Dynaco ST-70 Series 3 Tube Power Amplifier

A Classic Resurrected

Equipment report
Tubed power amplifiers
Dynaco ST-70 Series 3
Dynaco ST-70 Series 3 Tube Power Amplifier

Six months were spent evaluating prototypes of all three designs using the same output transformers and power supplies. During the listening phase of the testing process, the pentode-triode design ranked lowest. And although both triode designs were comparable in midrange and treble performance, the Quad Triode nudged ahead by virtue of its punchier low end. It’s worth noting that the lower gain 12AU7 was selected instead of the common 12AX7 as the driver tube to reduce dynamic distortion levels. It also offers a lower source impedance and thus less interaction with the capacitance of the EL34’s control grid. Many years ago, David Manley opined that he could tell the IQ of a tube-amplifier designer by his choice of driver tube; a 12AX7 would score poorly in this context. Another advantage of the Quad Triode circuit is that a second feedback loop is no longer required as the cathodyne phase-splitter is buffered from the power tubes by the driver stage.

Many consider the ST-70’s A470 output transformer to be a primary contributing factor to its sonic excellence. And this weighed heavily on the design team, which purchased various ST-70s, ranging from the original to updated versions, and began by taking these apart and then scouring the planet for possible solutions. They tested various “off the shelf” output transformers, but none could be found that matched the high-frequency response of the original. The A470’s design as detailed in Hafler’s U.S. patent shows interleaved primary and secondary sections, but what is most unusual is that some sections were wound with reverse sense, or backwards if you will, and then connected in parallel. I don’t envy the transformer winder who had to keep track of all this. Working for many months with their transformer vendor, a final design emerged that nearly matched the rise time of the original without waveform distortion or overshoot but featured much better bass response. The final output transformer is a point of pride for the design team, which feels that it may well be the finest output transformer anywhere near this price point.

The original power and bias supplies have always been weak links, and these were completely overhauled. A bigger dual-primary power transformer was designed that by virtue of a larger iron core features better AC line regulation. Solid-state rectification is now used instead of the 5AR4 tube rectifier. The filter capacitance has been greatly increased, basically using up available space while observing required safety clearances. The EL34 bias supply now uses a modern solid-state rectifier with additional supply capacitance, and its reference voltage is derived from an IC voltage regulator for improved bias-setting stability. Individual bias adjustments are now provided for each EL34 power pentode. This is done by tweaking a set screw so that that each pair of indicator LEDs is at equal brightness. A small Phillips-head screwdriver is helpful in reaching the recessed trimmers. Of course, any bias adjustment should be performed after a warm-up period—I would suggest 20 minutes—and with no signal playing.

Most of the hand-wiring has been eliminated by using a military-grade, double-sided, epoxy-fiberglass printed circuit board for the internal components and output tube connections. The PC board has been optimized using circuit development software to ensure that ground return currents from the power supply remain separate from the filtered power for the driver and output stages. The goal was to reduce channel crosstalk and lower overall noise. Output impedance taps of 4 and 8 ohms are selectable on the back panel via a switch. There is no longer a 16-ohm tap. In the original ST-70, feedback was always taken from the 16-ohm tap regardless of which tap was actually loaded. In the new ST-70, the feedback loop is tied to the impedance tap in use by the double pole impedance selector switch. One side of this switch connects either the 4- or 8-ohm tap to the output terminals. The other side connects the feedback loop to the tap selected and changes the feedback divider ratio so the overall gain remains the same. The payoff is reduced high-frequency distortion.

The original ST-70’s input stage incorporated filters designed to limit and roll off the amplifier’s frequency response. The subsequent loss of open-loop gain at the frequency extremes resulted in reduced feedback and thus additional distortion in the bass and treble. The new ST-70 implements a passive high-pass network, located after the input volume trimmers, to contour the bass response. The roll-off is a gentle 6dB per octave, and there are two selectable –3dB points at about 35Hz and 70Hz. The low-cut filter takes the form of a three-position switch on the front panel with the top position being a bypass. It’s perfectly safe to switch positions during play. I found this feature to be extremely useful in controlling the Quad 57 ESL’s excessive midbass. And I imagine that it could possibly be useful in taming room modes, which are typically below 200Hz in most listening spaces.

Prior to shipment, each amplifier is burned-in, power tubes are matched and individually biased, and a critical listening test is conducted for audio fidelity. A product report accompanies each amplifier that includes Audio Precision plots of frequency response from 10Hz to 20kHz and total harmonic distortion as a function of output power. The passing grade for THD plus noise is less than 0.03% at 1W. Dynaco anticipates a life span of 10,000 hours for the 12AU7 tubes and about 3000 hours for the EL34 power tubes under normal conditions. Since the new ST-70 sounded terrific right out of the box, my advice is to forget about tube substitutions and simply enjoy the music.

The ST-70 was my first tube amp. It replaced a Harman Kardon Citation 12 solid-state amp that had pretty good credentials and excellent benchtop specs. The resultant sonic transformation gave rise to the first “Wow” moment of my audio career. The second such moment, by the way, was hearing voice reproduced through the Quad 57 ESL. These events opened my eyes and ears to a conceptual world well beyond that captured by mundane engineering specs. That’s about when I started subscribing to both Stereophile and The Absolute Sound, the so-called audio alternative press that has made a living off the notion that not all amplifiers that measure alike sound the same.