It was a cold January Saturday night in Chicago and we had big plans. Buddy Guy’s Legends bar was packed. We setup directly under one of the PA speakers less than 15′ from the stage. Time to celebrate. Skip the glass, one pitcher each and keep them coming. We’re about to make bootleg recording history. Conversation evolved into bloviation on what our cover art would look like, certainly it would be a photo of our battery powered tube mic pre-amp recently created in my basement lab. We had four hours to kill before Buddy’s appearance. Our rate of Goose Island and Guinness consumption would put us at three-sheets to the wind by 11. Must focus. It’s time, Buddy was on. Much fumbling about and forgetting how to turn on the Japanese-made 24 bit digital recorder with its nested LCD menus, cryptic buttons, and late 90’s firmware. Make it work. We did, just in time for the bouncers to notice the boom mike and battery packs. Wait, wait… maybe we should talk about why tube amps are worth this kind of trouble first.
Yes, vacuum tubes do sound better than transistors (before you hate in the comments check out this scholarly article on the topic). The difficulty is cost; tube gear is very expensive because it uses lots of copper, iron, often point-to-point wired by hand, and requires a heavy metal chassis to support all of these parts. But with this high cost comes good economic justification for building your own gear.
This is one of the last frontiers of do-it-yourself that is actually worth doing.
Vacuum tubes work by thermionic emission, meaning that electrons are emitted by something really hot under vacuum. In this case a lightbulb filament or a lightbulb filament that heats up a cathode metal plate encircling the filament. These electrons are accelerated from the cathode to the plate, thereby flowing current from the cathode to the plate. This current is controlled by a control grid, literally a small wire mesh between the cathode and the plate that looks like window screen. For practical purposes the grid is similar to the gate of a field effect transistor although the physics are completely different. There are some enthusiasts who actually make their own tubes from scratch.
It is easy to get sound out of vacuum tube audio circuitry. Tube circuits lend themselves to self biasing and simple first-order approximations. Buy parts today and listen to Jimmy Hendrix by Sunday evening.
How a simple single-stage tube preamplifier works
The audio input is fed directly into the grid of V1, which is also shunted with resistor R1. R1’s job is to both pull the grid of V1 to 0V potential and also provide a termination impedance to the audio input (in most cases 50K or 100K). R2 and R3 set the gain and bias point. A simple approximation for voltage gain is R2/R3 assuming the tube has a ‘high mu.’ mu is a measure of transconductance or the tube’s ability to amplify, in other words current output/voltage input. Current flowing through the tube sinks a current across R3. This voltage across R3 while this circuit is idling (or not amplifying) is equal to the bias voltage. This configuration is known as ‘self biasing.’ Old-salt engineers will tell you that designing with tubes is easier than transistors. They’re right.
The output is AC coupled with C1 to block the plate voltage from feeding into whatever this circuit is wired to. Like a single-stage transistor amplifier the output is inverted, a rising voltage at the grid results in a falling voltage at the plate.
You will need 50-150V of B+ (or plate voltage, because it is connected to the plate of the tube via R2) and about 1 mA to power this preamplifier. You can achieve this with the following circuit that supplies a regulated 120 VDC using zener diodes.
It is important to note that tubes are high voltage low current and from a conventional/modern circuit design perspective the resistors seem very high, caps low, and current regulation approaches amateur at best. This power supply works because we only need 1 mA.
Connect the inputs to your audio source and the outputs to your favorite solid state amplifier. We now have a ‘hybrid amplifier,’ containing a tube front-end and a solid-state back end. Professional versions of this amplifier are for sale by a large consumer electronics manufacturer.
Take it on the road, a true story from my college experience
So back to the story I started before. My bootleg conspirators and I were well lubricated and about to get the best recording of a live concert to date when we were stopped by the bouncers. After a short interrogation we were booted out, back onto the cold streets but with a feeling of satisfaction that, although we did not record the show, we had the best audio gear in town:
In this design, two pre-amplifiers described above are used with a battery powered high voltage supply. The high voltage supply works by creating a square wave with a 555 timer. This square wave is above audible frequency, around 40 Khz. The square wave is fed into a small audio power op-amp. The output of this op-amp is back-fed into the secondary of an audio output transformer, generating high voltage AC 170V at 40 Khz. This signal is rectified, filtered, and regulated in a conventional sense. The entire system runs on 8 AA batteries and should operate for approximately 4 hours continuously. Everything you need to make your own is here in this PDF.
Eventually you may want to create a power amplifier. There are a number of classes and variants of tube audio power amplifiers:
- Class B are push-pull amplifiers, similar to conventional solid state amplifiers, where there is a pushing tube and a pulling tube being driven 180 deg out of phase.
- The class AB is biased so that the output devices are on all the time, thereby reducing or eliminating cross-over distortion which occurs when one tube hands the load off to the other.
- Class A amplifiers are basically a high power version of our pre-amp above, where instead of resistor R2 we have an audio output transformer. Lots more power is burned needlessly in Class A amplifiers because they are always on, they either let go of the load or pull it closer to 0 potential. When not doing anything they sit half-mast, burning lots of power. Many prefer class A because there is no cross over distortion and they are very simple to design.
- More obscure types of amplifiers include transformer-less, where lots of tubes are in parallel in a Class A configuration to directly drive a 4-8 ohm loudspeaker load.
Almost all tube amplifiers require a transformer to match the high impedance of the tubes (3K or so) to the low impedance of modern loudspeakers (4-8 ohms).
Feedback can be applied to all amplifier types, where just like an op-amp circuit some of the output signal is fed back to the input, allowing the amplifier to compensate for non-linearities. Feedback provides a much cleaner signal and improved performance. This is not always desirable depending on your goal. Without feedback expect on the order of 7% total harmonic distortion. With feedback and high gain expect 0.5% total harmonic distortion.
A typical tube power amplifier is a class AB with feedback. In this, there is a gain stage that also functions as the differential amplifier when feedback loop is closed. This is followed by what is known as a ‘phase splitter,’ which is basically a buffer or small amount of gain using two triodes providing both in-phase and out-of-phase output. The two outputs drive the output power tubes at 0 and 180 deg phase respectively. The output tubes push and pull on the output transformer. Tubes can only pull, so the output transformer is powered from a center tap on its primary winding. Finally, the output is fed both to a loudspeaker and through a feedback network to the first stage.
A working example of this is shown in the schematic with details on its implementation (PDF).
Tube home theater
With these pre-amplifier and power amplifier circuits you can scale your design to a complete home theater system with 5.1 sound. For technical details of its implementation, check out this PDF. You can also check out my project which was featured early last year.
Let them shine
Just like the reality TV shows where skilled craftsman make motorcycles or hot rods you too can customize the look of your tube amplifier. Place the tubes on the top of the chassis so you can watch them glow. Co-locate the output, power, and chokes on top as well. It’s fun to watch them glow in the dark while playing ‘Dark Side of the Moon.’
Where to find parts
Audio tubes are easy to find and continue to be manufactured in Eastern Europe, Russia, China, and others. Transformers of all types including audio output and power are readily available through Hammond Manufacturing and other sources. A new movement in the tube audio community is to find obscure pairs of output transformers and make single-ended non-feedback amplifiers. Recently, manufacturers have started to remanufacture high voltage axial leaded caps and other old-school parts. You can find this stuff at Antique Electronics Supply, Just radios, and even Mouser and Digi-Key (less expensive for same stuff but requires a lot of filtering and digging through the catalog).
There’s a lot of material on tube design, but these are my personal favorites:
- Radiotron Designers Handbook.
- Principles of Power.
- Where some of the tube audio community gathers, audioXpress Magazine.
Try vacuum tubes in your next audio project. Keep those filaments lit.
Gregory L. Charvat, Ph.D is author of Small and Short-Range Radar Systems, visiting research scientist at Camera Culture Group Massachusetts Institute of Technology Media Lab, co-founder of Hyperfine Research Inc. and Butterfly Network Inc., editor of the Gregory L. Charvat Series on Practical Approaches to Electrical Engineering, and guest commentator on CNN, CBS, Sky News, and others. He was a technical staff member at MIT Lincoln Laboratory from September 2007 to November 2011, where his work on through-wall radar won best paper at the 2010 MSS Tri-Services Radar Symposium and is an MIT Office of the Provost 2011 research highlight. He has taught short radar courses at MIT, where his Build a Small Radar course was the top-ranked MIT professional education course in 2011 and has become widely adopted by other universities, laboratories, and private organizations. Starting at an early age, Greg developed numerous radar systems, rail SAR imaging sensors, phased array radar systems; holds several patents; and has developed many other sensors and radio and audio equipment. He has authored numerous publications and received a great deal of press for his work. Greg earned a Ph.D in electrical engineering in 2007, MSEE in 2003, and BSEE in 2002 from Michigan State University, and is a senior member of the IEEE, where he served on the steering committee for the 2010, 2013, and 2016 IEEE International Symposium on Phased Array Systems and Technology and chaired the IEEE AP-S Boston Chapter from 2010-2011.