should still be no difference. Problem is, the 2003 model had injectors that weren't quite big enough to allow full compensation under DaimlerChrysler's test conditions.

For 2004, injector flow rate has been bumped up 10 percent, leaving the headroom for better altitude and temperature compensation. Because DaimlerChrysler measures injector flow differently than the aftermarket, getting a useful flow rating is difficult. In DC world, the new injectors are 577 cc/min. In our world, with lower fuel pressure (43 psi is the older standard; DC uses 58.) and different test fluid density, we think they may be considered 378cc/min. We'll bench test some when we get our project car.

With the new calibration, we've been told to expect about 5 hp more under our test conditions, and about 15 in more severe heat.

But power isn't enough to satisfy some of you. Since our first test of the car, we've received approximately 6,000 letters saying exactly the same thing: "Sure the SRT-4 is fast, but it's still a Neon. It'll break down before the end of the quarter mile, blah blah blah..." We really wanted to shoot back some knowledgeable defense of the poor Neon, but ignorance held our tongues. Sure, the 70,000-mile powertrain warranty suggests some level of engine strength, but we'd never actually seen inside one.

To close this glaring knowledge gap, we went to Michigan with a camera and a list of geeky questions about rod bolt diameters, valve angles and combustion dynamics.

What we found is a ridiculously strong-looking engine that should survive many a quarter mile. You can thank the PT Cruiser for the SRT-4's powerplant, as it was a desperate need for a horsepower-based PT sales boost that justified turbocharging the 2.4-liter engine and gave the development team the signal to go overboard.

With a short deadline on engine development, there was no time to prove the durability of low-cost, marginal-strength parts. Where typical engine development teams would spend time whittling away pennies finding the least expensive design that could meet durability targets, the turbo 2.4 team just threw in stronger parts and more expensive materials, cost be damned.

Along with its beefiness comes heft, however. The SRT-4 engine weighs 339 pounds (with the alternator, but without the A/C compressor, power steering pump, or any oil in the pan). That's enough words, look at the pictures. n

	The cast-iron block looks extremely stout, with a closed deck on top and a very strong split crankcase. Rather than individual main bearing caps, or even caps joined by a main bearing girdle, the block simply extends past the crank and is split right down the crank centerline. The bottom half (called the bedplate) incorporates all the main bearing caps. In addition to improved strength, this design offers reduced noise and vibration, so it's becoming fairly common in modern engine design.	The crank is cast, but the use of higher-hardness steel and careful attention to detail makes it quite strong. The bearings are large, at 60mm for the mains and 50mm for the rod bearings, and naturally it's fully counterweighted. It's the details like the cold-rolled fillets at the edges of the bearings that really matter, though. The corner where each bearing meets the crank throw is an area of high stress concentration, so the corner is rounded (filleted) to spread out the stress and prevent cracking. By cold rolling this fillet instead of cutting it, the steel gets work hardened, just as it would in a forged crank, making this most critical area stronger. The SRT-4 drag teams are rumored to be making between 800 and 1200 hp on the stock crank (and stock block and bedplate), so we figure it's strong enough.
The connecting rods are forged in one solid piece, machined, and then broken in half. That's right, the rod-bearing cap is created by simply snapping off the bottom half of the rod bearing. These "cracked" rods fit together perfectly, thanks to the jagged interface, ensuring a perfectly round rod bearing when the rod is assembled. This cracking technique is common on flimsy powdered-metal rods, but relatively rare on forged parts.	The rod bolts are the same 9mm in diameter as the naturally aspirated 2.4, but use a slightly different thread shape (called MJ9 if you insist on knowing everything) with a slightly more rounded chamfer at the bottom of the threads to reduce stress concentration. Since the base of the threads is the bolt's weakest point, this simple change strengthens the entire bolt.	The extra heat of turbocharged combustion makes for some hot pistons. That heat is carried away in part by a spray of oil from an oil squirter at the bottom of each cylinder. The high volume of oil flowing from these squirters can rob valuable oil from the rest of the engine, so each sprayer has a small ball valve to shut off oil flow at idle and low rpm when the oil pump can't flow enough to supply everything.
Of course, those low oil flow situations will be rare, since the SRT-4 oil pump uses a high-flow gerotor set (that's the set of goofy-looking gear thingies that actually pump the oil) stolen from Chrysler's 4.7-liter V8.	The balance shaft assembly hangs under the front three cylinders. While it could be removed in much the same way we removed the balance shafts on the QR25DE in our SE-R Spec-V rally car (being sure to plug the oil feed hole and add a windage tray), the SRT-4 engine has a longer stroke, more reciprocating mass from its beefier rods, and worse rod/stroke ratio, which causes more violent piston acceleration. Removing the balance shafts would probably free up 8 or 9 hp like it did on the Nissan engine, but engine vibration without the shafts would be far more severe than it was with the relatively lightweight QR25. Besides, with a turbo, the extra 9 hp would be more easily achieved with a hair more boost.	The structural cast-aluminum oil pan hides some interesting details. The gasket, for instance, has some extensions that stick between the counterweights to scrape oil from the crank (no, they don't actually touch it). With the balance shafts in front, these scrapers are only possible on the rear of the crank. Oil passages in the pan carry oil to the oil/water oil cooler and oil filter. The lack of baffling around the oil pickup does make us a bit nervous about high-g cornering, but remember, the balance shaft assembly sits right above the oil pickup, acting like a big, thick windage tray. This should help keep oil down in the pan.
The turbo pistons, cast by Mahle from a eutectic aluminum alloy, have a shorter pin height (the distance from the wrist pin to the top of the piston) than the naturally aspirated piston on the left, lowering compression to 8.1:1. The turbo pistons also use full floating wrist pins, where the naturally aspirated pins are pressed into the piston. Full floating pins cause less friction and better handle the higher cylinder pressures. The skirts are coated with Mahle's Grafal low-friction coating for, well, less friction. The top compression ring groove is hard anodized (that's the gray band) to prevent the hot top ring from microwelding itself to the piston. As with all modern engines, the top ring land has been made as short as possible to reduce emissions caused by unburned fuel hiding in the crevice between the ring land and the cylinder wall. The smaller ring lands (4mm in this case, compared to 8mm on the old 1990 2.5-liter Chrysler turbo) can't survive as much detonation, but modern engine management makes that detonation much less likely anyway.	The shorter piston never reaches the top of the bore; this is actually top dead center. The "ski ramp" piston dome encourages tumble in the cylinder, which improves idle stability and light-load performance. The best wide-open throttle performance, though, usually comes from a piston that comes as close as possible to the flat edges of the combustion chamber (called quench pads). The gas trapped between the piston and the quench pads gets squished out, flying directly into the flame kernel where combustion is starting. This added turbulence speeds flame propagation and reduces the chance of knock. Of course, a taller piston would raise compression, so the ski ramp would have to be replaced with a dish, but that's do-able. If you add up the crank throw (half the stroke, or 50.5mm) the rod length (151mm), and the pin height (28mm) you'll find the top of the piston 229.5mm above the crank centerline. That's 8.64mm short of the top of the block (238.14mm). If it were us, we'd make the rod 159.64mm, which would improve the rod ratio from 1.5:1 to 1.58:1. Then we'd use a dished piston with flat areas to match the quench pads in the head. Of course, we don't care about idle stability and emissions, and we don't have the simulation software, math skills, or test engines to prove that this is a good idea, but it's a hunch. If we ever manage to blow up a stock one...	The combustion chamber is classic pent-roof, with the spark plug in the middle where it belongs. The flat area extending between the intake valves (the bigger ones), and the smaller area hiding behind the exhaust valves would be the quench pads, if the pistons were designed to use them. Engine geeks take note: The intake valves are 34.8mm, the exhaust valves 28.45. For comparison, that's smaller than the 35.65mm intake and 30.65mm exhaust valves in Nissan's QR25. The included angle of the valves is fairly steep, at 48 degrees. This steep valve angle was common in the high-output Japanese four-valve until 10 or 15 years ago, but the newer engines have flatter combustion chambers on the order of 25 degrees. There's good and bad with both designs. Lastly, get this: The exhaust valves are made of Inconel, a hyper-expensive high-nickel steel superalloy designed for the high-stress, high-temperature world of gas turbine blades.
The valvetrain uses rocker arms, which increase valvetrain mass compared to a cam-on-bucket design, but with the long stroke, valve float won't be the thing limiting redline. Rocker arms allow the cams (and their big cam gears) to sit closer together, making the cylinder head physically smaller, and, more importantly, they allow more aggressive opening and closing ramps on the cams. Valve lash is hydraulically adjusted, and the cam hits the rocker through a low-friction roller. All good stuff.	Here's an important detail when you're trying to cool 223 hp. Water pumps are made up of several vanes arranged around a disc-shaped rotor, and the water between these vanes gets flung to the outside as the pump spins. Typical water pumps have exposed vanes, and the trapped water is surrounded by a vane on each side, the rotor, and a machined face in the block or front cover where the water pump is mounted. There is a small gap between the vanes and this machined face, though, so water can leak past into the next chamber. The old Chrysler 2.4-liter pump was a six-blade stamped steel piece of junk. Switching to a fully shrouded seven-blade plastic pump increased pump flow by 24 percent and output pressure by 48 percent. That's better.	Things start getting strange here. Tight packaging in the PT Cruiser forced some creative thinking on the turbo. The exhaust manifold and turbine housing are cast in one piece by Mitsubishi from high-nickel Ni-Resist steel. The one-piece design improves flow, reduces size, reduces thermal mass for quicker cat light-off, and makes it a pain in the ass to upgrade the turbo.
Things get stranger. The turbine discharge is also part of the manifold/turbine housing casting, and it loops back around and hits the manifold again on its way to the catalytic converter. Where they meet, there's a wastegate valve. Keeping the wastegate valve away from the turbine housing improves flow where it matters most. Problem is, flow where it matters secondmost, in the discharge pipe, looks like it will be fairly upset about all this wastegate flow jumping into the party. One of these days, we may try an external wastegate that flows back into the exhaust further downstream. Such a modification would require some fairly substantial fabrication on the manifold.	A few more interesting things about the turbo: Packaging constraints dictated a reverse-rotation turbo that spins counter-clockwise, unlike most turbos (but like the EVO VIII's). The Mitsubishi TD04 compressor has a compressor bypass valve built right into the compressor housing. The silver can on top is the diaphragm that opens the valve to vent boost from the volute (the outside of the snail) back to the compressor inlet. Both the wastegate and bypass valve get their boost signals via ECU-controlled solenoids.All together, the turbo is a Mitsubishi TD04LR-16Gk with a 6cm2 turbine inlet. This will only mean something to you if you're familiar with Mitsubishi turbos, and even then maybe it won't. Mitsubishi wouldn't cough up the compressor maps for this turbo, but we did learn that peak compressor efficiency is a very healthy 77 percent.	OK, engine management geeks, that's a 32-tooth crank trigger wheel (34-2). The engine management is model based, which means it constantly calculates the appropriate fuel and timing outputs instead of just looking them up on a table. The throttle position is interpreted by the ECU as a demand for a certain amount of torque, and the ECU does whatever it can to deliver. If it's hot out, or you're at high altitude, the ECU will try to satisfy your demand by increasing the boost to match the power output under normal conditions (presumably a dreary day in Detroit). Helping with the math are a few unusual sensors feeding the ECU valuable information about underhood conditions. The Air Charge Temperature (ACT) sensor measures the temperature of the air exiting the intercooler, and the Throttle Inlet Pressure (TIP) sensor measures pressure before the throttle body. These are used to predict the turbo speed, so the ECU can avoid overspeeding the turbo. The expected Throttle Position (TPS) Sensor and Manifold Air Pressure (MAP) sensor are also there, but there's no Mass Air Flow (MAF) sensor. Mass flow is calculated from the other values. The TIP sensor occasionally pokes its head out and measures ambient air pressure, too. Trick daddy. Both the MAP and TIP sensors have a 2.25-bar range, meaning they'll read from absolute vacuum to just more than 18 psi. Sensors with larger ranges will be included in Mopar's staged upgrades.