Vacuum tubes are almost obsolete. Nearly the last holdout, the cathode ray tube (CRT), is rapidly being replaced by the LCD and other new technologies. Despite this trend, the vacuum tube has seen a big revival in the field of guitar amplifiers, and to a lesser extent, hi-fi amplifiers. Vacuum tubes and related parts have become more readily available in recent years as numerous companies have tapped into this market.
The reason for the popularity of tubes in guitar amps involves the nice tones that are produced when tubes are driven to the point of distortion. For some background on this, follow some of the links on The Strat Monger. There are numerous solid state "modeling amps" that try to simulate vacuum tube amps with digital signal processing (DSP) techniques, but in the end, that method is never more than a simulation. It just ain't the same as the real thing.
One can spend a large amount of money and time building a tube amp from scratch. Hammond organ ampifiers chassis are available on the surplus market for a reasonable price, they make a good starting point for a guitar amp. The difficult job of cutting chassis holes for the tubes and transformers is already done, one just needs to drill a few holes for the potentiometers and connectors. This project started with the amplifier from a Hammond M2 organ, chassis model AO14-1B. The newer and more common AO-29 (B3 organ) chassis may also be used, the cathode bias resistor values on the 9 pin preamp triodes would need to be changed.
The output stage of this amplifier resembles a fusion between a Fender Princeton Reverb, Fender Vibroverb and ham radio transmitter. With 6V6 output tubes, it puts out approximately 18 watts of audio power. The 17" reverb tank provides a deep echoey sound. The "simpler is better" philosophy was used in the design, extra inputs were intentionally avoided to reduce hiss. The design is an attempt to use the minimum number of tube stages. It is plenty loud, and the sound quality is excellent. The amp has worked well driving both 12" and 15" guitar speakers. The 1.0 version of this amp no longer exists, it has evolved into the Hammonator 2VRT.
There are a few unique features with this amp design, and some slight deviations from the aforementioned simplicity goal. An optional fluorescent EM87/6HU6 "magic eye" tube is used for an output level meter, it is fun to stare at while playing. There is a reverb send control (Dwell) that can be used to expand the variety of reverb sounds. Most Fender amps send only a full-strength signal to the reverb spring. By turning the reverb send signal down a bit, a less "clangy" and more "spacey" reverb sound results. There is also a control for negative feedback. With the feedback control turned all the way to the left (max feedback), the amp distorts a bit more and the waveform peaks are reduced. With the feedback control turned all the way to the right, the sound is cleaner, louder and less compressed. The feedback control could also be called "Clarity".
This amp uses four octal base 6SN7 dual triode tubes for all of the low level signal amplification instead of the more common 12AX7 or 12AU7 tubes. This was done because the chassis was already set up for the octal sockets. Boutique amp enthusiasts will probably like this feature since the 6SN7 tubes are older and may have more of a vintage amp sound. Fortunately, the 6SN7 is still easy to acquire. This amp has been "tuned" for good sound, the bias settings of all of the tube stages were tweaked while a guitar was plugged in. This process was used to optimize the musical qualities of the amp. Not all vintage 6SN7 tubes are the same, quieter Sylvania tubes were used for VT1 and VT3 to reduce the hiss, nosier RCA and GE tubes were used elsewhere.
This is a fairly advanced-level project. It takes a lot of technician skills to deconstruct and reconstruct the ampifier circuitry. Also, there are plenty of lethal high voltages inside of this amp including 120 VAC and 400 VDC. The project should only be taken on by someone who has experience working with high voltage circuitry. The power should always be removed when working on the amp, the circuitry is designed to discharge the capacitors when power is removed, but it's always a good idea to short out the electrolytic capacitors before working on the amp.
Power Input - grounded 120VAC Guitar Input - High Impedance Reverb Send Reverb Return Speaker Output - 8 ohms
On/Off (on the back) Input Volume Bass Treble Reverb Send (Dwell) Reverb Return Feedback
The AC power input circuitry was modified from the original Hammond circuit. The power transformer is old enough that it was designed to run on 115V mains instead of the 120V mains found today. Running the stock amp on 120V produces higher filament and B+ voltages than desirable. This problem can be easily fixed by putting the 5V rectifier filament winding in series with the AC primary winding. The 5V phasing must be correct, the easy way to test this is to try both orientations and monitor the 6.3V filament winding, use the lower wiring that produces the lower voltage. When the tubes are connected, the filament voltage is very close to 6.3V.
A grounded plug was used, this is critical for safety. A 2 amp fuse and switch are used to provide a standard fused disconnect. The varistor on the transformer primary protects against line voltage transients, those can get multiplied on the high voltage output winding and cause damage.
The transformer high voltage winding is sent to a center tapped full wave rectifier consisting of two 1N4007 diodes. The high voltage DC is dropped through a typical chain of resistors and capacitors to produce the voltages used in the amp. The first resistor (500 ohms/10 Watt) is used to set the initial B1+ voltage that drives the power output tubes. One could probably reduce this value a bit, but the amp is plenty loud as-is and higher voltages may reduce the power tube lifetime.
There is a lot of misinformation on the net about tube rectifiers vs solid state rectifiers and the effect on amp sound. This probably derives from the more efficient nature of solid state diodes and the resulting higher voltage when a direct substitution is done. Putting a resistor after the diodes drops the B+ voltage to a level that is closer to that achieved with a 5U4 rectifier. The diodes have the advantage of better efficiency due to the lack of a filament, the power transformer will run much cooler using diodes. The 1nF/1KV capacitors across the diodes protect against high voltage transients and eliminate RF rectification issues.
The 5H inductor choke is used to reduce hum in the preamp stages, the value is not especially critical. The 220nF capacitors in the power supply are fairly unique to this design, they improve the high frequency response of the amp. This is a trick that was borrowed from solid state circuitry. If you don't have any 220nF caps, a pair of 100nF caps should do the job.
The Vbias- negative voltage is derived from a half wave rectifier and a resistive ladder. The 25K bias control can be adjusted to set the idle bias level on the power tubes. Bias levels for both 6V6 and 6L6 tubes can be generated.
The guitar input stage (VT1b) is a standard class A triode amplifier. The 1K bias resistor was chosen to put this most important amplifier stage into the "sweet spot". The tone controls use the Baxandall tone stack configuration. This circuit has a much more distinct boost and cut operation when compared to many of the traditional Fender circuits. A guitar player friend had the amusing suggestion that the "Bass" and "Treble" labels should be changed to "Balls" and "Grit". The post-tone amplifier stage (VT1a) is another class A triode amplifier. Again, the 2.7K bias resistor was chosen for the best sound.
The reverb send amp (VT2) gets its input from the post-tone amplifier. The 500K linear pot is used to adjust the reverb send level from half way to full. An audo taper pot was tried here, the linear pot had a better response. Both halves of VT2 are run in parallel, the 390 ohm bias resistor was chosen for the best tube drive level. VT2 runs slightly warm, with a bit of blue glow showing. A standard Fender "Twin Reverb" reverb transformer can be used to drive the reverb, I used a Buddy MC500 transformer. The reverb return signal goes to two class A triode stages formed by VT3b and VT3a. The reverb return level is set with the 100K audio pot and mixed into the phase splitter stage (VT4) via one of two capacitors. The reverb return control does not turn the reverb signal all the way off. I wired the circuit this way since (1) I always like a bit of reverb, and (2) this allows for a slightly higher maximum reverb volume.
The balanced phase splitter circuit is formed by VT4a and VT4b. This stage combined with the power tube stage is fairly close to the Fender Vibroverb circuit. The two opposite-phase drive signals are sent to the control grids of the 6V6 power output tubes. There are a few tricks borrowed from RF power amp designs here. The 10nF capacitors from the screen grids to ground reduce potential radio frequency oscillations. These caps should not be confused with the unpopular tone-deadening control grid caps added to Post-CBS Fender Twin Reverb amps. The coil/resistor arrays on the power tube plates also help eliminate any RF oscillations. While experimenting with the circuit, some nearly dead power tubes were used, the tubes tended to oscillate when biased to a useful setting. These additions reduced that problem and improved the sound, RF superimposed on audio does not sound good.
Another unique feature of this amp is the anti-resonance feedback (ARF) loop consisting of the 300 ohm resistor, a series-wired modem transformer and a 1.32uF stack of capacitors. The amplifier has a natural resonance point of around 70 Hz, this is presumably caused by the output transformer. Audio at that frequency is amplified to about twice the level as other frequencies, resulting in an exaggerated bass response and distortion. The ARF loop is set to resonate near this 70 Hz frequency, it produces a lot of negative feedback at that frequency and flattens out the response of the amp. A resonance of around 67 Hz was found to make the flattest overall response.
A fairly heavy modem isolation transformer from a 300 baud vintage of modem was wired in series to make the ARF inductor, the capacitor was made by putting several smaller capacitors in parallel. The ARF resonance was set by by temporarily arranging the parts into a parallel LC network, an audio tone generator was fed across the LC through a 10K resistor, the peak frequency across the LC was observed with an oscilloscope while the tone generator's frequency was changed.
The 6HU6 eye tube circuit gets its control signal from the output transformer. The signal is rectified, low-passed and sent to the tube's control grid. The 10M bias resistor opens the tube's display farther during quiet operation. The 5K trimmer should be adjusted so that the eye tube display closes completely when the amp is played to maximum power.
If you want more than 18 Watts of power, it is possible to replace the 6V6 tubes with 6L6 tubes, simply re-adjust the bias control. The bias is set by putting a DC volt meter between the Imon1 terminal and ground. The Imon2 terminal can be checked to see if the power tubes are well matched. Both Imon1 and Imon2 should have similar voltages. The 6V6 tubes work well with a bias of around 0.17V (17 mA) and 6L6 tubes work well at around 0.35V (35mA). Tube bias setting is a trade-off between loudness and tube life. Generally, the bias should be set so that the tubes don't feel too warm when there is no signal going through them.
The stock Hammond amp that this project was built on was dirty, rusty and filled with mostly useless parts. A wire brush was used to scrape off the rust and dirt. Leave the transformer filament wiring alone. You will need to rewire the filament wiring to some of the 6SN7 tubes (formerly other tube types). The high voltage leads can be left connected to the 5U4 socket. The ground wires that connect all of the tube sockets should be left intact. Just about everything else can be clipped off, leave all of the transformer wires as long as possible. There were two plug-boards in the center of the amp. All of the wires between the plug-boards and the tube sockets were clipped at the tube sockets. The wires to the screw terminals were also clipped.
The tall silver electrolytic capacitor was removed. The capacitor's hole was filed out and drilled to fit the 6HU6 eye tube socket. The black electrolytic capacitor was left in, but disconnected. The plug-in boards were removed and a replaced by several multi-point solder terminals. All of the tube socket terminals were cleaned with a "Soldapullt" tool and the remaining wire bits were removed with a dental pick. This step takes a long time and may be eliminated if you feel lazy, you just have to work around the old wires. The volume pedal tower was disassembled, the contents were removed and the tower was put back in place empty. It allows the amp to sit upside down without resting on the tubes. In the next version of this amp, I plan on turning the tower into a "doghouse" full of all of the electrolytic capacitors.
Holes were drilled for the various input/output jacks, the solder terminals, the potentiometers, the reverb transformer and the filter choke. Use a drill press and start with a small pilot drill before making the larger holes. Sharp drills and cutting oil are essential for making good holes. The on/off switch, fuse and power connection were installed under the small metal cover on the back of the amp using existing holes. All of the various parts were installed between the tube sockets and the solder terminals. All new resistors and capacitors were used, old resistors tend to go up in value and old capacitors are sticky with wax. All ground connections should go to the tube socket ground wires. A star grounding system would be best for eliminating ground loop hum. Practical considerations with this chassis limited making a true star ground. The ground wire between the old 5U4 and last 6V6 tubes was chosen as the center of the ground.
Plug the amp into a guitar speaker. Plug an electric guitar into the input. Tweak the knobs for a good sound, start playing, enjoy the amazing sound. Effect pedals are not required with this amp, true "pedal weenies" are free to disagree. Remember, with all tube-amps, putting a solid state pedal between the guitar and amp loses much of the benefit of the all-important tube preamp stage with its wide dynamic range.
If you liked this project, take a look at its evolution, the Hammonator 2VRT. The 2RVT design solves the problem of what to do with that pesky unused 7 pin tube socket. By adding one 6AV6 tube and an LFO circuit board, the amp now has the much coveted pitch-shifting vibrato and tremelo effects.