ENCYCLOPEDIA OF RADIO ELECTRONICS AND ELECTRICAL ENGINEERING Acrobatics of lamp cascades. Encyclopedia of radio electronics and electrical engineering Encyclopedia of radio electronics and electrical engineering / Tube Power Amplifiers Everyone who is at least a little familiar with tube circuitry knows that tube amplifying stages are usually distinguished by their extreme simplicity and a small number of elements. This factor, along with the natural linearity of the tubes, is usually cited as an argument when trying to explain the phenomenon of the superiority of tube sound over transistor sound. It must be admitted that such an explanation is very convincing from the point of view of common sense. In addition, it is so often confirmed in practice in circuit analysis of the best tube audio components that few people think of trying to challenge it. The main motto of the developers of lamp technology is this: the simpler, the better and more reliable (unfortunately, the concept of "cheaper" is not included here, although logically it seems to suggest itself). So, let's look at a conventional low-power resistive amplifying stage on a triode with a common cathode. The anode load resistor, the cathode auto-bias resistor, the grid leakage resistor and the triode itself - that, in fact, is the whole cascade. More precisely, its basic version (Fig. 1).
The rest is either elements of communication with other stages, or blocking of local negative current feedback (shunting the cathode resistor with a capacitor), or a divider in the cathode circuit for a more complex organization of bias, or decoupling filters for power circuits, or correction circuits. Usually, even the presence of all these additional components does not make the tube amplification stage much more complicated than what we see in fig. 1. Everything is extremely clear and simple (at first glance). It is known that the stage gain in the middle of the frequency range is (in the absence of local negative feedback): K=-Ra/ (Ri+Ra) (taking into account the input resistance of the next stage Rinlet 2 instead of Ra R is usedn.eq=Ra||Rinlet 2, and the output resistance ZO=RiWhere=SRi - voltage amplification factor of the lamp; S - steepness; Ri - internal resistance of the lamp; Ra - anode load resistance. It is known that for such a triode stage, the real gain is usually (0,6-0,8) and depends on Ra, as well as other stage parameters: quiescent current, bandwidth, slew rate, linearity, maximum undistorted output voltage, maximum output current. Usually Ra several times greater than Ri, while it is possible to obtain acceptable values of the listed parameters. But the possibilities of a cascade on a triode are limited, and since, in pursuit of one parameter, others, no less important, usually suffer, the degree of freedom for varying the resistance values of the anode load and cathode auto-bias is small. The same can be said for the anode supply voltage and quiescent current, since almost all lamps "sound" best at the edge of the anode power dissipation (although not always). However, even within these relatively narrow "limits of creativity" it is not so easy to find the optimal mode of operation of a particular lamp in a particular cascade, taking into account the previous and subsequent cascades. In this case, the optimal mode is understood as the mode that will provide the best sound, and not record parameters or beautiful oscillograms. Perhaps it is the mutual contradiction of the various parameters of the amplifying stage and the ambiguity of their dependence on the same factors that are the reason for the weak correlation between the digital values of these parameters and the sound quality. So, if you are chasing maximum linearity, you have to increase the value of the anode load, which, starting from a certain value, will adversely affect the bandwidth, the dynamic properties of the cascade, and the gain, which, with an exorbitantly high load resistance, begins to decrease, since it decreases quiescent current and lamp slope. In addition, the overload capacity of the cascade also drops sharply. Thus, the price for ultra-high linearity is also prohibitive, since you have to pay with the sound quality of the device as a whole. It turns out that we pay with sound quality for linearity, and not vice versa, as it should be. This is reminiscent of Krylov's fable "The Swan, the Crayfish and the Pike", only the swan in this case is not a bird (and not a general), but the amplification factor, the cancer is the linearity of the cascade, and the pike ... In a word, things are still there. Where these intractable characters are in relative peace and harmony. Therefore, if one stage on the triode cannot provide the necessary amplification, a second one has to be installed. And in order to obtain good dynamic properties, sometimes one has to be content with a modest gain, reducing the anode load and increasing the quiescent current of the stage. Even in the simplest amplifying stage, a lot of subtleties and phenomena that are difficult to explain come up when it comes to the "last judgment" - listening. So, let's summarize: in the amplifier stage on a tube triode, various parameters, each of which has a tangible effect on the sound quality of the entire device, are in mutual contradiction, and excessive zeal when "pulling out" one of these parameters inevitably leads to deterioration of others. However, there is a way to break out of this vicious circle. After all, so far we have been talking about the amplification stage on a single triode. And if you combine two triodes in the same stage? This, of course, goes against the concept of maximum simplicity, but sometimes, instead of going for an increase in the number of simple cascades, you can solve the same problem by complicating (and not very significant) one cascade. Depending on what kind of task is set, you can choose one of the options for such a complicated cascade on two triodes. I must say that there are quite a lot of them and they were invented a long time ago. For example, a cascode (Fig. 2) allows a sharp increase in gain and at the same time broadband, and therefore, along with pentodes, it has found wide application in television and radio receivers. Some world-famous High End companies use cascodes in audio frequency amplification devices (for example, Sonic Frontiers).
One can argue about the advisability of using cascodes in audio equipment, and opponents of this usually refer to the fact that the output characteristics of cascodes degenerate from triode to pentode. Yes it is. But after all, pentodes are not always bad - the question is rather not what to use, but how and where. Undoubtedly, in most cases, the triode is preferable, but in individual circuits (most often auxiliary ones), the pentode has no equal. For example, thanks to the high and Ri the pentode is unmatched in stable current sources, except for insulated gate field-effect transistors. But this is a completely different world, and although companies like Audio Research have had some success in developing and implementing hybrid topology, I personally have no doubts that if pentodes were used instead of MOSFETs, many of their products sounded much more musical. And let's remember the professional tape recorders of the golden era of magnetic sound recording of the 50s and 60s (for example, Telefunken). Many of them had an EF86 pentode (similar to 6Zh32P) in the first stage of the playback amplifier. But let's return from attempts to amnesty pentodes sentenced to life by many audiophiles to pure triodes. The next cascade we'll look at is a lot like a cascode. These are also two triodes, one of which is "perched" on the shoulders of the other. Yes, this "tube circus" causes a skeptical smirk in many, and, probably, it can be followed by a stream of moralizing remarks like "a man - I'm sorry, a triode - must walk on the earth!" But one way or another, this cascade deserves attention, since it provides a simultaneous tangible improvement in several important parameters: mode stability, linearity, output impedance, broadband, overload capacity and sensitivity to interference and ripple of the anode supply voltage. As far as sound is concerned, everyone knows that the Audio Note and Saga Audio Designs amps don't sound that bad at all! It is these companies that are most often used as an input or driver stage, shown in Fig. 3a. It is most often called SRPP (SRPP - Shunt Regulated Push Pull).
Let the decoding of this abbreviation not mislead you: the "push-pull" here is expressed only in the anti-phase signals of the upper and lower triodes. With the same success, a classic circuit of two triodes connected in cascade could be called a "push-pool" - there is also an anti-phase signal. Thus, SRPP is not a completely correct name that has taken root in the literature. You can also see the abbreviation TTSA (Two Tube Series Amplifier - a two-tube series amplifier), although it can rather serve as a general label for all stages of a vertical configuration, including cascodes. In Russian, our cascade is called simply and clearly: an amplifying cascade with a dynamic load. And it is this name that most accurately reflects its essence (that rare case when the Russian language turned out to be more concise than English). There is also a more exotic Russian name - a cascade with "electronic resistors" in the anode load circuit (TV Voishvillo. Amplifying devices. M., Svyaz, 1975). So, instead of the usual anode load resistor, the SRPP cascade has a second triode in the anode circuit, the grid bias of which is set by the resistor Rк2. When a positive half-wave of the signal appears on the grid V1, the current of the lower triode increases, which leads to an increase in the voltage drop across the resistor Rк2, and this, in turn, reduces the current of the upper triode V2. There is a trend towards stability of the anode current, which is now less dependent on changes in the input signal than in a conventional resistive amplification stage. Combined load - triode V2 and resistor Rк2 - in terms of its properties, it begins to approach a source of stable current. What's good about that? It is known that a stable current source has a high internal resistance, which is equal to infinity for an ideal current source (this, of course, is a mathematical abstraction). And now remember that the triode cascade is the more linear, the higher its load resistance. It is not possible to solve this problem head-on, as mentioned above (by arbitrarily increasing the anode load), since other equally important parameters of the cascade suffer. It remains only to "deceive" the gullible triode V1, while its load resistance "doubles": for direct current it is small and equal to (Rк2+Rivk2), which ensures the normal mode of the cascade without increasing the voltage of the anode supply, and for alternating current (or dynamic load resistance) it can be much larger, and is determined by the value of Rк2 and voltage gain of the upper triode: Rn. din.=Rк2(1 +)+Ri(V2). This makes it possible to obtain a slightly higher gain of the SRPP cascade compared to a conventional amplifying cascade. And since the output signal is taken from the cathode V2, the output resistance is much lower. In reality, in the case when such a cascade operates on a relatively low-resistance load, a very significant gain can be obtained both in gain and in bandwidth. Yes, and the dynamic properties, provided there is a sufficient quiescent current of the cascade, can be obtained very impressive (here it is important to take into account not only the speed of the cascade, but also how large the signal current can be given to the load). For these reasons, the SRPP cascade found application in video amplifier circuits, where it was necessary to ensure the maximum value of the product, as well as in high-speed flip-flop circuits (A.P. Lozhnikov, E.K. Sonin. Cascode amplifiers. M., Energia, 1964), probably long before someone had the idea to try it in amplification circuits audio frequencies. Its advantages are especially pronounced when operating in circuits where the parasitic load capacitance is quite large (this category includes some driver circuits that operate on a large number of parallel-connected output lamps or on single lamps with a high dynamic input capacitance). On fig. 3b shows the dependence of the gain of the SRPP cascade on the 6N3P double triode (=35,Ri\u5,8d XNUMX kOhm) from the equivalent load resistance at various values of Rк2 (curve 1 corresponds to a conventional cascade with a common cathode, the rest - SRPP: 2 - at Rк2=360 Ohm; 3-Rк2=560 Ohm; 4-Rк2\u820d 3 Ohm) In fig. XNUMXc shows the dependence of the output resistance of the SRPP cascade on the value of Rк2. On fig. 3d are given for comparison the transient characteristics of the SRPP cascade (top) and the conventional cascade (bottom) on 6N3P (curve 1 - at Cн\u5d 2 pF; XNUMX-Cн\u15d 3 pF; XNUMX-Cн\u30d 4 pF; XNUMX-Cн=55 pF).
However, SRPP is not the ultimate dream. And for this reason: although the combined anode load of the cascade, as already mentioned, acquires some properties of a stable current source, but due to the relatively smallcharacteristic of triodes, V2 lacks the "amplifying power" to sufficiently compensate for the voltage drop across Rк2caused by a change in signal current. There are two ways to solve this problem: either use a pentode rather than a triode as V2, or increase the signal level on the V2 grid. The first path leads to the circuit shown in Fig. 4, and the second - to the so-called "reinforced SRPP", which is also obtained more complicated (Fig. 5).
The fact is that to significantly raise the signal level on the grid V2 simply by increasing the resistor Rк2 fails, since the position of the operating point of the cascade also depends on the value of the same resistor, and if you get carried away in this way beyond measure, you can lose all the advantages of the SRPP cascade (in the first place, the overload capacity will deteriorate). But you can go further along the path of deceiving gullible triodes, now "fooling" V2 as well: arrange for it the required grid bias using a divider (Rк2 Ra), which will replace Rк2, which will give more freedom in varying the signal level on its grid (which will be proportional to the lower resistor of the divider), and apply this signal through the capacitor Ca. The gain of such a cascade can already be made quite close to the lower triode (we must not forget that it is he who remains the main "actor" that determines the operation of the cascade, and everything else serves only to create the best "working conditions" for him). Therefore, the amplified cascade of SRPP in foreign literature is called "Mu Follower" - "repeater". And again, this spectacular name is somewhat arbitrary, since the enhanced SRPP, although it is selected quite close in terms of gain to the value lower triode, but still does not "repeat" it. In addition, it leaves the possibility, by using a pentode as the top lamp and further complication of the circuit, to further reduce the distance between the actual gain and the value lower lamp, while lowering the already low output impedance and expanding the dynamic range. This cascade (Fig. 6) on the pages of the magazine "Glass Audio" is called "(-cascade" (Allan Kimmel. The Mu Stage//Glass Audio, 1993, N2).
The structural features of this cascade provide ample opportunities for choosing the quiescent currents of the upper and lower lamps. The currents in this case can be different, since the pentode bias is set by a separate divider (Rк2, R'к2), which also contributes to a further decrease in the output resistance (and, obviously, to equalize it for the positive and negative half-waves of the signal of a sufficiently high level, when the "push-pull" effect can manifest itself, i.e., the steepness of the leading and trailing edges of a rectangular pulse in the general case can be different). The value of the anode load of the triode Ra can also be varied within certain limits. The pentode, on the other hand, can be considered as a cathode follower with a transmission coefficient very close to unity. Thus, any change in the instantaneous value of the voltage at the anode, or lower terminal of the resistor Ra, is tracked with high accuracy by the cathode follower on the pentode V2, appearing at the top terminal Ra, and therefore the voltage drop across Ra almost constantly and does not depend on the signal - this is the real (not ideal, of course, but very close to it) source of stable current. Of course, those who suffer from pentode allergies can also use a triode as V2, but they will get more modest parameters. A triode cathode follower typically has a K gain of about 0,9, while a pentode can easily provide 0,995 or more. Now let's take Ra equal to 6,8 kOhm and calculate the dynamic resistance of the anode load of the cascade: Rn. din.=Ra/(1-K). In our example Rn. din. triode.\u68d XNUMX kOhm, and Rn. din. pent.\u1,36d 20 MΩ. The difference is XNUMX times! Cathode followers, by the way, also enjoy a far from impeccable reputation among technically literate audiophiles. But, nevertheless, according to the same Allan Kimmel, in such a scheme, a cathode follower on a pentode is just what you need. In general, pentodes in cathode followers give much better results both in terms of parameters (lower output impedance and attenuation) and in sound. In addition, Allan Kimmel writes that he experimented for a long time with all the tube cascades described above in all possible ways, and all of them, being correctly implemented, sound very good, and best of all - precisely-cascade. It is especially good as a driver, "swinging" output triodes with a smallrequiring a large signal voltage swing. The parameters obtained by Kimmel-cascade (Fig. 7) are very, very impressive: output impedance 100 ohms, output signal swing 215 V at a harmonic coefficient of 0,7% and an anode supply voltage of 300 V, frequency range by level (-3dB) 0,28 Hz - 1 MHz.
The triode is the well-known 6DJ8 (analogous to 6N23P), both halves of which are parallel, which favorably affects the output resistance (according to Kimmel, he also did this because he could not come to terms with the fact that one half of the triode "dangled around idle" ). Pentode - 12GN7 (analogue is unknown, but this is hardly important: any pentode with a sufficiently high, capable of operating at the required quiescent current, which is easy to determine based on the recommended current mode 6N23P; 6Zh9P will certainly show itself well). But that's not the end of the story. In Glass Audio's N5 1996, Allan Kimmel published an article titled "A Direct-Coupled Mu Stage" (-cascade with direct connection), in which he brought an even more perfect work of circuitry art (Fig. 8).
It is difficult to say whether the idea of creating this cascade belongs to him, or whether he borrowed it from the old lamp literature (after all, it often happens that many innovations actually turn out to be twice as old as their "inventors"). Be that as it may, the idea is very original: if the previous cascades resembled a "living pyramid" in a circus arena, then this one draws on aerial acrobats with a flying trapeze. Lost Capacitor Ca, the connection between the anode of the triode and the control grid of the pentode is now galvanic; at the same time, a floating stabilized screen grid power source is introduced, and the anode of the triode also receives power from it. Initially, in this scheme, the goal was to exclude the chain R that "loads" the output of the cascadeэ Cэ, although her influence was not in any way dramatic. One way or another, the records of the parameters of the previous stage (Fig. 7) were broken: the output impedance decreased to 80 Ohms, the maximum swing of the undistorted output voltage reached 269 V with a harmonic coefficient of 0,9% and the same anode supply (300 V), the frequency range was over due to the absence of a transition capacitor Ca now starts with Fн(-3dB)=0,15Hz, Fв(-3 dB) remained the same: 1 MHz. In order not to rewind the power transformer, Kimmel found a very ingenious solution for organizing a floating source: he installed a small incandescent transformer and turned it on "back to front", applying an alternating voltage of 6,3 V to the secondary winding, and connected a rectifier bridge and a simple transistor stabilizer to the primary winding , from which the required 75 V is removed. This non-standard method is also good because such a compact power supply can be placed in close proximity to our cascade, thereby preventing the signal from "roaming" along the long connecting wires leading to a common power source. Although, if there is a good decoupling, this issue can probably be resolved in the traditional way - by using a power transformer with a separate winding. So, we have considered several tube circuits, each of which is characterized by a vertical configuration. There are other vertical cascades, most notably complex cathode followers (such as White's cathode follower). Since in this case we were talking about voltage amplification stages, we will not touch on cathode followers in this article. This is a separate life with its own sores and medicines for them. In addition, the considered types of amplifying cascades in many cases eliminate the need for cathode followers altogether, combining the properties of an amplifier and a buffer (just like the famous Pantin Pro-Vee shampoo with conditioner - two in one!). As often happens, each subsequent cascade has better parameters than the previous one, but at the same time it becomes more difficult. Further into the forest - more details. Therefore, I would like to advise those readers who decide to try something from this article "by sound" not to be maximalists and not to immediately aim at the "coolest" version of the above schemes, but to start simple. Who knows, perhaps, in a particular design of an amplifier or other device, some intermediate circuit in terms of complexity and parameters will sound best. Personally, at first glance, the closest thing to me (so far only speculatively) is the SRPP circuit with a pentode. Author: Artur Frunjyan; Publication: cxem.net See other articles Section Tube Power Amplifiers. Read and write useful comments on this article. Latest news of science and technology, new electronics: Machine for thinning flowers in gardens
02.05.2024 Advanced Infrared Microscope
02.05.2024 Air trap for insects
01.05.2024
Other interesting news: ▪ Wind Turbine Challenergy Resistant to High Winds ▪ The brain is more important than the muscles ▪ The robot fuels the spacecraft ▪ Hydrogel for flexible electronics News feed of science and technology, new electronics
Interesting materials of the Free Technical Library: ▪ section of the site The most important scientific discoveries. Article selection ▪ article Nabokov Vladimir Vladimirovich. Famous aphorisms ▪ article How much does a person eat and drink? Detailed answer ▪ article Head of the postal car. Job description ▪ article Telephone line controller. Encyclopedia of radio electronics and electrical engineering ▪ article 160m transceiver (UA1FA). Encyclopedia of radio electronics and electrical engineering
Leave your comment on this article: All languages of this page Home page | Library | Articles | Website map | Site Reviews www.diagram.com.ua |