Specifications

3. As Screen Grid voltage increased between the tubes, OLG decreased accordingly between the tubes, in keeping with theory.
4. The tubes with the most even split of voltage across the plate and cathode resistors and tube (HH Scott), produced the least
distortion.
Recently, some attention has been paid to the basic topology of ST-70 type driver stages, and more specifically, how tolerances
in the screen grid of the pentode section can produce significant changes in its plate voltage, which by way of the direct
coupling, can then bias the triode section into an unfavorable condition.
The use of a large series dropping resistor to supply the screen grid with operating power works to largely overcome this
concern. But still, there are tubes whose tolerances are such that they will still cause the triode section to become over-biased.
With the poor performing Zenith tube having the lowest pentode plate voltage of the bunch, it just may be one of those tubes.
To produce optimum operating conditions, the standard rule of thumb for phase splitter designs is that 1/3 of the B+ supply
voltage should be dropped across each of: (1) the plate load resistor, (2) the tube, and (3) the cathode load resistor. To correct a
tube example that over-biases the triode section, the screen grid voltage of the pentode needs to be reduced, allowing its plate
voltage to rise, causing the bias to be reduced in the triode section, and optimum performance to return.
Accordingly, a variable voltage supply was then added to the screen grid circuit of this tube, allowing the screen voltage to be
reduced even further than it already was. Result? When the adjustment in screen grid voltage was made to properly bias the
triode stage, the THD for this tube did not change one iota from that before the adjustment was made. Zero, Zip, Nada.
Since tackling that symptom didn't provide any resolve for the inconsistency, the next check involved making sure that the driver
stage was supplied with enough B+ during periods of elevate power output, to ensure its capability of providing the output stage
with plenty of low distortion drive signal during those times. Using a variable voltage power supply, the B+ to the driver stage
was varied to see at what point the supply voltage to it impacted the THD produced by the unit. Not until the supply voltage
reached a low of 200 vdc was THD negatively impacted.
So the results from both of these tests would indicate that the elevated distortion produced from the one bad Zenith tube is a
product of that tube itself, with the circuit not contributing to it in any substantial way, and there being no apparent way to readily
correct for the defects of such tubes. This is the most unfortunate aspect of the Dynaco driver board: While its design is capable
of very high performance, the consistency of the tube designed for it varies all over the map. I have found no easy way to
determine which tubes will perform best in this circuit (short of using a THD test set). However, it appears from these tests that
the tubes which inherently drop a similar amount of voltage across the splitter tube and its loads -- without circuit adjustment --
seem to perform the best.
So a simple correction can eliminate any concerns for LF oscillations, and 75% of my 7199 tubes perform very well in my
amplifier. In relating a discussion had regarding the design of the power supply to the design of the driver board, and the results
of its design versus the reality of practical use, it does what it needs to do very well. In the finished amplifier, it provides credibly
flat frequency response over the whole of the audio bandwidth, exceptional HF stability, a well damped square wave
presentation with minimal overshoot, and low distortion, assuming a properly performing 7199 tube is installed.
Still, for a variety of reasons mentioned, alternate driver boards may be desirable. The problem is, there is precious little
information on the performance these boards actually produce once installed, against the copious amounts of subjective
information offered about them. All you can do then is to do the best research you can in making a decision as to what is best for
your needs.
Final Summary
The Dynaco ST-70 was designed in an age when high accuracy sound reproduction was THE ONLY goal in high quality, high
fidelity sound equipment, and came to the party with an impressive list of credentials to backup just how well it could accomplish
that. As originally designed, with:
1. A frequency response from 20 Hz to 20 kHz that is flat within .25 db,
2. A power response that will allow full reproduction of the 20 HZ to 20 kHz bandwidth within 1 db of rated power with credibly low
distortion,
3. Ample power to handle virtually any speaker of the day with virtually any source material of the day with ease,
4. Excellent HF stability,
5. Excellent transient response,
11