nm9, I'm not trying to put you on the spot here, so don't take it that way. Being a valvetrain guy, when I see/hear certain things, it raises a red flag and makes me ask questions. Again, not to put you on the spot, but to provoke thought that leads to logical conclusions.
First, how does the rocker arm design relate to a particular valve lift? The rocker is rotating on an axis, so shouldn't it be capable of any practical valve lift, at least within the capability of the spring installed and coil bind height?
Second, was the cylinder head designed for .575" lift, and what design parameters make that case?
Third, and last for now, when you measured the sweep on your example, did you measure it with an indicator of the roller centerline, or was it the width of the wipe pattern on the valve tip? The rocker is supposed to roll outward, and then roll backward the exact same amount when lifting the valve, and then repeat the process when closing the valve. Four cycles, out-in, out-in. In most cases, the valve is mostly sweeping out when opening, and back in when closing. That drastically increases the sweep and causes stability issues.
When I ran the numbers for the Edelbrock heads, with the PRW rockers, and .550" lift, the math shows .089" sweep, which is is within a thou or two of the numbers I have found in the past with that combination.
Just for pleasant arguments sake, .040" would be correct for .660" lift with that combo, and with .550" lift it should only be .028". With the .089" sweep, the valve lift would have to be .975", and you would have to do it with a standard length valve, which would be impossible. I've had to move the rocker shafts as much as .275" to get the geometry right with this very combination.
If it was just about sweep, I suppose one could live with the accelerated guide wear, but there is more to it than that. When the rocker is not doing the proper in-out motions, it screws with the velocities during the lift cycle. It should be, seat, out=accelerate, in=decelerate, full lift, out=accelerate, in=decelerate, back to seat. The decelerations, approaching max lift and the seat, minimize the risk of valve float and bounce and the resulting failures. The accelerations help get the valve open quicker and dwell longer at higher lifts where there is more curtain area for the heads to breathe.
One last thing, it isn't the rocker arm brand that causes this problem. It is the fact that it has a roller tip. Any brand roller rocker will need to have this effective geometry corrected because the roller as well as some other specs like net valve lift play a role. I'm actually a dealer for PRW, and they work very well in a lot of applications when properly setup. I used them for an example, because that is what the OP was using. Other brands, from the cheapest to the most expensive single shaft systems, have very similar results when just bolted onto the cast stands.
I'm just trying to get the word out that you can't use a lot of the information about mopar valvetrains that has been around for fifty years. It simply defies science, math, and the dyno.
