Goldmund Mimesis 8 power amplifier
To judge from the $6400 Mimesis 8, Goldmund walks its own way when it comes to power amplifier design. High-end solid-state amplifiers from US companies like Krell, Mark Levinson, Threshold, and the Jeff Rowland Design Group marry massive power supplies to large numbers of output devices (these often heavily biased to run in class-A), built on chassis of such nonmagnetic materials as aluminum. By contrast, the Mimesis 8 has a magnetic (steel) chassis, and uses a relatively modest power supply, that for each channel based on two main 4700µF reservoir capacitors. The 8 offers just two pairs per channel of complementary output MOSFETs (Hitachi K134/J49). These carry a modest bias current of around 80mA total.
Goldmund's Michel Reverchon has explained in the past (footnote 2) that he feels a small, "fast" power supply to be optimal in that power is taken from the AC line, "not the power supply," while a publicity sheet for the Mimesis 8 mentions that "low-value filter capacitors are used to avoid the detrimental slowing effect of most power supplies." We shall see from the auditioning whether this unorthodox approach better serves the music, whether the amplifier has that essential "swing."
Looking inside the Mimesis 8, AC power enters at the left rear of the chassis, with a separate ground post tied to the ground pin on the IEC AC receptacle (the main chassis ground is also adjacent), and is taken via a fuse to a small vertical printed circuit board attached to the two power switches. To turn the beast on or off, both switches must be pressed simultaneously, which activates a relay on this pcb and illuminates a green front-panel LED. AC power feeds a shielded module behind the front panel which contains two toroidal transformers, one for each channel. This module feeds rectified DC voltages to the single main circuit board, this double-sided and carrying the circuitry for both channels, physically well separated. There appear to be two separate sets of voltage rails per channel: ±65.2V for the output stage, smoothed by the pairs of 4700µF capacitors mentioned earlier; and ±76V, which then appears to be regulated to ±72V, for the input and driver stages, this filtered by six 470µF capacitors.
Looking at the signal path, each channel is carried on a loosely twisted pair from its RCA/XLR socket to the main board, where what looks like some kind of hybrid circuit provides the voltage amplification. There are said to be no capacitors in the signal path, and no electrolytic caps anywhere near the signal path. (A pedant, of course, would point out that the power-supply reservoir caps appear electrically in parallel with the signal path.) The two pairs of output devices per channel are mounted on an aluminum bar that runs the entire width of the pcb and is in thermal contact with the exterior heatsink fins. (This bar is drilled to take a third pair of devices for each channel.)
Again in contrast to American "muscle amp" philosophy, where solid bus bars are used to define the ground point between the positive and negative filter caps and to take both DC to the output stage and the output signal to the sockets, the Mimesis 8 relies on what appears to be 16-gauge wire for the latter, with the ground point defined by a wide pcb track. The output sockets themselves are good-quality 5-way binding posts. Again contrary to "audiophile" practice, a small air-cored inductor appears to be in series with the loudspeaker output.
The input sockets comprise both RCA and XLR types, but this shouldn't be taken as meaning that the Mimesis 8 has balanced inputs. Pin 2 of the XLR is wired in parallel with the RCA jack, while pin 3, which would normally handle the antiphase input signal, is connected to ground. If a balanced source is connected to the Goldmund, the amplifier will only process the "hot" signal, tying the "cold" to ground, meaning that the benefits of balanced workingsuch as the rejection of common-mode noisewill not apply.
In essence, the Mimesis 8 conforms absolutely to Commander Harrison's mid-'50s dictum (footnote 3) that an amplifier be "simple and straightforward in design in the interest of minimizing performance degradation." Does it offer anything in the way of sophistication?
The Mimesis is extensively protected against fault conditions. The two voltage rails for each channel have a series fusewhen one opens, its red "Fuse" LED lights on the front panelwhile protection circuitry mutes the output if either a DC offset or ultrasonic oscillation appears on the input, also turning the green LED red. A thermal sensor on the heatsink also mutes the output if the heatsink temperature exceeds 90°C (194°F).
The Mimesis 8 incorporates Goldmund's thinking regarding "Mechanical Grounding." The 15kg power-transformer block and the output stage and heatsink are mounted on two and one carpet-piercing cones, respectively, which are said to provide a "vibration evacuation path" for spurious mechanical signals. The chassis itself is mechanically decoupled from these feet by Teflon insulators. Much attention has also been paid to the amplifier's electrical grounding arrangements. The benefit of these techniques is the reduction of noise to astonishingly low levels (as was revealed by my measurementssee later).
All things considered, the enigmatic Mimesis 8's design is a mixture of the sophisticated and the mundane.
The Mimesis 8's cone feet penetrated the rug to the saltillo tile-on-concrete floor beneath so that the benefits of Goldmund's mechanical grounding would not be obscured. Goldmund points out in the user's manual that the sound of the amplifier is very dependent on the polarity of the AC plug. I experimented with this, using a two-pin cheater with the large pin filed down, and as I seemed to get the more musical sound with the power amplifier grounded and the preamp floating, this ground was achieved by running a separate ground wire from the Goldmund's grounding post to the wall socket.
It proved impossible to use the Mimesis 8 with The Mod Squad Line Drive, due to RF interference, so all auditioning was done with the YBA 2 preamplifier. The sound of the Mimesis 8 seemed extraordinarily subject to setup vagaries. Often I would embark on a listening session only to find that the sound was more threadbare than I had previously experienced. It turned out that I had made a minor change somewhere in the system and knocked the Goldmund away from its optimum position. In general, whenever the Mimesis sounded too lean or threadbare, then, as sure as amps is amps, something would not be right. The following descriptions of the amplifier's sound only apply to those times when everything seemed to be optimum. They also apply when the amplifier was fully warmed up; it seems to take about an hour to reach a sonic plateau after first turn-on. (Goldmund says that the critical circuitry needs to reach about 55°C for the best sound.)
In the paraphrased words of J. Gordon Holt, "get the midrange right and all else is gaslight." (I think that's what he said.) Well, the Mimesis 8 certainly gets the midrange right. The CD rerelease of Jacqueline du Pré and Daniel Barenboim performing the Brahms cello sonatas (EMI CDM 7 63298 2) virtually sang with the amplifier driving Wilson WATT/Puppies, the cello's dark woodiness being reproduced with vivid presence yet without being thrust forward into the listener's lap. The feeling of two instruments hanging in the air, the cello slightly in front of the piano, was uncanny. The Wilson speakers nearly always manage to disappear sonically, but with the Mimesis 8, there wasn't even a Cheshire smile left to show where they'd been.
Footnote 1: Goldmund, which also owns the Stellavox tape recorder company, is perhaps unique in the high-end field in that it has no design staff of its own. Instead, it subcontracts the creation of its components to highly talented outside consultants.
Footnote 2: Talking to Harry Pearson in The Absolute Sound No.65, May/June 1990, p.98.
Footnote 3: "A high-quality amplifier must be capable of passing rigid laboratory measurements, meet all listening requirements, and be simple and straightforward in design in the interest of minimizing performance degradation..."Cdr. Charles W. Harrison Jr. Audio, January 1956, From "High Quality Dual-Channel Amplifier," reprinted in Audio Anthology: When Audio Was Young, Vol.4, published by Audio Amateur Publications, Inc.