I bring to the attention of interested readers the second version of a telephone tube amplifier, this time with an output transformer. If the previously described amplifier was intended to work with headphones with a resistance of 100 to 600 Ohms, then this amplifier can work with loads from 15 to 600 Ohms.
Which amplifier is more advisable to make?
The main advantage of a transformer telephone amplifier is that it can be quite easily adapted to a wide range of loads, while simultaneously providing good damping - this makes its use universal. The advantages also include a lower harmonic distortion in the main operating frequency range, achieved due to a fairly light load on the output triode (however, at the lowest and highest frequencies, the harmonic distortion increases due to some fundamental properties of the transformer). Speaking of lightness of the load, I mean that the load brought to the anode of the output lamp turns out to be very high-resistance, much greater than the output resistance of the lamp, and the load line on the output characteristic of the lamp goes at a slight angle, ensuring operation with minimal distortion (for a tube triode in this In this regard, the ideal load is one with an infinitely large resistance - this will be a horizontal line on the output characteristic). For the same reason, there is no need for a push-pull output stage and, accordingly, there is no need for a paraphase stage to ensure its operation. Thus, it would be natural to use a single-ended output stage on a triode operating in class A. In this case, the quiescent current of the lamp will be biasing for the output transformer, and its core must have a non-magnetic gap that prevents saturation of the magnetic circuit and brings it to the most linear region of the hysteresis loop. The low distortion and high damping factor of such an output stage do not require the introduction of any feedback, and this has a beneficial effect on sound quality.
Let me move on to a description of the circuit diagram of the proposed telephone amplifier. It uses only three lamps: one 6N23P-EV (6N23P) and two 6N6P (6N6P-I). Each amplifier channel (see Fig. 1) is two-stage, with galvanic coupling between the stages. There are no coupling capacitors, which significantly affect the sound, in the amplifier.
Amplifier sensitivity is 0.5 V at maximum output power. The upper limit of the bandwidth at the -3 dB level is at least 60 kHz at the lowest resistance load and about 100 kHz at the highest resistance load. It was not possible to measure the lower limit of the passband; in any case, at a frequency of 17 Hz (the lowest in my GZ-102), no decrease in amplitude was noted. Nonlinear distortion is determined primarily by the second harmonic and amounts to 2-3% at maximum output power at a frequency of 1 kHz (for the third harmonic - approximately 0.3%). At normal volume, distortion in the second harmonic is an order of magnitude lower (falls in proportion to the decrease in the signal) and is very small in the third (the amplitude of the third harmonic drops in proportion to the square of the decrease in the output voltage).
Capacitor SZ (Fig. 1) is the output element of the power supply stabilizer installed on the amplifier board (or in close proximity to it). The anode power source of this version of the telephone amplifier (Fig. 2) has a constant voltage stabilizer, which can be very useful if the stability of the supply network leaves much to be desired (in my home, for example, the mains voltage constantly fluctuates from 180 to 230 V !).
The stabilizer consists of a current source on transistor VT2, resistors R4, R5 and diodes VD8, VD9. The source supplies stabilized current to series-connected zener diodes VD2-VD7. In this case, five zener diodes are identical, type KS551A, and the type of the sixth must be selected in each specific case (due to the spread of the rated stabilization voltage of the zener diodes) to obtain a total voltage of +(300 + 10) V. The stabilized voltage from the zener diode chain through the RC filter R3, C2 is supplied to the base of the composite transistor VT1, from the emitter of which a voltage of +300 V is supplied to both channels of the amplifier to power the anode circuits. A reverse biased diode VD1 is connected between the emitter and collector of this transistor, which protects the transistor from electrical breakdown when the amplifier is turned off. The power supply rectifier consists of a VD10 diode bridge and a storage capacitor SZ. Elements R1, R2, R6, R7, C1 serve to supply a positive potential of +52 V to the lamp filament circuit, which reduces the background noise resulting from the supply of alternating current to the filaments.
When manufacturing an amplifier, the main attention should be paid to the output transformers of the left and right channels (see Fig. 3).
Magnetic cores USH 16 x 24 with plates 0.3 mm thick and coil frames can most easily be taken from unified TV output transformers TVZ-1-9. In this case, the transformers will need to be carefully disassembled by straightening the legs of the magnetic core mounting clips. Then the coils are removed from the magnetic core, the frames are freed from the wire and rewound, after which the transformers are assembled in the reverse order. TVZ-1-9 have the required gap in the magnetic core, and it just needs to be preserved during assembly. The fastening clips placed on the transformers must be tightly pressed onto the magnetic cores with a vice (but not with a hammer!). The coil of each output transformer is sectioned, this is necessary to increase the bandwidth. There are seven sections: three in the primary winding and four in the secondary. The section numbers correspond to the order in which they are wound onto the frame. Sections 1, 3, 5 and 7 belong to the secondary winding and contain 150 turns of PEV-2 wire with a diameter of 0.3 mm (two layers), wound turn to turn. Sections 2,4 and 6 belong to the primary winding: sections 2 and 6 each contain 1,500 turns (6 layers), and section 4 contains 2,000 turns (8 layers) of PEV-2 wire with a diameter of 0.08 mm. The leads of the beginning and end of each section are passed through holes in the coil frame, the primary winding being brought out on one side and the secondary winding on the other, and they are marked accordingly. Between the sections of the windings, insulation with a thickness of 0.1 mm is laid from five layers of mica paper or one layer of varnished cloth. The last section is covered with insulation twice as thick. Having wound the coil, it must be thoroughly boiled in molten paraffin, stearin or ceresin. Do not impregnate the coil with compounds! Connect the sections of the primary and secondary windings on intermediate spacer blocks glued to the transformer in accordance with the diagram (see Fig. 3). The resulting two halves of the secondary winding (II and III) can subsequently be connected either in parallel (beginning of II with beginning of III, end of II with end of III) for load resistances from 15 to 100 Ohms, or in series (end of II with beginning of III) for resistances loads from 150 to 600 Ohms. Using a switch to switch half of the secondary winding is convenient at first glance, but this will introduce unnecessary nonlinear contact resistance and can worsen the sound.
The power stabilizer transistor VT1 must be installed on a radiator with an area of about 100 cm2, isolated from the case. It is better to place a mica insulating gasket 0.05-0.1 mm thick under the transistor body, otherwise the radiator will be under a voltage of +350 V.
It would be advisable to place the left and right channels of the amplifier on one board (not necessarily printed, wall-mounted mounting can also be used) 1.5-3 mm thick, preferably from getinax. On another similar board, mount the power supply elements (except for the power transformer). When arranging the amplifier elements in the housing, try to place the transformers away from each other, especially the output ones from the power one. It is also better to rotate them relative to each other by 90° to reduce mutual magnetic interference.
General recommendations for the design of an amplifier - shielding, layout, selection of elements, including wires - have already been given in the design of a transformerless amplifier. In the author's copy of the transformer telephone amplifier, the following types of elements were used: all resistors, with the exception of R10, which was a wire resistor in a ceramic case, were South Korean carbon ( The power of one such resistor is 0.25 W. When more power dissipation is needed, multiple resistors are used. They are connected in series to increase the maximum permissible voltage applied to them.); volume control - discrete, RP-1-57; capacitors in cathodes - "Philips"; Korean rectifier storage capacitor "Samhwa"; telephone jack - "Neutrik". The installation wires were made from the cable "Recoton Road Gear OFC Speaker Wire 10GA": the cable is unraveled into strands, which are then enclosed in varnished fabric ( Under no circumstances should they be made of polyvinyl chloride!) tube, and the direction of the wire is opposite to the direction of the inscription on the cable.
Setting up the amplifier comes down to selecting the type of sixth zener diode to obtain a voltage of +300 V at the output of the stabilizer when the amplifier board is disconnected, selecting resistors R7 in the cathodes of the halves of the first lamp until the voltage channels of +(42.. 44) V are obtained at the cathodes of the output lamps and balancing the channel gain selection of resistors R1.
Before you start listening for the first time, leave the amplifier turned on for a day to allow the electrolytic capacitors to form. Before each serious listening session, allow the amplifier to warm up for about an hour. Do not forget to periodically wash all connectors with a cotton swab soaked in alcohol. The polarity of the power plug also affects the sound.
Good luck, DIYers!
S. Kunilovsky
Magazine "Audio Store" No. 2 1997
Due to the large amount of information and photographs, the article will be divided into two parts. In the first part you will learn brief information that will help orient you to the upcoming work; in the second part I will describe it and also share my impressions after listening to it.
Scheme
The basis was a classic transformerless SRPP circuit using a 6n6p radio tube, the author of which was Oleg Ivanov. The diagram was slightly changed and reworked by me. We selected our own ratings of radioelements and changed part of the power supply circuit. Depending on the choice of the anode voltage rectification method, you can use a rectifier on a kenotron or use a diode bridge.
The choice to use a diode bridge or kenotron in a rectifier is everyone’s business. The diodes have a minimal anode voltage drop, there is no such load on the transformer, and a separate filament winding is also not required. For most tube ULF circuits, 1N4007 diodes are quite suitable.
Kenotron voltage rectification is a classic method in lamp technology; many people prefer it due to aesthetic considerations and some advantages over semiconductor diodes.
Advantages of the kenotronic feeding scheme:
— Smooth supply of anode voltage, which allows you to extend the service life of the amplifier radio tubes (indirect heated kenotron);
— Almost complete absence of through and reverse current;
— Limitation of current surges at the moment of switching on due to smooth heating of the cathode and supplying voltage to the LC filter of the anode power circuit;
— Reducing the magnitude of current pulses for recharging filter capacitors.
The disadvantages of kenotron nutrition include:
— High internal resistance, due to which the anode voltage drops;
— Limited service life of the kenotron;
— To power the kenotron, an additional filament winding and the output of the midpoint of the anode winding of the power transformer are required;
-If the filter elements are incorrectly selected, the kenotron may fail due to an inrush current.
To eliminate anode voltage pulsations, a choke with an inductance of about 5 H is used (in a thorough approach, the inductance is calculated according to the ULF power supply ripples). In this circuit, a D31-5-0.14 inductor was used.
Layout
To check the functionality of the circuit, a prototype is usually made. While working with the layout, you can repeatedly add and change the location of radio components, change the layout, modify the circuit, and also solve issues that may arise when building a tube amplifier. The layout is easy to make. Layout of the circuit can be done by mounted mounting “on wires” or using mounting racks. The plywood base for the model is easy to machine, holes can be drilled well and is pliable to a file. The main thing when desoldering the circuit is to make a good ground (negative) bus.
Mounting on a breadboard differs from final mounting on a chassis. When assembling a finished tube amplifier, long wires and the placement of loose circuit elements on the chassis are not allowed.
Chassis and housing elements of a tube amplifier
The chassis must be made of iron; protective casings for transformers are also made from this material. Iron is a ferromagnetic material; its use will protect against various types of interference and eliminate the possibility of their occurrence.
You can cut the chassis yourself from sheet metal, for example, from roofing iron, use an old case from a computer system unit, or select a metal box of suitable dimensions. You should also not forget about the iron ventilation hoses (ducts).
Protective casings for transformers are made by analogy with the chassis, or they use ready-made solutions (various metal boxes, stainless steel glass jars). Ventilation holes should be made in the protective casings to remove warm air.
At the chassis design stage, you should think about the concept of the overall appearance of the finished product. The paint must be applied to the chassis before anything is bolted to it. If various decorative overlays will be used, you should think ahead and make holes for their installation.
Radio components
To prevent failure, overheating and saturation, we select a power transformer with a power reserve. Electrolytic capacitors in the anode power circuit filter are also taken with a 20% voltage margin. To reduce the influence of temperature and external atmospheric factors, we choose Soviet resistors with a small power reserve. Input-output signal sockets and capacitor housings must be isolated from the chassis. Shunt capacitors are preferably film ones.
Before installation, select radio components by measuring with a multimeter close to the nominal value, according to the diagram. It's also a good idea to check the power transformer. Often, to save copper wire, transformers were initially not wound up at factories, which led to a large no-load current in the primary winding, and this in turn affects the hum of the transformer.
Tools for work
For convenient work when building a tube amplifier, all plumbing tools are suitable. The dielectric handles of the tool must be without damage to the insulation. Much, if not almost everything, has to be modified with a file and needle file.
In order to drill holes in the metal chassis, use a cone-shaped step drill. You can also use several methods to make a large hole for the lamp socket. For example, use a compass to draw a circle of the required diameter and drill holes tightly along the line, then use a needle file to grind down the jumpers between the holes. The ideal method for drilling is to use a drill press, but most lamp makers make do with a regular drill or screwdriver.
To solder circuits, use a powerful soldering iron to tin thick wires and wires; radio components are soldered with a soldering iron of lower power so as not to overheat. A sharp utility knife or scalpel is suitable for stripping wire insulation and varnish insulation on wires (when stripping, try not to grind off the copper wire itself). A good pair of tweezers will make installation work much easier and can be used as a heat sink.
A caliper will help with accurate determination of the dimensions of parts, and will also help determine the diameters and holes for them. Use a ruler and compass to mark the holes. Having a micrometer in your amateur radio arsenal, you can easily determine the diameter of the wire.
Location of radio components on the chassis
We place the power transformer on top of the chassis - this will protect the output circuits from interference coming from the transformer. Radio tubes and audio signal input/output jacks are placed away from the power transformer. The sockets to which the audio signal will be supplied and removed, as well as the variable resistor of the volume control, are located close to each other, preferably on the front panel closer to the output lamps.
It is better to place the radio panels on the chassis so that the amplifier does not have a three-story installation of radio elements. Moderate free space in the basement of the amplifier will allow you to quickly make adjustments to the circuit and facilitate accessibility to radio elements during repairs.
Circuit wiring
Almost all lamp designs use wall mounting. With this connection method, the use of wires is minimized; all connections of radio components are made with their own terminals. Part of the circuit is soldered onto the petals of the lamp panels.
The circuit is grounded to the chassis body at only one point; the point is chosen experimentally, away from the power transformer. The negative bus is made of thick copper wire and is grounded at the same common grounding point that was chosen for grounding.
Before soldering a wire, carefully inspect the integrity of its insulation. It is not recommended to tighten the wires of the anode supply (anode circuits) and control grids into bundles, lay them parallel or close to each other.
The cross-section of the conductor wires must correspond to the power consumption of the filament current and the anode of the lamps. For example, if your lamp, according to its passport data, consumes a filament current of 600 mA, then the diameter of the wire should be selected in accordance with the maximum permissible current value. For a current of 600mA, according to the table of permissible values for wire, the diameter of the wire will have a diameter of 0.56mm. For several lamps, the total current should be summed up and a suitable wire of the required cross-section should be selected accordingly. In the same way, the permissible current value that the winding of a power transformer or inductor can withstand is determined.
To eliminate background and additional interference, the filament wires are twisted (two filament wires are twisted along their length like a “pigtail”). The background and interference are eliminated due to the fact that the alternating component of the interference currents flows through the filament conductors in antiphase directions and, accordingly, are mutually compensated.
Also, to eliminate background noise, the filament winding is grounded through an artificial midpoint using two resistors of the same resistance value. Resistors of the order of 100 Ohm-200 Ohm are sealed together with the incandescent wires onto the lamp socket. Some ends of the resistor terminals are connected to each other, the other free terminals are soldered to one and to the second filament blade of the lamp socket. The point at which the resistors are connected is grounded to the negative bus. If the transformer has a middle terminal at the filament winding and the voltage on it is equal to half the total voltage, then it is grounded without using resistors (the same middle point).
The filament wires can be made in parallel from socket to socket, rather than running separate wires to each. For the convenience of wiring the circuit, the filament wires are first soldered to the lamp sockets, and the sockets themselves are turned on the side that will ensure the most convenient installation of radio elements. The anode wires from the last electrolyte of the power supply branch with a “fork” to the lamp sockets.
A few words about headphones
The circuit used high-impedance Hungarian headphones FDS-26-600 with a coil resistance of each speaker of 600 Ohms. Headphones with lower impedance have not been tested with this amplifier; to achieve the best sound, you may have to install an output audio transformer (TVZ). Usually the TVZ is rewound under the load resistance; in our case, the load is headphones, whose resistance is ideal for this circuit.
On the Internet, on one of the forums dedicated to tube topics, I came across a table with data from an experiment carried out on an amplifier circuit (please write in the comments whose experiment was carried out and on which forum, so that the author can be indicated in the article). As I understand it, the author did not use TVZ.
Added: Site visitor Andrei pointed to the author of the experiment. The parameters of the radio tubes were taken by Ignatenko Yuri Vasilievich link to
In homemade retro designs, you can rarely see a transformerless push-pull amplifier with a phase inverter and with sequential connection of lamps, but it is this one, according to L. Kononovich, that is more perfect for use in high-end radios and radios (Radio magazine, 1959, No. 6; article "Woofer amplifiers without an output transformer"). But I came across this circuit a little later, in the early 70s, while still a schoolboy, I even assembled this amplifier. By the way, it worked well as part of a radio receiver, but with a transformer at the output. How long ago it was. “High-quality amplifier” was the title of an article from a brochure dedicated to high-quality sound reproduction. But now, as I understand, no one is interested in high-quality sound, the main thing is that it be loud. So I have to plug my ears when visiting theaters and concert halls.
 |
| Scheme from the 70s. |
When I test a tube amplifier with a purchased output transformer, I observe a strong drop in the frequency response (up to - 10 dB) in the low frequency region (20 - 100 Hz). In an attempt to level out the frequency response, I increase the depth of the negative feedback and notice that the sound becomes dull (sound dynamics and transparency are lost), although the instruments say that everything is fine. I begin to increase the number of turns of the primary winding of the transformer or turn on two primary windings of the transformers in series to increase the gain at low frequencies, but then the top of the frequency response falls off due to an increase in the leakage inductance. To everything we must add nonlinear distortions, since the magnetic permeability of the core will change from the current flowing through the winding. I’ll keep silent for now about phase distortions, which are invisible to the ear.
My attempt to build a high-quality amplifier for a loudspeaker without an output transformer has not yet been successful, since everything depends on the acoustic unit, which is more responsible for the sound quality than the amplifier itself. In my opinion, a string of low-impedance series-connected loudspeakers in an acoustic box, matched to an amplifier, most likely will not sound. Even in my youth, Kononovich’s circuit worked for me with an output transformer, since it was designed for 5GD16 loudspeakers with a resistance of about 400 Ohms, which were in short supply at that time. And an attempt to increase the number of radio tubes to ensure operation of a low-impedance load turns the amplifier design into an electric stove.
But the construction of a high-quality stereo transformerless headphone amplifier was crowned with success. I assembled an amplifier using combined triodes in order to reduce the total number of tubes.
Tube high quality headphone amplifier.
Now that's what I would call this scheme.
 |
| Photo 3. Layout of the site. |
When I began to forget what constitutes high-quality sound, I assembled a circuit a simple single-tube monophonic headphone amplifier.
This amplifier worked on series-connected dynamic headphone heads solely because of its relatively high output impedance of 66 Ohms, and therefore was not suitable for stereophonic reception using household, relatively inexpensive headphones with a complex impedance of dynamic heads of 30 - 33 Ohms. A push-pull tube stage based on 6N3P triodes (the more widely used 6N2P radio tubes will not work in this circuit, perhaps due to their higher internal resistance), covered by feedback, already provides an output impedance of the order of 33 - 40 Ohms. True, through additional circuit solutions, I have so far reached the amplifier output impedance of about 25 Ohms. I could have gone further, but I stopped in time, so the process could throw me back a whole century (in 1921, experiments were carried out on radio communications between Kazan and the cities of the Volga region), at that time, to increase the transmitter power, up to 87 lamps were used in parallel. Despite the not-quite-cold weather outside, the heating was working at full capacity.
In the circuit in Figure 1, the first lamp operates in amplification mode. The second half of the double triode is a bass reflex. Two signals of the same amplitude, but shifted relative to each other by 180 degrees, are removed from the cathode and anode of this lamp. The gain of this stage is less than 1. The output is a push-pull sequential stage. The DC voltage is divided in half for each lamp. For this circuit, the 6N3P lamp, having a low anode voltage (100 volts), is well suited. Chain R C – negative feedback. As the resistor value decreases, the feedback depth increases, which in turn reduces nonlinear distortion and the output resistance of the stage, while at the same time reducing the gain. So, if the resistor in the feedback circuit is 10 kOhm, then the output resistance of the cascade is 33 Ohm. If R ooc = 18 kOhm, then R out = 40 Ohm.
However, as an option, you can reduce the output impedance of the amplifier without greatly increasing the depth of negative feedback by paralleling the lamps in the output stage.
I also tried this way of switching on the lamps in the final stage, slightly changing the circuit. Most noticeable are two parallel inclusions. When three lamps are connected in parallel, the effect on reducing resistance is negligible. Therefore, I did not build a chain of lamps to achieve loudspeaker output. It's a matter of choosing from three options.
I really wanted to listen to the vaunted warm tube sound. But to be honest, buying a device is overwhelming. Therefore, having found a circuit for a simple tube headphone amplifier, weighed my capabilities and calculated the costs, I realized that it couldn’t be better to start with.
What I present now is the second version of the amplifier I made. The first one was assembled using almost hinged installation. On it I played with the power supply for a long time and fruitfully. Due to the fact that the assembled power filter proposed in the original circuit could not suppress fifty hertz hum. Which disappeared only after installing the “electronic throttle”.
There are practically no differences from the diagram in the link above. But I lowered the anode voltage from 270 to 200 volts and increased the rating of capacitance C3 from 1 to 2.2 microfarads. Since I have a device assembled according to an authentic circuit, we can say that the changes I made did not affect the sound quality. At least to my ears.
Since I used, there is no need to talk about using 6N1P and 6N6P lamps due to the very high filament current (0.6...0.7 A per lamp). But by lowering the anode voltage, it became possible to use smaller electrolytes.
Due to the fact that to use 6N3P lamps you will have to make a different printed circuit board layout, only 6N2P and 6N23P remain. These lamps are interchangeable, but there is a catch. It is not possible to simply replace lamps one with another because... after replacing any of the lamps, the amplifier must be tuned by selecting the resistor Rk and achieving half the supply voltage at the anodes of the lower lamps. As for the rest, yes. First, you can make an amplifier using 6N2P tubes, as they are cheaper, and then reconfigure it for 6N23P and compare the sound.
A little about the details. All resistors must have a power of at least 0.25 W. Capacitors C3 and C4 must be designed for the supply voltage (I set them a little lower, at 160 V, not the ones in the box), but the capacitors in the cathode circuit C1 and C2 are voltage is 6...10 volts, but since they directly affect the sound, increased demands must be made on their quality and the higher the value of capacitor C1, the better.
A little about the case. After the problem with the anode voltage in lamp technology, there is another problem, this is the problem of the housing. It is almost impossible to choose a ready-made one for your needs, and making a body with your own hands is not so easy. Therefore, here I used the old proven method of making cases from foil PCB. And of course, where would we be without wooden case parts in lamp technology? :-) The overall dimensions of the required box are approximately 160X170X50 mm.
Because the lamps get very hot; I made special holes for them in the top cover, but after a short operation it turned out that they were sorely lacking and I had to drill out both the top cover and make holes in the bottom part of the case to enhance air convection.
In this simple way, all the giblets of the amplifier fit on the racks. After improving the convection of air to cool the lamps, the case does heat up, but not so much that you can’t easily touch it with your hand.
And finally, about my personal listening experience. Deep detail, soft, undistorted bass and that same tube sound. I’m not lying about “the one.” The difference in playing the same composition on a lamp and on transistors, through headphones, is very different, not in favor of transistors (I have had a Pioneer A305R since the end of the last century), and if you also turn off the timbre block, then in general everything is very sad. Yes, it is also necessary to add that in order to get proper listening pleasure, I had to eventually acquire high-impedance Sennheiser HD 280-13 300 Ohm headphones. Before that, there were inexpensive HD 180 and banana CX 215. But the bass on them was not expressive, and sometimes grunted when the music was turned up louder.
An unexpected sequel.
The fact is that one day a friend came to me to drink beer, listened to the device and said that he would not go home without it. I had to give him the device for a small reward. But since I can no longer imagine myself at a computer without this amplifier, I had to make another one. Board dimensions 95x95. Since I selected the case after I had made the board, I was unable to implement the idea of a side power connector, so I had to insert it in place of the tulips and move them to the edge. But it didn’t turn out bad either.
The case took a standard duralumin 120x95x35, screwed a transformer on top, and placed an amplifier and an anode power filter on the board.
Well, for greater importance, I covered the transformer with a small tin of green peas. I didn't even paint it. It is a little high, but the diameter is perfect.
Many radio amateurs who want to join tube sound, stops the presence in the design output transformer. The quality of this element largely determines the final sound quality of the entire amplifier. Ready-made industrial designs are very expensive, and it is not always possible to select a transformer for a specific lamp or operating mode. And to produce high-quality output transformer not every radio amateur can do it at home.
Therefore, almost together with the advent of the lamp, radio engineers began searching for ways to eliminate the output transformer from the circuit. To reduce the output resistance of the amplifier, cathode followers, parallel connection of several lamps, bridge and push-pull circuits were used. This topology is called OTL(without output transformer).
Similar OTL devices were even produced on an industrial scale, but, alas, only a few of them had decent sound. Therefore, interest in such schemes has noticeably faded recently.
However, keeping in mind that resistance(audiophile) headphones most often lies in the range of 32-600 Ohms, which, compared to the resistance of speaker systems of 4-8 Ohms, is several times, or even hundreds of times higher, radio amateurs do not give up attempts to implement OTL topology in low-power headphone amplifiers. The most common variations are on the theme of SRPP cascades and parallel connection of lamps. But there are other options.
One of the options was proposed by Morgan Jones in the not too distant 90s. He based his circuit on the EarMax amplifier circuit, which was produced by a well-known company and cost about $1000.
By changing some ratings and types of tubes used (the original was a 6N1P tube), Jones increased the load capacity of the amplifier and ensured relatively high-quality operation of the circuit for 32-ohm headphones. The amplifier circuit is shown in the figure:
Click to enlarge
The input stage is normal - with a resistive load. The quiescent current is 3mA. As R5, you can use two 1.5 kOhm resistors connected in parallel. The output stage for reducing the output impedance of the amplifier without using a general negative feedback is built according to the circuit White push-pull cathode follower. Its quiescent current is selected to be 10mA, while its output impedance is only 10 ohms. The total gain of the entire amplifier is 22.
As measurements have shown, the amplifier copes well with headphones with a resistance of 300 Ohms. Moreover, the lower the load resistance, the shorter the spectrum of harmonics at the amplifier output.
For a load of 32 Ohms, the amplifier at high powers (and the maximum was 32 mW) had an imbalance of the positive and negative half-waves of the signal.
Two enthusiasts, John Broski and Alex Cavalli, began studying this strange behavior of the amplifier under a low-impedance load (after all, the output impedance of the amplifier was quite low). As a result of their research, the resistor values (and, as a consequence, the operating modes of the lamps) of the output stage were changed. This made it possible to optimize the current distribution between the arms of the output stage:
Click to enlarge
As a result of seemingly minor changes in the values of some resistors, it was possible to increase the output power of the amplifier by 6 times and completely eliminate the asymmetry of the half-waves of the signal at a low-impedance load. The “tail” of harmonics was also significantly reduced (to the 4th, versus the 7th in the original version) and the frequency range in the low-frequency region expanded.
But you have to pay for everything. Due to the changes made, the overall gain decreased to 19, and the output impedance of the amplifier increased to 53 Ohms. Nevertheless, the amplifier coped well with headphones with a resistance of 32 Ohms.
For those who are not afraid general negative feedback, Alex Cavalli proposed a variant of the scheme:
Click to enlarge
Here resistors R3 and R10 form a general OOS circuit. Its depth was chosen as a compromise: on the one hand, to reduce the output impedance of the amplifier to less than 32 Ohms (with the indicated ratings it turns out to be 20 Ohms), on the other hand, the total gain should be enough to drive the signal from the output of a regular CD player. Alex advises experimenting with the values of these resistors to find the optimal sound. for specific headphones.
Not everyone loves the lamp 6N23P, so there were enthusiasts who modified the original version of the EarMax amplifier based on tubes 6N1P:
Click to enlarge
Here, the resistor values have also been changed to increase the load capacity of the amplifier; in addition, to optimize the modes, it was necessary to increase the amplifier supply voltage.
When repeating this option, I draw your attention to the fact that the 6N1P lamp in the filament circuit consumes almost twice as much current as the 6N23P. This should be taken into account when choosing a transformer and manufacturing a power supply. Considering that the internal resistance of the 6N1P lamp (11 kOhm) is significantly higher than the internal resistance of the 6N23P lamp (about 2.5 kOhm), the latter version of the amplifier is recommended for use with headphones with a resistance of 100 Ohms.
Due to its simplicity, all of the above structures can be easily assembled by surface mounting without the use of printed circuit boards.
The only capacitor in the signal path is C4. Here you should use a capacitor of the highest audiophile quality available to you! You should not use capacitors with an operating voltage lower than those indicated in the diagram, since before the lamps are completely warmed up, the voltage on them can reach supply voltage amplifier
The power supply diagram is not given, since here any radio amateur can unleash the full extent of his audiophile depravity. In the original, the amplifier was powered by a conventional bridge rectifier with a C-L-C filter. No circuits were provided to delay the supply of anode voltage until the lamps warmed up. In later amplifier models, constant voltage power was introduced to reduce the background level.
By repeating the design, it is possible to power the amplifier from kenotron rectifier, which will automatically protect the cathodes of the lamps, or, given the small current consumption of the circuit, from parametric stabilizer.
When listening to the amplifier, even the original (unfinished) version, everyone noted the clean and smooth sound without excessive tube softness, with clear elaboration of the high-frequency region and stunning “lows”, usually not typical of tube amplifiers. Despite the seemingly small output power, the volume level was more than sufficient, and sometimes exceeded reasonable limits.
Considering the availability of parts and the simplicity of the design, assembling this amplifier can be a good activity for a frosty winter evening, a chance to experience tube sound and an easy way to take the sound of your headphones to a completely different level.
When assembling and adjusting the circuit, remember that lamp structures contain high voltages that are life-threatening! Be careful and careful. Follow safety precautions when working with high voltage. Remember to discharge the capacitors before working inside the amplifier.
Happy creativity!
Editor-in-Chief of RadioGazeta.
