Building the McTube using a sub-miniature 6021 triode

For most every guitarist, the quest for tone never stops... it just rests sometimes! This leads us down many paths and the Mini McTube is the product of that quest combined with some history and the work of a great musician and engineer. First the history...

Tubes... it's very interesting to me how these devices can be so simple yet, even in this age of advanced electronics, still be the technology that delivers the best tone for amplifying a musical instrument. I was digging into tubes after my deep dive into FET usage to learn how they work and to actually get what I thought would be a better understanding of FETs. Instead, I encountered a little known device that caught my imagination...

The SubMiniature Vacuum Tube

The 40's and 50's were the heyday of tube electronics since that was all that was available for advanced circuits. Then as now, the military drove significant innovation and the need for miniaturization grew as engineers created more and more applications for "smart" devices. One such application was the proximity fuse. There were many important technology advances that helped the Allies win the war and the proximity fuse is held by many as one of the most important. Just imagine a tube-based circuit that was small enough to fit in a beer bottle and that could withstand the G-forces of being shot from a cannon or rocket. Non-trivial engineering requirements there! As a result, the sub-miniature tube was created; mil-spec, small, useful. This military technology was commercialized in the 50's, as was much of the war technology, by the auto industry. Car radios were another "perfect application" for the sub-mini tubes.

We all know the story from here... the transistor was invented, the world went solid-state, and tubes faded into a distant memory for most folks. That is... most folks who don't play the guitar and have a quest for tinkering with tone. Those other folks kept tinkering with tubes. Folks like Fred.

Fred Nachbaur and the McTube II

Talk about standing on the shoulders of giants! I first learned of the McTube project while surfing the web looking for tube amp schematics. The project intrigued me because it shared so many of the motivations that I have for my projects; namely, that of reuse of interesting but old componentry as well as frugality and the use of scrap/surplus gear. I filed that away in the memory bank and kept surfing. Then I came upon the whole sub-miniature subject discussed above and the Ahah! moment happened... why not build a McTube out of a sub-mini device?

As I browsed through Fred's sites I was more and more impressed. Here are Fred's own words describing himself in a Bio note: "I'm one of those mixed-brain types who enjoys both music and techie stuff, and relishes opportunities to merge the two -- sometimes successfully, sometimes not quite so. Classically trained in piano, but switched to guitar as main instrument some years ago. Musical tastes include rock, techno, folk, new-age, jazz, just about anything outside of rap and hard-core country. First and most abiding musical love, however, is classical and classical-romantic... and it shows in my music."

I then found out Fred had succumbed to cancer and was no longer with us. Building the McTube and continuing feed the fire that Fred's creativity sparked became my next project. Therefore, I present the Mini-McTube.

The Design

The original beauty of Fred's McTube project was that it used a standard triode tube in a fairly standard way but with a very non-standard power supply. The McTube is essentially the first two stages of a regular tube amplifier but with controls that allow a variable signal gain for each stage. These stages are in a simple grounded-cathode configuration. Controlling the drive output of each stage into the next allows the somewhat magical "tube overdrive" phenomenon to be controlled by the user. This over-driven signal is then fed into a standard amp and a very interesting an wide spectrum of tones are produced. The McTube can function as an ultra-clean signal booster that provides that much sought-after tube sound for most any amp. Turn the drive knobs up such that either the first McTube stage is over-driving the second or that the second stage is over-driving the amp and you've got the sweetest distortion/over-drive sound.

Why do tubes sound so good when over-driven like this? Here is a good discussion of what goes on inside the McTube as drive is increased to clipping levels. Basically, a tube amp never "clips hard" to the rail of ground or the positive power supply. Instead, the signal is gently "rolled off" such that intermodulation is decreased and desired harmonic distortion is increased.

Fred's twist on this project was to design the power supply section using surplus parts. This had the dual effect of making the project very accessible as well as portable and suitable for use in a stomp-box type application. The flash of insight was to use common "wall wart" transformers instead of bigger (and more expensive) power transformers.

One thing about the original McTube design did annoy me a bit... by Fred's own admission, "The component values in the two stages were chosen more or less experimentally to suit my own tastes." I wanted a bit more scientific approach to determining circuit component values. I suppose that's one of the differences between an artist and an engineer. For the artist, the goal is that of expression. For the engineer, the goals are more practical such as functionality, reliability, etc. However, both the artist and the engineer can produce works that inspire imagination. With Fred's help, I think we hit this one out of the park.

The Schematic

How it works

Stage One - The Power Supply

As mentioned earlier, Fred came up with a very neat twist on the power supply circuit for this design. Tube circuits have historically needed three different power supply voltages:

  • A+ = the heater filament supply
  • B+ = the plate voltage
  • C+ = the grid bias voltage
In old radios, these were actually supplied by three separate batteries, hence the A, B, C designation. These days, tube amps have a very specialized power transformer that steps the AC mains voltage from the wall plug down to the various voltage levels needed. This allows the generation of A+, B+, and where needed C+ directly from the wall current. I said "where needed" for the C+ supply because clever engineers soon learned how to self-bias tube such that the C+ supply was not needed. This is the case for the Mini-McTube. We only need to supply A+ and B+ However, bear in mind that for proper operation, B+ is typically about 150-350 volts! You can see why the standard tube amp power transformer is such a big-deal.

Fred's clever idea was to design the circuit such that a B+ of about 140 volts was needed and then to use two common step-down power transformers such that they operated in a two-stage manner. The first stage steps the mains power down to the A+ levels, and the second steps that voltage back up to the B+ level. By using the transformers from very common 110-120VAC to 6VDC "Wall Wart" power supplies, very workable A+ and B+ voltages are supplied. These are AC voltages at this point. B+ must be a DC supply and acually A+ is better (less 60 cycle hum) if it is also a DC supply. To accomplish this, two standard full-wave diode rectifiers and filter capacitors are used. In addition, a silicon voltage regulator is used to produce an accurate 6VDC supply for A+.

Warning: this circuit contains dangerously high voltage. Do not attempt this project without proper training and safety precautions. See our legal agreement and understand that assumes no liability for use of this information.

Click here for technical details

T1 is the step-down transformer that takes the 120VAC from the wall socket and steps it down to about 9VAC. Note that the transformers in use are the center-tap type. Bear this in mind should you want to build this circuit. Also, make no assumptions as to which pin is which on these transformers. Wall Warts are among the most cost-reduced components out there and there are no standards. Check for yourself to be sure.

Diodes D1 and D2 rectify the AC signal into DC and this is fed into the LM7806 DC voltage regulator to produce a smooth 6VDC. Note that the A+ specified by most 6*** tubes is actually 6.3VDC. At 6 volts, we are within the 5% tolerance and will be fine and this level may actually extend tube life a bit. Also note that the LED indicator power is taken from the unregulated 9VDC side of the power supply. This off-loads the regulator by a bit and the LED doesn't need regulated power anyway.

The original wall-wart in use was a 120VAC to 6VDC at 400maoutput. The diodes, capacitor, and voltage regulator components, as well as the circuit board to hold them all, came right out of the disassembled wall-wart unit. The result is an A+ supply that easily meets 300ma requirement of the tube that we're using.

T2 is where things get interesting. We normally think of transformers as step-down devices (such as T1 in this circuit). They may also be used as step-up devices and this is how T2 functions. It takes the non center-tapped secondary voltage from T1 and steps it back up to 120VAC. Note that the available current resulting from this step-up is very low. However, the Mini-McTube circuit is designed to only consume a few milliamps of B+ current so this is fine. What we really need is a solid, low noise, high-voltage DC supply. Diodes D3 through D6 (high voltage 1N4004 devices) rectify the AC wave into DC and C3 filters this signal into the needed supply voltage.

Note the voltage rating on this and other capacitors. They need to be much higher than we're used to in order to manage this voltage level. The diodes and caps are the only parts that need high voltage/wattage ratings. All resistors may be 1/8watt and all other caps may be 25v rating or so.

The basic formula for a full-wave bridge rectified power supply design is:

OutputVDC = Secondary OutputVAC RMS * √2 - (2 * VDC diode drop)

Since T2 has stepped up the voltage back to its original level (minus some loss due to transformer efficiency) we'll calculate the resulting DC voltage to be at a minimum:

OutputVDC = 110VAC RMS * √2 - (2 * VDC diode drop) ~= 154VDC

Put a little current load on this and it settles out to just below 150VDC. So, at the end of the day, we've accomplished exactly what a tube circuit needs, a groomed, high-current A+ for the heater filament and a high-voltage, low noise B+ for the plate voltage.

The Tube Amp

I chose the 6021 sub-miniature dual triode tube for this circuit. It has a medium amplification factor (mu) and appears to have been designed for audio amplifier applications. The McTube is a two-stage amp circuit and the dual-triode within the 6021 is perfect. Just wire one stage to the other and you've got it.

In my thinking, the two stages of the amp have independent purposes in the design. Stage one is to provide a clean, non-distorted but boosted signal to stage two. Stage two provides overdrive and harmonic distortion to this signal and drives the output. This arrangement allows for the Mini-McTube to span the tonal range from clean tube booster pre-amp to major tube break-up over-drive just by twiddling the two control knobs.

Click here for technical details

As I mentioned above, I wanted to the component values of the stages of the tube amp to be selected more as a result of science than Fred's trial and error. Given the individual requirements for each stage (stage 1 clean, stage 2 over-drive) the pathway to designing the stages was revealed.

A tube does not (nor FET for that matter) amplify in a strict linear fashion. Instead, when you plot amplification vs. plate voltage vs. grid voltage, it resembles a curve. These are called plate curve plots and are a standard part of tube datasheets.

The basic idea for audio amps is to pick your Q-point somewhere on the curve that maximizes the linear response region for your signal bandwidth. Sound too complicated? Not really. What it means is that you want your quiescent spot on the curve (the Q-point) to be in the flattest region of the curve. When signal is applied (that is, altering the grid voltage with the signal), and you move up and down the plate curve, the amplification factor is the same since you're in a region that's basically a flat line. This minimizes distortion of the amplified signal. OK, that sounds good for stage one. Using the plate curve plot, if we select a grid voltage of about -1 to -1.4 volts, we are in a pretty flat section of the curve and there's still plenty of head room for the input signal. What about stage two? By selecting a grid voltage for stage two at around -2 volts, we have placed it's Q-point in a more-rounded area of the curve. Signal applied to the stage will cause non-linear amplification as we follow the curved line of the plot. Also bear in mind that the input to stage two is the amplified output from stage one. This means we that we can swing much further across the plate curve arc because the signal can alter the grid voltage much more due to its amplification.

The basic circuit formula for a grounded-cathode tube amplifier is:

Rk = ( (Vb / Iq) - Ra - rp ) / (mu + 1)   where:
Rk = the cathode resistor
Vb = the B+ voltage
Iq = the quiescent plate current
Ra = the plate resistor
rp = the internal plate resistance
mu = the amplification factor

Our Vb is a dandy 150VDC thanks to our power supply. I selected Iq to be .3ma which is low but plenty enough for a tube based voltage amplifier. By selecting a value of 330kohms for Ra, we get:

Rk = ( (150 / .0003) - 330k - 6500 ) / (35+1) ~= 4.7kohm - where for a 6021, mu is about 35 and rp is about 6.5k in our operating range

With these values, the resulting grid voltage is (Rk * Iq) or about -1.4VDC. A perfect self-bias voltage for our clean linear stage.

For stage two, I selected a lower Ra. This has the effect of raising the plate voltage seen by the tube (since the voltage drop caused by Iq across a smaller resistor is smaller) as well as raising the required value for Rk. With a 220Kohm value we would nominally get:

Rk = ( (150 / .0003) - 20k - 6500 ) / (35+1) ~= 7.5kohm

However, even though I wanted to intentionally shift the Q-point downwards into the non-linear region, I didn't want to go too far. Therefore, I selected a 6.8kohm resistor which sets the grid voltage at about -2VDC which is starting to get into the curved region of our plate curve plot.

Each of the tube cathodes are decoupled with capacitors. These maintain gain by removing negative feedback due to cathode degeneration.

The input jack is capacitively coupled to stage one with a very high impedance resistor present to set the input impedance. Stage one's output is capacitively coupled to stage two with a level potentiometer. Stage two is then capacitively coupled to the output with a voltage dividing resistor and potentiometer to reduce the output signal to a reasonable level

All in all, a very simple machine. As I read through big tube amp schematics, I'm constantly amazed at how simple the schematics really are. A few tubes, a few resistors and caps makes the tone of a Fender Princeton!.

The Bypass Switch and Case

A standard stomp-box bypass switch provides the effect in/out function along with an LED indicator lamp. I managed to fit everything into a standard "Taiwanese 'C' Box" and all is good.

Click here for technical details

The switch is a triple-pole, double throw type where the LED, input, and output circuits share the same switch throw. In the bypass mode, the input jack is directly connected to the output jack and the LED is open circuit. In enabled mode, the input jack is connected to the first amp stage and the output of stage two routes to the output jack. In this mode, the LED circuit is completed and it lights.

Photos of the Build

The full circuit.

Starting from 12 o'clock: the LM7806 linear supply (w/ heat sink), the input jack, the two drive pots, the output jack, T1 (still using the wall plug and case), T2 (wired to the secondary of T1), and the circuit board in the middle.

Close up of the circuit board

That's a US quarter coin standing up there in front. B+ power supply on the left hand side, tube amp on the right.

The power supply

T1, T2, and the linear supply... all from two 6VDC wall warts.

Top view of the circuit board

Note that the 6021 is soldered into the circuit. These were not intended to be socketed devices.

The finished product

All tucked away in its box.