Furthermore, I just cannot believe that a string of zeners, even a string that is bypassed with a 47µF capacitor, can be all that quiet. Zeners are commonly used as noise sources. If I were to keep discrete transistor flavor of your regulator, I would arrange the zeners so that their noise generation would be further isolated.

    An excellent book that covers high-voltage regulators can be had in the Tektronix' Circuit Concepts series, titled Power Supply Circuits. This book by Kenneth Arthur is written from the perspective of a manufacturer that does want to see equipment returned for repairs. A second choice is the Kepco Power Supply Handbook, by Paul Birman. This book covers the issues surrounding the use of feedback loops in regulator quite nicely.
   On the other hand, floating type solid-state high-voltage regulators have satisfied me as being vastly, vastly quieter and more reliable. High quality Op-Amps are amazingly quiet, as are IC voltage references. By combining these two technologies into one regulator, we get a very low noise output. I have used the floating regulator in high gain phono stages and always  have had the regulator's noise swamped out by the tube's contribution. And I have seen an industrial floating high-voltage regulator that puts out 2500 volts with 10 mV adjustment increments (DAC generated voltage reference) with almost no noise due to the Op-Amp's extremely low noise and high gain.
    Even the lowly LM741 has a huge amount of DC gain compared to the single transistor amplifier.

   Having played with many transistor based high-voltage regulators, I can warn you that at some moment the regulator will probably die. Maybe a tube will momentarily arc-over or a fuse will blow or the wall voltage will spike for a few milliseconds or maybe a transistor will just give up facing high-voltage differentials at turn-on and turn-off. Short circuit protection can be added, but it will both slightly increase the output impedance and the complexity of the regulator. So the recommendation I make is to bathe the circuit in diodes. With the diodes in place, you have a small chance of surviving a direct short of the output to ground.
    I have built two transistor regulators similar to your design that have worked flawlessly for a year or so and then just died without explanation. The addition of the protecting diodes helps, but I am still nervous. This is where tube high-voltage regulators shine. I own two HP 711A high-voltage regulators. These marvelous units are tube based and are rugged. I have shorted the output more times than I care to admit and they keep on working.

< PREVIOUS

NEXT >