I doubt his RME was faulty. I'd bet it was being overloaded and/or the signal path was being pushed too hard.
It wasn't that long ago that Tony was seeking recommendations for comp/lims to control peaks going into his converters. Tony and headroom have never been comfortable bed fellows!
The one thing guaranteed to highlight differences between preamps is to overload them. With some, overloads are sudden, obvious and sound terrible. Others are more progressive and can sound fairly benign with musically pleasing saturation/distortion. Often people use or even rely on the overload behaviour as part of their desired sound. I think, based on previous conversations, that Tony often does....
I wouldn't describe RME's preamp overload characteristic as benign or pleasing, but some Focusrite preamps are certainly more forgiving...
Another likely issue in Tony's situation is the source feeding the interface — Tony was using a mixer. Knowing that Tony typically recorded with very little headroom at the time, it is quite possible that the RME interface was configured to need quite hot input signal levels (up to +19dBu, for the 'hi' mode), and consequently the mixer was being pushed pretty hard and sounded strained as a result — a common digital interfacing problem. Again, some focusrite interfaces expect lower line input signal levels (as do Behringer interfaces) thus reducing strain on the analogue mixer considerably and appearing to have a 'better' sound when, in reality, the better sound was due to a less harassed mixer. This would be particularly evident with a Mackie mixer....
In a nutshell, what I'm saying is that there are many potential reasons for a system acquiring 'a sound' . Some technical, some operational. And without a full understanding and isolation of individual elements its unfair and damaging to ascribe characteristics to one device absolutely.