When driving Mojo2 to my PMC MB2S studio loudspeakers, I tested a variety of amplifiers. Overall, I'm incredibly happy with my class AB Accustic Arts Mono-II's - both direct from Mojo2 or through my companion Preamp-II Mk2. However, I wanted to broaden my recommendations for more affordability and customer choice, so I tested class A (Allnic Audio), class D (AGD Productions) and various solid state class AB. All were 100W or more - sufficient to be able to drive my PMCs.
Mojo2 direct to headphones is the GOAT (Greatest of All Transparency) and will always be my reference for product development ...but many listeners (myself included) prefer a subject with some clothes - no matter how gloriously transparent a naked Mojo2 may be. So, in my tests, I started with a 100% absolutely clean Mojo2 output and then listened for all the ways that amplification and loudspeaker interaction affected the signal. Line level output from Mojo2 sounds the best, IMO, so I included my preamp in all tests. And I used battery based power (via Analytic Systems inverter).
Accustic Arts Mono-II
This monoblock amplifier uses a current mirror design and high quality components to provide a slew rate (rise/fall time) performance is 3.0 μs—sufficient to fully resolve a 20kHz square wave and hence essentially all PGGB definition ...so I don't think I am wanting for more resolution or refinement. One very nice feature is the fully balanced design of the preamp/amp - converting the Mojo2's unbalanced input to balanced for maximum performance.
AGD Productions AUDION
The issue with Class-D is their switching characteristic with the potential for lots of RF noise. The AGD design uses Gallium Nitride (GaN) semi-conductors for 500x more speed/resolution than class AB, ultra-low distortion, linearity to 50kHz and a Pulse Width Modulation (PWM) switching frequency of 800Mhz. So I became curious: would I hear more details? and what about RF noise audibility?
Unfortunately, even though I had a RF-STOP EXTREME SE Signal filter and DAC-WRAP isolating the Mojo2, the AGD took away some transparency - despite providing oodles of resolution. This could be i) RF noise drowning Mojo2, or ii) the out of band noise modulating the tweeter, or iii) the inherent nature of Class-D affecting small signals. This needs more investigation. I will continue to explore GaN designs since they are becoming very affordable. Stay tuned.
Allnic Audio A-2000 KT150
The issue with tube-based Class-A is their power consumption and higher distortion (relative to solid state). My battery based power could only drive the Allnic for 3.5hours versus my Mono-II's for almost 9 hours. Transparency greatly suffered from the tube design yet I could better discern low and mid-low range. So, the appeal of the Allnic is not ultimate resolution or fidelity but access to electrical current to grip loudspeaker drivers.
Solid State Class AB
My testing was the Benchmark AHB2 and Bryston 3B. The AHB2 is the more advanced design: linear class AB using class H power with patented THX signal correction in class A. Transparency was excellent - the only downside being the 100W power did not resolve fully the signal at the PMC transducers. My Mono-II's had more grip with equal transparency. The older Bryston model was an also-ran: lacking transparency and resolution from top to bottom.
Summary
My Mojo2 reference system is very useful for testing a system. There is so much information on hand that it may take huge reserves of power & current to resolve it all. I don't want to spend Dan D'Agostino dollars so I may be tempted to upgrade to the Accustic Arts Mono IV's. Alas these heavy amps may be antiquated in a few years. It's likely that the future of loudspeaker listening is class-D amplification combined with improvements in transducer resolution.