What one simple question can accomplish! This answer from Poland immediately put the power amp in an environmentally friendlier perspective. We would suggest that the published specs of the amp be revised to include sustained and not just peak power consumption. Another question we put to Marek was about the BQVA bi-quartet voltage amplifier mentioned in their spec sheet.

Marek replied as follows but was quite reluctant with his answer: "The voltage gain of the BQVA circuit depends on the value of its characteristic circuit parameter. This is the number of external transistors which provide voltage gain. These are controlled from the BQVA outputs and with the value of two resistors: R1 and R2. One of them is the load resistor of the voltage amplifier. Thanks to this ultra-linear voltage gain stage, the power amplifier is characterized by a very flat frequency response. The incoming signal is transferred from the input with virtually no loss at the R1 resistors.

"The output stage is fed from the voltage-amplified signal across resistor R2. The characteristic circuit's parameter can be expressed as a ratio of I2 to I1 which in our case is equal to 5. There is no negative feedback loop between the negative inputs of the voltage gain and current buffer stages."

About their output stage, Marek added: "In the output stage, the signal is generated with two sub circuits. One is an emitter follower working in push/pull, the other a bank of 10 controlled current sources again in push/pull. The emitter follower controls the amplifier's output voltage whilst the controlled current sources generate the output current. This solution is well known but the difference of Dr. Studnicki’s version is that in all documented solutions, only the emitter follower works in class A. This causes the ratio of the current sources' current to emitter follower's current—the current multiplication factor—to be quite small, namely below 10. In the AbysSound amplifier, both sub circuits work in class A."

"In our ASX-2000, the bias of the output stage is fixed at 2.5A and the current of the emitter follower is set to 50mA. The current multiplication factor is thus equal to 50, resulting in the amplifier’s very low internal resistance. Due to the fact that the emitter follower is providing only 1/50th the current to the load, its output resistance (RE=3.3Ω) is thus seen as 50 times smaller i.e. 0.066Ω. The damping factor is thus easily calculated as 8Ω/0.066Ω which is around 120."

Once on its stable trolley, the amplifier was rolled into its initial position between our giant Grecian Pnoe horns. With the Kemp polarity checker, we made sure that its Crystal Cable power cable was oriented correctly. Some time ago we had an electrical accident which caused DC running into our beloved AER drivers of the Pnoe horns. The results were two fried voice coils that had to be repaired. Since then, we are ever more careful about connecting equipment, especially unknown new equipment. One fate of doing reviews is getting all kinds of machinery into the house which must be connected to our own. With today’s tendency to use computers as digital audio sources and to avoid preamplifiers, the route between a PC and loudspeaker is often devoid of galvanic separation. In a worst-case scenario, a substantial amount of DC can happily sneak into your precious voice coils.

One possible remedy or protection if you like is to introduce a passive preamplifier with output transformers. That preamplifier should otherwise be as neutral as possible but also block DC. For this requirement we had earlier contacted Jonathan Billington of Music First Audio in the UK. We explained our wish to him and he came back with a proposal to wind special transformers that would efficiently block DC whilst preserving the highest musical quality. To our ears, this passive pre is free of any colouration or other negative influences whilst it gives us (and our bank account) the needed assurance that another DC incident won’t blow up our speakers again.

With DC safely out of the equation, the only other thing able to fry voice coils is driving an insufficiently powerful amp into clipping. Overload caused by playing too loud is one’s own fault and thus personal blame. Overload from a spike caused by violent turn-on transients is an amp-design flaw. We started our listening sessions with music streamed from an iMac running Qobuz Desktop into a La Rosita Beta. From there the signal went into the Music First passive whence it traveled to the AbysSound ASX-2000 [above]. Cables on duty were all Nanotec Systems except for the power amp’s Crystal Cable AC cord.

Powering up the ASX-2000 resulted in a confirmation pair of relay clicks. After a few minutes we neared the loudspeakers in search of hum, noise or whatever other pip squeaks wouldn’t be welcome to our party. With 106dB sensitive speakers, any tiny noise is audible. Our horns really form a merciless looking glass at any such artifacts. Yet we could hear nothing that would indicate that the amp was even working. The lukewarm temperature of the big black beast gave away that something was happening so the only way to prove proper operation was to turn up the preamp’s volume. There was music now from the first track in Qobuz’s play list. Point taken. The ASX-2000 is as quiet as a (dead) church mouse whilst idling. How would it perform with music?