but there is a decent bit of stiffness/binding when on the opposite side. It is actually stiff enough that the head will stay in the 90-degree position (counterweight bar parallel to the ground) with the clutch disengaged - but only on the one side.
if you have binding or stiffness, how accurate is the PE graph going to be? Solve the binding/stiffness issue, then look at the PE. I have an Atlas that I replaced "All" bearings in it because the previous owner had completely packed hard the bearings with "white lithium". I went through all the adjustments and still wasn't very pleased. I sent it off to Ed Thomas... Needless to say but I will; there is now NO binding or stiffness anywhere in that mount.
The Atlas is still a mass produces mount with the flaws of low QC when compared to the high end, hand built mounts. Assembly is per procedure, not as a single entity! My guess on the binding is that either something is adjusted too tight or that the internal spacers are incorrect.
If I could afford a higher end mount, I would have gone that route. I could even spend the money on the upgraded aftertmarket gear that is supposed to reduce PE in these mounts and I'd still be under by a lot of what it would cost to go to the high end mounts. A lot of people that get the Atlas basically look inside the box, close it back up and send it to off to get hypertuned. Atlas mounts are sensitive to ballance, I ballance mine so that if I crack the RA clutch loose, it gently falls to the home position.