Perhaps this will assure others who are as curious as I was. After I loaded the printer with Piezoflush and did an initial fill, I assumed that there should be no ink left in the system. To assure myself of this, I did first a nozzle check, which did not turn out the way I expected, so then I did actual printing. In theory this should yield a Piezoflush-pink output with basically no detail. However, this is what I get:
Obviously there is actual colored ink in the system. However, having been reassured that this is OK, I proceeded to convert to Piezography, meaning removing the PF carts and loading new B&W inks via another initial fill operation. Indeed, I don’t see any evidence of color ink once this is done, although given the above I rather suspect that a spectrometer would find some nonetheless. Still, the resulting prints look excellent, so I’m not bothering with whatever contamination might or might not be there.