Well, since I got more than two requests, I'll make this a post instead: I was trying to help MCrider out with Phaedrus II extra performance stuff, but don't want to divert his build thread, so I'll put this here.
This is really an excerpt from my upcoming CB750 book, from the "budget performance" section called "Performance I"...
The Cam Towers (or "Cam Bearings" "Cam Holders", or my favorite, "Rocker Towers") went through several production changes over the years. Each change brought less HP to the rear wheel and more wear into the rockers and their shafts. The shafts underwent several changes, too, none of which were a performance (or longevity) enhancement for these engines. It went like this:
In the K0 and K1 engines (and early K2 also), the rocker shafts were free to spin. The long 6mm bolts that hold them in location, in their wasp-waisted middles, were supposed to be lightly torqued so as to allow this spin while just retaining the shaft in position. A Honda engineer, probably a respected one in their community, designed in 8 additional spots on the castings where holes (and countersinks) could be applied to lube these shafts with a little extra splash oil. On Yoshimura racing versions of these parts (circa 1970-72), these holes were drilled and countersinked about .020" to .030" (hole size about 3mm, or 1/8") to capture this oil. The 6mm bolts only had about 4 or 5 threads in the heads, because they were only to be torqued to 6 ft-lbs or so, and the shafts had straight slots in one end (the outboard end) so the tech could twist the shaft with a screwdriver while tightening the bolt, not too tight. This slot was also used for the lathe tooling where the shafts were ground to their 12.00mm diameter in production.
In 1971 Honda decided the bike needed to be quieter, which led to a number of changes, the most noticeable being the switch from fiberglass-packed mufflers to the more restrictive, baffled HM341 pipes. As one of [many] quieting changes inside the engine, the rocker shafts each received a pair of 5mm tapped through-holes and the Rocker Towers each received 4 new bosses and 6mm holes where a pair of 5mm bolts could lock the shafts in place. This removed a small rattling sound that could be heard behind fairings: the 750 was the world's premier tourer at this point in time, and the hiways were crawling with them, most sporting fairings. The rattling noise was the rocker shafts, twisting back and forth as the rockers operated the valves: the tolerances in the Rocker Towers where the shafts rotated were not real tightly controlled at Honda. Some were as loose as .003" from the factory: this locked-down shaft then made the engine run quieter.
But...the rockers are splash-oiled, and immediately the shafts were starved of oil where the rockers press up from below, and we saw severe wear in as little as 8000 miles. So, starting in the second month or so of K3 production, grooves were added to the rocker shafts (at Honda) to try to move some oil from the rocker oil holes to the bottom of the shaft. Problem was, on most rockers the holes were randomly drilled and they did not line up with the grooves on the shafts: the wear problem was only improved on those that did line up. Except, the rockers now developed raised ridges in those shaft grooves, which would then seize and cause a different problem. The short life remained, and the friction got worse.
Honda was obviously worried enough about the shafts breaking (and the image that would produce) to keep bolting the shafts down. Several aftermarket cam builders persisted in making high-lift cams (Action Fours comes to mind) that could bind the valve springs and break these shafts if the builder of the engine did not pay close attention to this issue, so that probably added to Honda's worries. The Japanese mindset of honoring their elders kept the oil-hole markings on the Rocker Tower castings (in my opinion), but with non-rotating shafts, they were of little use unless the engines were changed (by someone like me
).
In our shop, we just removed the 5mm bolts when replacing the shafts (and worn rockers), and let the shafts spin. My own K2 (which has a K1 engine, an early K2) now has 130,000+ miles on those shafts, with less than .002" clearance in the rockers and towers, with spinning shafts. On racing engines, the bosses on at least the outboard ends of the shafts were drilled with 3/32" or 1/8" bits and countersinking applied to funnel some extra oil in, and the rocker holes were ovalled so as to actually REACH the grooves in the rocker shafts, so the oil could make its way down to the rocker-shaft bearing area. This resulted in very quiet engines, but in Honda production this much work would have been expensive. This also resulted in noticeably less friction, more like the K0-K1 engines, which makes more RWHP...
In the K3 and later engines, the head received more threads for the long 6mm locating bolt (about 7 threads instead of the original 4 or 5) so the bolt could be torqued like all other 6mm bolts (up to 9 ft-lbs). This ensured non-rotation of the shafts, even if the 5mm bolts were removed: so the screwdriver slot on the ends of the shafts became very important for guys like me. We Loctited the 6mm long bolts at whatever minimum torque was required to allow shaft rotation while not having the bolts back out while running: typically about 5-6 ft-lbs. the Rocker Towers are held down by 4 other 6mm bolts at the cam bearings, so these really don't matter for holddown, anyway.
In the 1990s (or so), Honda changed their replacement parts: the shafts became notched on one side instead of wasp-waisted (i.e., cheaper to make). The shafts for the K0-K1 engines did not receive the 5mm threaded holes, but the K3-later versions did. These shafts will never rotate, with or without the 5mm bolts. In the last version of the K0-K1 shafts (circa 2002), the notch appears to be extended about 200 degrees around the shaft, to allow some rotation, at least. The K3-later versions have the oil grooves on the shafts, but again, if the rockers are not modified, these don't help much.
So...the rockers need to be modified to get the oil down into the grooves (if yours have them: optionally you can add some, about .010" deep, for the least friction), the 5mm bolts need to end up anywhere but in the engine, the shafts need a V-groove all the way around their middles, and then the Rocker Towers need 8 extra oil holes added to capture the splashing oil nearby to ensure free rotation of the shafts. The modified rocker Towers look like the picture, below. Compare that image to a standard rocker tower and you will see all the holes, and also the places where you need to add them, to remove the friction.
(I am making these shafts, starting this winter, as Honda has stopped. They will be available around November, 2009.)
In the early 1970s (it may have been 1970), there were a couple of "mystery bikes" that appeared in the Chicago area: these CB750s were real street stalkers, stock from Honda. They were, to all appearances, stock bikes, but had considerably more RPM and high-end HP than most 750s. One of them bent an intake valve one day and ended up at Mannheim Honda in Mannheim, IL, for repair. The valve was not standard 750: someone called me for ideas, and we found that a 1968 SuperHawk (305cc, big intake valve head) valve would fit.
The bike's bore was smaller than 61mm, and the stroke was longer, at 65+mm (IIRC), calcing out to 700cc or so. The cam was visually different, too. [American] Honda had no idea what this bike was about, yet Mannheim Honda had sold it as any other 750. The only outward clue of any difference was the little pad on the front of the cylinders: it did NOT show "736cc", but was blank. One thing about this bike was the "swiss cheese" oil holes in the rocker towers: they looked exactly like the picture below, with the 8 extra oil holes drilled and chamfered. That's where this whole thing started...
I think that an old Honda engineer in their engine department finally had his say on those 2 bikes....