This article will discuss seemingly different items, but I will tie them together in the context of a small, relatively simple amplifier. Old style tube radios (the "all american 5") were powered directly from the 120 volt line for both the filaments and the plate supply. With enough insulation, and nothing external plugged into these units, they were relatively safe, but one side of the power line is the circuit common. This is dangerous (see disclaimer below), in that if the hot side of the line is circuit common (plug reversed) , and if you make contact, it is a fire and/or health risk. However, with correct isolation and proper precautions, this becomes a very efficient powering scheme, since it avoids the size, weight and cost of a power transformer.
I also did some investigation of the Amveco toroidal power transformers on the Owens Inductance Bridge (as described on these pages. I found those transformers had high primary inductance (on the order of 100Hy) with low leakage inductance (on the order of 100uHy). That implied that they *might* be good for audio. As it turns out, they work quite well, are inexpensive, and perform quite adequately.
What emerged is a small (300cu in, and less than 10 pound) stereo amp, capable of 3.5 watts per channel.
The transformers used are currently available from Digi-Key for relatively cheap (<$20).
Direct line circuits are dangerous. Do not even think about building this if you're a beginner, are unfamiliar with line isolation techniques and troubleshooting procedures, and/or need specific guidance in circuit building. PERIOD.
For troubleshooting and getting things going, until you're very certain that everything is properly isolated, even if you're experienced, use an isolation transformer (mains to 120V).
This is one of 2 channels, except for som of the common "powering" circuitry. Notice that neither end of the "line" is connected to chassis nor to anything exposed. This amplifier isolates line through its audio input transformer and audio output transformer. The input transformer functions as the phase splitter, driving a push-pull 12AY7 driving push pull 50C5. Interesting thing. I tried triode, ultra-linear and pentode connections on the 50C5s. The pentode connection, with the help of a little local feedback sounded best. (See performance section).
Notice the funny little circuit at the right. That's a "magic eye" tube I use as combined peak level indicator and pilot light. As shown, about 2 watts exactly "closes" the eye.
The input transformer is the closest Digi-Key had to a 1:1 transformer. It's small, but works very well. If you want a level adjustment (I did - see picture), place the pad on the grid side of the transformer. This improves isolation. As a sidebar, one of the things I was testing was whether those transformers had enough isolation - they do! No evidence of hum and/or "grunge" from the line.
The output transformer is the 7+7volt 50 watter. On the bridge, at 50Hz, the "117volt" winding didn't saturate until about 126 volts. So as wired, the secondaries can support 16 volts into 8 ohms at 50Hz, or 6.4V at 20Hz into 8 ohms without saturating (5 watts). Indeed, at 3.5 watts, there is no evidence of transformer saturation. Not bad for such a cheap part! Notice that with a turns ratio of 240:14, the primary impedance plate to plate is only 2400 ohms. This is low, but it works!
The 50C5: In the mode we're using, this tube, as best as I can determine is distortion free. (See performance section). All the distortion in this amp is essentially driver distortion. Possibly due to the fixed bias, and very low load impedance. Of course, 3.5 watts from a pair is a little low, but...
The amp was built using 1/4" MDF and copper clad board. I use it "at the office" since it's so small. Here's what it looks like:
How many tube amps will conveniently sit on top of a SeeDee player, and look small in comparison?
1. Frequency response, driven from 600 ohms:-0.4dB at 10Hz, -0.1dB at 20Hz, flat midband, -0.1dB at 15kHz, +0.1dB at 20kHz, +0.5dB at 30kHz, +0.2dB at 40kHz, -3dB at 50kHz.
2. Sensitivity: 0.7VRMS = 1 watt into 8 ohms.
3. PO: 3.5W just at observable clip into 8 ohms, 5.3W into 16 ohms.
4. Zout = 14 ohms.
5. Distortion 0.2% at 1 watt, 0.06% at 0.5 watt. This thing is amazingly "clean".
6. Transformer LF saturation limit, 1 watt at 10Hz, 4 watts at 20Hz.
The distortion characteristics are interesting:
Also note the frequency response. Not too shabby for cheap transformers!
OK. So, How does it sound? Well, it's not anywhere close to the A2 841 amp on these pages, but not at all bad sounding. Has more of a push-pull sound than single ended, but guess what? It's push pull. Still detailed, open, non-fatiguing. There is nothing in the sound that I can attribute to the cheap transformers.
Next time you need to consider an audio transformer, you might want to consider one of these "power transformers".