Ed Hohenberg
May 20, 2006

Horse Sense: Ed Hohenberg is the main manat Wolf Performance Engineering (www.wolfperformance.com), whichprovides high-performance and mechanical engineering (yeah, realengineering) services. While Ed is obviously a pretty sharp guy, he'salso an experienced racer who's had success racing a Fox Mustang in NHRAStock Eliminator. He's also an experienced tuner that sells bothDiabloSport and SCT products.--Ed.

The '03-'04 SVT Cobras are amazing cars. Not only do they excel in allmanners of vehicle dynamics (acceleration, braking, and cornering), butwhen you aren't leg-pressing the stop-and-go pedals, they're alsocompletely sedate for day-to-day driving. What's even more amazing ishow easy and economical it is to add significantly more horsepower tothese cars--like another 100 hp--which we did one day last summer for lessthan $1,200.

Our recipe was the tried-and-tested combination of after-cat exhaust,low-restriction air intake, smaller supercharger pulley, and performanceflash tuner. In the end, we picked up almost 90 rear-wheel horsepower(equating to roughly 105 hp at the flywheel, using the standardSuperFlow dyno correction for driveline efficiency). What you reallyfeel when you squeeze the go-pedal, however, is the torque--we picked upan amazing (there's that word again) 120 lb-ft at the wheels (more than140 lb-ft at the engine).

Other than the exhaust system (not shown), here's more than 100 hp justwaiting to be installed. Parts include a K&N FIPK air intake, a2.80-inch pulley and corresponding belt, the two flash tuners we tried,and new copper spark plugs (recommended when running additional boost).

But there's more than just the before-and-after. In order to provide ourreaders the best data to use as smart shoppers, we systematicallychassis-dyno-tested each modification, so you can see which made themost horsepower per dollar. To make sure the dyno data weren'tflash-in-the-pan flukes, all comparison dyno runs were done on the sameday, and were run twice for each mod (hot-lapped for the second run toensure fair test comparisons--for example the engine, tires, gear oil,and such were warmed up). Several engine parameters were also loggedduring the dyno testing, again to ensure maximum consistency between theruns. In the end, the better of the two back-to-back runs was taken forcomparison data, but typically the difference was less than 2 hp betweenthe two runs.

The first step was to baseline our Cobra, and this particular '04 snakedidn't disappoint. Stone stock and with just over 2,000 miles on theodometer, our Cobra laid down 381.5 hp to the rear wheels at 6,000 rpm,revealing the factory's 390hp flywheel rating was a bit conservative. Onan inertia chassis dyno (we used the SuperFlow at Motion Performance inWinnipeg), these cars are typically in the 360-380 rear-wheel-horsepowerrange, so evidently we ended up with a strong one. One thing to noteabout the stock dyno run is the air/fuel ratio goes really rich beyond6,000 rpm. This is likely due to the exhaust catalyst monitor in thestock tune, which, when it senses the cats might be getting too hot,dumps extra fuel into the engine to cool them back down.

Tune Time

Up first on the stone-stock Cobra was the DiabloSport Predator flashtuner. From the comfort of our passenger seat, we simply plugged theunit into the OBD-II port, then followed the on-screen prompts toeffortlessly add the new tune.

A performance tune was our first modification on the stock Terminator.While the '04 EEC V will still take a chip, we stepped up to the latestand greatest flash tuners--the DiabloSport Predator and the new SCTXcalibrator 2. Both tuners can be found for around $400 by a smartshopper and are loaded with many other features, in addition to theperformance tunes.

Installing the tunes was as easy as plugging the units into the OBD-IIport under the dash, then following the on-screen prompts. The SCT tunerhad three tunes installed for differing fuel-octane numbers (we wererunning 94-octane pump gas), while the Predator had a 91-octane tune, ageneric "performance tune," and a "pulley tune," the latter tworequiring 93-octane fuels. We chose the 93-octane tune for the SCT, andthe Predator "performance tune" for the stock Cobra. Then it was dynotime to see how the tunes performed.

After dyno testing with the Predator tune we flashed the stock tune backin, then tried the SCT Xcalibrator 2 tune. Adding the tune was astraightforward deal, although the process was a bit quicker with theXcalibrator 2.

While both the SCT and Predator tunes made more torque and horsepower,both also ran the A/F richer than stock, which was already quite rich.In hindsight we realized the reflashing process would clear any "learnedcorrections" in the EEC, so the gains we saw on the dyno from the flashtuners may have actually been greater had we cleared the stock EEC oflearned corrections for the baseline run. The EEC had previously learnedto lean itself out--hence the leaner A/F ratio for the stock baseline,which would make more power than a richer baseline.

The Predator performance tune made the previous high-rpm rich situationeven richer and, as a result, gained only a few peak horsepower overstock, even though rear-wheel torque was up by 10 lb-ft until the A/Fwent rich. This doesn't sound like much, but when we installed the samePredator performance tune with the after-cat exhaust and cold-air-intakeupgrades later, horsepower gains were much more dramatic over the stocktune. With the SCT tune, torque improved all across the board, and thepeak power was also up significantly, likely due to the A/F being leanerand more stable at the higher rpm range, as compared with the stock orPredator tunes.

For all dyno testing, we monitored and datalogged several engine-controlparameters, such as engine temperatures, rpm, spark advance, mass airvoltage, throttle position, and fuel pressure, using an SCT Raptorhigh-speed datalogger and our trusty laptop.

While the Diablo tune still went rich at the top end due to the catalystprotection programming, SCT disabled the catalyst temperature models intheir tune, resulting in a consistent A/F ratio at high rpm. Of course,if you choose to drive at 6,000 rpm all day long at WOT, you'll likelydamage your cats from overheating, but for short blasts they'll be fine.In the end, the SCT strategy worked for maximum horsepower, to the tuneof 16 peak rear-wheel horsepower over stock, with a stout 28 lb-ft gainin peak rear-wheel torque.

Flash Tuner on stock Cobra--cost per rear-wheel horsepower: $25





The Pipes

First on our list of mods was to replace the stock after-cat system.Here's why: The tailpipe outlets fake you into thinking the stockexhaust is bigger than it actually is. In fact, the stock tailpipes areonly 2.25 inches in diameter and full of crimped bends that furtherreduce the flow potential.

Next up was the exhaust install. The original mufflers did a great jobsilencing the factory-blown Cobra down to acceptable EPA noise levels,but at a significant restriction to flow. Similarly, the factorytailpipes with their crimped bends and only 21/4-inch diameters alsodidn't do the best job flowing the exhaust out the back.

The cure was a MagnaFlow after-cat exhaust system for the Cobra. Thesmart-shopping cost for the kit is around $500. Installation is simpleand straightforward, without the need for specialty tools assuming theaverage hot-rodder installs all the parts used in our combination ontheir own. Installation costs aren't included in our calculations. Butwe weren't up to crawling around under the car for an afternoon, so wetook the car and kit to Keystone Ford/Mustangs by Matt, where the techsused their lift for a more comfortable install. Rick Clemenson fromKeystone Ford did the after-cat installation on our Cobra, while we gotin his way taking pictures or stood around doing nothing.

We replaced the stock after-cat with this system from MagnaFlow. Noticethe real 2.5-inch pipes, and smooth mandrel bends everywhere. The shinyfinish was also a welcome change from our rusty-looking stock pipes.

The stock after-cat system is attached at the muffler-extension-pipeflanges connecting to the H-pipe, has two rubber hangers per side, andbolts onto hangers for each exhaust extension. To remove it, undo theextension hanger bolts at the rear, the flange connectors up front, thenlube and pry the after-cat out of the rubber hangers to remove eachmuffler/tailpipe assembly. Each side is one complete welded assembly.Possibly the best thing about the IRS system on these Cobras is the easewith which the after-cat can be replaced, as the tailpipes do not haveto snake up and over the axle--they go below everything so they just dropout.





To install the new system, Rick first (loosely) mounted up the front extension pipes and mufflers for both sides

Installation of the new after-cat is only slightly more difficult, sinceeach side now consists of many individual parts. Reading theinstallation instructions is always a good idea, but even macho he-mantypes who refuse to read should be able to get it right the first time.After scanning the instructions, Rick mounted the muffler-extensionpipes loosely, slip-fit the mufflers, hung the tailpipes on the factoryrubber hangers, slipped the system together on each side, then alignedit all and tightened the U-clamps holding everything together. Last, thetailpipe extensions were slipped on, aligned, then clamped tight andbolted up with the factory hangers.

Just the looks and sound of the new after-cat were worth the money, buthow well would it actually perform--how much real horsepower versuspsychological horsepower? Back on the dyno at Motion Performance,MagnaFlow did not disappoint. On the otherwise completely stockTerminator, the after-cat alone picked up over 30 hp at the wheels, withanother 21 lb-ft of torque--all that while still having a respectablymellow exhaust note (i.e., your neighbors aren't going to hate you).

After-cat exhaust on stock Cobra--cost per rear-wheel horsepower: $16

Deep Breathing

This is the torturous path air must take to get to the mass air meter with the stock setup. It does a great job of silencing inlet noise (quite significant with a blower), but with a flow restriction penalty

Now that we had improved the flow of bad stuff out of the engine, weturned our attention to improving the flow of good stuff into it. Again,with a factory-blown car, much effort went into silencing the intakenoise for EPA reasons, resulting in a quiet, but somewhat restrictivefactory air-cleaner setup. An aftermarket, low-restriction, cold-airintake system is a must for performance applications.

Word on the street was good about the K&N FIPK setup, and the price wasright (you can Froogle it for around $150), so it would become oursystem. Installation took about 15 minutes, with the car still strappedto the dyno, using the most basic of handtools and following theincluded instructions.


With the filter element removed, here you can see the direct air path to the mass-air sensor on the K&N FIPK, and the gentle radius on the inlet to minimize turbulent flow.

The best we saw on the rollers was a 7hp increase at the wheels and asmall reduction of peak torque. From previous dyno-tuning experience, weknew that installation of many cold-air intake systems affects theaccuracy of the mass air sensor, so it's possible there is a sweet spotto hit during installation of the FIPK that may result in a largerhorsepower gain by changing the A/F ratio, but we didn't find it. Wedid, however, find a significant horsepower gain when dyno testing withthe hood open versus closed (see the graphs at our Web site), so thatmight have contributed to the gains seen by K&N in its testing.





Our K&N filter system took about 15 minutes to install, using simplehandtools you probably have in your kitchen drawer. The hardest part wasactually getting the trim lock in place around all the edges. With theFIPK in place, blower whine was noticeable, again compromising ourCobra's stealth.

Even though the K&N FIPK horsepower gains weren't huge on our car, theblower sound with the FIPK was nonetheless impressive, and probablyworth the cost for some people. In the end, the cost was minimal, andinstallation a snap, so 7 rear-wheel horsepower was still a good deal.

CAI--cost per rear-wheel horsepower: $21










Tune Time, Part II

With the CAI and exhaust installed, we thought we'd retest the flashtuners on the now-modified Cobra to see if the results would change. Andchange they did. The flash-tuner horsepower gains were greater on themodified Cobra--now at 20 rear-wheel horsepower (see the graphs). Andwhile the Xcalibrator 2 edged out the Predator "performance tune" again,this time the results were much closer. If the A/F ratio had not gonerich again at high rpm, the Predator could have come out on top.

Flash Tuner on modified Cobra--cost per rear-wheel horsepower: $20

Extra Pressure

Believe it or not, on '04 Cobras there's nothing holding the blowerpulley on other than the press-fit on the shaft. Where there used to bea bolt on earlier cars, now there's just a plastic "Christmas Tree" plugcovering the threaded hole in the shaft. After removing the belt guard,we used a door-panel-removal tool to pry out the plastic plug. We foundthat simply unscrewing the plastic plug worked well too.

When equipped with a blower and in need of more horsepower, it'sstandard practice to just crank up the boost whenever possible. So whyshould we be any different?

Blower pulleys for the '03-'04 Cobra are everywhere, in various sizesand dirt cheap in many cases. We picked up a brand-new, steel 2.8-inchpulley from eBay for $35. Heck, the shorter belt cost us more than theactual pulley. $100 is typical for belt and pulley. And while a2.76-inch blower pulley is the smallest available and will spin theblower the fastest, theoretically making the most boost, we chose tostick with a 2.8-inch, since we had no intention of running extra idlersto minimize belt slip.

Belt wrap is reduced with smaller pulleys, so belt slip can become a bigproblem. One way to prevent slip is by using additional idlers, thusincreasing the belt wrap on the smaller pulley. Another solution is tostick with the stock-size blower pulley, but go to a larger crankpulley. The problems with this approach are the increased cost of thecrank pulley, which is significantly more than a blower pulley, and theother accessories spin faster too, which you may not want.

Before placing the puller tool on the pulley, we protected the rear faceof the factory pulley and blower-housing snout with electrical tape toprevent any scratches or "incriminating" marks, since apparently Fordchecks for evidence of pulley replacement for warranty claims. With thepuller set up, it was simple to remove the stock pulley (although athird hand helps)...

To minimize our belt slip, we did some research. First, we found thatsteel pulleys gripped better than the aluminium ones. Then we learnedthe solid-rib belts such as the Gates K080740 belt we chose didn't slipas much as the Gatorback-style cut-rib belts, which apparently also makea whistling noise. We think our research paid off, since we have yet tosee any of the black dust characteristic of belt slip gather anywherearound our pulley, and the inside of our belt guard is still clean aftermany dyno pulls and several quarter-mile jaunts.

By the way, did we mention the pulley is press-fit onto the blowershaft? And that's it--there's no actual bolt that holds it on. Believeit. So installation of the new pulley is where our "kitchen tools"finally failed us. We needed a specialized tool to pull off the oldpulley and press on the new one. Luckily, we had a friend with a similarCobra, who was more than happy to lend us his pulley puller, and a thirdhand if we needed. With the correct tool and good instructions, thepulley replacement was simple and quick. We had the old one off and thenew one on in about 15 minutes, while the car was still strapped to thechassis dyno.

...and install the 2.80-inch pulley in its place, using the puller as apusher. With the new pulley and shorter belt on, we reinstalled the beltguard to conceal our performance mod.

It would be more than just a little foolhardy to still use the stocktune when you crank up the boost with a 2.8-inch pulley that adds about4-5 psi on top of the stock 8-10 psi. So, in the interest of preservingour engine, we would first dyno the pulley with the SCT tune to be safe.Replacing the detonation-prone factory platinum spark plugs with somegood, old-fashioned copper plugs was another safety measure. For ourinstallation, we chose NGK TR6 plugs, gapped at a boost-friendly 0.030inch.

Rear-wheel horsepower gains were moderate at 20, but peak horsepoweroccurred at a significantly lower rpm, showing the loss in compressorefficiency of the Eaton at the higher speeds it was now spinning. If youlook at the gains in torque another 70 lb-ft at the wheels, you see amuch brighter picture. Torque is the real hero here.

The factory platinum plugs may last a long time, but when turning up the boost they act as hot spots, just asking for preignition and damaging detonation. To be safe, we installed some NGK copper plugs, and tightened up the gap to .030-inch so they'll reliably fire under the higher combustion pressures we expect.

Not to leave the Predator out of the pulley picture, after baseliningthe pulley with the SCT tune, we installed the Predator "pulley tune."Finally the tables had turned--the Predator powered beyond the SCT tuneand picked up another 7 rear-wheel horsepower. Still going pig rich atthe top end, we wondered how much more power potential the Predator tunehad if the A/F would have stayed constant.

Pulley--cost per horsepower: $5. But a pulley shouldn't be attemptedwithout a revised tune, so...

Pulley and flash tuner--cost per horsepower: $11







Play Time

With our newfound power and torque, we hit the track after adding somemuch-needed drag radials, and cranked out an 11.79 at 118 mph. The goodnews is we picked up over a second in e.t. and 12 mph from stock, butthe bad news is the smoke you see behind the car is from the death ofour stock clutch. With an upgraded clutch to handle the extra 140 ft-lbof torque, we expect even better times at the track.

To improve on the out-of-the-box flash tunes, we further custom-tunedthe Cobra. Using the SCT tune as the baseline, we tweaked the fuel andspark using SCT Advantage III software until the horsepower exceededthat of the Predator, yet we still felt the A/F and spark advance weresafe for pump gas. While we were at it, we also programmed in a two-steprev limiter we could use at the dragstrip. The cost for the custom tunewas nothing for us, since we had the software and knowledge, but forregular folk custom-tuning prices vary depending on the tuner. As such,we couldn't calculate a cost per horsepower on the custom tune.

The End

After almost $1,200 and over 20 dyno runs, what did we learn? The bestbang for the buck was the combination of flash tuner and pulley. Not abig surprise. The after-cat exhaust was the sleeper in our test,however, which did surprise us with a solid 30 rear-wheel-horsepowergain. The biggest lesson we learned was you can't add 140 lb-ft oftorque and expect the stock clutch to survive. D'oh!

On the Dyno

Well, finally, here it is--the official "before and after" dyno chart.Total horsepower gain at the wheels was 88, but more importantly thewheel torque gain was 121 lb-ft, all using an A/F ratio similar to stock(safely rich).