This article is an attempt to translate the original published on 06/11/2011 in Spanish. I can well understand a text written in English. My language study is limited to three years of elementary school English. These three years are not enough to learn to think in English so that when I attempt to write something in English, I think in Spanish. That leads to the result is something like: "Me Tarzan, you Jane." Unfortunately, neither the automatic translator thinks. The machine and I have worked together and the result is what follows.
Please, if the red key trigger nuclear missiles is within your reach, try to understand that whoever wrote this is a good guy who just wanted to talk a little about an audio amplifier. I hope not to trigger World War III.
When I was a child the transistor existed but wasn't popular. It was initially booked to a few applications, mostly military. In the fifties, the kingdom of thermionic valve or vacumm tube still continued ruling.
With regard to audio equipment were two different bands: the people who was fanatic of the triode output and the others that did the same, but with pentode output. First ones were always the audio purists and they didn't take into account the power and performance. For them, if the sound was better, any expenditure was good. On the opposite side of the street stood the latter group, who praised the best use of energy and the availability of more usable power, though, at these era, 1W RMS audio power was adequate for a home party, 8 W at most. Couples danced and talked. Fifteen or twenty watts were a musical orgy!
With monophony there were only push-pull (Class A) equipments or balanced outputs (classes AB and B). Class A single outputs were more in radios and televisions.
When the stereo sound appeared was no longer economical to have two push-pull amplifiers up to 10-15 W RMS. Many manufactures of consoles (I mean: a cabinet for a radio, television set, or phonograph, that includes the speakers in a single furniture designed to stand on the floor) returned to the single power output and there was even some who used a monofonic bass center channel and two stereo single outputs less powerful for midrange and treble. Then the transistor began to advance and the commercial world forgot the old tubes. Until one day, when the nostalgics and some explorers returned to the past, but with many memories of the future. They return with better understanding of some phenomena, improved components and maturity to reject fashion and choose what fits.
Today almost no one wants negative feedback, once considered a refinement. Some purists refuse to equipment with push-pull triode and defend strictly triode single output. It is also noticed that the transformer is simple and easy to achieve in a single class A there is nothing to balance. But there are other reasons: a push-pull amplifier in strict Class A even harmonics are canceled, fully in theory and in practice almost entirely. The musicians argue much the product of a power amplifier containing second armonics that aren't in the original sound. It's very nice to hear two notes simultaneously an octave apart. As for the timbre of an instrument, the musician is the owner of do what he wants, he is the artist. As for playing, especially in the reproduction of quality should not be added or removed anything from that is playing.
However, there are music lovers who appreciate the single amplifier second harmonic content that adds to the original program. There is no law or moral principle that opposes and can do as they please. I am personally in favor of introducing as little distortion as possible, I want to hear the original program, as left the instrument and the performer who gives life.
Memories of the future help with the understanding that the passive components are not at all, an innocent resistance or a capacitor can significantly change the sound of an equipment. What never suffered changes was always accepted that any amplifier is better than your output transformer. With this in mind, let's see the circuit before further commenting on other construction aspects.
We use three valves 6CG7. These double triodes are identical or very similar, not to offend the purists, that the oldest 6SN7. Are valves that were designed for television, especially for the oscillators of the vertical and horizontal amplifiers. They are inexpensive and little sought for audio. However with a good construction microphonic are as little as 5692 because the smaller size helps a lot in this regard.
If you plan a cheap amp, the only big expense that should not stop doing is an
excellent output transformer. After you choose what output pair to put. This circuit is suitable for driving a push-pull with a 6AS7/6080 or other similar valves. It can also handle tetrodes, beam tetrodes, or even pentodes. The valve I mentioned specifically is a double triode with important plate dissipation (about 20W), low mu, low plate resistance. It were used in serie regulated sources, in the function of pass serie regulators. Today there are several commercial equipments with these valves and others like as the 6336-A. Nothing is against that you use 300B tubes, but the cost shoots up.
Building the power supply with silicon diodes will result in savings of: power
transformer, rectifier socket and high vacuum rectifier. If you want a guitar amp, so good, the sound is harder, perhaps that's what you want. If you want a playback amp, you notice that the sound changes. For audio is essential to use high vacuum diodes. However, don't worry if you can't. With an appropriate output transformer, even with silicon diodes will enjoy a very good sound, not the best but good.The same goes for other components. If the output transformer is excellent will work reasonably well with any passive component that use.
You'll notice I drew two lines of mass. Thus, the midpoint of the high voltage winding of the power transformer is connected to the chassis and then go two thick wires of pure non recovered electrolytic copper, one for each voltage amp stage.They are soldered into the input connector, on the flap that goes to mass with a very good chassis welding only at that point. The input stage has all the components soldered to a single point in the first bar, including the decoupling and filtering capacitor of the source that isn't drawn. In this section is best if you can remove the bar and welded directly all in the input connector. If not possible, then the thick bar and a single spot weld on it. This bar ends at the connection point and not going to chassis except in the input terminal. With the second bar is the same: both the constant current source transistor, such as cathode resistors for the third 6CG7 and the corresponding filter will be on the same point of the bar. Why two bars? Here are divided oppinions. Some believe that one bar is enough. Just that all components from one stage to go to the same point in it. Have come to extend the single bar to the mass connection of the output stage. Others believe that in a single bar, across the piece of mass bar of the lower signal stage flows a current equal to the sum of all individual streams of the sections that make up the equipment. This would create electronic noise in the ground circuit of the stage most vulnerable to noise.
Separating the paths, we reduce electronic noise on the input. But there's also a partition noise in the soldering point because of the different paths of the currents.
The first and third valves 6CG7 have their cathodes potential with respect to the filament that exceed the maximum allowed value of 200V. The cheapest way to eliminate this risk is by putting the AC power of the filaments without reference to ground, connecting the two terminals on the transformer directly to the respective terminals of the sockets without any grounding. If need be referenced, or opt for a supply voltage of 6.3 V DC, the midpoint of the wound or the mass of the source of 6.3 V must be connected to a potential of 150 volts.
The calibration is done by connecting the amplifier and adjusting the cathode of the second valve 6CG7 to 8 volts by potentiometer of 500 ohms, which can be a turn cermet or more accurate, if possible to make the expense (there are 10 turns presets). After about fifteen or twenty minutes of warm places a signal generator to input a 1 kHz sine wave. If you have a spectrum analyzer, regulates the 500 ohms to obtain a minimum of harmonic distortion. If isn't possible, adjust the voltage to 7.9 V with a digital voltmeter. The 22 K potentiometer is used to dynamically balance the two outputs. With an audio voltmeter measure AC voltages are equal in the 68 K resistors in the cathodes of the two cathode followers. If you do not have the appropriate instrument, we settle for having equal DC voltages on both plates (resistors of 120K), or both cathodes (68K resistors).
The two output capacitors, which coupled with the grid of the outlet valves, can have values from 100 nanofarad to 1 microfarad depending on the configuration of the output stage.
The 1 microfarad capacitor, which connects the SRPP input to the next stage, figure connected to the cathode triode which makes of load. You can test the connection to the bottom triode's plate, sometimes lowers harmonic distortion.
IMPROVING PROJECT (If your pocket is full of money)
If it's used for audio, I recommend triode and a high-quality output transformer -brands: Amplimo, Tango, Tamura or Audio Note- suited to the selected valves. This circuit can handle a pair in class A with a grid to cathode voltage of -120 V at most.
For the ultimate in quality, first thing to think about is placing porcelain sockets and change valves 6CG7 by 6SN7 General Electric tubes of the fifties. Many people agree that gives the best sound.
Isn't a bad idea to make the chassis of copper and weld with eutectic solder (with 4% silver). Some advocate using all the wiring with pure silver, except the filament wiring. My pocket was never full enough to wiring in silver, I have no experience with silver wiring.
The best coupling capacitors are the Audio Note silver foil - paper in oil. To expensive but proud. Wima is a good choice if you do not go that high. Also the brand behave Jensen, Houland or MIT.
All resistors in the grid and cathode circuits should be of pure tantalum oxide, also brand Audio Note. Rodenstein is a good choice.
The plate resistors, as the two 120 K 2 W and the 1K8, it would be were Allen Bradley carbon composition. Not been manufactured since 1997. Some are of World War II and would have to be measured very well. They behave audio excellence and support better than any other power requirements. They have the defect of being unstable in their values because they are hygroscopic. Carefully remove the cover; there are those who warm to 80 ° C for one day and then measured. (The mechanical scraping is very delicate and if it hurts the body carbon resistor becomes noisy. In Argentina it's possible to try to dilute the protective coating by dipping resistor for three or four days in diluent-lacquer Petrilac MELACRIL. This proved very effective solvent to dissolve polyester resin of sealed PCB or transformers, after the time mentioned, and clean out unscathed. I never tried diluting coverage of Allen Bradley resistor, ignore the material that is made) After losing moisture regain their original value. Then it should be sealed with a layer airtight and impervious to moisture. If you can find many of them, it is faster to measure and choose these ones that serve, then sealed with a material to prevent future changes.
The best filter capacitor or bypass can be reached using today is the Black Gate. It 's based on a theory designed to achieve a capacitor similar to an electrolytic capacitor, but with very low internal resistance, almost zero response time and a number of advantages. In the dielectric are carbon particles. It has three disadvantages: very high price; were discontinued and we have to fight for them; should be used every day, because if you spend more than 24 hours without use, you have to recharge for 24 hours. The sound is excellent with what is available today, but if you hear nothing but the weekends, choose Elna Cerafine (instead of coal used ceramic particles), reaching 98% of what makes a Black Gate at a lower price or -possibly- by some oil-paper capacitor of very good quality.
Finally, with so much care in the components, I must say that more than one authority on the subject object to the constant current source with a transistor. For them the motto is: no silicon. Another objection would come from the SRPP input stage, many prefer classic two-stage cascade.
Anyway, it's a circuit to test and learn. Get a good output transformer, a suitable pair of power tubes and do experiments. One more warning: is addictive.