There are at least five different patents that cover the PD 9100 and 9200 series converters. I've read them all and they really tell us a lot about how they work. The low PF is the main concern. The input needs to keep a high DC voltage on the input capacitors. Those caps store the energy that gets chopped and sent out through the output transformers under the control of the Unitrode 3846 chip. I've read the specs for that chip and it's really flexible - allowing pulse-by-pulse current control.
If the PD has a weakness it's going to be in the input stage, as previously commented. The low PF comes from the fact that it can't "refill" the stored energy in the input caps until the input AC voltage has risen sufficiently. Then it wants lots of current. Fortunately, I tend to run my gen at relatively high voltage levels.
I've got a second (non-RV use) Onan that uses an electronic voltage regulator (that I had to repair when I bought it by tracing the circuit). That VR circuit is used in many Onan models - I studied a half dozen when fixing mine. The design of that VR has a delay in the voltage regulation, much longer than a single cycle, that could cause problems with a low PF load like the PD converter. It makes sense to me that putting another resistive load on the Onan would provide improved performance.
I'm hoping my older Onan CCK, which sets the voltage purely by engine RPM will not have that same delay issue (there are other issues, I just don't know what dominates). Between that and the fact that I tend to run at a relatively higher AC voltage, which means it will start "refilling" the capacitor bank sooner and fill for longer, I'm hoping for better performance with the PD, but we'll see.