A Really, Really Tiny Tube Amp

After building his first tube amp from a kit, he set to work on his next amp build. Since tube amps are a much more experimental endeavor than their solid state brethren, [Jarek] decided to make his next amp unique with military surplus subminiature tubes, and in the process created the smallest tube amp we’ve ever seen.

Instead of bulky 12AX7s and EL34s tubes usually found in tube amp build, [Jarek] stumbled upon the subminiature dual triode 6021 tube, originally designed for ballistic missiles, military avionics, and most likely some equipment still classified to this day. These tubes not only reduced the size of the circuit; compared to larger amps, this tiny amplifier sips power.

The 100+ Volts required to get the tubes working is provided by a switched mode power supply, again keeping the size of the final project down. The results are awesome, as heard in the video after the break. There’s still a little hum coming from the amp, but this really is a fabulous piece of work made even more awesome through the use of very tiny tubes.

[youtube=http://www.youtube.com/watch?v=m5T-cXi09xQ&w=580]

37 thoughts on “A Really, Really Tiny Tube Amp

  1. Switch mode PS in a tube amp!? Theres your hum… thats like putting a Picaso in a frame from wal-mart. Probably has an AC filament supply rather than DC too, even more hum. You could probably find a small enough HV transformer with a filament winding and then rectify both and still fit, just wind your own or take a couple of very small 120V step downs and hook up the secondaries 1:1. And doesn’t pignose make a tube amp not much bigger than this? And while I’ve seen many tube amp rat nests (in full production models no less), he could probably help his noise issues a little just by addressing that layout a little. Still very cool to see someone having a go with tubes. Which are actually a lot more alive and well than most people think. Just outside of the direct consumer world. If you listen to AM/FM or watch TV via an Antenna, its probably being brought to you via a very large and expensive CPI (Eimac) or Thales tetrode.

    1. From the featured article:
      “That said, if you’ve made any changes to it and it sounds better, let me know!”

      Sounds like he’s open to comments and adjustments. Why don’t you tune up his schematics and send them to him?

    2. Switchmode supply can produce a lot of EMI that would whine or buzz but hardly hum. Hum the way I understand it is low frequency/low harmonic content buuuuuuu noise like your typical 50Hz. In the video I did not really hear anything that would call my attention.

      Maybe the author could record the hum at high volume, without music. That way we would at least know what he’s dealing with.

      1. A simple check would be to run it off batteries. Or at least the filament. Tho 100v isn’t unachievable with $11 of 9 volt batteries from a dollar store. Current should be within their capability if you run it quietly, probably even if you run it loud.

        If the hum disappears then you’ve found the problem.

    3. I’ll give my next build a shot with transformers, since I can now calculate the exact amount of current these things take and apply it to finding the smallest transformer I can safely use.

      If you guys would like, I can take a video or find a way to hook up the output to my sound card and take audio samples. Email me if you have some time to help :) I’d really appreciate it.

  2. Been 35 years since I’ve worked with tubes, but isn’t the signal from the second triode stage c0oming off the B+? Is the the signal using B+ as the path for modulation? That doesn’t look right.

      1. It’s a consequence of the unfortunate way the internal components of the tube are laid ‘upside down’ relative to the pins are broken out in the schematic. If you zoom in, you can see the ‘internal wires’ leading to the correct elements inside the tube.

  3. Beautiful! I especially love the case and HoneyTone grille.

    I agree that a switchmode power supply seems a bit out of place. The B+ needed is only milliamps. I built a little tube FM radio, and used a 60hz transformer with dual 120/120v primaries and a 6v secondary. One primary was powered, the other primary was rectified to provide B+, and the secondary powered the filament.

  4. It’s a nice looking amp, and using a switch-mode is a novel way of applying modern technology where it fits. Should sound better than a simple linear PSU since all the noise and “hum” is well past the audio spectrum.

    What I’d like to see though, is the tubes visible outside the case! It looks good when bigger amps do that, many expensive hifi tube amps make a show of them. In this case, the fact they’re so tiny is even better!

    1. I’m not an audiophile at all, so I’m not sure if it sounds ‘better’ but the sound is definitely ‘different’ from my regular receiver.

      My next build will definitely showcase the tubes a little more :)

  5. Fun stuff! As to all these SMPS nay-sayers, don’t mind them- switch mode supplies are used for audio applications all the time. You do want to make sure that the switching frequency is high enough that it’s a) not audible and b) far enough away from the audio band that it won’t get folded-back back crappy digital audio equipment with poor out-of-band filtering. As for the hiss, it could be layout, the speaker, the original recording, etc. Some switch-mode supplies will also revert to a pulse-skipping or low-power mode under low load. You might want to look up the controller chip datasheet and make sure that it will never go into any sort of low power mode as they almost invariably result in audio-band artifacts.

    1. Yup, I’ll be making my own HV driver for my next project, so I can play around with filtering out the switching frequency.

      (HaD mods, sorry for the comment report, I accidentally hit that trying to reply)

Leave a Reply to DougCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.