I shot a fair amount of HCC and H&W Control in the late 60s up to the mid 70s and found, like just about everybody else, that it could be fabulous but was so touchy it wasn't worth the trouble.
I liked Panatomic X quite a bit and generally developed it in Edwal Super 20 which worked wonderfully with it. I have a couple bulk rolls from the 80s which were kept frozen from new, but regret to say that no matter what I process it in, it is substantially grainier than my old negatives. I have not had any issues with fog or reduced speed and the tonality is very close to that of the old negatives, but the grain is much more pronounced on even semi-close examination of a16X20 print from a 35mm neg. This differs from what others are reporting but is my experience.
Edwal Super 20 contained paraphenylenediamine which had a very strong solvent action. You are not going to get similar solvency from today's commercial developers. I am not surprised that your older negatives are finer grained. What has changed is not the film but the developer used.
You might want to try mixing D-25. In both the Altman and Henn (Kodak) study and Richard Henry's tests Panatomic-X produced higher acutance, lower granularity, and higher resolving power in D-25 1+0 than a fairly wide variety of other developers, from fine grain to acutance types. Granularity was the same as Microdol-X, but acutance was significantly higher. In addition, with Panatomic-X the speed penalty tradeoff for these image characteristics with D-25 was found to be relatively minor. There are unconfirmed reports Panatomic-X was optimized for D-25. Only Kodak knows for sure.
Results such as these might seem counterintuitive at first, until we remember most of what we think we know about developers is based on bad information. It is much more difficult to generalize regarding the working characteristics of developers than many think. It is more appropriate in many cases to look at a specific developer/film combination. Here we have a developer that should not produce good acutance (and it did not when paired with Tri-X for example), yet with Panatomic-X it produced entirely different results.