There’s just something magical about vacuum tube stereos. For me, it’s the combination aesthetics and sound quality. These two considerations almost appear to be interlinked; the glowing filaments seem to add a feeling of life to an otherwise bland appliance. The warm, vibrant tones of tube-equipped amplifiers make even inexpensive amps interesting to own. After a false start attempting to build a simple single-tube amp a number of years ago, I thought I’d give it a second try. This is the result:
This amp features miniature 7-pin tubes, which were made in the millions from the 1950s through to the 1970s. 50c5’s and 12av6’s were used – one each per channel – partly because I had them on hand, and partly because they only required line-voltage (120V), removing the need for expensive step-up transformers. A smart-looking wood case showcases all the tubes and transformers while hiding the ugly electronics. Plus, using wood means easy machining when cutting all the required holes.
Tubes Mean Experimenting! Unlike solid state amps, tube amps let you swap out individual tubes in order to experiment with different models/manufacturers. This is one of the main reasons a lot of people buy or build them; experimenting is fun!
On hand, I have a few different compatible output tubes: 50EH5s and 50C5s. It is interesting to note that swapping a 50EH5 tube for each 50C5 seemed to improve bass performance. Tonality of the middle and mid-low end was also bumped up. I can only surmise that this is because the 50EH5 is specc’d for a lower harmonic distortion.
The amplifier takes most of its basic audio design straight off the tube datasheets – with a little experimenting to get values tweaked. What’s this mean? Well, vacuum tubes aren’t exactly like transistors. They have a lot of the additional circuitry needed to make a transistor work already wired inside the glass envelope – so wiring and design around tubes is really a very simplified, off-the-shelf job. Manufacturers from the era did most of the heavy lifting, published a specsheet with a few gain and voltage limits, some I/V curves, and a bias circuit for typical operation.
The designer was left to tweak the “typical operation” for their specific use case, and incorporate the tube into the rest of the design. In this way, tubes operate more like op-amp packages than individual transistors.
The amp uses cathode feedback to bias the tubes, and negative feedback from the output transformer in order to reduce hum. Removing the feedback altogether boosts gain, but also boosts hum noticeably. The power supply has 2 1H chokes, which are inexpensive and easy to source, and many capacitor filters to remove as much hum as possible for the preamp stage.
Since this set uses the 2*50V + 2*12V = 124V heaters lined up in series, there is no need for any filament transformers. The filaments are wired to the power transformer – but there is nothing preventing it from being wired straight to the mains. This arrangement would in fact have the benefit of reducing the size of the transformer needed (and thus reduce price). As it stands, the transformer voltage-sag on a cold start allows a “soft-start” for the filaments, reducing the chance of a burnout.
The schematic below (done with gEDA) is the final evolution of the prototype and should provide a great bedroom amp for around $40 to $60 in components.
As mentioned earlier, the schematic makes use of 12AV6 (mislabeled as 12SQ7) and 50C5 tubes. The 12AV6 is an interesting tube: it is electrically equivalent to half of a 7025 or 12AX7 tube, if you have no need for the ‘sensor diodes’ (pins 4 and 5). If you want to have a real selection of tubes to play around with, substituting the dual 12AV6s with a 12AX7 is a good bet. Since 12AX7s are popular both as “NOS” items and as new manufacture, there are many tubes to experiment with.
As always, this design features an isolation transformer to prevent electric shocks. I made my isolation transformer out of 2 back-to-back 120V:24V 30-Watt transformers. Although it had a higher voltage sag than a ‘real’ isolation tranfo, this arrangement worked exceedingly well at a fraction of the cost of the real McCoy.
How Not To Wire for Hum: I did point to point connections, and since I didn’t care much for precision in this first prototype, I left a lot of connections with extra slack. I didn’t much care for wire routing either, as can be seen in pictures below! Since the picture taken below, I have gone back and rewired many connections to reduce slack and generally fix some of the craziness in wiring. In a proper set up, of course it’s best to shield all the wires to prevent cross-talk and noise by forming twisted pairs. In my set up, simply re-routing the power wires farther away from the input stage made a great deal of improvement.
The single best improvement for hum was going back and rewiring to accommodate additional filter components in the power supply. Much of the ‘hum’ was coming from the input stage – which is sensitive to the power supply ripple on the B++ voltage. Adding and extra choke and a few more capacitors, as in the final schematic above, fixed things up nicely.
Don’t forget to align your transformers properly either! That’s another thing you learn from experience. Some 60Hz power line hum was being induced into my output transformers because the PS tranfo and OPT tranfo cores were aligned the same direction. By rotating the OPTs 90 degrees so the cores are perpendicular to the PS’s cores, the mutual inductance between the two is minimized and no hum is induced.