death. I write with one serious issue that I thought that I must bring up and several observations. In the January 2000 issue you wrote the following in an OTL article:

    "With the heaters, I am much less nervous. The heater must be insulated from the cathode to prevent the heater voltage from contaminating the audio signal. Furthermore, in spite of dealing with hundreds of tubes, I never have found a tube that went bad because its heater shorted to its cathode. Open heaters, yes; shorts, no."

   I worked in a TV repair shop in the late 60's and I remember finding an occasional tube with a H-K short, or leakage. This usually caused a hum bar in the picture or hum in the sound. In an amplifier with a series filament string the result could be a nasty electric shock since the line voltage could become shorted to the system ground thus energizing the CD player or turntable if the chassis is not properly grounded. I would stress proper  grounding if a series filament string is used. I connect the power line ground directly to the chassis at the power input, and connect the signal input to the chassis at the audio input connector.
        Your last issue addressed low cost or "free transformers". I think I have found another source of cheap power. My latest amp used 4 36LW6 (AES has them on sale) sweep tubes. The filaments are connected in series and driven with a bridge rectifier and cap directly off the line. You can find "industrial transformers" on ebay for about $20 for a 500 watt unit. You need one with a 480 volt primary and 120 volt secondaries. Hook it up backwards. This results in about 450 - 500 volts with a choke input filter and 600 volts with a cap input. 600 volts will drive the 36LW6's to over 100 wpc. I am currently experimenting with cathode feedback and direct coupled screen drive with these tubes.
    You asked for experiences with under voltage heaters. I have made amplifiers that used 7805 regulators to provide heater power, primarily to use cheap 5 volt tubes (5U8). I have powered 6

volt tubes with them and never noticed any ill effects with one exception. There are some audiophiles that claim that an amplifier sounds better after it has been on for an hour or so. This effect seems to be more pronounced with under fed heaters.
       On the subject of audio testing, I devised a push pull pentode amplifier (6550) with an adjustable amount of negative feedback. I used a four gang pot to reduce the feedback and the input level at the same time such that the volume remained relatively constant. The added control was unmarked. Bench testing showed a THD of .28 % at 20 watts with maximum feedback (~12 db) and 2.4 % at 20 watts with no feedback. I then loaned the amplifier to several of my audiophile friends and asked them to play with the mystery knob. The results were nearly identical, on most material (excepting dance music, techno and other music with accented bass) everyone preferred ZERO feedback on a pentode amplifier. After I explained the function of the mystery knob I wound up selling the amp to one of the listeners and removing the feedback from the CAD75i of another's.
       Keep up the good work - I look forward to reading it, I will let you know how the screen feed goes (an amp with no caps in the audio path)

George

  Thanks for information and warnings. I too once built an amplifier with a mystery knob that varied the ratio of feedback and input signal. To my ear both extremes, no feedback and total feedback, had virtues. The bass was best with feedback, but the rest of the sound was canned; the openness, uncompressed, live sound quality of the amplifier blossomed with no feedback, but the bass was fat. The comprise position I settled on was 4.2 dB of feedback. I was surprised to find this value so close to the result of a mathematical analysis I performed on the limit of safe feedback (6dB) in that amplifier.