Hi DaveyF, M625 and M725 are high bias class A/B designs. The rather significant amount of heat they dissipate incongruent with this fact. All Rowland amps take an inordinate amount of time to stabilize, and an audition of an immature Rowland amp is often disappointing, which may explain your experience of a "digital hardness" on an amp that uses a linear power conversion circuit.
The Rowland knowledge base confirms M725 class A/B of operations at:
http://jeffrowlandgroup.com/kb/questions.php?questionid=592
The features article on M725 reports:
"Precision machined from two single blocks of non-resonant aircraft aluminum, the Model 725 features a unique composite circuit architecture incorporating
separate Class AB voltage and current gain blocks with no overall negative feedback. The Model 725 features mil-spec ceramic circuit boards, h ...
"
Robert Harley's highly favorable review of M725 in TAS 228 at:
http://www.theabsolutesound.com/art...-corus-linestage-and-725-monoblock-amplifier/
confirms that "All the audio circuitry, including the output stage,
are linear (non-switching) circuits. "
Roy Gregory , on HiFi Plus #83 at:
http://www.avguide.com/review/jeff-rowland-corus-preamp-625-power-amp-hi-fi-83
states of M625: "XLR sockets feed an input transformer,
separate voltage and current gain stages and a class AB, bipolar output stage in a dual differential, dual mono configuration. "
I have fully broken in (1200 hours at least), and have been immensely enjoying M725 as my principal amp for the last full year, because of the refined musicality that it delivers to me, regardless of its particular class of operation. And yes, I do concur wholeheartedly with Harley's glowing findings on this amp... Although, I only wish I could write like Roy Gregory does.
My M925 is showing already signs of greatness, all be it within the inevitable morasses and shortcomings of early break-in (approx 70 hours). It already projects stage and images significantly larger than M725... Whether or not it will be capable of greater musical refinement than the M725 monos, in a manner commensurate with the 2X price difference and 3X weight difference, is something that I hope to discover during the next 4 months… But Jeff Rowland has already warned me that -- because of the oversized input transformers, -- M925 break-in may extend well beyond what I am used to. Oh yes... M925 runs pretty toasty to the touch.