Would pairing the shafts on the outer cylinders, and adding caps under the bolts for additional clamping force, be enough to eliminate the center bolts?
I ask, because I would think the termination of the shaft immediately after the end bolts of each shaft provides a significant leverage point against the shaft and bolt. The entire load is taken up by the bolt, in a very focused point. There's enough area in the middle of the head to extend the shafts a bit and use a cap to add clamping force (even if 2 shafts were used, instead of 4). Same goes between the outer cylinders. If the shaft were one piece on the outer adjacent cylinders, a significant amount of force could be added with a center cap, transferring the load over a greater area. The only area where it would be limited to a small cap would be the outer ends. I'd mention a bridged cap in that area, but think you'd run into valve cover clearance issues then.
Great minds think alike
Actually, my original design did use caps to hold the shafts in place, and no bolt in the middle. You can see a picture of that design in the thread in the Vendor Classifieds, at the link below:
http://fepower.net/simplemachinesforum/index.php?topic=7336.0This was the design before any FEA analysis, and although it appeared to be pretty robust, it would have had the same problem with flexing of the shaft in the middle. I really want to avoid that, because of the length of the intake rocker. It's over two inches long, and the clearance between the shaft and the rocker is only a couple thousandths, and if you put a five thousandths bend into the shaft at max lift, you are going to have an interference issue.
I spent some time looking at the T&D race rocker setup for the FE, and that one uses bolts, no caps, and also a bolt between the rockers. I've gotta believe that T&D knows more than me about this, so I took some cues from their design. The other consideration is the additional material and machining time to make the caps; it is not a trivial issue, and would add significantly to the cost.
I don't know if using a single shaft for four of the rockers would help with the flex or not, but I don't like that idea because it means that removing the shaft would mean backing out the bolts while the shaft is under valve spring pressure. Kind of like removing the rocker shaft on a normal FE valvetrain. With individual shafts you can get the engine at TDC firing and take off the rocker pair without any spring pressure, and when you are running big springs I think that's important. Plus it's just easier to service the valvetrain if it needs to come apart. So I'm going to stick with a single shaft for each rocker pair.