HF power amplifier with hybrid input. Hybrid umzch without environmental protection

This circuit of a tube-transistor headphone amplifier has been repeated by many amateurs good sound and is known in many versions, both using bipolar transistors at the output and field-effect ones.

Anyway this is Class-A. It attracts with its simplicity and repeatability, which I was also convinced of, at the same time having a desire to hear the music “performed by him.”

I bring to your attention the concept of building a hybrid single-ended circuit, the development of which was prompted by the articles “Pocket Ugly Duckling, or Pockemon-I” by Oleg Chernyshev and “Tube-semiconductor ULF” (zh. Radio No. 10, 1997).

The first article describes a tube amplifier whose output stage is covered by a parallel negative feedback (NFE) circuit. The author complains about possible criticism for the lack of modernity of such a circuit solution (OOS and even on the first grid). However, such solutions were widely used during the golden era of tube sound engineering. See, for example, the article “Radiola Ural-52” (zh. Radio No. 11 for 1952).

I like the simplicity of implementing such an OOS: there are only two elements in the feedback circuit, these are resistors and one of them, as a rule, serves as a load for the driver stage. Such OOS does not require adaptation to the type of output lamp used (within reasonable limits). But! In the same article, the author, citing calculation formulas, says that it is necessary, depending on the output resistance of the driver stage, to adjust the ratings of the feedback circuit resistors.
So many “opportunities for creativity”! I installed another lamp and re-soldered a couple of resistors. It seemed wrong to me.

In my article I propose a solution to this “problem”.

They asked me to make an amplifier for sounding a room of 50 m 2, a kind of “village club”. It must be said that there is already some kind of industrial amplifier there, which is used for all kinds of events such as “disco”. That is, it plays loudly, but at the expense of quality. An amplifier was needed specifically for more or less high-quality listening to music, 30 watts per channel.

I couldn’t make a tube amplifier of such power, so I turned my attention to hybrid amplifiers.
We have it on Datagor. Let me remind you that “Corsair” is in a fan-powered configuration with a tube buffer at the input. I decided to study reviews and opinions on the Internet.

What remained was a working prototype of the SRPP on 6N23P.
It was a shame to throw it away. There was a desire to finish the amplifier to the end. In the previous craft, we had to apply some simplifications related to the size of the case, for example: common power supply for both channels, not exactly the capacities that I would like to try.

It was decided to make a new SRPP headphone amplifier on the 6N23P without these simplifications.
The result was suddenly this kind of hybrid.

Greetings, dear Datagorians!
I present to your attention a hybrid headphone amplifier based on a 6AQ8 (6N23P) tube and IRF540 field-effect transistors.

Printed circuit board drawings, installation details included, no background.

04/29/14 changed by Datagor. Amplifier circuit corrected

I have long wanted to listen to how a lamp and a stone sound in tandem. I decided to build a hybrid headphone amplifier. I looked at several diagrams. The main criterion for choosing was the simplicity of the circuit, and therefore the ease of its assembly.
I settled on two:
1) S. Filin. Tube-transistor amplifier for stereo phones.
2) M. Shushnov. Hybrid headphone amplifier. (Radiomaster No. 11 2006)
In general, these schemes are not much different from each other and without major changes you can try both one and the other. I decided to put together a diagram of M. Shushnov with field workers.

Another failed experiment led to the idea of ​​​​a lamp buffer for and it turned out when I conscientiously filtered the power supply to the lamps.

It took me a long time to come up with the idea of ​​a tube buffer, but all the failures are in the past and the idea justified itself. Not only op-amps can match resistances - a cathode follower on a suitable lamp is also suitable for this task.

The plane was confidently descending along the glide path, as if following an invisible thread; the runway was quickly approaching. The turbines smoothly switched to idle, the plane hovered over the runway and a second later rolled, counting the joints between the concrete slabs. The reverse flaps shifted, and the silence was cut by the sound of air being turned away by the flaps...

Alas, I heard it many times, but the reproduced sound of reverse by the flight simulator through Genius tweeters did not impress me. And listening to music without headphones did not bring any pleasure. And then I decided it was time to get decent acoustics for my computer. Without thinking twice, I wrote a message to Sergei (SGL) asking what I could buy that would please my ears. To which I received the answer, the best speaker is a self-made speaker!
Let's say. And then I received a link from him. That's how I ended up on Datagor.

It started a month ago with Alexander’s good-natured provocation at the Datogorsky forum, when discussing indicators.
At the output I had a debugged final stage and I remembered that there were some indicators in the junk. And it was Guntis’s successful attempt to play with the indicator that “started up”.

Then everything developed into what can be seen in the photo, and what my wife calls a nightmare, and what I call “a sweet-voiced creative disorder.”
If you wish, you can even see how the indicators glow, but they do not blink in time with the music, as Alexander hinted at.

Sorry about the photo, I only have a multimedia camera.

Few people have tubes left, but they can still be purchased, so tube audio equipment is of constant interest to radio amateurs. You give us that same warm tube sound that has long become a meme that people like to put into place and not so much. Now let's try to combine old tube audio equipment with a more modern element base. You can get simply magical sound.

The amplifier is assembled according to a classic single-ended circuit. During the setup process, I changed some resistor values. So it was necessary to select R23, R34 so that the voltage at the anodes of the 6p14p lamp was 190V. Then, by selecting R45, we set the anode voltage on the 6n3p lamp 90-110V.

I used a BA3822LS circuit as a tone block. This microcircuit has good technical parameters and is not expensive. The main advantage of its use is the absence of a huge number of shielded wires and screens; in the absence of a signal, background noise was not heard. Connect the assembled tone block to the input of the tube ULF through 100k tuning resistors.

When making the power supply, I used a ready-made TS270 transformer and added a little more turns on top of the windings.

One rectifier is used in both channels. The output transformers are completely homemade, type TS-20.

We wind them as follows: the primary winding contains 94 turns of 0.47 wire and 900 turns of 0.18 wire; in short, in the end it should be 94/900/94/900/94/. We connect the primary winding in series, the secondary in parallel.

For the body I took sheets of three millimeter aluminum. I took the adjustment knobs from aluminum furniture handles, drilled holes to the required diameter and put them through heat shrink directly onto the variable resistors.

The power supply for the lamp stage is supplied from an unstabilized source of 300…350 volts. The filament voltage of 6.3 V does not need to be rectified or stabilized. The filament lamps of the right and left channels of the amplifier can be connected to one winding of the transformer, but it is recommended to make the anode circuits separate.

The amplifier passed the hearing test excellently - crystal clear sound, especially in the middle and top of the sound range.

The input amplifier is made in pairs field effect transistors 2SK68A and on high-voltage bipolar 2SC1941 forming a cascade that performs the function of a phase inverter for the output push-pull stage on the EL34 in triode connection. This hybrid power amplifier circuit using field-effect transistors and tubes is a very high-quality sound reinforcement equipment of the highest class, so installation and soldering must be done with the utmost care and attention.

Static balancing of the amplifier is carried out with a 5 kOhm trimmer in the circuit for supplying a fixed bias to the control grids, and dynamic balancing with a 2-kOhm trimmer in the power supply circuit for the collectors of bipolar transistors. Despite the fact that the circuit contains transistors, the amplifier is made without OOS and has a clear “tube” sound.

Hybrid UMZCH 70 W

This hybrid UMZF provides full power bandwidth from 30 Hz to 100 kHz and small-signal frequency response from 10 Hz to 170 kHz. The function of a voltage amplifier and phase inverter is performed by a cascade based on composite transistors Q1Q3, Q2Q4 with a current generator Q8 in the emitter circuits and an improved current mirror Q5Q6Q7 in the collector circuits.

The fixed bias on the control grids of the radio tubes is adjusted using resistor R15 so that the initial anode currents are about 40 mA. The output toroidal transformer VDV3070PP Amplimo was purchased at an online auction. Its primary winding has a resistance of 2757 Ohms, its rated power is 70 W

This hybrid amplifier circuit delivers 80 W of power into an eight-ohm load with a THD of 0.04%, a bandwidth of 5 Hz - 35 kHz (20 W, -3 dB) and a signal-to-noise ratio of over 100 dB.

The only voltage amplification stage in the circuit is built on a 2SC2547E bipolar transistor with a dynamic load on an ECC88 triode.

The output stage is designed as a push-pull source follower based on a complementary pair of powerful field-effect transistors IRF640, IRF9640. Their operating point is set by trimmer PR1 during adjustment.

Capacitor C2 and resistor R9 are used to form a voltage addition circuit familiar to transistor amplifiers. In this circuit, it helps radio tube V1 to ensure normal swing of the output stage at a relatively low anode voltage.

The audio signal, through the volume control on resistor R1, enters the VL1.1 triode (control grid) of the amplifier and is amplified. The negative bias potential slightly blocks the triode formed on its control grid with the help of the anode current, which passes through resistors R3 and R4 located in the cathode circuit. The voltage will drop across these resistances, therefore, relative to the negative bus, a positive voltage of approximately +1.7V will be present at the cathode of the lamp.

On the control grid of the amplifier tube, if compared with the cathode, there will be a negative bias potential, since the grid has a common contact through resistor R1 with ground. To reduce the effect of feedback, the tube amplifier circuit has a resistance R3, which is shunted by the electrolytic capacitance C1. Resistor R2 plays an important role as a load for the anode circuit of a tube amplifier. The voltage of the amplified audio signal generated on it is supplied through the isolation capacitor C2 to the control grid of the lamp pentode. Through the first output transformer, the signal amplified by it goes to the amplifier's loudspeaker.

Resistor R8 and capacitor C7 perform the same function as similar elements in the first stage. C6 and R6 are designed to change the timbre of the sound. Using resistor R9, a second negative feedback circuit is obtained. By capturing both stages of a tube amplifier, it reduces nonlinear distortion and creates the smoothest amplification of the audio signal across the entire audio frequency range.

The second transformer of the tube amplifier is wound on a magnetic core with a cross-section of 10 cm (W22 x 40). The primary winding is PEV-1 wire 0.2-0.25 mm 1040 turns. The secondary winding has 965 turns of the same wire, the third has 34 turns wound with PEV-1 wire 0.6-0.8 mm.

The first transformer of the TVZ21 type. It is allowed to use any output transformer from a tube TV.

Transcript

1 Circuit design of a hybrid amplifier. E. Vasilchenko, Kazan. June 2002 In this article, I decided to abandon the generally accepted rule of writing technical, scientific and pseudo-scientific articles that require presentation in the third person. Reflections on the role of sound-reproducing devices in our lives led me to the conclusion that the creative, emotional aspects of this problem are no less important than the technical ones (though not so much as to replace one with the other). In the world of technology, which is 100% formalized, there is no place for the author’s emotions. The scientific world has much more degrees of freedom; serious passions boil in it, and sometimes the academic lines “it has been studied, it has been shown” cause a storm of delight or indignation among initiates. This tradition, carried over into popular technical publications, played a cruel joke on low-frequency radio amateurs, largely predetermining the modern situation. While magazines in recent years have been talking about a vinyl and tube renaissance, it's time to wonder where we were all looking before? After all, there were people who never put signal transformers on the shelf and never threw lamps in the trash. I keep on my desktop, unknown how, a clipping from Radio magazine with an editorial article from 35 years ago with the subtitle “From the XI Scientific and Technical Conference at the IRPA” that came to me. Without comment, I will quote an excerpt: In the reports and speeches of the conference participants, the heads of individual enterprises, which still continue to produce receivers and radios, the cost of which is higher than the selling price, were sharply criticized. Radio industry enterprises face great challenges in the current five-year period. First of all, production volumes must be increased. If during the period 21.5 million radios and radiograms were sold, then in it is planned to sell 30 million. But the sharp increase in production volume and the task of selling products put forward demands for continuous improvement of models, increasing reliability and sound quality, improving their appearance, design, architectural forms, colors, ease of use, and reducing costs. This means that it is necessary to organize production in such a way, to find such technical and organizational solutions that would facilitate the rapid introduction into production of models that are in all respects at the level of world standards. The work carried out at IRPA and design bureaus of leading factories, as well as the experience of production activities of all enterprises in the industry show that these problems are solved by transistorization and unification of broadcasting equipment. During the period from 1966 to 1970, it is planned to convert all radios of the first, second and third classes to transistors. The only exception will be monophonic and stereophonic radios of the highest class, which will continue to be produced on tubes. Transistorization of household radio broadcasting equipment will significantly reduce its dimensions, increase reliability by 1.5-2 times and achieve significant savings in energy and materials. It is estimated that as a result of transistorization, savings due to reduced costs for materials per year will amount to 2.5–3 million rubles. In addition, 170 million kWh of electricity per year will be saved. Radio, 1966, 8, p. 21. “The focus is on transistorization and quality,” writes an author unknown to me. Every time I share my experience with readers or interlocutors, I remember this article. The creation of sound equipment is a unique area of ​​human activity, where almost any person who knows how to handle a pin and a metalwork tool, to the best of his qualifications, can appreciate the value of the ideas embedded in the design. That is why the description or presentation of the plan must be personalized and separated from the opinion of the editors or fellow employees. The impersonal formula "one can conclude" must give way

2nd place to honest “I think”. To the best of my ability, I will try to implement the decisions of the mentioned conference with detailed comments. The history of the creation of the amplifier described here began quite a long time ago, more than 10 years ago. At that time, there was no domestic audiophile press, only a select few lucky people had access to the Internet, and libraries had already stopped receiving foreign magazines. The main and most popular source of information by inertia remained the magazines "Radio" and PTE (Instruments and Experimental Techniques). When almost all known transistor UMZCH circuits over the past 20 years have been repeated and tested by ear, the question arose: “What to do next?” It cannot be said that there was nothing worthy in the whole mass of schemes and designs. Each year brought a new leader. The first milestone in the mass transistorization of amateur designs was, undoubtedly, “High-quality amplifier”, S. Bat, V. Sereda. This was the first "people's" UMZCH. In essence, it was a high-power operational amplifier. The development of this topic now seems to me to be a dead-end branch. Not everything that is good for driving electric motors and other actuators is good for amplifying sound. This construction turned out to be unusually tenacious and was replicated in dozens of varieties, despite the poor sound. Transistor amplifiers of those years did not win the war with tubes. These lamps gave up key positions without a fight. Leafing through the “Radio” of tube times, one cannot help but be surprised at how well the authors implement the decisions of the mentioned conference. It’s just that high-quality tube amplifiers did not seem to exist, but “small-sized ULF”, “VLF with increased efficiency”, etc. were presented in abundance. The tube theme in mass publications was doomed, and a few years later young radio amateurs were perplexed when they encountered comparisons of this or that device with tube monsters. The deterioration of amateur transistor amplifiers of those years was no secret to anyone. But the developers worked tirelessly and at the end of the 70s there were already very decent-sounding amplifiers. Until 1965, most Telefunken, Grundig, Fisher amplifiers were made using tube circuitry: with interstage transformers, using germanium transistors of the same conductivity. After 1965, manufacturers gradually switched to silicon transistors. The characteristic circuit topology of that time is illustrated by Beomaster 3000, Uher CV-140. With the advent of powerful complementary transistors in the 70s, amplifiers began to be built using symmetrical circuits. One of the first representatives of this trend was the JBL amplifier, released in 1967. Subsequently, such circuitry was used by SAE, McIntosh, Hafler. At the same time, circuits with differential amplifiers appeared. It is curious that experts note the better sound of European amplifiers, which did not use differential drive of the output stage, unlike amplifiers from Japanese and American companies. By the mid-70s, integrated circuits (Braun A301) began to be widely used. The amplifiers mentioned deserve detailed analysis and even repetition. However, let's return to the schemes that domestic amateur designers could see and repeat. This is Quad-405, the diagram of which was published in Wireless World in 1978 and is familiar to us from O. Reshetnikov’s article in the December 1979 issue of Radio. Without a doubt, the most famous amplifier is Michael Wiederhold, first described in 1977 in Radio fernsehen electronik. This scheme is still published in various variations ("Radio" 4/78, "Radio" 6/89, "Radio" 11/99). Thanks to the work of M. Otal and Marshall Leach over the years, amplifiers have gotten rid of one type of specific TIM distortion caused by the limited speed of transistor stages. Around the same time, works by A. Mayorov and P. Zuev appeared on dynamic distortions in amplifiers. Many people remember A. Vitushkin’s good, although not the simplest, amplifier from the July 1980 issue of Radio. A. Syritso's bridge amplifiers sounded very good (especially from Radio 11/82). Many interesting circuit solutions were published in PTE collections. As high-voltage and high-frequency pnp transistors appeared, amplifiers became increasingly broadband, powerful, and linear. However, the problem of unsatisfactory sound in general remained

3 solved. The article “The phenomenon of transistor sound” added fuel to the fire. Let me remind you that the authors compared various amplifiers in a fairly good environment (studio sound reproducing devices and a professional spectrometer) and, based on their observations, made conclusions that are presented here: From all that has been said, the following conclusions can be drawn: - “transistor” sound is not a mandatory property of transistor amplifiers LF; its nature seems to lie in the imperfection of these amplifiers; - “transistor” sound disappears when the harmonic distortion decreases to 0.03 0.04% over the entire operating frequency range; - with a modern element base, the specified limit of the harmonic coefficient is achievable only with a sufficient depth of overall environmental protection. Now, when the number of our own amplifier developments has exceeded the second ten, it is easy to blame the authors for the incorrectness of the problem statement, but 20 years ago, it seemed to me, like so many amateurs, that a recipe for good sound had been found. You can simply, without paying attention to the long tail of distortions, suppress them with deep OOS plus some additional measures. The "race of zeros" began. The eighties became a dark period for sound circuitry. In order not to be unfounded, I will comment on the above quotes. The authors were looking for the “transistor sound phenomenon” in amplifiers with deep feedback, and this is akin to searching for a black cat in a dark room. It seems that if the kit had additionally included one amplifier, tube and transistor, without a common OOS, the results of the experiment would not be so unambiguous. The first comparison leader was the amplifier assembled by M. Leach. This is not surprising, it is truly the best in this class (that is, in the class of high-power operational amplifiers). In addition, M. Leach himself especially noted the role of the amplifier’s power supply, or more precisely, its ability to provide high current. Nobody took into account this most important feature of his amplifier. And a few more points that few people paid attention to at that time. Such a sound characteristic as “transistority” is subjective, and it is simply incorrect to extend the experience of one’s own perception to all listeners. And most importantly, the absence of a feeling of “hardness” or “transistority” is necessary, but not at all sufficient for a high-class amplifier. Readers of the modern audio press can easily name a dozen more signs by which sound quality is assessed. Rice. 1. Amplifier by Yu. Mitrofanova. The amplifier of Y. Mitrofanov, the circuit of which is given in the article, according to the authors, sounds better than all the others. It's not difficult to explain. The voltage amplifier (VA) of this UMZCH, Fig. 1, performed on V5, V6 has a small own SOI (0.15%) and a fairly large

4 power. The parallel OOS circuit has the minimum possible length; it is much shorter than in traditional amplifiers and is fed to the inverting input of the UN. The intrinsic nonlinearity of the output stage is also relatively small. This output stage was used in the famous QUAD 303 and in Brig. If we add to this a powerful low-impedance power supply, then these factors are quite enough to make the amplifier sound. And the SOI value of 0.02% is only a consequence of the topology features, and not the reason for good sound. Thus, the conclusions of the authors of the articles are, as in a joke about a mathematician, as accurate as they are useless. The race of zeros reached its peak in 1989 with the publication by N. Sukhov of the famous high-fidelity UMZCH, the basis of which is the M. Wiederhold amplifier, Fig. 2. Fig. 2. M. Wiederhold amplifier, 25-watt version. It has been repeated and continues to be repeated (using a modern element base) by thousands of amateurs. The range of reviews about its quality is very wide, and this is natural. How many people have so many opinions? Everyone's kits are different. Most are very pleased with them. Many claim that they have never heard anything better. I'm sure this is true, but what about those who are dissatisfied? And there are many of them; These are, first of all, owners of tube equipment, good acoustic systems, and simply experienced listeners. Let's try to figure out what's going on here. Undoubtedly, everyone’s capabilities are different, audio salons are not available everywhere, and modern “branded” equipment is not at all a role model (rather, on the contrary). The first thing that comes to mind is that all listeners have different speaker systems. N.E. Sukhov himself considers it his merit not so much to create the amplifier circuit as to equip it with a device for compensating wire resistance. It is possible that the influence of wire resistance on the damping of the AC cable PA system is relevant for amplifiers with zero output impedance, but not all amplifiers have a negative output impedance feedback. In addition, it would be a mistake to assume that the sound character of a complex is determined only by the damping coefficient. The main complaints of “listeners” about the sound of UMZCH VV relate to the accuracy of transmission of medium and high frequencies, where electrical damping of the speakers by the output impedance of the amplifier does not work. It is often said that this compensator “smears” the sound. At medium and high frequencies, nonlinear effects arise in speakers, which no device for generating output impedance can cope with. S. Ageev wrote about this in detail in. Differences in the design, packaging, and parameters of the power supplies of this amplifier also do not allow a correct assessment of the UMZCH VV through a “popular vote.” Those who want to get an idea about it can only resort to the most reliable way to listen to it themselves. This was done. The most promising amplifier of the 80s was

5 is assembled in metal in compliance with all rules for installing such devices. Comparison with other homemade products did not reveal any advantages of the UMZCH VV. Amplifiers by A. Syritso (N11/82) and according to the circuit from E. Gumeli’s article (N9/85), made by me several years earlier, sounded much more natural (and at the same time different in the midrange) with the same configuration and design. By the way, the measured SOI of all these amplifiers did not exceed 0.02%. Confidence in the correctness of the chosen path was shaken. New ideas were needed. First of all, it was decided to check the influence of OOS and various circuit topologies. Deep negative feedback was the first to be blacklisted. The prototype of the amplifier with number 6 was the “Amplifier without general feedback”. The authors used well-known components in circuit design: a push-pull emitter follower at the input, a push-pull scaling current mirror as a voltage amplifier, and a composite emitter follower at the output as a current amplifier. They very elegantly bypassed the problem of DC drift at the output of the amplifier, placed a large-capacity electrolytic capacitor at the output and used a single-polar supply. Perhaps, if I had had Black Gate or Elna Cerafine audio series capacitors then, this solution would have satisfied me. The best “electrolytes” then were K50-18 and I didn’t want to use them at all. Getting around this problem proved difficult. The amplifier was switched to bipolar power supply, the output capacitor was eliminated. To obtain more power, the voltage was increased to 2*30 Volts, the element values ​​were recalculated, and the bias circuit was replaced with a traditional one (Fig. 3). Rice. 3. Amplifier without general feedback (6) Along the way, it turned out that the amplifier works better (more stable) with conventional rather than composite transistors. The struggle with the zero offset at the output began. A voltage amplifier assembled using a current mirror circuit is very sensitive to all disturbing factors: instability of power supplies, temperature and its gradient inside the structure, variation in the values ​​of the elements and, most importantly, to the parameters of the transistors. If we calculate the overall gain of the voltage amplifier using the approximate formulas given in the article, it will be equal to approximately 7 (for an output power of 25 Watts). In fact, with this coefficient, the power ripple or, in the case of bipolar power supply, the difference in ripple of the positive and negative poles is transmitted to the output (not counting the useful signal, of course). It is for this reason that the authors of the circuit used the R19C5 filter in the power supply. Let's consider the cascade on VT4 (6). Its gain is approximately equal to the ratio of resistors in the collector and emitter, that is

6 R15 K u = 100. Therefore, the slightest drift in the emitter-base voltage of any of the transistors R 12 included in the cascade will lead to a significant change in mode. If this drift is caused by a general change in temperature in the case and the temperature of all transistors changes simultaneously, then the change in current VT4 and VT6 will be the same in magnitude, opposite in direction and will not lead to a change in the output potential. This is only possible in the ideal case, when transistors VT4 and VT6 are completely identical. In practice, there are no two identical transistors, much less with different conductivities. The difference in the values ​​of h 21 E and U BE of the transistors of the cascade will lead to a significant difference in the collector currents, and, consequently, to a zero shift at the output. If you use the transistors recommended in the article without selection, then most likely the bias will be about 0.5 1 Volt at best. Moreover, when the temperature inside the case changes, the bias will also change due to different temperature drifts of the transistor parameters. In addition, the gain and output voltage AC shoulders will also be different. To some extent, this difference in gain can be compensated for by trimming resistor R9. Balance the UN according to DC Changing the resistors included in the cascade is not possible, since this will change the gain for alternating current. The cascade load consists of two parallel-connected branches, linear and nonlinear. Resistors R15 and R17 form a linear low-resistance (about 5 kohm) branch. The gain, efficiency and output impedance of the cascade are determined by them. The input impedance of the final stage is very nonlinear, but much higher (at least 100 kohms). Therefore, the component of the voltage output current that goes into the nonlinear load branch is relatively small, a few percent, and can be ignored. Let us examine in more detail the operation of the voltage amplifier stage. The DC operating mode is set by the resistance value R10. The current through it U is approximately equal to 1.2 mA: I R 10 =. The properties of the scaling current mirror are such that R10 IVT 3 R12 = = 3. Therefore, the current through transistors VT4 and VT6 is 3.6 mA. The magnitude of the quiescent current IVT 4 R 11 must be selected in such a way that when the current through the transistor changes under the influence of a signal, its gain remains, if possible, unchanged. The dependence of h 21 Oe on the emitter current is one of the two main reasons for the occurrence of nonlinear distortions in transistor cascades. Therefore, when choosing transistors and their operating mode, the corresponding characteristics should be taken into account. Unfortunately, then, more than 10 years ago, documentation for transistors was practically inaccessible to amateurs. Therefore, the mode had to be selected approximately to minimize distortion at the output of the entire amplifier. The maximum output voltage of the cascade is close to the supply voltage. Consequently, the alternating voltage at the output of the UN can be about 20 volts in our circuit. In practice, after 15 volts, soft limitation already began. This is due to the insufficient quiescent current of VT4(6), but it was fully consistent with the power of 50 Watt speaker systems. By increasing the quiescent current to 5 or even 10 mA, the power and linearity of the amplifier should increase, but such a goal was not set. The gain of the cascade on VT4 is about 100, which means 0.15 Volts are applied to the base of VT4. Let's check: 15 V at load R15 = 10 com will be at a current of 1.5 mA. This means that the alternating current VT4 is 1.5 mA, and the signal voltage drop across R12 = 100 Ohm will be 0.15 V. To find out what part of this voltage is applied directly to the base-emitter junction, remember that the volume resistance of the emitter of the transistor is directly proportional to the temperature and inversely ϕt to the current: rе =, where ϕt is the so-called temperature potential, at room temperature IE is approximately equal to 26 mV. With a constant current through VT4 equal to 3.6 mA, its emitter resistance will be 7 Ohms. An alternating current of 1.5 mA will create a voltage drop across it

7 10 mv. Another useful relationship is that each millivolt of AC voltage applied to a pn junction adds 1% of the second harmonic level to the output current. With such a signal at junction VT4, the output current will contain 10% distortion. Local negative feedback is created through 100 Ohm resistor R12. Its depth is equal to the ratio of resistances R12 and r E, that is, 100/7=14. This OOS reduces the level of the second harmonic by 14 times. That is, transistor VT4 in this mode introduces 0.6% distortion. In push-pull cascades, even harmonics must be compensated, provided that the cascade is completely symmetrical. In reality, shoulder reinforcement always varies slightly. Therefore, we can assume that the level of the second harmonic is from zero to 0.3%, depending on the degree of symmetry. The level of the third harmonic with such a signal value at the transition is usually several times less than the level of the second and it is not compensated. You can expect its level to be 0.03-0.06%. At high frequencies, the asymmetry of the arms increases and compensation of even high-order harmonics is not as effective. The second source of distortion is the nonlinearity of the base current VT4. It can also be estimated from a graph of gain versus current. Since we do not have the required data, the domestic industry is not very kind to developers, we will use typical values ​​​​for imported transistors general purpose. For example, take a pnp transistor 2N3906 from ROHM. In terms of parameters, it is approximately equivalent to (or better than) KT3108 and KT313. According to graphs from the company's website, when the emitter current changes from 1 to 4 mA (that is, by 300%), h 21 Oe changes from 110 to 140 (by 25%), Fig. 5. This is a significant nonlinearity; modern transistors for audio applications have much better characteristics. Rice. 5. Dependence of the gain of the 2N3906 transistor on the collector current. Typical for small-signal cascades, the change in the emitter current is % of the quiescent current. In other words, during the period of the signal, the base current transfer coefficient changes by 0.5 1%. The base current also changes accordingly. In our case, the base current is I E 3.6 I B = = = 30 μA. The nonlinear component of the base current, equal to 1%, will be 0.3 μA. h21e 120 The alternating current of the VT4 base, flowing through the output resistance of the previous stage, creates a voltage drop across it applied to the base, and this voltage contains a nonlinear component. The output impedance of the previous stage is determined mainly by the R8R9 circuit. The output resistance of the composite emitter follower VT1VT2 is a few to tens of ohms and can be ignored. The nonlinear component of the base current VT4 flowing through circuit R8R9 will create a voltage drop on it of 0.3 µA * 3.3 kom = 1 mV. This is the amplitude value, peak to peak. The effective value is less by 2 2, or approximately 3 times, i.e. 0.3 mv. As we remember, the useful signal based on VT4 is 150 mV, therefore, the base current already contains 0.3/150 = 0.2% distortion. Everything that was said about compensating for even-order distortions also applies to base currents.

8 A quick analysis of the operation of this voltage amplifier gives us the opportunity to draw some conclusions. The first and obvious one: in the author’s (magazine) version of the amplifier, the transistors operate in a non-optimal mode. To increase linearity, the quiescent current of the cascade should be increased several times, because even at 10 mA the dissipated power will not exceed the maximum permissible. The second conclusion concerns the choice of transistors for such a circuit. These must be modern high-linear transistors. KT313 and KT3117, and even more so KT502/KT503, are not complementary pairs. With them it is almost impossible to obtain an acceptable SOI. Complementary pairs must be carefully selected according to h 21 E and U BE. Only in this case can the stability of the operating point be ensured and low level distortions. Additionally, the thermal stability of the operating point of the voltage amplifier can be ensured by design measures. The printed circuit board had to be positioned so that all four transistors were nearby and could be covered with a cap. Without it, any breeze blowing on the board would cause the output zero to drift. I managed to increase the potential at the output of the amplifier channels to 25 and 50 mV without using additional balancing. The third conclusion may seem somewhat unexpected, but we should not forget that this small study was launched with the aim of understanding the influence of OOS on sound. In my opinion, introducing general OOS into such an amplifier not only makes no sense, but is also harmful from the point of view of sound quality. Feedback can cover cascades that are initially linear, and then it will fulfill its purpose. Namely: it will ensure the stability of the circuit parameters over time and under different operating conditions. In the analyzed circuit, this stability is ensured parametrically, that is, by using components with precisely specified parameters. If the component parameters are chosen at random, the circuit will become unbalanced and become a source of distortion. Using OOS to correct this curvature only leads to a change in the spectral composition of distortions towards increasing the number of harmonics, but not to their elimination. The higher the degree of symmetry of the original amplifier, the less “work” there will be for the OOS. To realize all the capabilities of this voltage amplifier, I had to re-wire the printed circuit board several times and change the design of the amplifier. In intermediate versions, the UN was even placed in a thermostat. The most difficult thing was to select four pairs of complementary transistors. After futile attempts to select such pairs from KT3117, KT313, KT3108, KT502, KT503 using a simple stand and tester, I took 50 pieces of unknown Korean transistors S8050, S8550, also known as S8050, S8550. It was not possible to find their characteristics, so I looked into the incoming control department of one of the factories. Armed with an automatic transistor tester, I checked the maximum permissible voltage between the collector and emitter and sorted them by h 21 E and U BE. An increase in the reverse collector current began at voltages above 110 V. The base current transfer coefficient turned out to be within the limits for both n p n and p n p transistors. When the emitter current changed within 1 10 mA, h 21 Oe changed slightly. After that, selecting pairs and finishing the amplifier turned out to be a very simple matter. I did not specifically configure the output emitter follower with a shunt compensator, limiting myself to selecting the quiescent current of the output transistors to minimize distortion. At a current of 300 mA, the automatic nonlinear distortion meter S6-11 showed a minimum of about 0.1–0.15%. Each amplifier channel was powered by a parametric stabilizer, Fig. 6. The heating of the stabilizer transistors is insignificant, so it turned out to be possible to attach the corners on which they are installed directly to the duralumin bottom, through a mica gasket. The amplifier's printed circuit boards, measuring 70 x 80, are screwed directly onto the heatsinks of the output transistors, which have an area of ​​600 square meters. see the channel. The radiators have good thermal contact with the bottom and massive front panel. The heating of the amplifier during operation does not exceed 60 70

9 degrees. 80 Watt toroidal power transformers are separate for each channel. Rice. 6. Amplifier power supply 6. Listening to the amplifier showed that the time spent searching was not in vain. The amplifier had an extremely soft and delicate voice. The mid-frequency range was especially good. The resolution and sound detail were higher than all of its predecessors. He softened the highest registers, while the traditional ones, “from the Radio,” simply turned them into “sand.” Despite the packaging, which is completely unsuitable by today’s standards (K73-17, K50-18 and not the best transistors), this amplifier still has no competitors in sound quality among the so-called “affordable high-end” and delights its owner with the opportunity to listen to his favorite recordings , not test discs. It must be admitted that the experiment turned out to be very informative. The experience gained during the design of amplifier 6 without general feedback set the direction for further developments. The analysis of the circuits and the listening results are quite consistent with the modern unspoken rules of audio circuitry. In recent years, as the Internet has transformed from a symbol of incomprehensible luxury into necessary tool, do-it-yourselfers received an excellent opportunity to communicate and exchange experiences, both among themselves and with professional developers. The specifics of using transistors in sound circuits are gradually becoming available to a wide range of DIYers. There has never been a single recipe for building a good amplifier, but there are some general principles, which most designers come to sooner or later. All developers assess the importance of a particular principle differently; This scale of values ​​is not linear, constant and absolute, because it depends on many subjective factors. Therefore, I present my own list, based on more than 20 years of experience in building amplifiers, of the most important requirements for the design of UMZCH in descending order of importance. Of course, no one is stopping the designer from sacrificing any of the items on the list to some additional idea. A) The power source must provide the final amplifier with a current that is both powerful and clean. In modern interpretations, the amplifier is often represented as a current modulator. Therefore, the quality of the current supplying the output stages must be as high as the development budget allows. The power supply is a full-fledged participant in the audio path with all the ensuing consequences. Any secondary power source contains reactive elements that form filters. For filters, parameters such as transient response, quality factor, and characteristic impedance are defined. The influence of these factors on sound is practically not considered in the literature. But these are well-known, easily measured and at the same time parameters that greatly influence the sound. B) One of the most important components is the voltage amplifier. Perhaps this point is not as obvious as the previous one, and not all amplifiers are built according to the UN UT circuit, but many designers note that both tube and transistor output stages are “transparent”

10 for sound, and the “voice” of the amplifier is determined by the driver stage or NA, respectively. Human hearing, especially trained hearing, has an extremely high sensitivity to the spectral composition of distortions. Small differences in the power of even and odd harmonics, differences in the rate of decrease in spectral density, and the presence or absence of dominant harmonics are perceived as a change in the character of the sound. In an ultrasonic amplifier, the dynamic range of the amplifier element is usually fully used and the operating point covers the largest portion of the amplitude characteristic. Its nonlinearity manifests itself most clearly here. Therefore, all elements have their own spectrum of distortion, a kind of barcode by which they are unmistakably recognized by ear. C) The number of cascades should be minimal. It doesn’t matter whether it’s transistor or tube, but each additional stage introduces additional nonlinearity. There are many reservations in this point, as well as in all the others. Getting maximum gain from a stage can degrade stability and with it linearity. This means that there is a certain balance between the depth of the local OOS and the magnitude of the cascade gain. The designer's task is to find a compromise. D) The quality of components, both active and passive, must be adequate. An absolutely indisputable point. The only question is what is considered important and what is secondary. Most often, this question is closely related to the degree of hearing training and the thickness of the wallet. D) Thoughtful design and temperature conditions. We are talking primarily about vibration isolation, since most radio elements have a noticeable microphone effect. Calculating sound fields in devices is very complex, so designers usually use empirical data and their own experience. The temperature inside the case not only affects the service life of the elements, but also significantly affects the sound. The formation of these principles for me began precisely with the experiments described above. In the next development, I decided to test the effect of the principle of minimalism on amplifier 8 (number 7 was a tube amplifier corrector for a vinyl player). The assembled UMZCH VV boards remained from previous work, and they became prototypes for studying the nonlinearity of various cascades. The first test subject was the output emitter follower; it then entered the circuit of the proposed amplifier without changes, Fig. 7. Fig. 7. Amplifier output stage 8.

11 Circuit design analysis. The quiescent current of all three stages is set by resistors R3, R4, and is regulated by a variable resistor R2. The VT7 transistor is traditionally mounted on the heatsink of the output transistors and performs the function of setting and thermally stabilizing the quiescent current. Resistors R6, R7 are added to ensure stability of the amplifier during tuning when the length of the connecting wires is long enough. Sometimes the same resistors are required in the bases of the output transistors. Typically, the output stage is connected to the voltage amplifier either by the upper (according to the circuit) or the lower shoulder. The first stage of the repeater always operates without cutoff, in class A. The same signal current flows through VT1 and VT2, the voltages at their emitters must be exactly equal in amplitude. Therefore, it is considered acceptable to excite the output stage in one arm. This is correct only for traditional circuit design - when the transistor that sets the bias (VT7) is located in the collector circuit of the voltage amplification stage. The UN has a large output resistance, especially when connected with a common base, and is usually (if the circuit is asymmetrical, that is, excited only on one side) loaded onto a current source that has an even larger output (megohms). Therefore, there is practically no current through transistor VT7. We had to replace current sources with resistors. Under these conditions, a noticeable alternating current flows through the stabilizing transistor VT7. Therefore, its dynamic resistance and its nonlinearity can no longer be neglected. The direct current through this transistor is approximately equal to 1 mA (the current-setting resistors are 43 kΩ from the 44 V supply). The transistor itself is turned on with a gain of 6 times, since it sets the bias to 6 p-n junctions. Therefore, its dynamic resistance in such a connection is 6 times greater than the resistance of its emitter. As already mentioned, at this current the emitter resistance is 25 ohms. We find that the AC resistance of VT7 is 150 Ohms. This means that the signal is supplied to the second arm slightly weakened, by 3.5% (150 Ohm/43 kOhm = 0.035). This gives about 0.17% even harmonics. Capacitor C2 is turned on to bypass the dynamic resistance VT7, and this significantly reduces the THD. It would be more correct to send a signal to both shoulders at the same time. In conventional amplifiers (i.e. DC op amps), shunting also improves performance, but this comes at the cost of improving the symmetry of the RF base circuits. Blocking the phase difference in the halves of the push-pull stage suppresses distortion caused by unequal delays in the arms. When the output stage is powered at 44 V, the maximum amplitude value of the output signal will be approximately 4 volts less. This drop is the sum of the saturation voltage of the output transistors (about 1 1.5 V), the drop across the emitter resistors R9, R10 (also about 1 V). In addition, 0.65 V will remain at the emitter junctions of all three stages: after all, the signal voltage based on VT1 should not be higher than the supply voltage in order to avoid breakdown of the collector junction. The amplitude value of the output voltage of 40 V into a 4 Ohm active load will give 10 A collector current. This is a lot for the selected type of transistors. At this current, the cutoff frequency and gain of the transistors drops significantly. Transistors remain relatively linear up to a current of 2–3 A. Even the best imported transistors, specially designed for audio applications, lose their amplification and frequency properties when the collector current increases above 5–6 A. In addition, when the collector-emitter voltage decreases to several volts, the capacitance of the collector junction increases ten times or more. Therefore, it is undesirable to use this stage in this mode due to high distortion. The output power 2 U m will be P = 200 W, if the power supply allows. Each transistor in this 2 Rn 2 1 U pit case dissipates Pdiss = 50 W (in class B), which is quite acceptable if there are 2 π Rn sufficiently efficient radiators. But still, the amplifier works much better at an 8-ohm load, this is confirmed by measurements. If the load has a reactive component, then the dissipated power and collector currents increase.

12 The base current transfer coefficient of high-quality output transistors is usually in the linear region and up to in the high current region. For domestic transistors these values ​​are slightly lower, 1.5 2 times. For calculation purposes, the minimum values ​​are usually taken, because in the production of equipment, the selection of components is usually not allowed. Nobody will stop us from selecting transistors based on their gain and setting standard, not minimum, values. Despite the fact that the transistors in the emitter follower are covered by 100% negative feedback, it is better to ensure symmetry by design measures. The amplitude of the base current will be Ib = Ie / h21e = 10A/30 = 0.3 A. The pre-output transistors must supply this current. In real operating conditions, the current amplitude of transistors VT3, VT4 does not exceed 100 mA, but this is also a lot. At this current, few medium-power transistors can operate in the linear portion of the characteristic. Among the domestic transistors there are no such ones that would have an extended section with a constant h 21 Oe, have good frequency properties and would be complementary. Therefore, it is necessary to use either very low-frequency and nonlinear KT850/KT851, or, when power is reduced, KT940/KT9115 or KT639/KT961. Both of them are not complementary pairs, since they have significant differences in gain factors and cutoff frequency. Looking ahead, I note that transistors for output stages of TV or computer displays have good frequency properties and high linearity, such as 2SA1380/2SC3502 from Sanyo. They will be very good in the first emitter follower stage. If this amplifier were being made now, I would put available imported pairs 2SC1837/2SC4793 or 2SB649/2SD669 into the second stage. The output could have been Samsung TIP41C/TIP42C, Toshiba 2SA1302/2SC3281, Mospec or SanKen 2SC2922/2SA1216, Motorola MJ15003/Mj15004, etc., but at that time they were not available. In addition, I was interested in the contribution of each component, so the transistors were not selected according to parameters, only those with low gain or noticeable leakage were rejected. Power was supplied from an unstabilized source of sufficient power. The first question that had to be resolved was what quiescent current to set. To do this, a signal from the G3-118 generator, which has fairly low intrinsic distortion even without additional filters, was supplied to the input of the emitter follower. The amplifier was loaded with a resistive load equivalent of 4 or 8 ohms, and the signal was monitored by an oscilloscope and an automatic nonlinear distortion meter S6-11. Most measurements were made at a frequency of 1 kHz. At a quiescent current of 100 mA, the current amplifier showed a stable SOI result of about 3% over almost the entire power range. And only for a small signal, when the output transistors operate without cutoff, in class A, the harmonic distortion drops to 0.5-0.6%. By increasing the quiescent current to 3 A, we get 0.6–0.7% of the output power up to W. It’s worth making a big digression here regarding crossover distortion. On a small signal, while the signal current through the transistors (or lamps) is less than the quiescent current, the transistors of the arms work on the load simultaneously, then when the level increases, one of the transistors closes. This is equivalent to doubling the output impedance. That is, the dynamic characteristic has a sharp break. You can “see” crossover distortions this way: by connecting and disconnecting the load at a low level, use an oscilloscope to detect the “drawdown” of the output signal. Then increase the level and do the same operation. While the reinforcing elements work simultaneously, they practically do not notice the load change; when moving to class B, the drawdown is more noticeable. In practice, the mechanism is somewhat more complicated, since the output resistance of the transistors depends on the current through them, in addition, stabilizing resistors R9, R10 are connected in series with them. The value of these resistors greatly affects the amount of crossover distortion. There is some resistance, which at a given quiescent current provides a minimum of distortion. The optimum is obtained when the output impedance of the entire amplifier changes least during the transition from a small signal, when both arms are active, to a large one, when one arm is closed. That is, it is necessary to calculate the output resistance for a small signal (output voltage is near zero) and for a large one, when the emitter current is greater than the current

13 rest several times. For powerful transistors the simplified formula for calculating the resistance of the emitter body is not applicable; domestic transistors have never been accompanied by such data, so we will use data from the Internet. The website of the Danish company LCAudio provides a description of the amplifier The End Millenium. This is an amplifier without general feedback, so everything said above also applies to it. The output stage uses 200 watt SanKen 2SC2922 and 2SA1216, one of the best modern output transistors. I will give a table of the dependence of the emitter resistance on the load current, taken from there. The main feature that distinguishes these transistors is the relatively slow decay of output resistance at high currents, which is very useful for reducing distortion. Other high-power transistors have much lower output resistance (as well as gain and cutoff frequency) at high currents. Table 1. Load current Resistance, Ohm 100 ma 0.2 500 ma 0.10 1A 0.09 5A 0.08 10A 0.07 At a small signal, the output resistance of the amplifier will be m 1 1 Rout = (Rtr + R9) = (0 .2 + 0.1) = 0.15 Ohm, 2 2 B On a large signal R = R + R9 = 0.09 + 0.1 = 0.19. The difference, although not twofold, is there. out tr Consequently, there are also nonlinear distortions caused by a break in the dynamic characteristic. Let's calculate other combinations of quiescent current and resistance of stabilizing resistors. The linearity criterion will be the relative increase in output resistance during the time the current increases from zero to maximum: drout = (rb-rm)/rm in percent; The transistor resistance is obtained by interpolating the tabular data: Table 2. Current, ma R9, R10 Rm, Ohm Rb, Ohm drout, % ,1 0.15 0.17 0.1 0.12 0.17 0.2 0.17 0, 27 0.1 0.1 0.17 0.1 0.17 0.18 0.1 0.1 0, As can be seen from the table, stabilizing resistors greatly influence the nonlinearity of the output resistance. Their influence is greater, the higher the quiescent current is selected. The output resistance of the amplifier changes the least without these resistors (line 6) and The End Millenium (line 1). In the article “Current dumping: does it really work?” (Wireless World, 1978) Vanderkooy and Lipshits especially emphasized the advantage of amplifiers operating in class B - they have no crossover distortion. I think that a simple Current dumping amplifier (Radio N9, 1985), like the famous Quad 405, is not bad. sounds precisely for this reason. Concluding the analysis of this part of the circuit, I note that “seamless” joining of half-waves is possible if the transistors have ideal (that is, logarithmic) current-voltage characteristics, and the emitter and base resistances are equal to zero. If the voltage at the base junction of one of the transistors increases by 100 mV, the emitter current will increase 10 times. In this case, the voltage at the second junction

14 transistor will decrease by 100 mV and its emitter current will decrease by 10 times, but will not stop. The overall characteristic is not linear, but there is no sharp break leading to the appearance of high-order harmonics. In real conditions, the resistance in the circuits of the transistor electrodes has a non-zero value, so the decrease in the emitter current of the closed arm occurs faster than according to the logarithmic law. Therefore, the switching of the arms occurs faster and, most importantly, with a complete cutoff of the current of the arm being closed. If no additional measures are taken, switching distortions are of a high order and are practically not attenuated by the OOS circuit. The consequence of all that has been said is the presence of a certain region of the optimal regime. This is intuitively guessed without any thought experiments. However, most often amateurs make the wrong conclusion, believing that the quiescent current should be as high as possible. In fact, the optimal quiescent current of the output stage depends on many factors, among which the decisive ones are the resistance of the emitter resistors and the parameters of the transistors used. Of course, if the entire amplifier operates in gain class A (that is, the current through the transistors never stops), many of the problems described are automatically eliminated. But still, true class A is quite difficult to implement in high-power transistor amplifiers. One problem is replaced by another. An indirect indicator of complexity can be the almost complete absence of such amplifiers on the market. The only monsters that come to mind are Mark Levinson, AM audio, Accuphase A50, Nelson Pass single-ended amplifiers, and the old 12-watt Sugden A21. Many manufacturers, declaring amplifiers as “Pure class A”: Plinius SA100, SA102, SA250, Musical Fidelity A2, etc., are clearly wishful thinking. Just look at the dimensions, weight, radiator area and power consumption to be convinced of this. Most likely, they operate in class A up to watts of power, like the top models from Pioneer, Sony, etc. The problem of thermal stabilization and energy supply of the cut-off mode at output powers W is solved quite simply. When trying to obtain more power, the designer is faced with the task of ensuring the normal operation of all components over the entire temperature range of operation, as well as with a sharp increase in the cost of the entire structure. Therefore, the vast majority of industrial amplifiers operating with high quiescent current have a break in the amplitude characteristic in the region of medium powers. As has already been shown, the higher the quiescent current, the more the output resistance changes when switching. This change is a prerequisite for the occurrence of distortions. All efforts of designers are aimed at optimizing the switching speed of transistors. In this case, the spectrum of distortions moves to the low-frequency region, where they are quite effectively suppressed by the negative feedback. An abundance of trademarks “class A+”, “AAA”, “economical A”, etc. indicates the marketing attractiveness of the “Class A” badge, but even the simplest calculations indicate that the least problems will occur with a reasonable choice of the quiescent current at the ma level. Let's return to our diagram; the smallest integrated SOI of the final amplifier was obtained at a quiescent current of ma. Without a weighing filter it is about 0.5%. Most likely, by selecting the value of the emitter resistors and the quiescent current, this value can be further reduced. The pre-output cascade operates with a quiescent current of 35 mA. Signal cutoff in one of the arms occurs when signal currents are close to the maximum, that is, most of the time the cascade operates in class A. Of course, switching the transistors of the pre-output stage also changes the output current and causes distortion. Typically, designers try to transfer the switching moment to the region of statistically rare amplitudes. The first stage of the current amplifier has a quiescent current of 4 mA. This is enough to ensure that the current through the transistors is not interrupted over the entire range of signals and loads, including when short circuit loads. The mode of this stage is selected as usual, in the region of a stable gain of the applied transistors. Before moving on to the analysis of the input stage, I will note the role of the Bouchereau chain R11C3. Its task is to ensure a favorable load of the output stage at frequencies above audio frequencies, that is, more than 50 kHz. At HF, the load (speaker systems with cable) always has a reactive nature with random

15 module and phase. Therefore, various RLC circuits are used to match the amplifier and the RF load. The best results are provided by a two-link chain like . As already mentioned, the composite emitter follower VT1-VT7 has a sensitivity of about 35 V rms. Its input resistance is almost completely determined by resistors R3, R4, connected in parallel with alternating current. Thus, the input resistance does not depend on the signal amplitude (which has a beneficial effect on the linearity of the amplifier) ​​and is com depending on the value of R3, R4. Power consumed by the final U input stages from the voltage amplifier: Pc = = 0.06 W. Rin 20k The choice of an ultrasonic electronic tube as an amplification element is justified mainly by the simplicity of the solution and the predictability of the result. It would be possible to use semiconductors, but, firstly, this was already tested in previous work, and secondly, the microcircuit-transistor UN, with which this output stage previously worked, did not work well. To check the linearity of the voltage amplifier, we will assemble a rheostatic stage on a triode with common cathode, rice. 8. Fig. 8. Rheostatic triode stage. A signal from a sinusoidal generator with a voltage of 1–3 V is supplied to the input of the cascade. Resistor R4 is a load resistor. The voltage from it is supplied to the nonlinear distortion meter. The purpose of the experiment is to select a lamp that allows you to obtain the highest output voltage with minimal distortion. The anode resistance to alternating current in this circuit is less than 7 kohms, so the internal resistance of the lamp must be much less than this value, otherwise it will not be possible to obtain sufficient gain. To study the cascade, the input voltage is gradually increased until the level of nonlinear distortion begins to sharply increase. The peak output voltage (as measured by an oscilloscope) and the SOI level are recorded. Table 3. Lamp Quiescent current, mA Uout.max. (Peak)V SOI, % 6N6P N23P N1P Table 3 shows the measurement results with some widely used lamps. As you would expect, low-resistance ones allow you to get higher voltage. Therefore, 6N23P was chosen, which also has a relatively high gain. Despite

Fundamentals of circuit design FUNDAMENTALS OF CIRCUIT DESIGN...1 1. BASIC PROVISIONS...1 2. Amplification OF WEAK SIGNALS...6 3. Amplification OF STRONG SIGNALS...14 4. FUNDAMENTALS OF AMPLIFIER CIRCUIT DESIGN...18 1. Fundamentals

Lecture 6 Topic Amplifier stages based on bipolar transistors 1.1 Power supply of amplifiers. Applying a bias to the input of the active element The position of the initial operating point is determined by the polarity and voltage value

Lecture 8 Topic 8 Special amplifiers Direct current amplifiers Direct current amplifiers (DC amplifiers) or amplifiers of slowly varying signals are amplifiers that are capable of amplifying electrical

Chapter 5. Differential Amplifiers 5. Differential Amplifiers A differential amplifier is a symmetrical amplifier with two inputs and two outputs, used to amplify voltage differences.

Lecture 9 Topic 9 Output stages 1.1 Power amplifiers (output stages) Power amplification stages are usually output (final) stages to which an external load is connected, and are designed

Lecture 7 Topic: Special amplifiers 1.1 Power amplifiers (output stages) Power amplification stages are usually output (final) stages to which an external load is connected, and are designed

Lecture 8. Power amplifiers Feedback in amplifier stages. Cascode circuits. Plan 1. Introduction. 2. Power amplifiers 3. Feedback in amplifier stages 4. Cascode circuits. 1. Introduction.

Lecture 5 Topic 5 Feedback in amplifiers Feedback () is the transfer of part of the energy of the amplified signal from the output circuit of the amplifier to the input circuit. Figure 4 shows the block diagram of the amplifier

K548UN1 Integrated dual multi-purpose pre-amplifier. This technical specification is for informational purposes only and cannot replace a registered copy of the technical specifications.

Lecture 11 Topic: Analog integrated circuits (Continued). 1) Operational amplifiers. 2) Op-amp parameters. 3) Op-amp circuitry. OPERATIONAL AMPLIFIERS Operational amplifiers (op-amps) are called amplifiers

Lecture 6 Topic 6 Temperature stabilization of amplifying elements Dynamic characteristics of an amplifying element In real circuits, a load is usually connected to the output of amplifying (active) elements

TOPIC 6 ELECTRONIC AMPLIFIERS. An electronic amplifier is a device that converts a low-power electrical signal at the input into a higher-power signal at the output with minimal distortion. By functional

Class A pseudo-push-pull output stages Push-pull or single-cycle voltage followers are most often used as an output stage. The construction of push-pull output stages is hampered by the lack

Nizhny Novgorod State University named after N. I. Lobachevsky Radiophysics Faculty Report on laboratory work 5 Aperiodic amplifier Performed by students of group 430 Nizhny Novgorod, 2018

84 Lecture 9 VOLTAGE STABILIZERS Plan 1. Introduction 2. Parametric stabilizers 3. Compensation stabilizers 4. Integral voltage stabilizers 5. Conclusions 1. Introduction For the operation of electronic

LABORATORY WORK 3 RESEARCH OF CHARACTERISTICS AND CALCULATION OF PARAMETERS OF A DIFFERENTIAL AMPLIFIER CASCADE (MA) The purpose of the work is to become familiar with the principle of operation of the AC; familiarization with the circuit and principle of operation of the source

303 UDC 621.3 POWER AMPLIFIER CIRCUIT STRUCTURE WITH CURRENT FEEDBACK Kolosha I.S. Scientific supervisor Mikhaltsevich G.A., senior lecturer Let's consider a simplified block diagram of a power amplifier with

0. Pulse signal measurements. The need to measure the parameters of pulse signals arises when it is necessary to obtain a visual assessment of the signal in the form of oscillograms or readings measuring instruments,

Lecture 8 Topic: Integrated amplifiers 1 Direct current amplifiers Direct current amplifiers (DC amplifiers) or amplifiers of slowly changing signals are called amplifiers that are capable of amplifying electrical

Amplifiers AMPLIFIERS WITH FEEDBACK Feedback is widely used in a variety of semiconductor electronics devices. In amplifiers, the introduction of feedback is intended to improve a number of

Electronics Direct current amplifiers (DCA) Purpose: amplification of signals slowly varying in time, including a direct component. In the UPT, elements cannot be used as communication elements,

4. BASIC CIRCUIT CONFIGURATIONS OF ANALOG CIRCUITS AND DC AMPLIFIERS 4.1. Differential amplifier stage, its main properties and circuit implementations Features of analogue design

3. FEEDBACK IN Amplification Paths 3.. Block diagram ideal controlled source with single-loop negative feedback (NFE) and its use to analyze the influence of NFE on parameters and

“Electronic choke” Evgeniy Karpov The article discusses the features of the operation of an electronic power filter and the possibility of its use in sound-reproducing equipment. The motivation for writing

MINISTRY OF GENERAL AND PROFESSIONAL EDUCATION OF THE RUSSIAN FEDERATION REL 2 NOVOSIBIRSK ORDER OF THE RED BANNER OF LABOR STATE UNIVERSITY Faculty of Physics Department of Radiophysics Bipolar

109 Lecture CIRCUITS WITH DIODES AND THEIR APPLICATION Plan 1. Analysis of circuits with diodes.. Secondary power supplies. 3. Rectifiers. 4. Anti-aliasing filters. 5. Voltage stabilizers. 6. Conclusions. 1. Analysis

Lecture 7 Topic 7 Pre-amplifiers, their basic and equivalent circuits Dynamic characteristics of the amplifying element In real circuits, the output of amplifying (active) elements is usually

3. Transistor amplifier stages (calculation for alternating current) Introduction The tasks given below are related to the calculation of the parameters of amplifier stages, the circuits of which were calculated for direct current in the previous

Chapter 4. Operating modes of amplifying elements 4.1 Mode A This mode is characterized by the fact that the rest point is selected in the middle part of the load current-voltage characteristic (load straight line) of the amplifying element used for operation

SPECTRA - II Evgeniy Karpov The article discusses the results of studying the parameters of various types of lamps in a cascade with a current source in the anode circuit. The parameters of the electrical modes of these lamps are given,

Protection of the power supply from overload. (with modifications) Consider the original circuit shown in Fig. 1. And let’s take the GT404D transistor as an example as VT1. According to reference data, the static coefficient

STABILIZED SINGLE-CYCLE CASCADE ON A VACUUM TRIODE Evgeniy Karpov The article provides a diagram and discusses the operating principle of a tube output stage with increased linearity. This article is a logical

MINISTRY OF EDUCATION OF THE RUSSIAN FEDERATION MOSCOW STATE INSTITUTE OF ELECTRONICS AND MATHEMATICS (TECHNICAL UNIVERSITY) Department of Electronics and Electrical Engineering Guidelines to implementation

280 Lecture 27 CIRCUIT DESIGN OF OPERATIONAL AMPLIFIERS Plan 1. Introduction. 2. Operational amplifiers based on bipolar transistors. 3. Operational amplifiers based on MOS transistors. 4. Conclusions. 1. Introduction Operating

LECTURE 13 BIPOLAR TRANSISTORS Dynamic and key modes of operation of a bipolar transistor Lesson plan: 1. Dynamic mode of operation of a transistor 2. Key mode of operation of a transistor 3. Dynamic

MINISTRY OF GENERAL AND PROFESSIONAL EDUCATION OF THE RUSSIAN FEDERATION REL 3 NOVOSIBIRSK ORDER OF THE RED BANNER OF LABOR STATE UNIVERSITY Faculty of Physics Department of Radiophysics Field

5.3. AMPLIFIER CASCADES ON BIPOLAR TRANSISTORS In an amplifier based on a BT, the transistor must operate in the active mode, in which the emitter junction is biased in the forward direction, and the collector junction in the reverse direction.

5.12. INTEGRAL AC VOLTAGE AMPLIFIERS Low frequency amplifiers. ULF in an integrated design is, as a rule, aperiodic amplifiers covered by a common (direct and alternating current)

Draft patent application Compound vacuum triode and method of its use in low-voltage amplifier stages A known method of using a conventional vacuum triode in amplifier stages with low-voltage

TOPIC 7 Temperature stabilization As the ambient temperature rises, the transistor current increases and its characteristics shift upward (Fig. 1). Fig. 1 Emitter stabilization. Is to use

Laboratory work # 2 (19) Study of the characteristics of a bipolar transistor and an amplifier based on a bipolar transistor. Purpose of the work: Study of the current-voltage characteristics of a bipolar transistor and amplifier

POWER AMPLIFIER Oleg Stukach TPU, 30 Lenin Avenue, Tomsk, 634050, Russia E-mail: [email protected] Power amplifier A characteristic feature of power amplifiers is the high absolute value of the output

Ultralinear pentode mode in preliminary amplification stages Evgeniy Karpov When designing tube amplifiers, the problem of obtaining a given clock transmission ratio often arises.

ZHUSUPKELDIEV SH, TUTKABAEVA B. [email protected] STUDYING THE DIFFERENTIAL CASCADE OF AN OPERATIONAL AMPLIFIER IN AN ELECTRONICS COURSE FOR CHEMISTRY STUDENTS Kyrgyz National University

Topic 4. Inverters and batteries (2 hours) An inverter is a device that converts direct voltage into alternating voltage. The need for inverters exists to solve the problem of powering devices for household

Work 4.7. Study of multistage power amplifiers Single amplifier stages, as a rule, cannot provide the required voltage, current and power gain. To get the necessary

7. Basic elements of digital integrated circuits. 7.1. Diode-transistor logic A transistor cascade operating in a switching mode can be considered as a two-state element, or logical

1 Lecture 7. AMPLIFIER CASCADES ON FIELD TRANSISTORS. MATCHING PROPERTIES OF AMPLIFIER CASCADES ON BIPOLAR AND FIELD TRANSISTORS Plan 1. Introduction. 2. Amplifier stages based on field-effect transistors.

Nonlinear distortion coefficient meter S6-1 The device (Fig. 8-5) is designed to measure the nonlinear distortion coefficient of audio frequency voltage in the range of 50 Hz...15 kHz with a symmetrical input

SIGNAL AMPLITUDE MODULATORS WITH POWER 10...100 W RANGE 10...450 MHz (Electrosvyaz. 2007. 12. P. 46 48) Alexander Titov 634034, Russia, Tomsk, st. Uchebnaya, 50, apt. 17. Tel. (382-2) 55-98-17, E-mail:

Moscow State University named after. M.V. Lomonosov Faculty of Physics Department of General Physics Laboratory practice in general physics (electricity and magnetism) V M. Bukhanov,

Mordovian State University named after N.P. Ogarev Institute of Physics and Chemistry Department of Radio Engineering Bardin V.M. RADIO TRANSMITTER DEVICES, POWER AMPLIFIERS AND TERMINAL CASCADES OF RADIO TRANSMITTERS. Saransk,

Chapter 5. AC VOLTAGE AMPLIFIERS 5.1. PRINCIPLE OF ALTERNATING VOLTAGE Amplification Purpose and classification of amplifiers. AC amplifiers are the most common type of electronic

K.V. Kireev (student), V.M. Tchaikovsky (Ph.D., Associate Professor) DEVELOPMENT OF A FUNCTIONAL GENERATOR OF LINEAR VARIABLE VOLTAGE Penza, Penza State University Depending on the electrophysical

(attention, a typo has been noticed: in the previous part, the reasoning about Motorola is slightly incorrect. I hope to correct this in subsequent editions) Again, catching up. I wonder how the amplifier will work

Separation filter Evgeniy Karpov, Alexander Naidenko The circuit and design of a separation filter for implementing a two-way playback system is considered. The filter is implemented as a separate, autonomous

AMPLIFIERS Most passive sensors have very weak output signals. Their magnitude often does not exceed a few microvolts or picoamps. On the other hand, the input signals of standard electronic

The amplifier is designed for linear amplification of single-sideband, telegraph and AM signals in the ranges of 10...80 m. When amplifying telegraph and AM signals (in carrier mode), the input power is 200 W, when amplifying single-sideband signals, the average input power (when pronouncing a long “a” in front of the microphone) is also 200 W, while the peak input power can reach 400-500 W. The amplifier efficiency is 65-70% depending on the operating range. The amplifier uses four G811 lamps connected in parallel according to the OS circuit (Fig. 1).

A. Jankowski (SP3PJ)
Despite the general trend of using semiconductor devices in all technical devices However, we must not forget that tube HF power amplifiers (with an output power of more than 100 W) are much simpler to manufacture and more stable in operation. Experimenting with transistor devices is an expensive pleasure, because as someone aptly noted, no one dies so quietly, so quickly and as surely as a transistor.

Who needs power amplifiers? Few amateurs work with QRP, but most sooner or later begin to dream of increasing the transmitter power. However, you must be aware that in order for the correspondent to notice a change in signal strength of one S scale point (6 dB), the output power of the transmitter must be increased fourfold, and it does not matter whether it is a local connection or a QSO with DX.

Vyacheslav Fedorchenko (RZ3TI), Dzerzhinsk, Nizhny Novgorod region. Many radio amateurs construct short-wave power amplifiers using direct filament lamps, such as GU-13, GK-71, GU-81. These lamps are inexpensive, easy to use, have a highly linear characteristic and do not require forced cooling. Main positive quality These lamps are ready for use within one to two seconds after power is applied. According to the proposed description, more than a dozen structures were manufactured, which showed excellent technical specifications, good repeatability, ease of setup and operation. The design is designed to be repeated by averagely qualified radio amateurs.

V. Gnidin UR8UM (ex,UR4UAS) I took the amplifier circuit from the article by V. Drogan (UY0UY) as a basis. “HF power amplifiers” I simplified the circuit a little, adapting it to fit the parts I have, so to speak, into a budget option. I present for review what happened.

Oleg Platonov (RA9FMN), Perm
This amplifier operates on the amateur bands 3.5-28 MHz. At power input signal 25...30 W its output power in SSB mode on the 3.5-21 MHz bands will be at least 600 W and at least 500 W on the 24 and 28 MHz bands. The input impedance of the amplifier is 50 Ohms.

It is made on two GMI-11 pulse generator tetrodes, connected in parallel according to a circuit with a common cathode

Using a hybrid amplification circuit and impedance matching with an input P-circuit, we pump the signal up to a power of 150-160 W at an anode current of two GU-50s - about 300 mA in key-press mode. It is also advisable to control the current of the screen grids and not exceed its value more than 40 mA for two lamps. 250V x 0.02A = 5W - the maximum permissible level of power dissipation on the screen grid for one lamp. A protective diode will protect the stabilizer transistor in the event of a possible lamp shooting through the grid.

Typically, a power amplifier for a radio station or HF transceiver is built on "GU..." type lamps or on powerful high-frequency transistors. Both of these options may not always be acceptable. GU series lamps are relatively scarce, and powerful RF transistors, although they can be purchased, are prohibitively expensive. In addition, to build an output stage with a power of more than 100 W, several such transistors will be required, plus labor-intensive high-frequency transformers. The power amplifier described in this article is built according to a hybrid circuit using two relatively affordable transistors (KT610A and KT922V) of medium power, and one 6P45S lamp, which was widely used in the horizontal scanning output stages of tube TVs and, in this regard, is also relatively accessible and cheap.

I. AUGUSTOVSKY (RV3LE), Smolensk region, Gagarin The idea of ​​​​building a push-pull amplifier using electronic tubes is not new, and the circuitry of this amplifier, in principle, is no different from the circuitry of construction push-pull amplifiers on transistors. It should be noted that current lamps work best in this circuit, i.e. lamps with low internal resistance, which are capable of providing a significant pulse of anode current at a low supply voltage. These are lamps of the 6P42S, 6P44S and 6P45S types. However, I was able to build an amplifier with good characteristics using a GU-29 type lamp.

Hello everyone.

I will continue about the final cascade of Alexander Pavlovich Deriya.

At the beginning of 2017, I published the circuit of Alexander Pavlovich’s completed amplifier on this site, and at the same time, to discuss this circuit, I published it on AP and on diyaudio.ru

During the discussion at the AP, many questions were raised, and these discussions were not in vain.

DIY has a lot of manners and vomiting, like give me an amplifier with a transformer ass

or oh, it’s a pity that I’m standing in line at the hospital right now. Otherwise I would have taken a picture with a glass. So take a picture. You don't have to drink. Although it's a pity... In general, moderation on this forum “ordered to live.”

Yes, the sad and vile is also present, and happens, on some forums.

This is a classic ITUN with all that it implies. If you include resistances of 0.5 ... 1 Ohm in the emitters of the output transistors (and the corresponding resistors in series with the bias diodes), the distortion will decrease significantly. And the thermal stability of the quiescent current will become much better.

Alexander Pavlovich drew conclusions and decided to experiment with complementary pairs at the output and field-effect transistors at the input.

The main idea belongs to Alexander Pavlovich. and to describe it briefly - “then you don’t need to be afraid of high output resistance”

We all love numbers, and this is also very necessary and good. As they say, a fact is a fact!

But the fact must not be disguised. It happens that the amplifier’s numbers are fine, but there is no sound.

And recent measurements have shown that the amplifier is linear from 20Hz to 20kHz and even higher. By -3dB 75kHz!!!

Personally, I was glad that it was possible to shoot from 10 parts, and up to an undistorted sine wave of 1000Hz 65 watts in the hybrid version.

The lamps used were 6Zh11P, 6Zh43P in triode and 6F4P in standard mode.

6P9, 6P15, 6E5P, 6E6P and IL861 and El861 were also tested

(I would like to note that the IL861 lamp is 20 volts)

The only thing that can be considered a “fly in the ointment” is the high output resistance from 6Om to -20 Om from Alexander Pavlovich’s prototype, and from 30 to 50 Om for my hybrid version, depending on the lamps used. The output impedance of the amplifier depends on the choice of driver.

Many people think “and know” that the high output impedance of an amplifier has a bad effect on the damping of the acoustics, but part of the small population still believes that the acoustics, moving mechanically in the opposite direction, create a field that also affects the amplifier no less than the amplifier affects the acoustics and, accordingly, the sound generally!

Some literature says that with an output impedance of 18 Ohm, acoustic damping is already a fact.

But the majority will not agree with this statement, since the closer to “zero” the output impedance of the amplifier, the more correct it is.

There is another opinion - that an output resistance within 10-20 Ohm has a beneficial effect on the final picture as a whole. The sound is not compressed, “cut off from the ground”, the panorama is expanded, it is easy to perceive, there is no fatigue even after several hours of listening.

Triode and pentode amplifiers also have different output impedances, but both have the right sound and have their pros and cons. There are so many ears, so many opinions.

The following photographs provide a rectangle at 1000Hz at 10kHz and at 20kHz. Load 5Om. From these it is clear that the amplifier is in perfect order. These are measurements of a purely transistor amplifier assembled by Alexander Pavlovich Deriya.

Amplifier sensing 1.5v

Power supply +- 24 volt transformer - overall power is only 80 watts (from the Radiotekhnika -101 amplifier)

29 Watts of undistorted sine wave!

0.dB - 20Hz - 20KHz

We couldn’t measure the low at -3dB, the high at -3dB -75KHz

Output impedance 20 ohms.

Looking ahead, a tube hybrid amplifier with the same circuit design produces 65 watts at 0.75v when powered at +- 38 volts

20Hz -0.25db 20kHz +1db 45kHz-3db

The output stage of the amplifier is provided in the following figure.

It can be organized with both common emitters and common collectors. IN latest versions we settled on the version with common collectors.

It is very convenient to mount transistors on a radiator without mica plates.

Below are two versions of the driver, 1988 and 2018

The field-effect transistor KP901 can be replaced with a conventional composite transistor KT972, this does not affect the sound quality, this transistor acts as a repeater. Resistors R11 and R12 can and should be replaced with 0.6 Ohm, the stability of the output stage will increase and distortion will decrease. It is advisable to install a Tsobel circuit at the output and install 56 Ohms parallel to the speaker, which will reduce the output impedance by 10-15%.

The quiescent current of the transistors and the zero level are set by resistors R7 and R10 when the values ​​decrease, the currents decrease, and when they increase, they increase. The quiescent current is set from 100 to 200 mA, it all depends on the size of your radiators. For example, in the hybrid version I generally set it to 280 mA, and this is not the limit.

IMPORTANT! It is imperative to install a selected complementary pair; if this is not done, the modes may “float away”.

When assembled correctly, the amplifier works immediately

Below is the hybrid version of the amplifier. Power supply +- 38 volts. Anode 200 volts. EL861 driver tubes.

Ctr of the transformer 12.5/1/1 The primary winding is wound with wire 0.25-0.33 3000 turns Secondary 2X240.

I wound on OSM 0.063. Winding was carried out in the following way.

900 turns first — 120 turns sec. — 1200 turns first. — 120 turns sec. -900 turns first

The secondary wire is wound with a double wire from 0.33 to 0.51. Each layer was laid with graph paper.

The transformer is not phase inverted. The role of the bass reflex is performed by the output stage. This is a big plus in this circuit design. I also think it’s a plus that the transistor collectors are screwed directly to the radiator without mica spacers.

The amplifier is assembled in a 6mm plywood case. Plywood dampens hum from transformers well, vibration is not transmitted to the lamp grids. At 65 Watt output, the background noise is minimal. At 100 dB acoustics it is barely audible if you stick your head into the speaker.

Metal top and bottom.

I will provide an additional photo and video report when I finish the installation.

Sincerely, Evgeniy Vilgauk Chelyabinsk