If our goal is an universal headphone amplifier that can handle all available headphones, then it must be able to deliver up to 35 mA and 6.3 V at its output. Fortunately, the amplifier is not required to meet these two extremes simultaneously, as the product of the peak current against the peak voltage will always equal twice the RMS power output, 20 mW. The geometric center of these current and voltage extremes is found by taking the square root of the product of the extremes multiplied against each other:
    Igeo = ÖIlow x Ihigh)
    10.5 mA = Ö35 mA x 3.16 mA)
    Vgeo = Ö2Wrms / RLoad)
    1.89 V = Ö6.3 V x 0.566 V).
Which also serves to imply the geometric center for the two impedance extremes: 179 ohms. Thus we have the high, the medium, and the low points of operation mapped out. The next step is to decide on the output stage's configuration and mode of operation.
     Because even relatively inexpensive headphones ($350) can reveal more subtle detail than the most expensive speaker, headphones deserve better than design compromises (why else would tube headphone amplifiers be desired?). The no compromise design choice would be for a direct coupled, single-ended, current source loaded, Class A amplifier. This ideal would give the tube its best chance to dazzle, but it is also one of  the least efficient design choice.
   Next month, I will review the relative efficiencies of the possible topologies for a portable tube headphone amplifier, starting with the simplest output stage topology, one tube and one resistor in either the grounded cathode, grounded grid, or the grounded plate configuration (cathode follower) and ending with horizontal push-pull amplifier. I will also cover the BIGEST problem facing this project: the power supply. Until then, your ideas and suggestions will be happily received.

Subject: Power tube biasing
Highly interesting and informative webzine! Coming back to your commentary in the October 2000 issue, I'm wondering if you know or even have details about AudioValve's  "ABR" - an obviously innovative and unique circuit, which simply replaced the cathode resistor in fixed / auto-biased power amps.  More about this circuit on Audio Valve's website
    http://members.aol.com/audiovalve
"our revolutionary and proprietary ABR circuitry, which monitors, biases and controls each tube independently. This feature allows the use of any of the mentioned (KT-88s, 6550s or EL-34s) tubes or, indeed, a combination of any of the three. This redefines in a truly innovative way the concept of "plug in and play." It also makes tube replacement very economical, as there is no need for costly and exotic tubes which have to be matched in order to get the best sound from the amplifier."

-Jens 

   First of all, I couldn't find a description or schematic of AudioValve's ABR circuit at their website. But my guess is that it is some form of auto-bias circuit. Which if designed for Class A amplifiers, are neither difficult or innovative or unique. If designed for a Class AB or B amplifier, then it could be all of the above. You see in a Class A amplifier, the idle current equals the average current drawn through the amplifier; but in a Class AB amplifier, idle current is much less than the average current. Having said that I want to make clear that I have a lot of respect for AudioValve's products and I lament that we in the US have nothing comparable. I once tried to talk a US maker of electronic kits into bring out a tube line. The owner was enthusiastic, until he spoke to an attorney that is; he was terrified of the liability.
    If anyone has any information on this circuit, please send it to us.