But when the heater is still cold and the electrons haven't left the cathode, the high plate voltages can force pieces of the cathode to be ripped off and to fly up into the plate. Two approaches are usually taken to prevent this damage. The first is to let the heaters receive their voltage before the rail voltages develops. This can be accomplished by using indirectly heated tube rectifiers in the power supply, which slow the turn on, or by using a relay and a timing circuit to stall the B+ voltages. The second approach is to protect the cathode by presenting a large negative voltage to the grid so that the field that the electrostatic develops around the grid wires will shield the cathode from the high voltage present on the plate. The worst scenario is when both grid and plate are at high positive voltages.
  In this circuit, the output tubes are intrinsically  protected from cathode stripping by virtue of the lack of a path of conduction until both tubes are  conducting. This is a big improvement over the plate resistor loads, as at turn on, the full 600 volts potential is developed from the cathode to the plate. The Split Load phase splitter's triode is also better off, as the DC feedback does not come into effect until the output tubes begin to conduct, which allows the triode's cathode-to-grid to only see a 0 volt voltage, instead of a +120 volts.  Actually, just adding a diode that spans from the grid to the cathode will protect the cathode as well.

   In the same vein, the three diodes that span from ground to the cathode of the Split Load phase splitter at the input are there to protect that triode at turn on. Normally, the cathode is slightly positive to the grid so these diodes cannot conduct. Should the cathode become more than 2 volts negative, the diodes will conduct and prevent the cathode from falling anymore negative than -2 volts. At turn on, the triode will not see 0 volts on its grid and -300 volts on its cathode. Does not the inclusion of these diodes limit the maximum voltage swing of the cathode to 4 volts peak-to-peak? Yes, it does, but one volt of input signal will drive the amplifier into clipping anyway.

Cathode Follower Output Stage
   The output stage can be reconfigured to provide a much lower output impedance, but no gain. Understand, this may not improve the sound at the headphones, but then this article (and this journal while we are at it) really is not concerned with the single supposedly only correct way of designing a tube amplifier. Our aim is just the opposite: we want to explore and expand, not confine and constrict.
   Additionally, a balanced, direct-coupled, push-pull tube output stage can be used for more tasks than just driving electrostatic headphones. Maybe high capacitance balanced interconnect terminated by 600 ohm resistors is the intended load.

The output stage re-wired as a buffer, rather than a gain stage.

pg. 7

<PREVIOUS

NEXT >