[2005-08-19]

Oh great, yet another guitar amp schematic. Just what the world needs.

full size schematic

While this isn't exactly earth shattering, I think it addresses one of the major problems with guitar amplifier design in an elegant way.

Most guitar amp designs are direct descendants of hi fi designs of the 40s and 50s. The earliest amps that we love so much were taken straight from the recommended operation points of the tubes in question. This is a good thing, because the circuits are straightforward, and do not require any drastically unique engineering to make them sound good with guitar. It is also a bad thing because there are a few small differences between a guitar and a record player for example. The primary difference between a hi fi amplifier and a guitar amplifier is that a guitar amp spends a great deal of time heavily distorted, and must sound good when distorted. This sounds like an oxymoron - good distortion - but for those of us who have embraced the vacuum tube as an audio device, it has become clear that it is not the amount of distorion that causes an amp to sound good or bad, but the kind of distortion.

In Hi Fi circles good distortion tends to mean low order distortion, which is distortion that is only a few harmonics away from the fundamental tone, and even order distortion, which tends to sound more "euphonic" (a classic hi fi mojo buzzword) because it is a simple doubling or halving of the fundamental tone and therefore creates less dissonance. For guitar, this does not appear to be as well defined. Guitarists like many different tones, and at least 2 distinct kinds of distortion. Many players like Fender Champ style single ended amplifiers because of thier rich blues tone. Single ended amplifiers are heavy in even order distortion, especially second order. This is the nature of a single ended design. Some guitarists find this preponderance of second order distortion somewhat muddying to the sound. A guitar sound heavy in second order distortion can lack definition of the individual notes, and have a hard time cutting through the other sounds in a performance. These players tend to like push-pull amplifier designs. In a push pull amplifier, the signal from 2 opposing tubes is summed in the output transformer. This summing cancels out any even order distortion, so a push-pull amp should have primarily odd order distortion. Many guitarists prefer this sound.

The effect that this design attempts to address is that when you drive a tube way into distortion, the grid can swing positive, and grid current will flow. Most guitar amplifiers have a low current 12ax7 driving the grid of the output tubes, and at 1mA of current, and very high plate resistance, the 12ax7 is just not equipped to deliver any current whatsoever to the output tube's grid. As the output tubes require grid current that isn't available, distortion increases dramatically. As a result, the amplifier might sound good at a particular volume level, but the range of volumes at which you get desirable distortion is very narrow.

So what we need is a beefier driver tube. Something with lower plate resistance that idles at higher current, and can therefore deliver more current to the output tube when necessary. There are 2 ways to do this that come to mind:

1) Cathode followers. A cathode follower is a common solution for situations where you need a lower impedance output. This is commonly used in preamps to drive long interconnects for example. Many Marshall/Bassman style amps use a cathode follower to drive a complex tone stack. In this case, you would insert the cathode follower between the phase inverter and the power tubes. The only problem with a cathode follow is that they sound like crap. As a general rule of thumb, no hi fi amp or preamp with a cathode follower in it will sound good. There are exceptions, and in fact many legendary guitar amps use cathode followers to drive the tone stack. I even admit to creating a few designs that use cathode followers. To drive the output tubes, we need a better approach.

2) An additional driver stage between the phase inverter and the output pair. Keep the phase inverter the same, but use it to drive a higher current tube, which in turn drives the output tube. This is like seperating the tasks of phase inversion and driver into two tubes, each optimized for the specific funtion. This is a good strategy, and has been used by my guitar amp guru friend Woody with much success. There are 2 things that I don't totally like about this design, and neither of them are true criticisms, becuase they may not actually affect the sound in the end. I don't like the idea of having to add an additional tube just to drive the outputs. The fewer stages the better. The fewer tubes the better. This is my minimalist side. The second thing I don't totally like is that the signal passes through 2 stages before being summed. I think it is a good idea to sum the signal as soon or often as possible after splitting it in the phase inverter. Every gain stage where distortion is added to the signal whle it is split causes differences between the signals that will either be amplified or cancelled when it is eventually summed, and that leads not only to distortion, but the loss of detail.

So not totally buying into either of these solutions, I went in search of a way to cheat. I wanted a high current driver, but I wanted that tube to be the phase inverter too. So I played with building a long tail pair phase inverter (the most balanced phase inverter during heavy distortion, and therefore most liked by Woody, and I trust him) that used a higher current tube than a 12ax7. I like octals, so I chose a 6sn7 (which is quite a legendary hi fi tube and sounds great in guitar amps). The problem with implementing a high current long tail pair is that the balance of the pair relies on a large shared cathode resistor. Additionally, to get a good amount of gain and linearity, you need a fairly large plate resistor. Ohms law begins to cause problems for you when you start trying to use large resistors with lots of current flowing through them. You begin to drop large amounts of voltage across those resistors, to the point that your B+ supply isn't enough. So you are forced to compromise by using smaller than desired plate and cathode resistors and hurting the linearity of the phase inverter.

Short rant: Isn't engineering great - a solution always raises other problems. I get so pissed when I see ads that promote "uncompromising" amplifier designs. In my experience there is no variable for cost in Ohm's law. These companies would have you believe that there is a little known correlary that states that R = V/I$. As if when you use very very expensive resistors, you can bend the rules. There is always a compromise. The whole point of engineering is getting the job done with a solution that makes compromises you can live with. Tubes designed in the 30s and 40s are always going to sound better than later tubes because those tubes were designed to sound good. Period. They were also a tremdous pain in the ass to use. Check out old Western Electric theater amps to see what I mean. Later tubes were designed to sound good and be smaller. Or to sound good enough, once you add a bunch of feedback. Good with feedback will never sound as good as just plain good. End rant.

Back to the design - what we need is a device that has relatively low DC resistance so that it won't drop a lot of voltage, but high AC impedance so that it will provide a substantial load for the cathodes of the phase inverter, and give us the linearity we want. One such device is the poor misunderstood choke (inductor). Chokes are the first component eliminated from a design to reduce cost because a resistor is much cheaper than 50 feet of wire wrapped around a hunk of steel. Another Hi Fi truism is that anytime you replace a resistor with a choke, the sound gets better. Ideally, we would replace the cathode resistor and both plate resistors with chokes. Maybe even on all stages of the amp. But that would be a very expensive solution, and have other problems, like picking up hum in the hundreds of feet of wire wrapped around the hundreds of pounds of steel :)

So this amp simply replaces the cathode resistor with a choke, to allow the use of a high current 6sn7 as a phase inverter/driver for the output pair of 6v6s. This high(er) current driver allows more current to be deliverded to the grids of the output tubes, extending the sweet sopt well into the region where grid current flows. Phase inverters tend to have neat names: cathodyne, long tail pair, concertina. I nominate this phase inverter be the "choke tail" phase inverter.

The phase inverter no longer looks like a long tail pair because the increased AC impedance of the choke makes the "tail" unnnecessary. The bias voltage is generated by the DC drop across the choke, but the large AC impedance typically generated across the cathode bias resistor plus the large tail resistor is generated across the choke as well. A side benefit is that we get to lose the capacitor that grounds the undriven grid in a long tail pair. Removing capacitors is a good thing in Hi Fi, so why not in guitar amps too.

The rest of the design is pretty conventional. Gain stage -> tone/volume -> gain stage -> phase inverter -> output pair. I am using a "moonlight" tone stack, which is a signle control low loss tone control that I think sounds nice. I am building this using all octal tubes, so we end up with a 6sl7, a 6sn7, a pair of 6v6s, and a 5y3 rectifier. You could replace the 6sl7 with a 12ax7 and the 6sn7 with a 12au7 if you like. You would get a brighter tone, a bit more gain, and smaller tubes :)

[2006-01-20]

The BHL-15 circuit in a Fender Pro Jr.

I implemented this circuit in a Fender Pro Junior. The chassis, cabinet, and transformer make a nice compact package, and once you rip out the PCB, there is room to build some good stuff in there. The stock 10" speaker even sounds pretty good.

Bob, the official Boozhound Labs guitar playing amplifier evaluator and co-conspirator says this design has a very wide sweet spot, souding good from 3 to almost cranked. Sounds like the higher current driver is doing it's job.

Bob's official report:

"It has an awesome bluesy sound that starts to break up early (volume = 3), and just gets better as you crank it up. It doesn't turn dirty until some point above 9 (these amps go to 12, that's 2 more than 10)... so the sweet spot is really big. I tried this with a few of my 10" speakers that I've collected lately but it actually sounds really good with the O.E.M. ceramic magnet Eminence..."

Sound Clip.