Clearly more investigation is needed
My initial tests we admittedly flawed in that I had insufficient Fe(III) in the sensitizer, and I'd blamed this for my low DMax. Sandy saw this as well, but didn't give a formulation for the tested sensitizer (presumably this was in the U of S link that is now dead).
On Arches 140lb HP I did get fine detail and clean highlights in both 20% NaCitrate developer and a Borax/Rochelle developer (more or less from the Stevens book). Both of these were souped with K2Cr2O7 (about %0.02 final concentration).
I did have a milky suspension rather than a clear solution for the sensitizer, the higher oxalate concentration in a solution with sufficient Fe(III) will probably make this worse.
Guess it's time to pull out the step wedge and do some more controlled testing...