With preliminary numbers that looked encouraging, we allowed the engine to cool down and installed the inner spring to the valvespring assemblies. Our loads were now quite high for a flat-tappet application, just in the range of what would normally be used with a serious flat-tappet cam. The 930 is a conventional dual-spring with a 1.509-inch outside diameter, and has proven to offer good rpm potential with a typical high-performance flat-tappet camshaft. We also took this opportunity to make a jetting change, leaning the carb six jet sizes at the front and rear. The dyno was set to make a pull to 6,000 rpm. The results showed the jetting to have helped lift the power curve in the lower rpm ranges, and revealed some interesting insight on valvetrain control. The curves carried virtually parallel through 4,600 rpm, although the engine was making an average of 10 lb-ft more torque across the range with the more appropriate mixture. At 4,600 rpm, the single outer spring dropped off like a stone, while the dual spring with its increased loads carried on with the slope of the power curve as straight as a bullet. However, at 5,400 rpm, the gig was up, and the valvetrain became unstable with an immediate nosedive in power from there up. Peak numbers were now 551 lb-ft, at 3,500 rpm, and 560 hp at 5,500 rpm, right where the valvetrain went unstable. The low apparent peak torque is deceptive, since the engine's torque curve was extraordinarily flat. At 4,800 rpm, the engine was making 550 lb-ft, within one lb-ft of the torque reading at 3,500. In fact, looking at the average torque over the full test range from 3,100 to 5,500 rpm-the point at which valvetrain instability became apparent-the engine delivered an average of 544 lb-ft, which was very unusual.

We could see the engine in this configuration was going to come up short of our goal of 600 lb-ft of peak torque. However, the slope of the horsepower curve before the onset of instability showed the potential was there to meet the 600hp target, if only it would rev more cleanly upstairs. We had a set of Comp's high-tech Beehive springs (PN 26120), which have tested to provide unusually good valvetrain control in other applications. We made the spring change, installing the new Beehive spring at 1.880 inches, which provided spring loads of 155 pounds on the seat and 377 pounds over the nose. This is about as much spring as can be run with any reliability on a conventional flat tappet, and the Beehive winding significantly reduces inertia at the valve. We used a steel retainer (PN 964) and a modified lower locator to fit the Indy heads. Interestingly, the change to the Beehive spring did almost nothing to improve the rpm potential of the valvetrain. Peak torque improved slightly to 556 lb-ft at 3,500 rpm, with the same flat curve varying little and holding nearly constant to record 554 lb-ft at 5,100 rpm. Peak power nudged up to 559 hp at 5,400, at which point the valvetrain became unstable and power began to plummet.

It seemed as if there was little we could do to correct the valvetrain instability problem, but the lower part of the power curve at which the valvetrain was operating correctly seemed normal. We had run all of our tests to this point with the ported Indy 2D two-plane intake. We also had an Indy 440-2 single-plane intake and were curious about how it would compare. The 440-2 was port matched, but not otherwise modified. The single plane manifold added marginally to output above 4,500 rpm, recording a new high in peak power of 570 hp at 5,500 rpm, while peak torque moved to 561 at 5,100 rpm. The lower end of the torque curve dropped sharply for the 10hp gain at peak torque, dropping as much as 50 lb-ft off the bottom end. Outright top-end power, however, was clearly limited by the problems we were encountering at the higher rpm.