Dale Amy
June 1, 2004
Pedal to the metal on modular experimentation, after creating ported versions of each casting, Livernois Motorsports says the pre-PI head casting in the background gets an unduly bad rap in com-parison to its newer PI brethren. Yes, the first-generation SOHC modular was a weak-kneed little twit out of the box, but perhaps we've been laying too much blame on the heads themselves. That's not all the Livernois boys have been working on, though.

Horse Sense: To see how affordable it would be to score a set of non-PI (read: '96-'98 GT) cylinder heads for porting cores, we jumped on the Mustang Parts Specialties Web site. There we found used non-PI heads for only $100 each. Knowing how these would perform with the Livernois touch, we had found a real performance bargain.

Don't relegate those non-PI heads to bass-boat-anchor duty just yet. The ever-inquisitive minds at Livernois Motorsports have discovered that those much-maligned castings, as fitted to '96-'98 GTs, weren't quite the restrictive corks we all thought them to be. With a little touch-up work-and the proper sup-porting cast-they can support big power. More surprisingly, it turns out that even stock non-PI heads can do a decent job if you feed 'em properly.

You'll no doubt remember that the wheezy first-generation modulars could manage only 215 factory-rated horsepower, compared with the 260 ponies claimed for their newer PI (Power Improved) cousins. Through extensive dyno and flow-bench testing, Livernois has determined that a lot of the PI's improvement can be attributed primarily to its intake manifold, and secondarily to its revised intake valves, and camshafts, rather than to the heads themselves. Not to put too fine a point on it, the old, non-PI manifold was crap.

One of the first-generation modular's worst components was its flow-handicapped intake in the foreground. What would happen if we bolted the much-improved PI variant on in its place? But we can't because of the obvious incompatibility of port shapes between them. Or at least we couldn't unil now.

Livernois airflow specialist Rick Swain-a man whose rsum includes many years at Roush working on Winston Cup and other race cylinder-head programs-found the fatal flaw in the non-PI intake's design in seemingly non-scientific fashion. He stuck a hand deep inside, felt around, and discovered that the individual plastic intake runners simply jutted, rough-edged, into the plenum, causing flow nightmares as incoming air tried to transition into the runners. The PI intake, on the other hand, has a runner/plenum interface that is carefully finished and gently radiused to greatly enhance flow. Naturally, this had Rick wondering what effect a PI intake would have on a non-PI engine.

So he whittled out an exploratory set of 3/4-inch-thick aluminum plates with passages that smoothly adapted from the PI intake's rounded-rectangular port shape to the non-PI's near-circular shape. To Rick's gratification, sandwiched between a PI intake and pre-PI heads, these plates produced eye-opening results on the dyno. But hold the champagne-they proved just a bit too high to fit under a stock hood. The total height increase was 0.750 inch for the plate, plus 0.110 inch for the necessary extra gasket, totaling 0.860 inch. But Rick was faced with somehow cutting the total height increase down to no more than 0.400 inch to fit underhood. The obvious answer was slimmer plates, but a plate thickness of 0.550 inch was the minimum necessary to facilitate a smooth transition between the distinct port shapes.

Half of Livernois Motorsports' solution are these CNC-carved aluminum adapter plates that match up to the PI intake's port design on one side and to the non-PI head's round-port design on the other.

But, add the 0.110-inch gasket to that and you get a total of 0.660 inch-still too tall-so the only option was to mill the top of the manifold where the inlet adapter bolts, shortening it by 0.300 inch, for a total installed height increase of only 0.360 inch. Why are we reeling off all these boring dimensions? So you'll understand why you can't just buy simple adapter plates and bolt on a stock PI manifold. The bottom line is, Livernois has to sell the plates and the modified manifold as a package, but a look at the dyno results comparing this setup to a stock non-PI manifold will show that this is still an outstanding bang-for-the-buck purchase for you '96-'98 guys out there.

To test the effectiveness of its new PI manifold and adapter-plate combination, Livernois strapped a stock '96 Saleen on the Dynojet, made some baseline pulls, and then bolted on the PI manifold/adapter plates for comparison, with no other changes. When the fumes cleared, peak-to-peak gains were in the order of 22.7 hp and 1.8 lb-ft. Not much had changed below about 3,600 revs, but notice that, at 6,000 rpm, the car was healthier by almost 38 hp, while, at 4,500 rpm, torque had swelled by more than 35 lb-ft. If you don't feel those kinds of gains, it's time to find another hobby.