Tuesday, April 1, 2014

Techy Tuesday - What's All This Vacuum Tube Stuff, Anyway?

With apologies to Bob Pease.

Get a group of audiophiles in a room - especially a bunch of guitar players - and ask their opinions of vacuum tube amplifiers or solid state amplifiers.  And get ready for a really active discussion. 

Let's back up a second.  A certain amount of "WTF are you talking about?" is expected.  When we say "amplifier", we're actually referring to a system of parts: a cabinet with a speaker or two that takes input from an audio preamp or maybe a guitar and brings it up to room-filling volume.  But each one of those is based on an active amplifier device that takes small voltages (or powers...) and makes them into bigger voltages (or powers...).  The active device that does this - the amplifier in the amplifier - could be silicon: a bipolar transistor, or a Field Effect Transistor.  Or it could be a vacuum tube: metal electrodes, hot plasma and wire running in an evacuated glass envelope. 

I've got to tell you, I've been paying attention to technical details on this subject for years and it is one of the most complex subjects I've ever come across.  One of my frequent reads, Lou Frenzel in Electronic Design, dives into the topic.  He ends up sounding like me. 
I have actually compared solid state audio power amps to the vacuum tube equivalents several times and using the same speakers.  (It seems to me that the speakers would have more of an effect on the sound than the type of amplifier.)   I could discern a difference between the two.  I do not have the words to describe the difference.  It is akin to comparing wines in a tasting.  There are words for that but they are also vague and subjective to be sure.  So it is with audio sounds.  I have actually heard people say they can tell the difference between two different sets of speaker cables and connectors.  I still don’t believe it.
There's a big deal in there, when he refers to the speakers having more effect on the sound than the type of amplifier.  The guys who build these amplifiers frequently don't account for everything.  Speakers couple magnetically into audio wires; wires interact with each other.  Moving the input transformer with respect to the input circuit can change the way the system sounds.  Maybe it's all the effects of the transformer and has nothing to do with the active devices (transistor amplifiers tend to drive the speakers directly while vacuum tubes need the transformers). 

One of the most common things you'll hear is that transistors tend to odd order distortions - odd harmonics of the fundamental tones you produce by playing a string - while tubes tend toward even order distortions.  The theory goes that humans perceive even harmonics as sounding nicer than odd harmonics sounds.  I recall back in my days playing with electrical things as a kid running both a square wave and a smooth sine wave into a speaker: the square (odd harmonics) did indeed sound harsher. 

I'll be 100% honest, as always.  I don't know what to make of this.  Our ears, our "psycho-acoustic" processing is amazing.  But audio is ferociously complicated.  Everyone who has played around with stereos has heard of "8 ohm speakers", but there is nothing inherently 8 ohms about any speaker, and the actual impedance measured at the input varies with where the speaker sits in the room.  The back pressure from the speaker compressing the air in the room and that air pressing back will change the impedance.  Our ears can hear amazing things, but so can modern instruments.  I think the situation is that there is no simple, tidy little answer.  It's not odd harmonics or even; it's not transformers or direct drive; it's not transformers perpendicular to their wires or not.  It's all of this and more in a subtle blending.  It's complicated. 

Maybe the answer is get a box of tubes, some sheet metal, and the rest of the parts and put a few things together.


  1. "Back In The Day" I used to build custom amps, and repair anything out there.

    My "big" amps were solid state, would loaf along at 500 Watts RMS, and the guys playing large venues liked them. They had flat frequency response from 10Hz to 50kHz, separate split power supplies, and were really a "Power Op Amp" with a differential output rather than a single-ended design.

    I also built several tube amps using 813's as output devices, but nobody liked them for road use as they were too big, too heavy, used too much power, and threw off large amounts of heat.

    Had a lot of fun building them, and even made a few $$.....

  2. Moderately related: I know a guitarist who built his own tube amp. It goes to 12.

  3. My Dad, who was an electrical engineer, came home one day (circa 1960) with a new Harmon-Kardon amplifier, which he attached to our turntable and AM/FM receiver. Even as late as the 1990's, I recall him saying that the Harmon-Kardon vacuum tubed amp had yet to be equaled without spending 10 times the money.

    And, for all of those Gen X-ers and Millennials who don't remember going to the tube testers in the hardware stores, here is a good background primer video on vacuum tubes:


  4. Even MacIntosh - the home stereo hold out for tube amplifiers - has gone to solid state. You can still buy tubes from them - they have a 50th anniversary edition stereo amp available.

    But their single channel, 2000 amp monoblock is transistor based.

  5. Yeah, it's complicated. Still when you go from an acoustic signal to an electrical signal and back again you expect things to get changed in the translations.

    Personally I can't tell the difference between high quality digital sampling on solid state playback verses tube playback. Might be different where there is a "live" feed going into the system instead of a recording though.

  6. I should have pointed out somewhere in there that I don't play a tube-based amp. My own amplifier is a Fender Mustang I "modeling amplifier", which combines solid state amplification with digital signal processing to allow the modern amplifier to more closely copy the sound of the older amplifiers. I'm sure to some folks that would be "the worst of all worlds".

  7. I'm sure that at least half of those "some folks" couldn't tell the difference in a blind sound sample.

  8. "a square wave and a smooth sine wave into a speaker: the square (odd harmonics) did indeed sound harsher"

    Interesting - as I imperfectly recall AC drive controllers often provided a range of steps to model a sine wave, in an attempt to avoid unpleasant system harmonics. Seems reasonable that a speaker would make those harmonics more noticeable.

    Have you seen the youtube videos of the French guy making vacuum tubes by hand? Incredible!


  9. Itor - yes I have seen that guy. Simply incredible, like you say.

    And I've made motor controllers that do that. I have a picture that shows it here from my "Least You Should Know" series. The picture at the top is more like what I was listening to. The lower one is the motor drive with lots of smaller steps. I'm guessing that probably would sound quite a bit better than the pure square.

  10. Yeah, that's what I was attempting to communicate!

    Will you make additional posts in that LYSK series?


  11. I really intend to, but never seem to get around to it.

    It's a pretty daunting task, really. Electronics is a wide field and the least you should know varies with what you want to do. I could go down several roads. And that's just one field. Any preferences? Shoot me an email, address over in the right side bar.