[A Quiet Boom]

Nat,

while a good idea, I don't think it will work with stock cams. I don't know the hardness of stock Porsche cams but I can tell you that V8 roller cams are hardened as opposed to non-hardened when running flat tappets. The shell of a bearing is hardened, if the cam is not of equal hardness the bearing will tear it up.

Also as stated by someone else the primary reason for going to roller valvetrain is to handle radical lift. When I was drag racing I was running a custom Lunati solid roller grind with over 0.750" lift on the intake side. A roller follower or tappet allows the engineer to increase the area under the curve by setting up very steep open and close ramps which effectively puts the valve near peak flow for a longer duration than is possible with a flat tappet or follower. When comparing a typical roller vs. solid lifter racing cam the roller has an almost oval shape to it while the solid lifter cam has more of a teardrop which means with flat tappets the valves spend more time at low lift, hence lessor airflow. In order to get full benefit a custom hardened cam with "fat" lobes would be necessary. I doubt there would be enough HP gain from frictional losses to justify a roller rocker even if it would work with the factory cam. The real gains for roller valvetrain from a frictional loss standpoint occurs with a pushrod engine where rollers are used on the cam lobe as well as the valve stem. Rollers at the valve stem also allow for greater lift without causing heavy side loading.

Having said that if you could run a wide bearing (take a look at a V8 roller lifter) with a custom cam hardened to the same hardness as the bearing and with a ramp profile designed to take advantage of the roller some real gains could likely be had although they may slow air velocity enough to put them into the realm of "race only". It is also true that a roller setup is going to be heavier but this is easily overcome with stiffer valve springs that would be necessary anyway to compensate for the greater valve acceleration as well as potential for greater lift. My race V8 ran triple valvesprings with ridiculous open pressures (800lbs comes to mind but it's been years). This was necessary to prevent valve bounce as well as big lift. To do it properly you would need to look at the system as a whole. Valve lift and timing should be setup to keep the valves near peak flow during each open/close cycle, vehicle weight, intended use, and other engine mods would all come into play. It's really a science.

I'd build one just for the heck of it and using a degree wheel measure lift every 5 degrees of cam rotation just for the heck of it compared to the stock follower, map it out on a graph in excel. Then if you had a scrap cam weld it up and regrind and see if you can't improve upon that. Someone on the board is bound to have flowbench numbers for a given head, see if you could build a profile that maximizes those numbers then see if a camgrinder couldn't grind a set of cams from blanks you produce on the CNC. It could be a really fun project.