Z-Infinity Audio was founded in 2001 by Zsolt Mathe with the goal of offering perfectionist tube gear that is designed, tested, and hand-assembled in the U.S. Zsolt, who is of Hungarian descent, has designed audio and other electronic circuitry throughout his life. The connection with his European heritage is reflected in his focus on execution and quality. Distribution is currently limited to direct-to-the-consumer on-line sales, though there are plans to expand nationally via showrooms at select locations. The Z120 amplifier is available either as a 60Wpc stereo amp or as a 120W bridged monoblock. Optional configurations include minimal or no global feedback and either 4- or 8-ohm impedance taps.
The circuit is quite traditional in appearance, as befitting a design vision that can be summed up as the highest quality sound with the simplest circuits. According to Zsolt, “many other amplifiers have complex filtering, feedback and correction circuits; we don’t. Even when using negative feedback, it is done so sparingly. With that said, ultimately there is nothing groundbreaking about our actual designs, following the simplicity principle of not complicating the amplifiers unnecessarily.” The input voltage gain stage is a grounded cathode 12AX7, directcoupled to a 12AU7 long-tailed-pair phase-splitter. A pair of KT88 beam power tubes are operated push-pull, connected in ultra-linear (UL) fashion to the output transformer, which are sourced from Hammond Manufacturing. KT120s may be substituted for slightly more headroom but at the cost of slightly increased distortion.
UL remains a popular output configuration even 60 years after it was popularized by Hafler and Keroes. Numerous amplifier classics from Dynaco, Heathkit, EICO, and Marantz serve as a testimonial to its commercial and sonic success. Connecting the KT88 screen grids to taps on the output transformer obviates the need for screen bias supplies and may be thought of as the application of negative feedback to the screens. None other than Herbert Keroes noted in 1958 that in this paradigm, “the feedback is of the power type rather than the more usual voltage or current feedback…It can be demonstrated mathematically that when power feedback is applied to the screen grid of a tube, the linearity of the plate characteristic curves can be improved over and above the amount normally to be expected by a consideration of voltage feedback only.” Another important benefit of UL is a source output-impedance that is much lower than that of a pentode connection, being closer to triode in value. That makes for a much higher damping factor relative to pentode, even with only moderate global feedback levels.
The 4-ohm Z120 uses 19.3dB of global negative feedback, a reasonable level that is pretty much in line with what was used in many successful vintage tube amplifiers, and well below extreme levels in excess of 30dB. While triode amplifiers may be able to get by without any global feedback, feedback is essential for pentode and beam-power-tube output stages to maintain a decent damping factor, reduce distortion and noise levels, and increase bandwidth. Anecdotal evidence suggests that between 10 to 20dB of feedback may be ideal in most circumstances. Feedback levels much above 20dB, aside from stability issues, appear to constrain the soundstage as well as compress microdynamics. I asked Zsolt about the possibility of incorporating a feedback switch to provide a couple of options. His response was as follows: “The question has come up a few times internally. The reason why there aren’t switches to change feedback options, triode/pentode/UL operation, and the like is because it takes more than switching a few wires around to tune the amplifier for a particular mode. For example, stability, frequency response, and impedances change drastically when changing negative feedback, requiring several components also to be changed. We wanted to avoid making a compromise amplifier, and instead, we are making a tuned amp for the particular configuration. This is unfortunate for all the ‘tweakers’ out there, but we feel that quality over tweaking ability is the winner.”
The output stage is fixed bias, adjustable via individual bias pots for each KT88. However, there are neither test points nor a built-in bias meter. In order to monitor or adjust the bias voltage you will need to purchase a suitable bias meter. I use and recommend the Compu-Bias meter. It monitors both the cathode to plate voltage and cathode current for up to two power tubes and calculates and displays the total tube dissipation dynamically in real time. To use it, power down the amplifier, remove a pair of power tubes and insert the probes into the tube sockets. Plug the tubes into the probe sockets and power up the amplifier. Wait a couple of minutes and then adjust the corresponding bias pots as necessary to bring the bias to about 60mA. You then repeat the process for the second pair of power tubes. The location of the pots is not particularly convenient and requires the reach of a long screwdriver. Out of the box, I measured a bias current (plate plus screen) in the range of 60 to 63mA for both monoblocks and a cathode-to-plate voltage of 393V RMS for all of the KT88 tubes. The average idle plate dissipation works out to a very reasonable 24 watts. The power supply features a toroidal power transformer, fast-recovery rectifier diodes, and plenty of filter capacitance. The preamp tube-heater supplies are DC and are regulated using zener diodes. There’s also a relay-activated soft-start circuit that ensures tube filaments heat up before high voltage is applied—always a good idea that should significantly increase tube life.