Powerful tube amplifier. Powerful tube amplifier The principle of assembly of tube amplifiers

The 6P7S beam tetrode is an almost complete analogue of the 6PZS, 6L6G “sound” lamps, adapted for operation in horizontal scanning circuits of televisions.

It is characterized by improved insulation between the electrodes, a slightly larger anode current pulse, and increased electrical strength. The anode output is placed on the dome of the pump flask in the form of a metal cap (Fig. 1). At the same time, the current-voltage characteristics of the 6P7S tetrode are very close to those of the 6PZS and 6L6.

Rice. 1. Design and pinout of the 6P7S lamp.

The high quality of its sound approaches the sound of a generator tetrode of the G-807 type. The latter is noticeably superior to such generally accepted “classics” as 6PZS/6L6 and 6P27S/EL34.

When constructing the output stages of AF amplifiers, you can use the electrical modes adopted for 6PZS/6L6 or 6P27S/EL34 lamps without any problems.

• voltage on the anode Ua = 250 V, screen grid Uc2 = 250 V, cathode Ek = 14 V (automatic bias resistor Rk = 180 Ohm 2 W);
• anode current Ia0 = 72 mA, screen grid current Ie0 = 5.8 mA (quenching resistor Rc2 = 2.4 kOhm 0.25 W);
• excitation voltage on the control grid Ucl=10 V.

In this mode, the slope of the lamp is S = 5.9 mA, internal resistance R (= 32 kOhm, anode load resistance Ra = 2.5 kOhm, maximum (Kg = 10%) output power 6.5 W.

Filament voltage/current 6.3 V/900 mA, maximum permissible voltage at the anode 500 V, continuous power dissipation at the anode no more than 20 W.

Schematic diagram of UMZCH

An example of the practical implementation of a UMZCH with a single-ended output stage on a 6P7S lamp when operating in a circuit with automatic bias on control grids is shown in Fig. 2. The input signal goes to resistor R1, which acts as a gain regulator.

Rice. 2. Scheme of a homemade UMZCH with a single-ended output stage based on a 6P7S lamp.

Let us dwell on this element in more detail, since the input circuits largely determine the sound quality of the device. Let's start with the adjustment characteristics.

For volume controls, resistors with an exponential (inverse logarithmic) dependence of the resistance on the angle of rotation of the slider are generally accepted, i.e., a type “B” characteristic is required.

The design of the resistor must ensure reliable mechanical contact between the moving electrodes and the conductive element.

The explanation is very simple: the strongest degradation of the sound signal occurs in this zone, not to mention the fact that wheezing and crackling noises during the adjustment process simply get on your nerves.

For dual resistors, an important quality indicator is the imbalance of characteristics. Let's consider possible options choice.

We immediately reject the “extremist” option - the use of typically “high-end” components like Ricken Ohm - they are available to few people. Let's opt for the more common element base.

Among imported audio components of sufficient quality and not too expensive, we can recommend resistors from ALPS, Bourns, Spectroll. Among the domestic ones, volumetric composite ones such as SP4-1 or SPO work well.

Advice. Metal film and varnish film elements should not be used.

Among discrete regulators, it is possible to use domestic ones of the RP1-57E type. Those interested can try installing PTP-21 wirewound potentiometers.

The first stage of the amplifier is assembled on one half of a double “sound” triode 6Н8С (VL1.1). The 6N8C input amplifier uses both parts of this tube.

It is a standard voltage amplifier with a resistive load and a transmission coefficient of about 11. The operating mode of the VL1.1 lamp is set by the automatic bias resistor R4, the anode load is resistor R5.

The second cascade, like the first, is also typical amplifier voltage with resistive load R8 in the anode circuit. Its transmission coefficient is about 5.

Note. The only difference between the second stage and the “classical” circuit is that the automatic bias resistance R9 in the VL 1.2 cathode circuit is increased by an order of magnitude. This is caused by the need to set the correct operating mode with a large positive potential on the triode control grid.

High resistance in the cathode circuit determines a large depth of local feedback, which significantly reduces the gain alternating current. In addition, according to the concept of building High-End equipment, the presence of OOS is undesirable.

In this regard, resistor R9 is bypassed by electrolytic capacitor C2. Increased demands are placed on its quality, since this element has a significant effect on the sound of the device. Specialized audio electrolytes such as Elna-Gerafine of high quality have an equally high price and are not readily available.

Advice. You can use aluminum oxide-electrolytic capacitors such as K50-24, K50-29; a little worse than K50-35. Of two components of the same type with the same electrical characteristics, but different sizes, preference should be given to capacitors with large cases. The latter usually sound better, although in pre-amplification stages this sign is not always justified.

An attempt to bypass C2 with film or paper capacitors did not lead to a clearly defined desired effect. It is not recommended to use oxide semiconductor materials as C2.

However, we will return to the specifics of choosing capacitors installed in the cathode circuit of the lamp when studying the final stage. For alternating current, the second and final stages are connected to each other by separating C4.

This element affects the sound quality in the most radical way, so talking about the requirements for its quality deserves special attention.

Let us immediately note that an ideal component that would not spoil the sound at all simply does not exist in nature. These could include vacuum or air capacitors.

However, it is very problematic to imagine, let alone implement in practice, an amplifier with a “feeder” the size of a pair of tank batteries. Therefore, the choice of type C4 is always a compromise.

Of course, you can just note high quality specialized audiophile products from companies such as Jensen Capacitors or exotic “spill” Audio Note, and call it a day. But the prohibitive price of such components instantly transforms them into the category of equally prohibitive dreams for almost all radio amateurs.

Let's take a closer look at the actually available elements of general use domestic production. According to many developers of audio equipment, paper-oil and paper-foil products of types K40-9-5 (with 5th acceptance) are considered the best; K40-U9; K40A-2; CBG; OKBG; BM-2; BMT-2.

Slightly worse are metal paper ones like MBM, MBG, K42-.... The latter differ in that their covers are obtained by applying a thin, no more than 1 micron, layer of metallization to the paper (for comparison: the thickness of aluminum foil is 80 microns), and after rolling the package into The roll blank is impregnated with ceresin.

Due to such design and production-technological features, metal-paper capacitors, compared to paper-oil and paper-foil ones, have a reduced electrical strength, which decreases even more due to the diffusion of metallization ions into the dielectric during the aging process.

There is some “viscosity” in the sound of paper capacitors in the high frequency range. At the same time, “mica”, while providing clarity and transparency of the “top”, does not allow obtaining the necessary plasticity and relief of sound in the mid-frequency and mid-bass region, for which “paper” is so famous.

Note. After a series of experiments, the author was able to establish that the parallel connection of paper and mica capacitors, the capacitance of the latter should be 1-7% of the capacitance of the main one, allows one to combine the advantages of the sound of both types.

By selecting the ratios of capacitances, you can to some extent change the nature of sound reproduction. Practice has shown the following: for a decoupling capacitor with a capacity of more than 0.1 μF, in the case where the input resistance of the subsequent stage is at least 200 kOhm, the additional mica capacitor must have a capacitance in the range of 2-10 thousand pF.

Thus, C4 can be composed of a “wallet”, say, type K40U-9 or BMT-2 with a capacity of 0.22-0.25 μF with an operating voltage of at least 250 V and a mica capacitor, for example, KSO-5, KSO-11 , with a capacity of 3000-6800 pF with the same or greater maximum operating voltage.

Note. In the case of building a stereo version of the amplifier, the selection of capacitors that make up the “feed-through” C4 should be approached especially carefully.

First of all, from the existing stock of the same type of “wallets”, and it is desirable that they be from the same batch, using a digital device it is necessary to select two capacitors with actually the same capacitance.

The last requirement is more important than exact compliance with the nominal value indicated on the circuit diagram. Since the capacitance of the coupling capacitor is less critical than in correction circuits, C4 can lie in the range of 0.17-0.29 μF.

The need to use identical elements in both channels of the device is caused by the desire to obtain equal frequency response and phase response, the mismatch of which stereo systems are very critical to. And with single-channel sound reproduction, even very large phase distortions have practically no effect.

It would be useful to measure the coefficient of intrinsic nonlinear distortion of capacitors using the device and methodology proposed in [Lukin E. “Complex for measuring ultra-low nonlinear distortion” - “Radiohobby” No. 2/2000 p. 40]. It is useful to ensure that the capacitor's own mechanical resonance does not fall into the audio frequency range.

!!! Attention. Parts that have a “mechanical” resonance in the audio range are not suitable for audio equipment.

Having completed the selection of paper capacitors, do the same with mica capacitors. After this, they can be installed in the circuit. Of the film capacitors, the most suitable for the audio path are fluoroplastic types FT-...; K72-..., slightly worse polystyrene PM-...; BY; K70-...; K71-...; polypropylene K78-....

!!! Attention. You should not use polyethylene terephthalate (lavsan) capacitors like K73-... in the audio path, which seriously spoil the sound.

This feature allows you to select the most acceptable sound character of the device when listening to music programs of various genres and styles. For example, for hard rock music performed by groups such as ACDC, tetrode connection is most suitable.

For these genres, a slight deterioration in resolution and transparency is not very harmful, especially since it is fully compensated by the additional “drive” and aggressiveness of the sound.

The ultralinear mode is more suitable for chanson, including “Russian”, some styles of reggae and jazz, and pop music. In general, this inclusion is a kind of reasonable compromise, allowing one to obtain quite acceptable results for both not very aggressive rock and a number of classical works.

And finally, the triode connection reveals its capabilities to the greatest extent when listening to classical and some varieties of the so-called. "acoustic" music. However, these reasonings and observations should not be perceived as dogma, because who else but you knows what is best for you.

Mode switching is carried out by switches SA1.1 and SA1.2. It is best to choose a double biscuit, and a double-panel, in other words, a two-bill. This is due to the fact that an electrostatic screen must be placed between the biscuits.

Attention. Failure to comply with this requirement may lead to self-excitation.

In the stereo version of the SA1 device, you can make it as a pair of two-board switches, separate for each channel, or use one four-board switch.

Advice. It is necessary to install SA1 as close as possible to the final stage and connect it to the corresponding circuits with conductors of the minimum possible length. It is best if these are directly the terminals of resistors R12-R15.

The closest attention should be paid to the quality of the contact groups of the SA1 switch, since they can become a source of severe distortion. It is unacceptable to use products with contact groups made of phosphor bronze or copper, brass, or silver-plated metals:

• the first material has a high contact resistance;
• the rest are not suitable due to their low mechanical strength and tendency to oxidation, and in the atmosphere of large industrial cities also to the formation of various chemicals, primarily sulfur compounds, which are semiconductors.

For the first experiments, you can take components that have contact groups made of beryllium bronze or coated with a silver alloy with 40% nickel. All these materials:

• resists abrasion well;
• have good electrical characteristics;
• relatively cheap.

A more expensive option is to use switches with gold-plated contacts. “Elite” products include components that have contact groups coated with a platinum-iridium alloy or rhodium (the material used is indicated in the manufacturer’s specifications).

And finally, even the “best” material will be completely useless if the design of the product does not provide reliable mechanical contact, which should also not be forgotten.

In principle, SA1 can be assembled on the basis of a relay with sealed contacts, for which it will be necessary to organize a logical control system. Its circuit solution does not present any difficulties for an experienced radio amateur.

Briefly about the circuits associated with SA1. The first bib of the switch SA1.1 is connected to the screen grid circuit of the terminal lamp VL2. With its help, the desired circuit for constructing the output stage is selected:

• fixed contacts attached directly to the biscuit are connected to the corresponding terminals of the primary winding of transformer Tr.1 and the anode voltage source;
• a moving contact mounted on the rotating rotor of the switch is connected through resistor R15 to the second grid of the lamp VL2.

In a tetrode connection, R15 serves as a current-limiting element that prevents the danger of electrical overload of the lamp grid.

When operating in ultralinear mode, using R15 the voltages on the screen grid and the anode VL2 are equalized to some extent, and a local feedback loop of moderate depth is created, which increases the linearity of the cascade.

The second section of switch SA1.2 is connected to the cathode circuit of the same lamp. Automatic bias cathode resistors R12-R14 are connected to the fixed contacts.

Note. During the process of setting up the circuit, their resistance is selected in such a way that the anode quiescent current of the output lamp in all three inclusions lies within the range of 72-75 mA.

The schematic diagram shows the approximate values ​​of R12-R14. It is better to select them more accurately only after the new end lamps have been “fried” at idle for at least 20-30 hours.

The moving contact SA1.2 is connected to the cathode of the end lamp. The positive terminal of the electrolytic capacitor C5 is also connected to the same point.

This circuit element eliminates the occurrence of local negative feedback on alternating current due to the voltage drop across the cathode resistors. Initially, the capacitance of capacitor C5 can be taken equal to 1000 μF.

Its more accurate value depends on a number of factors, not least the characteristics of your acoustic systems. Of course, taking into account their influence in advance is a very difficult task, so we have to fine-tune the device based on the results of control auditions.

The generally accepted simplified formula for calculating the capacitance of a capacitor shunting a cathode auto-bias resistor is the following:

where Fн - lowest frequency specified operating range in Hz; Rк is the resistance of the automatic bias resistor in Ohms.

Substituting Fn = 10 Hz and Rk = 200 Ohm, we get Sk = 500-1000 μF. After increasing the C5 capacitance from 500 µF to 1000 µF, the bass becomes deeper and more voluminous, which, in principle, could have been predicted in advance.

But increasing it to 2000 µF gives a sharp negative effect. In the lower bass region, hum and a characteristic “boom” appear, and the mid-bass becomes “grainy”. In addition to everything, extremely unpleasant extraneous sounds begin to be heard in the mid-frequency register.

The quality requirements for this element have already been discussed when describing the preamplifier, but in this case there are a number of nuances.

Here the specificity is associated with the high power of the audio signal, which is developed by the final stage. The author tested small-sized, no thicker than a little finger, electrolytic capacitors Nippon, Rec and Rubycon with a capacity of 1000 μF with an operating voltage of 63 V, which sounded, firstly, differently, which is not surprising, and secondly, somehow “flat” "

Replacing them with K50-29 with the same denominations, but having several times larger geometric volume, led to a positive result. The much-desired depth and dynamics immediately appeared, and the bass itself became more concentrated, elastic and rich.

The explanation for this effect is as follows. In the final stages, an audio signal of significant power is applied to the cathode capacitor. Therefore, they begin to take their toll:

• and such a characteristic as the maximum permissible ripple voltage (it must also be taken into account when constructing pre-amplification stages)
• and permissible reactive power, i.e. the thermal processes of the component have a significant influence.

Note. All aspects of the selection of components discussed above are relevant not only for this design.

All stages of this amplifier are powered from a single anode voltage source. Interstage decoupling is made in the form of RC chains.

They include resistors R7 and R16, as well as electrolytic capacitors C1, SZ. Compared to circuits in which an audio signal operates, the requirements for the quality of filter elements are simpler. Here it is quite possible to use capacitors of types K50-20, K50-26, K50-27, K50-31, K50-32, K50-35. Components of earlier designs K50-3, K50-6, K50-7, K50-12 are also suitable.

At first glance, it does not matter where exactly a higher-quality component is installed in the power circuit, because it does not seem to interact directly with the audio signal. But this is far from true.

Let's study the influence of the final cascade on the previous ones. For a simpler understanding of what is happening, we will assume that there is no interstage isolation. In the process of signal amplification, the total anode current of the lamp is divided into two components: constant and variable. G

The generator of the latter is the lamp itself. If the anode power source had zero internal resistance, then the alternating component of the anode current of the output lamp would flow through that source completely “transparently”, without having any effect on the operation of the previous stages.

However, in practice, any power source has a certain, even small, internal resistance. Therefore, part of the alternating component of the anode current of the terminal lamp is branched into the anode circuits of the previous stages, assembled on triodes VL1.1 and VL1.2.

In this case, this part of the current passes through quenching resistors R16 and R7 (they are installed since the supply voltage of the preliminary stages is usually lower than the output stages), anode load resistances R8 and R5, separating elements R6 and C4, as well as leakage resistor R10.

The second stage of the amplifier has a similar effect on the first, and the situation here is aggravated by the presence of a damping resistor R16. Because of this, the equivalent internal resistance of the anode power supply increases significantly.

Note. The current amplitude in the anode circuit of the preliminary stage is many times less than that of the final stage.

Now let's consider the case when C1 and SZ, having good formal characteristics, have unsatisfactory “sound” properties.

Note. In such a situation, they are not only unable to effectively perform their function of short-circuiting interference to the common wire, but (what is much worse) they themselves can generate additional “dirt.”

Spreading along the power bus, all this “garbage” goes through the path described above, is amplified, and, mixing with the useful signal, is clearly not able to decorate the music program.

A very effective way to combat this effect is separate meals device components - ideally, a separate rectifier for each stage, widely used in luxury audio equipment. In more simple devices you have to make a compromise by powering all circuit nodes from one source.

Now let's draw conclusions. The more gain the entire circuit has when the OOS loop is broken, the higher quality elements should be used in the power circuit.

The first amplification stages are the most critical to the quality of components, and the output stages are less critical. Therefore, for the power decoupling elements of the first stage of the UMZCH, components of high, ideally “signal” quality should be used.

In addition, in some cases, shunting an electrolytic capacitor with a high-frequency one has a good effect, just as is done for a “feedthrough”.

Note. Particular attention should be paid to the details included in the interstage decoupling of the circuit in the case of using kenotron power supplies.

The latter have an increased intrinsic resistance compared to semiconductor ones.

A common and fairly effective means of reducing the equivalent resistance of a rectifier is to use a very large capacitance at the filter output, at least several times greater than that required to obtain a given ripple factor.

Pulse capacitors are especially good here. They are distinguished from similar products for general use by increased energy consumption, low series resistance (ESR) and the ability to deliver large pulse currents.

Of the domestically produced capacitors, the K50-23 has proven itself well in this application; the K50-17, K50-21, and K50-13 are slightly worse. You can use components from earlier designs - K50-ZF, K50I-3, ​​K50I-1.

Therefore, it is no coincidence that so much attention has been paid to highlighting the processes occurring in the power circuits of the circuit. It remains to add that the issues discussed here are relevant and valid not only for tube sound amplification technology, but also for semiconductor technology.

In the latter case, the situation is complicated due to the high currents operating here, which are tens, hundreds, and sometimes thousands of times higher than those in lamp equipment.

The remaining elements included in the power circuit of this design and shown in the circuit diagram (Fig. 2) contain switch SA2 and resistors R17, R18. Let's dwell on their purpose. Using SA2, the anode power supply circuit is broken. This is necessary in three cases:

• firstly, at the moment of initial connection of the amplifier to the network, when the cathodes of the lamps have not yet had time to warm up sufficiently. Applying full anode voltage at this moment is fraught with breakdown in the lamp and/or destruction of the cathode;
• secondly, you have to use switch SA2, and this must be done, at the moment of transition from one final stage circuit to another. Removing the anode supply sharply reduces the intensity of transient processes, which is guaranteed to protect. AC from failure during this operation;
• thirdly, this element is necessary for organizing the so-called. standby mode.

This mode boils down to the following. In the first seconds after the filament voltage is applied, the heater-cathode system experiences significant electrical and mechanical loads. The former are caused by the low resistance of the cold filament, and the latter by thermal deformations that occur during heating of the cathode.

Of course, turning the filament on and off negatively affects the longevity of the lamp. Therefore, during listening breaks of up to several hours, it is better not to turn off the amplifier.

On the other hand, keeping the apparatus fully prepared for 2-3 hours is unacceptable for economic reasons (unjustifiably increased consumption electricity and, again, a reduction in lamp life due to wear of the cathodes), and for safety reasons.

Therefore, during not very long pauses in work, only the high anode voltage is removed. Resistors R17, R18 in Standby mode form an anode voltage divider.

Its function is related to the fact that operation of the lamp with the incandescent switched on, but without current extraction, is considered a more severe mode compared to the nominal one and can lead to the so-called. "poisoning" of the cathode.

To eliminate this “scourge”, it is enough to apply a voltage of 7-15% of the nominal voltage to the lamp electrodes. There are no special specific requirements for R17 and R18 themselves.

The power supply for initial experiments can be a simple semiconductor rectifier with a capacitive filter.

It must provide an output current of at least 120 mA in the mono version of the device at a voltage of 290 V. In the future, it is advisable to assemble a power supply with a 4-fold power reserve.

Advice: To smooth out ripples, a CLC filter is best suited, and it is useful to increase the output capacitance to 1000-1500 µF per channel.

In the case of constructing a rectifier using semiconductor devices, preference should be given to high-frequency diodes with a large crystal area. The valves themselves can be bypassed with mica capacitors with a capacity of several thousand picofarads. It’s even better to assemble a kenotron rectifier. In the filament circuit, one channel of the amplifier consumes a current of about 1.5 A, although a reserve of up to 1.8-2 A, of course, would not hurt.

The power supply circuits for lamp heaters are standard, using conventional anti-background measures. Ideally, this is the use of constant stabilized voltage.

Manufacturing of transformers

The output transformer is made on the basis of a serial “networker” type TPP-286U produced by the Nikolaev (Ukraine) Transformer Plant. The products of the TPP 283—TPP 289 series have the same standard sizes, structural elements and dimensions.

All these transformers are assembled on the basis of a ShLM 25x40 magnetic core. His design characteristics: cross section of the central core - 10 cm2, average length of magnetic power line— 16 cm, window dimensions 15x45 mm, thickness art. tapes 0.35 mm. To avoid saturation of the core under the influence of constant magnetization, it is assembled with a gap of 0.25 mm.

Advice: When assembling a stereo version of the amplifier, try to find transformers from the same batch or at least with the same production date. This largely guarantees the identity of the electrical characteristics of the magnetic cores.

The transformer coil frame of the serial transformer is 39mm wide and 13mm deep.

Before you start winding using a file, you need to give it the correct geometric shape, first of all, draw out the right corners of the frame window.

Otherwise, the required amount of wire may not fit. After this, you should cut to the outer surface of the bottom those slots in the cheeks of the frame through which leads 1,2.a—2.6 and 3 pass. All that remains is to remove the burrs and slightly round the edges of the slots intended for winding leads to avoid wire breakage.

The anode winding contains 3000 turns, divided into 6 equal sections of 500 turns. Each section of winding I is made of 5 layers of 100 turns.

From the 1300th turn, tap 7 is made, which is used in the ultralinear mode and provides a switching factor of p = 0.43. The secondary winding consists of five single-layer sections of 32 turns, the total number of turns is 160.

Rice. 3. Layout of windings and electrical connections between their sections.

The layout of the windings and electrical connections between their sections is shown in Fig. 3. The specified ratio of the number of turns ensures optimal matching of the output lamp with a load of 8 ohms. IN

The choice of this option is not accidental, since most high-sensitivity acoustic systems have exactly this resistance.

Note. To obtain satisfactory sound, this amplifier must be operated with speaker sensitivity of at least 92 dB/W/m.

A characteristic feature of the coil design of this output transformer is its winding with two folded wires. Making the windings of signal transformers, especially input and inter-lamp ones, with a bundle of several wires folded together or Litz wire is not particularly new and is relatively common.

Much less often, such winding is used in powerful output transformers. The creator uses this technique in some of his models. brands Audio Note and Kondo Hiroyashi Kondo and Susumu Sakuma are the founders of the “cult” company Tamura.

In the design under consideration, the use of two parallel winding wires is explained as follows:

• on the one hand, the conductor has a directional property, so the sound quality is affected by the “polarity” of its connection;
• on the other hand, the output transformer coil is one of the very important and labor-intensive components of tube amplifiers.

Note. At the same time, it is almost impossible to immediately guess the correct direction of connecting the wire, much less be absolutely sure of it. A series of similar experiments is a lengthy, extremely labor-intensive and expensive task.

Considering that the amplitude AC voltage, operating in the anode winding of the output transformer, is commensurate with the magnitude of the anode supply, and the most critical to the direction of connection of the wires are small-signal circuits, in which direct current also operates at the same time, it was decided to use the proposal of V.I. Goryunov. This idea was published in [Goryunov V. Letter 1, “What if in parallel?” “Radio Hobby” No. 6/2000, p. 42].

An additional argument in favor of this design can be considered the fact that when using two wires, it is possible to save 7-10% of the area of ​​the core window compared to the case when one conductor is used with a cross-section equal to the total, but of a larger diameter. To perform the primary winding, 00.16 mm PETV-1 wire was selected.

Technologically, the winding of the transformer coil is carried out as follows. First, about half of the wire reel is rewound onto the empty drum, after which you can begin to work. Using this method, rather than using two ready-made bays:

• firstly, it ensures the deliberate receipt of counter-parallel inclusion;
• secondly, this guarantees the homogeneity of the chemical composition and crystal structure of the material of both conductors.

During the work, you must carefully ensure that the wires are laid in even parallel rows and in no case intersect anywhere. An example of correct coil winding is shown in Fig. 4.

Rice. 4. An example of correct coil winding.

On it, the wires that belong to one turn are highlighted with a white/black background. Between the layers of the anode winding there is insulation in the form of one layer of paper 10-15 microns thick from powerful so-called. "cosine" capacitors. The active resistance of a correctly constructed primary winding is about 220 Ohms between pins 1-14.

Note. The oil with which such paper is impregnated should not be a problem, since it is an excellent dielectric and, moreover, dissolves perfectly in paraffin and/or technical wax, without interfering with the normal progress of the “welding” of the coil.

Rice. 5. Layout of the terminals of the winding sections on a standard frame from the Chamber of Commerce and Industry: a - primary; b - secondary.

The secondary winding is also made with a double PEV-1 0.5 mm wire. Interwinding insulation is a combined three-layer insulation.

Bottom and top layer of wired cable paper 0.08 mm thick. It won't be a big deal if this paper is soaked in transformer or capacitor mineral oil. The inner layer is a fluoroplastic tape with a thickness of 50 microns.

The last section of the primary winding is insulated with two layers of fluoroplastic and one layer of electrical cardboard with a thickness of 0.3-0.4 mm. The layout of the terminals of the winding sections on a standard frame from the Chamber of Commerce and Industry is shown in Fig. 5.

The Roman numeral I indicates the initial direction of laying the wires, and II indicates the direction of rotation of the coil frame during the winding process. After winding the coil and completely assembling the entire transformer, it should be completely soaked in paraffin or technical wax.

Summary

When using an output transformer of the recommended design, the amplifier has the following characteristics: maximum output power of 4-6 W with a nonlinear distortion factor of 2.5-6%, depending on the operating mode of the final stage. The frequency range at a level of 1.5 dB is no narrower than 40 Hz - 22 kHz, regardless of the circuit of the output lamp.

The nominal sensitivity of the device is approximately 0.11 V when the final stage operates in tetrode and ultralinear modes; in the triode mode it decreases to 0.2-0.23 V. All parameters are given for the case when the circuit is not covered by the general OOS loop.

Pre-tuning an amplifier correctly assembled from known good parts does not cause any difficulties. It usually starts working right away.

It is advisable to check the DC lamp modes and adjust them if necessary. It is advisable (if you have an oscilloscope) to make sure that the circuit does not self-excite.

After this, the amplifier is allowed to “warm up” for 30-40 hours without sending a useful signal to its input. This operation can be divided into several stages; Here the total operating time is more important. During this procedure, the final formation of the components that make up the circuit occurs and should not be neglected.

This phenomenon is explained simply: the orientation of the magnetic domains of the transformer core material and the ordering of the structure of the conductors of its coil cannot occur instantly due to the presence of “memory” in metals.

After preliminary “warming up” of the device, the most interesting stage of the work begins - fine-tuning the product to the “highest limit” condition. That's why it's so detailed description requirements for parts, studying the methodology for their selection, etc. is not accidental.

The example of the proposed amplifier clearly shows that, despite the apparent simplicity of the circuit, there are many pitfalls when building audio equipment. Those interested can try to “play” with the operating modes of the triodes of the preliminary stages.

By maintaining the same value of the anode supply voltage, by changing the resistance of the resistors in the cathode and anode circuits, you can get the sound of the entire device from “terry-tube” to “flat-transistor”.

Advice.“Freshly baked” output transformers (this effect is especially pronounced in single-cycle devices) must be allowed to operate for at least 25-30 hours, only after which they begin to “wake up”.

At a certain stage of work, you will feel that each element and/or wiring has begun to “play”, you will begin to understand the influence of the materials used, you will see the dependence of the results obtained on general layout devices.

Summarizing the above, we can say: simple repetition of designs according to descriptions given in various literature provides the sound of only a certain “initial” level, which can be lower or higher. Using the full potential inherent in this or that scheme depends only on your abilities, taste and intuition.

Literature: Sukhov N. E. - The best designs of ULF and subwoofers with your own hands.

During my amateur radio career, I have assembled and tested more than a dozen different tube amplifiers - both push-pull and single-ended, including those with several connected in parallel. Most often, the good old ones were used. However, circuits with horizontal output pentodes - 6p45s, 6p44s and 6p41s - have repeatedly appeared on the Internet. I decided to stop at the latter, since despite the lower power than the 6p45, it does not have an inconvenient and dangerous pimp on top where the high-voltage anode wire is connected.Interest was further fueled by conflicting reviews on audiophile forums - from praise to complete denial of its sound parameters. As you know, it is better to collect it yourself, and then make a final conclusion. I took as a basis the circuit diagram of a single-ended amplifier by S. Sergeev, only slightly changing the ratings of the piping and the bias of the output stage.

The driver has the usual 6p14p output - here its role is secondary, pre-amplification. The output stage is 6p41s with automatic bias, which has proven itself to be excellent for its simplicity and stability of lamp operating parameters. The only difficulty - a powerful resistor - was solved simply. Since a search in boxes with 10-watt green ceramic resistors did not produce results (everything is available except the required 450-680 Ohms), I had to solder a garland of three MLT-2s on a small scarf, 180x3 = 560 Ohms.

The cathode resistor of the second channel is also assembled on it. Since the design power is 2 watts, these 6 are quite enough. You would still have to think about how to attach 2 powerful tubular resistors.

Power to the ULF comes from the mains transformer, rectifier and inductor. Transformer TSSh-170 is from a tube TV; you can also install TS-160, TS-180 here. In general, anyone capable of providing 250-300 V 0.3 A anode and 6.3 V 3 A filament voltage. Rectifier diodes - IN4007, choke - Dr-0.1. It has 1000 turns of 0.25 mm wire (this is if you don’t find a ready-made one and wind it yourself or take a network transformer to replace it).

Despite the significant voltage and current in the output stage - about 0.06 A, I took the risk of installing the relatively weak TVZ-1, which is more appropriate in 6p14p amplifiers. As it turned out later, I did the right thing :)

It would not hurt to take a metal case for our single-ended ULF, as I always did before, but I decided to take a risk here too, using an unnecessary Chinese front speaker from a 6-channel computer amplifier. This number also went with a bang :)

We will gut the acoustic system, design the future location of the radio elements and cut out the necessary windows.

Naturally, the lamps should be on top; we install them on a metal base - a sheet of two-millimeter aluminum, with cut-out round windows for the panels.

Then this sheet is covered with self-adhesive metallic color to match the main body. After gluing, the holes for the lamps are carefully cleared using a blade.

The lower part of the case is also reinforced with metal - so that the heavy network transformer does not fall out. It was also planned to install an electronic power filter on it, but in the end it was abandoned. The voltage at the power supply output is already not enough (only 260 V), so losing 20 V to the EF is wasteful.

At the back we cut out a rectangular window for a textolite panel of sockets and connectors - network, audio input and audio output to speakers.

We also cover this panel with self-adhesive tape.

Then we insert all the contact elements and screw it to the pre-cut AC window.

Large electrolytic capacitors were installed on a single aluminum base. There are 4 of these dimensional electrolytes - three for the power supply filter and oneat 300 uF 63 V, installed in the 6p41s cathode.

The case material - chipboard - turned out to be very easy to process, and electromagnetic interference from devices, which I was so afraid of, was absolutely inaudible. But this article is about assembly, configuration and testing of the circuit.

Let me make a reservation right away - this anthology in no way claims to be a manual on lamp circuit design. Schemes (including historical ones) were selected based on a combination of technical solutions, with, if possible, “highlights”. And everyone has different tastes, so don’t blame them if you didn’t guess right... In the old schemes, a number of denominations are brought to standard ones.

Skeptics claim that some schemes cannot sound at all “by definition.” Here is one diagram that gives just such an impression. But still it worked!

This diagram is taken as a starting point. The amplifier is made on the then new finger-type tubes, according to the classical design on pentodes without general OOS. The high-frequency tone control circuit has been interestingly designed, but it can actually work “upward” only with a high-quality output transformer. Since the amplifier was intended for an electric player, they saved on the power transformer. If you don’t connect anything else to it except the pickup, electrical safety is respected with some reserve. It's good to live in civilized countries - the sockets are correct. Here is the phase, here is the neutral, here is the zero. And for some reason it’s the same in all outlets. And in my apartment, for example, some of the switches were not in the phase wire, but in the neutral wire. What can we demand from sockets after this...

Pentodes in the first stage were abandoned quite quickly. Two triode stages coped with this task no worse, and the sound quality increased. Ultralinear output stage designs brought further improvement. In this connection, the screen grid is connected to the tap of the primary winding of the output transformer. The resulting local negative feedback significantly reduces the output impedance of the cascade and increases its linearity, while the gain does not decrease much. True, the ultralinear circuit was mainly used in push-pull amplifiers. Below is the circuit diagram of a typical single-ended amplifier with an ultra-linear output stage.

Fig.2

The values ​​of the parts in the tone control have been adjusted to meet modern requirements - in the original they only humped the frequency response at 5 kHz. However, HF boosting was rarely used at that time. Variants of this scheme flourished wildly in the era of economic councils, when the party and government decided to flood the country with cheap radio products. The ultralinear cascade disappeared, the tone control was simplified, and the power transformer was often abolished altogether or only an incandescent one was installed. They saved on everything, and it shows. Many people remember the sound of record players in cardboard suitcases - a good middle ground, but nothing else.

When repeating the circuit, you can abandon the tone control, and along with it eliminate the first gain stage. Then in the two-channel version the driver will only need one double triode. You can also introduce shallow OOS from the output of the amplifier into the cathode circuit of the first or second stage.

An increase in the depth of feedback in tube amplifiers is prevented by the phase shift on the coupling capacitors. To eliminate this drawback, interstage communication must be direct. And this diagram appeared:

Fig.3

Since the transconductance of the tube decreases at low anode voltage, a pentode had to be used to obtain the necessary gain. Triodes with the necessary characteristics appeared later. Another highlight of the circuit is the inclusion of a bridge tone control in the general feedback loop of the amplifier. The advantage of this solution is that with a maximum increase in the frequency response, input overload is eliminated. If the adjustment is made in preamplifier, there is a risk of such overload. Therefore, the inclusion of regulators in the OOS circuit of the power amplifier was used for a long time and in amplifiers on transistors and microcircuits. The sound quality, by the way, clearly benefits from this.

The direct heir of this circuit is the Gubin amplifier, a constant participant in Hi-End exhibitions. It can work with pentode and triode connection of output stage lamps. For complete happiness, you can also provide an ultra-linear option.

Fig.4

However, direct coupling circuits also have disadvantages. The first is the need to apply anode voltage only after the cathodes have warmed up. Otherwise high voltage on grids may damage the lamps or shorten their service life. To do this, you need to use devices to delay the supply of anode voltage, or make a rectifier on a kenotron with a large thermal inertia of the cathode. At worst, you can use a separate toggle switch for the anode voltage, but this is not very convenient.

The second drawback is the contradiction between efficiency and sound quality. When using automatic bias in the output stage, you have to either reduce the anode voltage of the driver, or accept an increase in the power dissipated by the resistor in the cathode circuit.

An interesting solution to this problem was found at http://www.svetlana.com/. You can apply a signal to the screen grid circuit of the output pentode; the constant voltage on it is usually close to the anode voltage of the driver. The automatic bias resistor may have a relatively small resistance. True, the slope of the screen grid is much lower, but the linearity is also better. In this case, the first grid is grounded, and the pentode turns into a kind of triode that works with the grid current (mode A2). But the driver will have to be reinforced with a cathode follower.

Fig.5

By the way, if the first grid of the output pentode is not directly grounded, it can be used to supply a local feedback signal, including a frequency-dependent one. And this is the path to creating a bandpass amplifier without a separate crossover.

A similar driver solution is used in another amplifier. It got here due to the parallel connection of the output tube triodes. However, there are many disadvantages, first of all - monstrous wastefulness. Of the total power consumed by the amplifier, almost a third comes from the bias circuits. It would be much more reasonable to use separate rectifiers for biasing, and in the driver - SRPP on a medium power dual triode.