Single-ended tube amplifier. Single-ended tube amplifier Push-pull HF amplifier 6p45s

The power amplifier for two 6p45s was designed for everyday on-air operation. In addition, it can be recommended for repetition by beginning shortwave radio amateurs. The amplifier uses 6P45S tubes, which are affordable, have good linearity and a huge operating life (5000 hours). They can be used even after many years of work in the line scan unit of televisions. The power amplifier for two 6p45s has output power 200 W on all HF bands with an input power of 30 W and is assembled in a housing available to the author with overall dimensions of 193x393x270 mm.

Often, beginning radio amateurs (and not only) purchase an inexpensive imported transceiver that does not have a built-in antenna tuner (automatic matching device). Based on this, a power amplifier based on two 6p45s uses a circuit for connecting lamps with a common cathode, in which the excitation voltage is supplied to the control grid. The amplifier allows you to “unload” the transceiver by decoupling it from the antenna. In fact, as they say now, it is an active antenna tuner. Among other things, the transceiver is protected from static electricity charges at the antenna terminals and other troubles associated with this (for example, a break in the antenna or a short circuit in it). In the event of a breakdown of the lamps (an incident unlikely when using 6P45S lamps), such a circuit solution is much safer for the transceiver than a circuit with common grids.

The schematic diagram of a power amplifier using two 6p45s is shown in the figure. The input signal through the RF connector XW1 and relay contacts K1.1 is supplied to two low-pass filters with a cutoff frequency of 32 MHz, which are made in the form of P-circuits, the input and output resistances of which are 100 Ohms. At the amplifier input, the P-circuits are connected in parallel, therefore, the input impedance is 50 Ohms. The circuit does not contain capacitors with a capacity of about 60 pF included in the low-pass filter. In reality, these capacitors are formed by mounting and other capacitances. The input capacitance of the low-pass filter is formed by the capacitance of the coaxial cable, through which the output of the transceiver is connected to the input of the amplifier, as well as the mounting capacitance and the capacitance of the K1.1 relay contacts, which totals 120 pF. The linear capacitance of the RK50-3-13 coaxial cable is 110 pF/m, therefore, the length of the cable connecting the transceiver to the power amplifier on two 6p45s should be about 90 cm. More precisely, the cable length is selected according to the minimum SWR when tuning the power amplifier on two 6p45s.

The output capacitance of each low-pass filter includes the lamp input capacitance (55 pF) and the mounting capacitance (approximately 5 pF), for a total of 60 pF. The use of low-pass filters is useful for several reasons. Firstly, to reduce the level of higher harmonics, and secondly, to compensate for the capacitance of the coaxial cable connecting the amplifier to the transceiver, the length of which should not exceed 0.1 of the shortest wavelength of the amplified signal, i.e. 1 m. When this condition is met, the cable represents a capacitance and does not transform the input impedance of the amplifier. Thirdly, the low-pass filter compensates for the input capacitance of the lamp, as a result of which the input impedance of the amplifier becomes frequency-independent, and the amplitude of the exciting signal does not decrease with increasing frequency. It is obvious that the use of a low-pass filter is justified.

The low-pass filter outputs are loaded with resistors (R7 and R10, respectively). From these resistors, through capacitors C7 and C9, alternating HF voltage is supplied to the control grids of lamps VL1 and VL2. The gain of each tube is 6.7 times the power (approximately 8.2 dB). This, of course, is not much and is comparable to the gain when operating lamps with common grids, but it is justified by the very stable operation of the amplifier. In addition, its input part is simplified. The task of filtering spurious oscillations at the amplifier input was not set, because The output circuits of the transceiver cope with this, although some filtering of higher harmonics, of course, occurs.

This construction of a power amplifier using two 6p45s has another advantage, namely that the throughput capacitances of the lamps are not summed up, which happens when the lamps are connected in parallel. Consequently, the stability of the amplifier is further increased.

The use of a switchable anode choke in combination with other measures taken made it possible to obtain the same output power (200 W) on all HF bands. The DrZ choke and capacitor C12 serve to protect the power supply in case of possible self-excitation of the VHF amplifier. An RF voltmeter is installed at the output of the P-circuit for ease of adjustment. In transmission mode, when the pedal is pressed, the electronic key, made on transistors VT1 and VT2. Transistor VT2 opens, and relay K1 - KZ, included in its collector circuit, is activated. The contacts of relay K3.1 (Fig. 2) are switched, and the power supply voltage is supplied to the lamp screen grids from a voltage stabilizer made on transistor VT1. A parallel type stabilizer, which provides protection for lamps during the dynatron effect of the anode or screen grid, despite its simplicity, works well. Resistor R9, which is connected to the output of the stabilizer, facilitates the thermal regime of transistor VT1 in receive mode.

Of course, it would be possible to use a parallel-series voltage stabilizer, which is more economical than a parallel one, but also much more complicated, because contains actually two stabilizers. In the opinion of the author, such constructive complication with not very significant savings is inappropriate. The operation of the stabilizer can be improved by using, instead of ballast resistor R5, a light bulb with the appropriate voltage and current, which will play the role of a barter, increasing the stabilization coefficient. In fact, a parallel voltage regulator is simply a powerful high-quality zener diode, the current through which (62 - 70 mA) is set using ballast resistor R5.

The power transformer Tr1 of the power supply is connected to the network smoothly through the limiting resistor R1, which, some time after switching on, is short-circuited by the contacts of the toggle switch B1 with the middle neutral position. Such a simple connection circuit significantly extends the life of lamps and power transformers, and the entire amplifier as a whole. It is known that the filament of a cold lamp has a resistance ten times less than a heated filament. Therefore, the inrush current of the lamp filament is ten times the rated filament current. A large surge of current at the moment the voltage is applied overloads the filament, destroys its structure and reduces the life of the lamp. Therefore the application smooth start more than justified.

At the input of the power transformer, a network filter is installed, made on two winding inductors Dr1 and capacitors C1 and C2. The anode power supply has protection against excess anode current. Resistor R11 (Fig.) limits the current during breakdown or short circuit of the output of the anode voltage source at 600/10 = 60 (A). The FR207 type diodes used in the power supply (Fig.) will withstand this current pulse and will not fail. The anode voltage source is composed of two, 300 V each, connected in series, which improves the dynamic characteristics of the power source.

On the rear wall of the case there is a power amplifier with two 6P45S, opposite the 6P45S lamps, an M1 fan for 24 V is installed, working for exhaust. It turns on when the power amplifier operates for a long time using toggle switch B2. To reduce acoustic noise, the fan is powered by a voltage of 20 V. The fan is secured through a soft felt pad. In addition, the screws that secure it to the back wall are equipped with polyethylene tubes and two washers each made of felt and textolite. Thus, the fan housing is completely isolated from the metal surface. In the case of using a fan with a plastic casing, this is desirable, but if the casing is metal, then such fastening is mandatory. 6P45S lamps are installed on a plate made of double-sided fiberglass, for which a 125x65 mm cutout is made in the chassis. All voltages are supplied to the lamps through pass-through capacitors (except, of course, the excitation voltage, which is supplied coaxial cable diameter about 4.5 mm with fluoroplastic insulation). Relay K1 is located near the input connector XW1 (Fig.).

All parts related to the high-frequency unit are interconnected by 20 mm wide busbars, which are cut from tinned tin from instant coffee cans. The cathodes of the lamps, the current collectors of the variable capacitors included in the P-circuit, the antenna connector, the “ground” terminal, and the blocking capacitors in the anode choke circuit are connected to the busbars. Particular care should be taken to connect the current collectors of the KPIs, the grounded terminals of additional capacitors connected to them, and the cathodes of the lamps to the bus. Considering that a large loop current flows between the grounding points of the KPI and the cathodes of the lamps, other parts going to the housing should not be grounded between them. Due to the large total output capacitance of two 6P45S lamps (about 40 pF), a significant part of the loop current (about half at 28 MHz, much less in the low frequency ranges) flows through the bus section between the anode KPI and the cathodes of the lamps.

Inductors L1 and L2 of the input low-pass filters each contain 12 turns of PEV-2 1.2 mm wire. Winding diameter - 10 mm, length - 20 mm. The winding is frameless. Both low-pass filters are enclosed in one common screen and located under the chassis, near the lamp panels.

All inductors of the P-circuit are wound in one direction, the taps are counted from the “hot” end. Coil L3 is frameless (diameter - 26 mm), wound with silver-plated wire 03 mm on a mandrel, winding length - 30 mm, number of turns - 4. Anode KPI, which uses one section from an old-style two-section variable capacitor with a gap between the plates not less than 0.5 mm, soldered to the tap from one turn of coil L3. This connection reduces the influence of the initial capacitance of the KPI on the resonant frequency of the P-circuit in the range of 28 MHz.

Coil L4 is frameless (diameter - 40 mm), has 4.5 turns of PEV-2 wire 02 mm, tap - from the 3rd turn, winding length - 27 mm. Coil L5 is wound on a 45 mm frame and contains 5+5 turns, the wire diameter is 1.5 and 1.0 mm, respectively. Winding pitch - 5 mm, winding length - 50 mm. The anode choke is wound on a fluoroplastic rod with a diameter of 18 mm, the winding length is 90 mm, the wire is 0.4 mm, the tap is from the middle.

Power transformer Tr1 is made on a magnetic core ШЛ32х40. Its skein data is given in the table.


The line filter choke is designed somewhat unusually. It is wound with a double network wire from a burnt-out electric soldering iron on a 08 mm ferrite rod from the magnetic antenna of the radio. The length of the rod is at least 120 mm. Before winding, the ferrite rod is wrapped in several layers of varnished cloth. At first, the inductor is wound as usual, but when the winding reaches the middle of the rod, the winding direction is reversed. To do this, the wire is bent in the middle of the throttle, the loop is secured with a strong nylon or silk thread. Then, if the winding was carried out clockwise, after the middle of the length of the rod it is wound counterclockwise. The inductance of the inductor remains quite large, but magnetization of the ferrite rod and its saturation due to a possible insufficient cross-section are completely eliminated. Consequently, all nonlinear effects and changes in the inductor inductance when the load on the line filter changes are completely eliminated.

The power amplifier on two 6p45s operates in class B. The quiescent current of the lamps (80 - 100 mA) is set using a variable resistor R13. The bias voltage is about -45 V. The use of additional resistors R14 and R15 completely eliminates the erroneous setting of the bias voltage and its loss when the contact in the variable resistor R13 is broken.

At the input of the power amplifier using two 6p45s, between the connection point of the lower (according to the diagram) terminals of the coils L1 and L2 and the common wire, a capacitor with a capacity of about 120 pF, made up of 3 KT-2 capacitors, is installed. The capacity of this capacitor is specified when tuning the amplifier in the 28 MHz range according to the minimum SWR in the cable connecting the transceiver to the power amplifier. It is advisable to carry out the adjustment with well-warmed lamps. The low-pass filter is adjusted by selecting the inductance of the coils L1 and L2, as well as the cable length.

The P-circuit must first be adjusted in a “cold” way. The stand diagram is shown in Fig.Z. When setting up the P-circuit, you should not, as some authors recommend, disconnect the lamps and the anode choke and replace them with an equivalent capacity. Firstly, it is difficult to accurately measure this capacitance, and not all radio amateurs have a capacitance meter, and secondly, the anode choke in the parallel power circuit is connected precisely in parallel to the P-circuit coils (via blocking capacitors C12 and C15). Consequently, a loop reactive current flows through it, depending on the magnitude of the alternating voltage at the anode of the lamp and the inductance of the inductor itself.

As is known, when two (or several) coils are connected in parallel, their total, total inductance decreases and becomes less than the inductance of any of the parallel-connected coils. It is clear that the greatest decrease in the inductance of the P-circuit will occur in the range of 1.8 MHz. In the 28 MHz range, the effect of the anode choke on reducing the inductance of the loop coil is insignificant and is within the error measuring instruments, and can be neglected.

If coils L3 - L5 are made exactly as described, setting up the P-circuit comes down to checking the resonance in the middle of each range. For this purpose, a heterodyne resonance indicator (HRI) is suitable, which, despite its simplicity, is a universal high-frequency device and is completely undeservedly forgotten at the present time. Don’t forget about the neon light bulb, which, when mounted on a long fiberglass stick, is an excellent peak indicator high frequency voltage and allows you to accurately determine the moment of fine-tuning the P-circuit to resonance, or, for example, the appearance of self-excitation. By the color of its glow, you can approximately determine the frequency of self-excitation. At the operating frequency, the glow of a neon light bulb has a yellowish-violet color, and when self-excited on VHF, its glow takes on a bluish tint.

The anode current of the lamps with a detuned P-circuit should be about 600 - 650 mA, with a tuned P-circuit - no less than 535 - 585 mA, i.e. The “dip” of the anode current during the P-circuit setup process should not exceed 65 mA, because in this case, a redistribution of the anode current occurs “in favor” of the current of the lamp screen grids. Therefore, a higher current of the screen grids will cause them to be overloaded with power, which is undesirable.

You should not strive to achieve an output power of more than 200 W. However, by increasing the anode voltage to 900 - 1000 V and accordingly changing the P-circuit data, an output power of 300 W can be obtained in SSB mode. But the reliability of the amplifier decreases, because the maximum permissible power dissipated for a long time at the anode of one lamp is only 35 W. Therefore, using this mode is not recommended, and the difference in the level of emitted signals is not so great.

“Sir, why are they monsters? They are heavy, huge, and very hot.” Let me start by saying that the magazine you are reading is not an audiophile magazine. What is audiophilia? This is a passion for canned (in a good way!) sound. The power switch clicked and... enchanting sounds poured out.

Not from an Edison roller, not from a gramophone, not from a gramophone, but from yours, precisely your acoustic systems. But how to achieve magic, or enchantment with sound? Of course - by using the appropriate components of the sound reproduction system. Let's not talk about turntables and speaker systems, especially about gold-plated cables and silver chassis.

Let's turn our attention to the amplifier circuitry. In the past, in our huge country, all efforts were spent on “defense.” Issues of high-quality sound reproduction were dealt with by individual enthusiasts. There were few publications. The main achievements were obtained not here, but somewhere overseas.

The main sources of information are also located there. Who among us has heard before about Cucing’a triode amplifiers, the famous D.T.N. Williamson’e or that local transformer OOS in the pentode cathode was proposed by Peter I. Walker on f. Acoustical manufacturing, which produces products under the “Quad” brand? Something has appeared in our country in recent years. Although there is still not enough information.

• Firstly, these are lamps.
• Secondly, these are triodes.
• Thirdly, this - (God forbid!) - do not use negative feedback (NF) and class “B” (only “A”!).

Fourthly, the simpler the scheme, the better it is. “Single stroke” is better than “two stroke”.

Unfortunately, I was not able to hear the real “Ongaku” work. Among my friends there was no owner of this wonderful device from Audio Note. And all kinds of “Priboi” and even one “Luxman” sounded somehow equally “dull” on the tubes, and did not make an impression. But one day, an audiophile friend complained that the tube amplifier that he had assembled with his own hands over the course of a year did not live up to expectations, did not “sound” and did not even provide the required power.

I helped him adjust the lamp modes, reduce the background and get an output power of 6 W per channel, and also introduced a switchable OOS from the output to the input stage, i.e. covered three stages with it, which is often done in tube amplifiers. In addition, I added an RC circuit at the output (Zobel circuit) to eliminate RF oscillations at idle. Based on the instruments, it turned out to be approximately the same settling time as without OOS, and the same exponential.

And here we are listening to this amplifier. Sounds great! Deep, without being tied to speakers, surround sound is simply mesmerizing! Instead of this tube “monster” we turn on the American “Harman Kardon” (NK-1400) - transistor with OOS (“inexpensive”, only $700). The sound is noticeably worse than that of a homemade one - there is no such volume and depth. We are launching the domestic tube “Priboy 50 UM-204S”. The sound is even drier.

Finally, the most decisive experiment. We turn on OOS in a homemade lamp. At the same time, the bandwidth is expanded from 30 kHz to 100 kHz, the output power increases to 12 W with the same harmonic distortion (about 3%), and the output impedance decreases. Everything seems to be fine, but the effect is amazing! The sound becomes the same. like “Surf”. The charm has disappeared, the sound is “dry”, there is no volume. not to mention the small details.

I don't want to listen. We remove the OOS - and the “magic” is restored! Again, I don’t want to turn off the amplifier. so I would listen and listen... Then we compared its sound with the sound of the Orbita UM-002 Stereo amplifier, copied from the Quad-405, and established that the Orbita is in the same place as the NK-1400, but this place is much lower than the home-made lamp.

It should be noted that the listening was carried out in the same room of 16 m², with the same acoustic systems, with the same CD player, on the same discs (test, jazz, choir, vocals, symphony orchestra).

A homemade amplifier is an I. Morrison amplifier, adapted to our configuration by A. Bokarev. I bring this simple diagram(Fig. 1) with the OOS circuit, which improved the objective technical parameters, but “spoiled” the sound. The amplifier uses a housing and transformers from the Priboi 50UM-204S ultrasonic unit.

The supply voltages turned out to be slightly less than indicated in. The output power was also less. What are the benefits of using triodes instead of pentodes in the output stage? More precisely, 6P45S lamps in triode connection, in class “A” and without environmental protection. In class “A” the output power at the same supply voltage is significantly reduced compared to class “B”.

But for high-quality sound in small rooms (16...18 m²) and with speakers with high output, 6...8 W per channel is quite enough. Triode connection gives a lower harmonic coefficient than pentode connection, by a factor of 2-5% and 10%, respectively (without OOS) at optimal load, and even less with an increase in the load applied to the anodes, but at the cost of a decrease in output power.

The internal resistance of the triode (Rj = ∆Ua/∆Ia) is significantly less than that of the pentode. This can be seen from the given anode characteristics of the GU-50 (P-50, LS-50) pentode (Fig. 2). In triode connection, GU-50 and 6P45S have almost identical output characteristics. For 6P45S in triode connection they are given in.

The use of an output transformer designed for a pentode and having a large inductance of the primary winding makes it possible to greatly expand the frequency response towards low frequencies, because The Ri of a triode is several times smaller than the Ri of a pentode. For the same reason, the current capacitances of the windings are recharged faster, and the frequency band expands towards high frequencies.

The small Ri of the triode gives a low output resistance even without negative feedback, although the low frequencies are somewhat emphasized. And finally, the most important thing. The absence of negative feedback gives a purely aperiodic transient process, without delays and oscillations (tyst = 10 μs to the level of 99% of the steady-state value of Uout). The introduction of resistive feedback with a depth of 20 dB (only resistor R7 is turned on) leads to large fluctuations transient response(PH). The oscillation amplitude reaches 60% of the pulse amplitude, and the oscillation period is 6...7 μs.

Turning on capacitance C2 = 1500...2000 pF eliminates oscillations, the process becomes similar to exponential, tyst 5 μs. Oscillations with a period of 6...7 μs indicate the presence of a resonant maximum or dipole on the Bode diagram at a frequency of about 150 kHz, which can cause the PH to tighten and “spoil” the sound. So choose! Either the efficiency is like that of a steam locomotive and great sound, or good performance and the desire to turn off the amplifier as soon as possible. Audiophiles are not afraid of low efficiency. Their slogan: sound quality - at any cost!

The article is devoted to some features of building an amplifier using the most powerful and relatively small-sized lamps from the series known as “television”. The text contains a significant amount of reasoning on related topics. Oddly enough, it is the adjacent areas that are extremely important for ensuring the resulting quality of the amplifier. For example, it is the matching transformers that radically influence the sound, and not the lamps at all. Serviceable tubes have virtually no effect on the sound amplification characteristics. However, the lamps look beautiful and glow in the dark. And this is probably why the names of the lamps give the impression of decisive signs of the quality of the product. Already by appearance The solidity of the pot-bellied glass cylinders 6P45S is noticeable. Taking into account the power reserve traditional for Soviet lamps, it is possible to build a push-pull amplifier in which the anode dissipation can be increased to 45-50 watts. With such a large dissipation, the heat release will be enormous. This is of course a drawback. But, according to the GURU, the sound quality in modes close to A can be obtained excellent. My attitude towards such extreme is cautious. I am not a supporter of mode "A" in a tube amplifier. The second inconvenience of the 6P45S can be considered the upper location of the anode terminal. In addition, the filament current is 2.5 amperes and the lamps heat up very much, which is also inconvenient. Therefore, a structure with a mesh-covered top, or at least with crossbars, should be provided. For heat removal, we can recommend the use of low-noise computer fans at +12VDC, with automatic switching on when the case heats up above 50 degrees.

Given the considerable power of the selected lamps, you should pay close attention to the design of the power supply. It should be noted that the traditional frivolous attitude of many TV viewers towards the power source of a tube amplifier is not suitable. The amplifier's power supply is its power plant, the core of its design and the source of all its success. The power plant must be created extremely thoroughly and precisely according to the block principle. And novice lamp manufacturers need to learn how to quickly and accurately calculate the required power of power transformers. It is better to focus on the maximum consumption mode and approximately calculate the total power of the transformer windings. First you need to calculate the power dissipated at all anodes. In maximum mode, 4 lamps can dissipate 40x4 = 160 W. Small lamps dissipate 4-6 W in the anodes. Then you need to add to the heap the power that is planned to be sent to the load, for example 50x2 = 100 W. The incandescent circuits of powerful lamps consume 2.5x4x6.3=63 W. Small incandescent lamps will consume 12-14 watts. In total, the resulting consumption is 260+75=335 W. The design efficiency of a two-channel amplifier does not exceed 30%.

The power of power transformers can be reduced somewhat, since the maximum mode is used extremely rarely. When designing transformer power supplies, the large overload capacity of transformers is taken into account. For this reason, this is what they usually do when creating serial amplifiers, reducing the installed power of the power supply by 20-30%. This solution is quite permissible, but for high-level amplifiers manufactured in single copies, it is better not to do this. In addition, the filament power cannot be reduced, since heat losses cannot be deceived. You should not overestimate the installed power of transformers, since this unjustifiably increases the weight of the product. Remember, with the calculated power value of the power supply transformers, the resulting ratings correspond to high operating temperatures. Therefore, transformers heating up to 60 degrees should not be a surprise to the designer. If the viewer has the idea in his head that the iron should be cold, then all powers should be doubled and prepare for the fact that the weight of a 15 W amplifier will become 35-40 kg.

In my opinion, the most promising circuit solution for push-pull tube amplifiers of high energy efficiency should be considered a matching stage on a differential pair of transformers. The advantages of such a scheme completely cover its disadvantages. I attribute any discussions about hand-to-hand winding of matching transformers in a tube amplifier to perfectionism. To me, this seems to be one of the designer's self-consolation methods or one of the marketing steps in justifying the extreme cost of the amplifier. Self-winding is a harmful excess and stupidity. In itself, hand-to-hand winding of transformers into a push-pull amplifier is not a technically difficult task. But making a symmetrical pair is no longer an easy task. Manually manufacturing identical four transformers for differential series connection is an unimaginably complex project. For single-ended amplifiers, the creation of symmetrical trances is possible, since using Ignatenko’s technology, it is possible to use hammer tapping when adjusting the air gap on the glue at the joints of the cores. Characteristics of iron for transformers with a gap special significance they do not, since the gap dampens the magnetic properties of the core by 1000 or more times.

An example of a first level diagram is shown below. Here the anode voltage is quite high, and the grids are connected according to an ultra-linear circuit to the symmetrical 42% taps of the transformer windings, relative to the center of the +330 volt anode supply. This is not good, since according to theory, the second grids should have a lower voltage than the anodes. But in practice, such inclusion, along with the advantages of the ultralinear scheme, may have a drawback - the occurrence of additional distortions described by Ignatenko. Therefore, you can consider an alternative version of ultralinear inclusion according to a different scheme, shown in the article below. A special feature of these particular circuits is the inclusion of an output stage driven by cathode followers. Fans know that television lamps are low-sensitivity. Therefore, you have to resort to additional tricks, use preliminary stages with dynamic loads, or install additional powerful drivers. The use of a circuit with direct connections somewhat complicates the setup, but allows you to avoid the use of decoupling capacitors. Practical repetition of the circuits shown here should be performed using 6N1P lamps, with carefully selected halves according to the condition of symmetry. Yes, and the output lamps in this version must be selected according to the bias voltage. There are general recommendations for building high-level push-pull amplifiers. You need to use symmetrical lamps, and there will be significantly less hemorrhoids. And in these specific schemes this is no longer a wish, but a requirement.

There are no gaps in highly efficient matching transformers, so the result depends only on the quality of winding, the equality of turns, the quality of assembly and the nonlinearity of the iron characteristics. The last two conditions are extremely difficult to achieve in reality. Here, right off the bat, you need to assume a discrepancy in operating parameters of approximately 10%. And this discrepancy can be established in practice only by measuring the finished product. And when a discrepancy is discovered, the finished trance can be safely thrown into the trash, since such a discrepancy will not allow building an energy-efficient amplifier. To require pinpoint precision, you can take the path of selecting symmetrical pairs from a bunch of bourgeois output transformers, but it’s hard to even imagine how much money this will cost. You need to understand that a very good result in an amplifier is given by a discrepancy in the load characteristics of transformers of no more than 2-3%. Moreover, it is curious that such a difference in currents XX does not at all guarantee the equality of the EMF of the windings when connected in series! This feature is described in my method for selecting transformers, here on the website. As a rule, out of 4-5 transformers with approximately the same no-load current of 10-12 mA, only two products produce a symmetrical pair. The rest differ by 8-10% and you have to select a pair from adjacent values ​​of 8-10mA or 14-16mA for XX currents.

The explanations presented here show the depth of the abyss on the way to building a high-quality and energy-efficient amplifier with a differential pair of matching transformers at the output. If the requirements for symmetry are somewhat roughened, for example to 15-20% divergence of the EMF, then the selection of pairs is much easier to perform. At the same time, at the stage of setting up the amplifier, the curvature of the operating system with respect to alternating current must certainly be corrected by hand-to-hand adjustments to the instruments. It will not be possible to find a direct connection with the quality of sound amplification here, since there is none. Don't think that an amplifier with crooked transformers will sound much worse. You won't notice this by ear, even at medium power. Tube circuits, as a rule, are self-balanced and easily tolerate curvature. And the adjustment allows you to equalize the characteristics of the sound path. You just need to be aware that the maximum operating parameters of such a design will actually be lower. For example, a car with the inscription Bugatti will not go at a speed of 299 on the highway to Abakan. The available speed limit will be only some 150 km/h. I declare with full responsibility that blind listening to amplifiers with lamps operating in different areas, even very different operating characteristics, will not be reliably identified by experts. There are no such people who distinguish between different spectrums of harmonics, beautifully mixed within the musical range. Using instruments, it is certainly possible to determine the difference in spectral composition. But only by instruments. Therefore, for experts, all that remains is to smack their lips and shake their heads, saying they like this and don’t like that. Moreover, it is not a fact that specific people with a damaged worldview will like a more even frequency spectrum, without outstanding harmonics.

Beginning designers should remember that in reality the situation is even simpler. If the requirements for the product are reduced even further, then when setting up the amplifier, it will be possible to straighten the more significant curvature, or at least smooth out its consequences. In this case, the lamps themselves can also be crooked. But even using clumsy light bulbs, you can push them to different performance characteristics. At the same time, being in curved modes, the lamps will be able, within reasonable limits, to deliver the power of an undistorted signal to the load, quite sufficient for comfortable perception of sound. The difference is easy to see in the comparison shown below. A beautiful and compact Chinese jack with the inscription 12 tons, made without tricks, will easily lift the Kukuruzer, but it should not be used for Kamaz. After all, he will lift the Kamaz only once. And if so rigorous testing not to do so, then the Kukuruzer driver will be satisfied with the small Zhiguli dimensions of the jack and the inscription 12 tons and will never know the reality. This is ordinary marketing, oh, a typo in the text, this is ordinary deception.

An example of a second level circuit is shown below. The division into levels is, of course, conditional; the output transformers are exactly the same. The number of windings is fixed. And whether to adapt these windings for cathode OS or grid windings is a matter of taste. The main thing is to perform error-free desoldering, for which there is the usual “scientific poking” method. A properly assembled and functional amplifier is quite sensitive to transformer feedback, therefore, any incorrect activation of them is fraught with a sharp deterioration in the mode. And the option correct inclusion there is only one winding. This is what you need to detect when setting up an amplifier with OS.

In general, we can conclude that the 6P45S lamp is an excellent motor, suitable for building a dynamic and almost omnivorous amplifier. It is absolutely possible to double the tetrodes to increase power. We must be very careful with the authors of pictures in which, instead of a classic tetrode, a 6P45S lamp is depicted as a pentode. This is the wrong image. This is where we should proceed in assessing the reliability and resulting authority of the circuitry and the author’s reasoning. In continuation of this article, another article is planned on the site - about the features of selecting 6P45S lamps.

At the end of the presentation, I dare to assure you that all the hardware described on the site can be purchased for rubles. In order to buy a 6P45S amplifier at a price starting from 45K, the buyer simply needs to negotiate with the seller, preferably in Russian. The algorithm for fulfilling obligations under supply (purchase and sale) agreements is as follows. The interested party calls me by phone at a reasonable time in the Krasnoyarsk time zone. We are actively discussing the details of the contract. The buyer then credits my phone number payment of 1% of the purchase price. This serves as a sign that the buyer is serious and allows me to promptly call back if necessary. After discussing over the phone, I send to email partner Commercial offer, with product characteristics, warranties and delivery times. Next, negotiations are completed through correspondence and the buyer transfers 20% of the purchase price to my account. The remaining 79% of the amount is transferred to the supplier’s account after the buyer receives notification of readiness for delivery. Please remember, prepayment for glands is 100%. Therefore, the buyer can immediately transfer the entire amount, already at the first stage of correspondence, but only after my written approval. There are no movements on my part without advance payment. Advice is free. Delivery of pieces of iron by Russian Post or transport company at the buyer's expense. Pickup is possible by arrangement. If the buyer refuses the transaction, no refunds will be made.

Evgeny Bortnik, Krasnoyarsk, Russia, November 2017

I propose a well-developed UHF circuit for 6p45s, with a five-band tone block. The amplifier is made according to a classic single-ended circuit. The circuit of A. Manakov was taken as the basis. The diagram does not need a description of the operation.

Tube bass amplifier

During the assembly and setup process, some resistor values ​​were changed. During the setup process, you will need to select R23, R34 so that the voltage at the 6p14p anodes is 190V. Then, by selecting R45, we set the anode voltage to 6N3P 90-110V. You may have to select the resistance R22, R33; the voltage at pin 9 is set to 90V. The negative voltage on the 6p45s control grid can be from 45 to 70V, it all depends on the lamps used and the degree of wear. For me this value is 54V. This completes the setup.

Tone block

I used a BA3822LS circuit as a tone block. This microcircuit has good parameters and is available for sale. Our price is 69 rubles. The advantages of such a circuit solution are the absence of a bunch of shielded wires and screens; in the absence of a signal, no background noise or hiss is observed. It is advisable to connect the finished tone block to the ULF input through trimming resistors of 100 kohms, since the microcircuit has sufficient high level gain.

Initially, instead of a microcircuit, I used a similar circuit with two 6n3p lamps, but in the end I abandoned it due to the impossibility of getting rid of interference and background due to weak shielding of the lamps and the entire circuit due to insufficient space in the case. I will note that the control unit on the lamps still sounds warmer, it seems to me. For those who are interested in this option, the diagram is also attached.

Amplifier power supply

Now about the power supply. We took a ready-made TS270 transformer, just wound a few turns on top of the existing windings. The throttles were taken ready-made from... I don’t remember what. Or from a b&w TV or receiver... it is advisable to organize the power supply for each channel separately to reduce interference and distortion between them.

I used one rectifier for both channels. There was no particular desire to wind another winding, just like the wires in particular. Paid more attention to capacitors instead. Nothing like this was noticed; I made do with one step-up winding. The output transformers are homemade, such as ts-20 ts-30, whoever has them, with a horseshoe-shaped core.

We wind it this way: the primary is 94 turns with 0.47 wire, then 900 turns of the primary with 0.18 wire should turn out like this: 94/900/94/900/94/. We connect the primary to the secondary in parallel; we do not place any paper gaskets between the halves of the iron. We apply supermoment (second glue), assemble and put a bandage on top of the iron if there is one, if not, then we clamp the iron until the glue dries completely.

The advantage of this solution is that it does not make noise during operation (provided the iron is good and the windings are tightly laid), the iron holds securely and, if necessary, can be easily disassembled - just hit it lightly with something heavy at the gluing site. For the body I used 3mm aluminum sheets. The adjustment handles are decorative duralumin handles for furniture doors; the holes are drilled to the required diameter and put on via heat shrink directly onto the alternators.

The body is painted with auto enamel and half is covered with wood-like film. I made the power supply transformer remote in order to reduce its influence on the power supply. The trans was packaged in a case from an old power supply unit, connected to the amplifier with a 6-core cable through a connector on the case of the amplifier. The cable is assembled by hand. There is an inaccuracy in the diagram: R40 AND R29 ARE USUAL MLT-2. BUT R28 R39 must be five-watt!

Comments (28)

Uv. Sam! Can you indicate the output power of this ULF?
Thank you.

Yes, of course. In this version there are 12 watts per channel. in triode mode 24 watts per channel. pins 3 and 6 are connected directly to the power supply positive. and select the bias voltage. Set the quiescent current to 100 mA. in triode mode, nonlinear distortions increase. And 12 watts is enough for 100 watt acoustics

12 watts is enough for 100 watt acoustics

You want to say that such an amplifier will only power my S90 by 12 W, something I don’t understand...

If your acoustics are not S-90, then all the others will do just fine. I wish I could demonstrate it. Of course, the main and main condition is correctly wound output trances.

s-90, although they are considered, if not one of better acoustics domestically produced, but their sensitivity is very low

but if there are no other columns, then you can slightly adjust the number of turns of the secondary.

Good day to the author
There are a couple of problems with the diagram
1 The left-right input goes to R52 and R53, are these the central cores, and the common core sits on the ground?
2 there is no information about the 6.3V filament power supply, how many are there for each type of lamp?
3 in the diagram shows where to supply power -70 and +350, and do +70 and -350 go into the circuit or not? some of the pins are unlabeled (end R29 R40)

Good day! I was interested in your article, I want to make the same amplifier, but I didn’t understand how to wind the output signal transformer from lamps (I have no experience in this).
As I understand from the description, the power transformer and the output transformer are combined or what?
I used one rectifier for both channels. There was no particular desire to wind another winding, just like the wires in particular. Paid more attention to capacitors instead. Nothing like this was noticed; I made do with one step-up winding. The output transformers are homemade, such as ts-20 ts-30, whoever has them, with a horseshoe-shaped core.
Quote:
We wind it this way: the primary is 94 turns with 0.47 wire, then 900 turns of the primary with 0.18 wire should turn out like this: 94/900/94/900/94/. We connect the primary to the secondary in parallel; we do not place any paper gaskets between the halves of the iron.
From the description, I didn’t figure out where the secondary is and where the primary is, since there are two of them in the description. If possible, please write a little more about the output trans or correct where the error is. Thanks a lot!

P.s. Yuri
Where it says 70 and 350 V, conclusions. Pay attention to which of them are grounded, this is not a minus, but the author meant the dash.

I answer for Yuri. first: yes, the braid, aka the common core, sits on the ground. second: the power supply is made on a ready-made transformer from a black and white box of type TS-270. (270 watts). All windings on this transformer can be left. You just need to wind up the secondary winding so that the changes at the output are 300-320V. Those that are on it give 220, this is not enough. For incandescence we use the original winding. It will quite draw 5 amperes. All the filaments of the lamps are connected in parallel. third: 70V winding, plus connects to the ground common wire. -350 is supplied to the ground wire according to the diagram. I didn’t sign, I thought there would be no questions, everything seemed obvious.

in the description above I mentioned the TS 270 and that it was taken from a b&w lamp box. It has all the windings, if there is no 45 volt winding on this trance, we wind it ourselves

70 is served where -70 is on the diagram, this is not a dash

for ALEXEY.. there is only one power transformer; it is not in the photo, it is packed in an old case from a computer power supply, I wrote about this. and two output trances. For the output trance, it is best to take a trance with a horseshoe-shaped core, the same as the ts270, only smaller in size. If there is no such transformer, as I wrote above ts20 or ts30. then you can take any other with an approximate power of 30-40 watt, with W-shaped iron. Why is horseshoe-shaped iron better? Because four ready-made horseshoes have already been assembled and glued together. Sh-shaped is bad because if the plates have burrs, rust, crumpled, then when assembled they will short-circuit with each other, due to the fact that the insulation is broken. such a trance will not work as it should, and will bring nothing but disappointment. A prerequisite for such iron is unbroken plates. If there are no vehicle transaxles, then you can take ready-made factory ones like TVK-30. iron is the same. but it will take some time to disassemble them, because they are filled with paint and, in the worst case, epoxy. Now about the winding. First, 94 turns of the secondary winding are wound, the one to which the acoustics will be connected, then a layer of insulation, then 900 turns of the primary winding, then a layer of insulation , in the end you should have three secondary windings independent of each other and two primary windings. You draw conclusions from each winding so that you can then connect them as needed. secondary windings connect in parallel, just don't confuse the beginning and end. beginning with beginning end with end. and you connect the primary windings in series, starting one with the end of the other, in the end you get one primary winding. and one secondary. Well, it seems that it couldn’t be clearer. there is nothing complicated even if you have never done anything like this

hint: if you can’t get rid of the AC background in the speakers, then look at the diagram with the name: entry-level tube amplifier, there is a circuit in the circuit design that completely eliminates this problem. You can easily find this diagram through Yandex.

In reality, this lamp produces 6-7 watts of high-quality sound, the higher is slag.

for Vlad. most often the slag is obtained from the fact that they use low-quality elements, there is no knowledge, skill, and even more often by hand..

There is a significant flaw in the scheme. The anode load 6P14P is 22 kOhm, which is much more than optimal. This significantly increases the output impedance of the driver, which makes it less pleasant for the output tube. Driver distortion also increases greatly.
The 6P45S lamp has a huge steepness. This means increased distortion.
And indeed, with a power of more than watts, these distortions are both audible and visible even on the oscilloscope screen.

for sound.. everything you wrote was in theory.. here everything was selected manually and it was the 22k anode resistor, try to assemble it yourself first, check and then criticize. for each 6p14p the value of the anode resistor turned out to be different. and everything was set up using an oscilloscope.

Sergey, all the answers to your questions are shown in the following picture (click to enlarge the picture):

Hello, please tell me what the problem is: I assembled such an amplifier, but one channel is louder than the second. It turns out that the channel in which the 4.7k resistor is installed is louder.

Hello. Try swapping the amplifier inputs connected to the 1x250 pass-through capacitors (C1, C26) - if now another channel becomes louder, then the reason must be sought in the signal source, signal wires or preamplifier circuit on a 6N3P lamp.

The circuit diagrams of the LF power amplification channels on 6P14P and 6P45S lamps are completely identical. If they play at different volumes, then there may be faulty or damaged electronic components installed somewhere.

For example, to select R23 and R34, which are 22K each, you can take a variable resistor of 36-56K, set its slider to the position to obtain a resistance of 22K. Then turn on the circuit, measure the voltage at the anodes of the lamps, as indicated in the article, slowly rotating the variable resistor knob to achieve the desired readings from the device. In place of the variable resistor, solder a constant one with the same resistance.