Hello and welcome to the third video about the Echolette NG-51S. This time we want to discuss the Echolette Oscillator. Yes and I think we can roughly say that with this video we will be halfway through this series. I reckon it will take two more episodes before we have fully covered all deatils of the device. Yes, and even though it's been seven months since the last video: As you can see, it's worth staying tuned in! Visit here regularly, because there will definitely be more content on this topic and others. And if you don't have time to check back regularly, you can of course also subscribe to this channel - then you won't miss any more videos. Alright, so let's get right into it and take a closer look at the "echolette oscillator". Of course, we will first have a look at what the background is: what is an oscillator? What do we need it for, especially in tape applications? And of course we will take a relatively close look at the circuit. I have to say very briefly that we will skip some aspects of the circuit in this video a little faster and then not look into it in such detail, but there is an absolutely detailed description on my homepage and this description will be in German and English and the links to these PDF files that you can download can be found in the video description below this video. Well what is an oscillator? We can get closer to that via Wikipedia. I have reproduced here an entry from the German and English Wikipedia and it says in principle that an oscillator swings back and forth between two variable state. "Oscillates", from Latin "oscillare" - meaning "to swing". And this swinging back and forth between two states usually takes place around a central point, i.e. the point of rest of such a system. Yes, and if you naturally think of a swing when you're 'oscillating', you're not entirely wrong. A swing can swing forwards and backwards, and when nobody is swinging on it, it hangs straight down by the ropes. This is the resting point and you already have a certain idea of ​​what such an oscillator can be. Let's stay with such a mechanical oscillator for a moment and look at two other examples to perhaps clarify the principle a bit. We are now looking at two video examples of mechanical oscillators or things that are also oscillators: First, a metal disc is suspended between two springs and when the springs are at rest, the metal disc is in the middle. And we can now deflect that in either direction with one hand and let go, and then this disk swings back and forth on the springs. And the other example of a mechanical oscillator is the pendulum in a table clock. In principle, both examples of a mechanical oscillator describe such oscillators with a damped oscillation. This means that if we deflect this metal disk between the springs once - that is, practically putting force into it by pulling it in one direction and letting it go - then this oscillation will die down over time; it is a so-called damped oscillation, because of course there are also friction losses between the springs. We also have some losses because the spring is also moving through the air. So this oscillation will eventually come to a standstill again if we don't keep putting energy into it. For example, by deflecting it again by hand. Or with the clock, the older ones may remember that - you have to, of course, so that this pendulum continues to swing forever, you have to wind up a spring from time to time with a large key. Otherwise this pendulum would of course stop at some point and the clock would no longer move . Yes, and if we now take a look at the oscillator in electrical engineering and get a little closer to our actual topic, then of course there is also a whole range of wave generators and these can first of all be differentiated according to the type of oscillation they generate. And in audio technology you come across different oscillators, which usually produce four larger different waveforms. We see this here: a square wave generator. that creates a rectangle. Sawtooth generator creates a sawtooth. Triangle generator creates a wave that has a triangular shape and lastly the sine generator that creates a sinewave like the one we see at the bottom. What is now important to keep in mind for later understanding of the Echolette oscillator is that only the pure sine tone consists of a single frequency. All other oscillations that we see here: square, sawtooth and triangle - they consist more or less of a broadband frequency mixture with a more or less pronounced overtone spectrum. Yes, and the range of oscillators and oscillator circuits in electrical engineering is unfortunately huge. You can see that here in a small table, which I also have from Wikipedia. There are a lot of oscillator circuits and principles out there, but for the Echolette oscillator we're looking at today, very few of the keywords you see here on Wikipedia are actually at all relevant or of any content interest. I've underlined those in red that we could actually pull something out of for the understanding of the Echolette Oscillator. Other keywords that would help us to understand the oscillator circuit in the Echolette NG-51 S a bit better are, amusingly, missing from this Wikipedia listing. That would be, for example, this "Abraham-Bloch Multivibrator", which could definitely help us to understand it if it were to be found here in Wikipedia, for example. But one to one you won't find the Echolette oscillator here at all and that's why it's worth it if we take a closer look at the whole thing. I have given you a few practical examples of what an e-technical oscillator could look like - although it's not entirely true, you see two oscillators here, oscillator circuit boards, and the rest are basically just the coils from oscillators. For example, the top left is a tube oscillator. I assume it comes from a reel to reel machine, I can't remember exactly. Below left is the oscillator circuit board from a Dynacord Echocord, the transistor version of course. It can be found in more or less this form or in a modification in the Echocord Mini, in the Echocord Super 75 and 76 and in the Echocord 100 and so on and so forth. And at the top right, as I said, different oscillator coils and you can already see the coil and capacitor attached to it. There is a reason for this, which we will of course explain in detail later. But these pictures are just to illustrate how different such oscillators can look in tape recorders and tape echoes. Yes, and why is this oscillator needed now? Let's take a very simplified look at a few physical principles. Everyone may remember this from school: If we have a magnet, for example a bar magnet, and we bring an iron nail or an iron screw close to it, then it will be attracted by this magnet. But there is a second effect with this iron nail or this iron screw, which has itself become magnetic when we have brought it close to the magnet. Yes, and if we consider this situation or this connection, then we can also think up the following little thought experiment: Let's assume we have a magnetic field that is emitted by a coil. This coil emits a magnetic field because a direct current flows through it. For example we connected a battery to the coil. And we can either strengthen or weaken this direct current excitation of the coil, for example by increasing or reducing the voltage, and you can also change the direction of this direct current excitation, for example by reversing the polarity of our battery, i.e. connecting it the other way round. And if we now bring this iron nail or an iron screw close to this magnetic field, as in the school example just shown, then we can measure a) the strength of the magnetic field as it appears around this iron nail and we can b) measure the magnetic flux density, i.e. the magnetic flux that comes about because the magnetic field flows through this iron nail. Yes, and if we enter these two relationships in a graph on a piece of paper, then something always comes out that looks like this diagram on the left. Namely something like a loop. These loops are called "hysteresis loops" and in principle they look a little different for each magnetizable material, e.g. an iron nail, iron screw or even a piece of tape, which is coated with such a magnetizable layer. They look a little different for each of these workpieces, but always in the same loop shape. The first thing we can see from this loop is the relationship between the magnetic field strength and the magnetic flux that this field strength causes in the workpiece. This relationship is not always linear. This means that we have areas here in which the connection is linear. So I increase the magnetic field strength and the magnetic flux or the magnetic flux density increases linearly. But we also have areas here where this linearity no longer exists. We can increase the field strength here as much as we want, it no longer increases linearly here. Rather, the magnetic flux density seems to run into a kind of saturation here. So the ascent becomes flatter and at some point there is no more ascent at all. Then it just doesn't work anymore. So you can no longer magnetize an iron nail more and more, because at some point it will be fully magnetized. Then all magnetizable particles that are in this workpiece, are aligned in the magnetic field and more is simply not possible. It's exactly the same with tape; you can fully magnetize it and that's it. Yes, and as I said, this connection between H and B is not necessarily always linear. A double magnetic field strength does not always result in a magnetic flux density that is twice as strong. Instead, saturation occurs and that is the area marked in red here. If the magnetic field strength decreases, then B does not go back to the point of origin. So we have our point of origin right here in the middle, where the whole thing started. Instead, a residual magnetization remains. You can imagine it like this in the illustration. If we magnetize the workpiece, then it follows this curve here up to saturation and if we switch off the magnetic fields then it doesn't go back to the point of origin on this curve, but it goes back on this hysteresis loop on this upper leg and there remains a certain residual magnetization, the so-called "remanence". And that is of course the reason why music remains on a recorded tape when we switch off the tape recorder or take the tape out of the device and put it in the closet. So the music stays on here. In order to eliminate this remanence, a new magnetic field with the opposite field direction and a corresponding strength is required, this is the so-called "coercive force", which we have to use so that the residual magnetization in the material is at some point here in the center or at the intersection of the coordinates and so is zero. And the workpiece is no longer magnetic. Yes, let's take a very short break here and go back in history: Direct current or the magnetic fields generated by direct current were used very early on to erase first wires and then also magnetic tapes, but erasing a magnetic tape with direct current or with constant fields adds an audible background noise on the tape, which is then also heard in the recordings. Because it is precisely this recording alternating current that is recorded on the tape during recording that is normally not strong enough to completely overwrite this remanent magnetization by the DC field. This means that you always hear a diffuse noise from the deletion in the background. Yes, and the first US patents relating to AC quenching for wire recorders, which interestingly existed as early as the 1920s, but they are being forgotten again because this wire recording process has never really caught on. So you have to say that in the home area there were also wired audio devices in Germany up until the 1950s , in which music was actually recorded on such a small magnetizable iron wire. And these devices were also in competition with the tape recorder, but the whole thing got lost, it didn't catch on across the board. And in 1940/1941 this HF process, i.e. this erasing process using high-frequency alternating current, was rediscovered in Germany. And that was rediscovered by a Dr. Hans-Joachim von Braunmühl and Dr. Weber in Berlin. That's quite interesting if you look up this story on the internet. Apparently, as with so many discoveries, chance played a big part. But that would now lead too far in this video, but the story is very interesting to read up on. Back to the physical basics: Let's summarize the features of the high-frequency erasure process. So the deletion, the demagnetization is no longer done by a constant field, but by a rising and falling alternating field. How do you have to imagine rising and falling now? The alternating field increases as the tape approaches the erase head. So we have a tape transport, the tape moves towards the erase head and when it's past the erase head, it moves away from the erase head again. And it is rising when the tape approaches the erase head and falling when it moves away from the erase head accordingly. And the distance to the head is always very important. Yes, and an alternating field: we have a high frequency in this alternating field, i.e. many reversals of magnetization per period of time and the magnetizable particles in the tape material are magnetized and demagnetized several times to saturation and thus disordered again. So this disorder, which in principle is the initial state of the tape. And the strength of the alternating magnetic field must, of course, be greater than the field previously used for the erasure magnetization. That means you have to use a lot more energy to delete it than you did when you recorded it, because otherwise you wouldn't be able to completely remove the recording. Yes, and the falling strength of the alternating field opens up a narrower hysteresis loop with each oscillation run, until there is complete demagnetization at the point of intersection of the axes. Yes, how can we now visualize this again? So, as I said, we have these hysteresis loops here again and we see that hysteresis loops are becoming ever narrower. And that's what happens because we have magnetization and magnetic reversal here and the tape slowly moves away from the erasing head, which opens up an ever narrower hysteresis loop. It gets tighter and tighter the further the tape moves away from the erasing head, and at some point the whole thing actually comes out again with complete demagnetization here at the intersection of the axes. Yes, and that's also the reason why you should slowly remove one of these demagnetization chokes from the heads before you switch them off . This is also exactly the same. An ever narrower hysteresis loop is opened up until the workpiece - in this case the head - is completely demagnetized. Yeah, and so now we're going to play a high frequency on this tape. Isn't that a problem? Nope! Actually not, because the high frequency is of course not audible. We're moving somewhere in the range of 50 kilohertz. So we're going to see that a little more closely now and we humans don't normally hear that anymore - maybe Batman can still do that, I don't know. However, due to the so-called skin effect , the force fields of this high frequency do not penetrate very deeply into the tape anyway, and they do not remain on the tape for very long. This means that this high frequency does not remain when we take the tape out of the tape recorder, but after a few moments it is no longer there. Yes, and how do you actually have to dimension this high frequency that the oscillator generates? On the one hand, the pre-magnetization frequency should be selected so low that the HF losses occurring in the recording head remain small. So we have something similar to what we discussed back then with the transformer, we simply have certain losses in such a coil and the higher the frequency, the higher the losses are normally. And that's why you have to make sure that the pre-magnetization frequency is not endlessly high, because otherwise the losses would also be relatively high and we simply have to put more in to achieve the same effect. But on the other hand, this bias frequency should be so large that there is no combination formation between the harmonics of our music that we are recording and the fundamental wave of this bias frequency. This means that if we record a guitar on tape today, for example, or send it through our tape echo, this guitar tone does not only consist of the base frequency, i.e. the string does not only vibrate with its base frequency. But the timbre, which is what we actually want, is created by the fact that there is a characteristic spectrum of overtones, i.e. multiples of the base frequency, of the base frequencies. And if we were to choose this HF frequency too low, then it could well be that some overtone, perhaps the third, fourth, five overtones of a tone that we play on the guitar, comes into the range in which the base tone of the HF is also lies. And then it would come to a combination and then there may be very unwanted effects where the two signals weaken or over-emphasize each other. Yes, and that would also distort what we record a lot, and that's why it's always a balancing act. The so-called "rule of thumb" says that the bias frequency is five times f0 the highest cut-off frequency transmitted by the device. What does that mean now? For example, if we assume that the highest cut-off frequency that our tape echo will still transmit is 12 kilohertz, then we should choose the HF to be 5 x 12 kilohertz. Then the whole thing doesn't affect each other and we have exactly the right HF frequency that we need for recording and deletion. Yes, and for the NG-51S that would mean what I just took as an example: Here the frequency response is specified that echo/reverberation should be transmitted up to 12 kilohertz. So 5 x 12 kilohertz, there we are at around 60 kilohertz, which should be the HF frequency for the NG-51S. Yes, more on that later in this video, but much more on that in a later video where we take a closer look at the tape heads. Because this topic with the hysteresis loops and the HF is also important again there, but we will then really discuss that in detail. Yes, and at the beginning we talked a lot about using the HF for erasure. But not only for erasing, it is also used for recording. Why? In principle, the pre-magnetization by high-frequency also ensures a better signal-to-noise ratio, a significant improvement in the reproduction of higher frequencies and, as a result, less distortion. That is why it is not only interesting for erasing, but also for recording. Mr. Braunmühl and Weber also found out that things simply sound better if you mix in the HF for the recording. Yes, and the high frequency and the low frequency, i.e. what we want to record, are simply fed to the same line and thus mixed. Yes, how do you have to imagine that they are simply recorded on the same line? You have to imagine it like the picture on the left. So we have the curve for the low frequency current and below that for the bias current. And the low-frequency current, that becomes the envelope of the bias current, so to speak. So you see it there, here is the low frequency current. These are practically the outer lines and this will be the envelope within which the high frequency is found. That's how you have to visualize it. Yes, and the HF pre-magnetization, that sets the working point, it is also known under the term "bias" on the hysteresis loop, which is valid for the special magnetic tape material. .And it just sets the bias for recording our music that we want to record on tape or that we want to transmit via tape. And what does "bias" mean in that point? The bias is set within the linear range of the hysteresis loop, which is marked in red here. This means that the HF ensures that we have these areas here that distort heavily, i.e. where the sound reproduction is no longer linear, where the physics of this tape would practically distort our sound relationship to one another. Where that's just no longer true, with the treble, with the ratio of treble to bass and so on. We want to leave these areas out and the HF helps us that we can no longer get out of this linear area here. But more on that later in another video when we take a closer look at the tape heads. Back to the topic again: If we look at the circuit diagrams of the Echolette devices, the Echolette tape echoes, we find, interestingly enough, a single circuit diagram for a potential predecessor model "NG2" on the Internet. I've never seen the NG2, I'm not sure if anyone has seen it at all in the past few decades. In any case, this one circuit diagram has been preserved and there is a completely different oscillator circuit than the one we are dealing with today. So we see it's built around a single pentode, an EL84. But in the next models, NG-4 and NG-41, we already find what I would like to call the "echolette oscillator", i.e. the familiar circuit. The NG-41 was a direct contemporary model of the first NG-51, they overlapped, the sales were it's the same oscillator. Funnily enough, what changes from time to time is the value of this cathode resistor. Here is 500 ohms, there it is 470 ohms. So if you consider that the resistors normally have plus/minus 5% tolerance, then this difference of 30 ohms is not really that much. I'm also not sure if the 500 ohms resistor is a value that was available for sale widely. 470 ohms in any case. Well, I don't know if that was really 500 ohms or if it was always 470 ohms. As I said, it doesn't matter, so these differences can be completely ignored. Yes, with the NG-51A, i.e. the facelift model, now with the circuit board, the circuit diagram of this oscillator has changed somewhat in a broader sense. A resistor R81 and a capacitor C47 were added. Yes, they are more or less for spark quenching, or in other words for avoiding switching noises, because the reverberation on/off switch switches almost 300 volts and an inductance. Yes, and when the switch opens and closes, sparks can jump and sparks fly. And you can hear that too and of course it's not too good for the switch and I assume that's why they built it into the facelift. Yes, the successor to the NG-51, the E51, also the last tube tape echo that flashed the name "Echolette". In principle, this is like the NG-51 in the second model series with the circuit board, but the circuit diagram has been completely redrawn. It looks like this. Yes, it also makes sense to take a look at the competition from that time: What did the Dynacord people do, for example? Here's a snippet of the Dynacord Echocord circuit diagram, and as we can see, it's more or less exactly the same oscillator - with a few minor differences. Dynacord added a cathode bypass capacitor, which the Echolette doesn't have. The voltage is slightly lower than with Klemt and we see up there in this so-called resonant circuit, i.e. the coil and the capacitor C37, they have different values. So the capacitor definitely, we have that on the circuit diagram, that's 600 picofarads instead of 1000 picofarads on the Echolette. But just because this capacitor value is only a little more than half as high, the coil inductance would have to be different if Dynacord was aiming for the same HF frequency. Because, as we shall see, the two are closely related. So here coil and capacitor are probably a bit different. Yes, and where does it come from that they did it the same way? The resemblance is probably no coincidence, because if you look to the right, you have the very first Echocord model. I extracted the picture from the publications by Hans Ohms. And you can see here quite clearly: the very first Echocord Super was a license construction that was carried out by Klemt. And it was actually nothing more than an NG-51S with a black Dynacord front panel. And yes, inside was probably 100% NG-51S too. Dynacord then went a different ways, so the later Echocord has a completely different circuit than the NG-51S. But at least they seem to have adopted the oscillator with a few differences. Before we get into the circuit and the details of the circuit, I would like to point out two things: I just worked out a few test setups for this oscillator, where I then also carried out the measurements. And you can theoretically recreate it at home. The Echolette oscillator is very well suited for experiments with reduced operating voltage and I had such a setup on a breadboard. You can do that, but you it is not always preferable. I'll also say a few words later on where the problems lie with these breadboard setups when you try to implement high-frequency things with them. And in another test setup, as you can see on the bottom right, I soldered the capacitors and resistors directly to a tube socket. And I ran these two experiments with a significantly reduced plate voltage, namely only 30 volts. And I simply pulled the heater voltage of 6.3 volts AC out of my variac isolating transformer. But all oscilloscope images in this presentation are practically taken from these two test circuits. Yes, here is a detailed view of the breadboard assembly. So of course I got myself a tube socket for breadboards, which you can then simply plug in. Yes, and the green-yellow cable, which you can see on the far left, is used to connect the oscillator coil externally, because of course you can't plug it onto the breadboard. Yes, and the structure with the tube socket, it's basically the same. We have everything on the tube socket, but the coil is also connected externally here, and then of course you have to use a few cables, even longer cables, to connect this coil. And that has an impact on the experimental setup, as we will see in a moment. Yes, let's take a quick look at the oscillator tube, that's the ECC82. Let's start with the nomenclature of these tubes: ECC, i.e. the "E" stands for parallel heating 6.3 volts. A "C" stands for a triode, so we have a "CC", a double triode, there are two triodes in a glass bulb. The "8" means that this tube has a noval socket and the "2" is another type designation. So ECC82, you already know exactly: parallel heating 6.3 volts, we are dealing with a double triode with a Noval socket. Yes, and the designation in the USA is completely different, it's called 12AU7. But this is basically the same tube. The 30 volt application is unfortunately not in the manufacturer's data sheets, which is why we cannot rely on standard values ​​here. We have drawn an example for 20 volts here and we can see what plate current would flow there with a grid voltage of -2 volts. But our application is not quite there, but the comparison with the ECC83 alone shows that experiments with a plate voltage of 30 volts and a grid bias of about -1 volts, as I had in my experiments now, do not lead to anything with an ECC83. So you can see it here, with the ECC83 there is almost no plate current in the area I used in the test. With the ECC82 it is. And if plate current flows, then you can also use it for experiments. Again, of course, the application case is missing in the data sheet, how it behaves at -1 volt grid bias and at approx. 30 volt plate voltage, but it would probably be somewhere in the range here. Between these two curves and you can see: Some plate current is already flowing here, where almost nothing is flowing there with the ECC83. Such low-voltage experiments with the ECC83 would lead to nothing, so that's only possible in this special case. The ECC82 is well suited for use in an oscillator because it has a much higher transconductance, abbreviated with "S", than an ECC83. So we simply have a larger plate current change per voltage change on the grid than with the ECC83 and we will see why, but so much in advance: linearity is not really decisive when used in the oscillator, since it is expected that the resulting output signal will be a rectangle with many harmonics. That means: Distortion is definitely wanted here and high steepness promotes the swinging back and forth of the two tube systems in the oscillator, as we will see. And that's why the ECC82 with the properties for this oscillator is actually best suited. The voltage amplification, see the no-load amplification "µ" in the tables below, of the ECC82 is far lower than that of the ECC83. This means that with the 82 we have a plate voltage of 19.5 at 100 volts. And with the ECC83, this is no less than an open-loop gain of 100. So there is a clear difference, but a high voltage gain is also not important in the oscillator, but rather the level or the rapid increase in the plate current is the decisive factor here . And that's where the ECC82 comes out on top. Yes, and last but not least, the 82 was also specifically developed as an HF tube, the 83 more as an NF tube. So you can find them a lot, the 83s, in the preamplifiers and the 82s more for oscillators or multivibrators. So now it's on to the circuit, or now we're going to look at the circuit. And to do this, let's separate this oscillator circuit into its two assemblies. And now I'll drop the names, we have the astable multivibrator on the left, you've heard that before, and a parallel LC resonant circuit on the right. The two together make up our Echolette oscillator, but let's look at the two assemblies separately for now. Yes, at this point I would like to point out again: I have written a very, very detailed summary that you can download as a PDF - in German and in English - what exactly happens step by step in this astable multivibrator. To chew it up in this video would simply go beyond the time frame I have set myself. And yes, I don't have to read it to you, you can read it yourself. So here is just a very rough summary. Yes, let's move through the following slides, which describe exactly how the astable multivibrator works, maybe a little faster, I won't go into every detail here, but you can read every detail in the two PDFs that are available for download in German and in English. And there you can read it all again and understand it at your leisure. So we now have our simplified structure of this astable multivibrator here on the left and let's see what happens here from time to time at the moment when we switch on the device, i.e. apply voltage to this circuit, marked B+ above. The moment the operating voltage B+ is switched on, the two capacitors C1 and C2 are charged first. And the charging is done through the respective charging circuits. That means a charging circuit, that's the path between the supply voltage B+ and the ground point down here, ie the circuit ground. That means I always need + and - for the current to flow. And these capacitors charge through their respective charging circuits: this one around here and the other one around here. I've shown it here on the right. For the sake of completeness, the resistance that exists between the grid and the cathode of a tube is also drawn as half a tube, since it is of course parallel to the grid bias resistor. So it is there in terms of effect, you could also omit it, but for the sake of completeness I have drawn it in. So these are the charging circles. C1 charges over here the moment we turn on B+ and C2 charges over here. Due to the charging of C1, a positive voltage peak drops at the grid leak resistor RG2 and the same happens for the capacitor C2 at RG1. This impulse ensures that both tubes open, i.e. the plate current increases and both plate voltages VA1 and VA2 drop correspondingly briefly at the plate resistors, equivalent to this positive voltage impulse that is at the input. Yes, you could now think: Okay, so now both tubes start to work at the same time and they could now go into a state in parallel where they either drive fully: so here the full plate current then flows in both tubes. Or are somehow half open, i.e. at least follow some exactly identical course. But that is not the case, because there are usually the tiniest asymmetries between two tubes, even between two systems in a single one, ie between the two triodes of a single tube. And these asymmetries can be: Unequally strong tube noise, unequally distributed emissions along the cathode, etc. And that ensures that both tubes don't always go into exactly the same state at the same time, because then we wouldn't have any oscillation either, that's all we had two tubes somehow both doing the same thing in parallel. So let's take a look at what happens instead. Let's assume for example that the plate current IA1 here simply increases faster than the plate current in the other tube subsystem due to some asymmetries between these two tube systems. This means that the voltage VA1 then also drops more than the voltage VA2. Capacitor C1 now transfers this negative voltage change from VA1 to the grid here. So there you can see a small voltage drop. Yes, and this voltage drop at RG2 is now amplified by tube 2; up here the plate voltage, which increases accordingly and this change in voltage is now transmitted via C2 to the grid of tube 1. And it is also amplified there again, so we now have a voltage drop VA1 and that is also transmitted back over here and it is amplified again - and you can see where the whole thing is going. That is, what was initially a relatively small voltage dip, transmitted by the charging of the capacitors, is amplifying more and more and one of these tubes is now starting to really kick off. And amplifies into saturation. Yes, and the summary at this point in time is: tube 1 in our example is fully driving, i.e. the plate voltage VA1 is at its minimum. Of course, the minimum does not mean that it is at zero, but it is the minimum that can now be achieved in this circuit. So this is the lowest point of this plate voltage that is reached. Yes, VG2, so the grid of the second tube is so negative in this state that this tube closes completely. And VG1, i.e. the grid voltage of the first tube, is close to ground potential, since there are no longer any voltage changes from over here. And it has ground potential via the grid leak resistor. So one tube is completely open, the other is completely closed. And the two capacitors are still fully charged. What happens next when this tube 1 is completely switched on: then its internal resistance drops very sharply and the capacitor C1 can discharge through the tube to ground. Analogously to the charging circuits, there are of course also discharging circuits. It looks more or less like this. The capacitor can discharge through the tube, but this discharge circuit also includes the RG2 grid leak resistor and the discharge current through the tube also causes a negative voltage drop across RG2. That means what happened now, when this capacitor C1 discharges via tube 1, then it ensures that there is also a negative voltage drop here at RG2. And the grid voltage of tube 2 is shifted very, very much negatively, far from allowing that tube to open. But of course this discharge current decreases again after a short time and the grid voltage slowly migrates back towards ground potential and towards the area where the tube can be driven. And that now ensures that when we have arrived here at point A, tube 2 slowly opens again. Plate current starts to flow again and C2 of course passes this negative voltage change back to tube 1 or the grid of tube 1 and tube 1 closes a little and so this process slowly reverses. Precisely because tube 1 is fully open or fully driven and tube 2 was completely closed and because this capacitor C1 is now discharging, the tide is turning again. That is, tube 2 can slightly open again and thereby tube 1 closes slightly again. And as you can imagine, the following happens now: The whole thing slowly switches over until tube 2 is fully driven again and tube 1 is slowly closed. This means that the oscillation of this multivibrator ultimately ensures that one tube is always fully driven and the other is completely closed. And the next moment the whole thing reverses, so that, as in the example here, tube 2 is fully driving and tube 1 is completely closed. So you always alternately have one tube open, one closed, one tube open, one closed. And it keeps changing back and forth as to which tube is currently open and which is closed. Yes, and because that goes back and forth, but within a closed circuit , both tubes show the same waveform at the anodes, but are 180 degrees out of phase with each other. That means you see this here, it's already the same waveform, but we have a phase shift of 180 degrees in there. And this phase shift of 180 degrees is essential for the oscillation of the oscillator, because the whole thing is called "positive feedback" - more on that in a moment. Now we see here on the right an oscilloscope picture recorded from my test circuit and the waveform that we see here was recorded from this breadboard test setup. They're not 100% ideal, as we'll see in a moment. But that's because there are always stray capacitances on this breadboard, especially since we're doing the whole thing with a high frequency. This means that we simply have to contend with capacitances here on the breadboard that do not come from the components that we put on there themselves, but rather from the breadboard itself. Or as in this case: I had to connect the coil externally and it is then connected to the breadboard with two cables. These cables also have capacitances, because of course they also have a length - maybe only 20 centimeters, but such a 20 cmcable also has a capacitance that is parallel to our astable multivibrator. And of course this is reflected in the waveform. Yes, and this slightly distorted waveform, slightly distorted from what it should theoretically look like. This is also in the areas of the wave that are generated by the capacitive components. That hump and that dip down and that hump, it's from charging and discharging the capacitors. So are dependent or shaped by the capacitances that we have in the circuit and because we have stray capacitances right here, those areas are not exactly as pronounced as they theoretically should be. But I really recommend that you look at the PDF files again, because what we just skimmed over a bit, how one tube always closes and the other opens and the whole thing swings around again, that's all there is of course explained in detail. And also to understand in individual steps. Yes, and this oscillation, i.e. this swinging back and forth, that one tube is always open and the other closed and this is how this square-wave signal is formed. In order for there to be a permanent and stable oscillation, there must be a number of prerequisites that are formulated in the stability criterion according to Barkhausen. So I'll just take a small excerpt. Of course, if you want to delve deeper, you can read about it. Yes, the most important criteria: The phase shift must be 0° or 360°. This means that the amplified signal must be looped in again in phase. This principle is called "positive feedback". Otherwise there is phase cancellation and the whole does not oscillate. Yes, and if the loop gain of this system is at least 1, then the circuit will produce an oscillation. With a loop gain of less than 1, the amplitude becomes smaller and smaller and there is actually no oscillation. And with a loop gain greater than 1, the amplitude would keep increasing until the gain saturates. This is the principle we want in this astable multivibrator. A loop amplification greater than 1. After all, we want a square-wave signal, i.e. a signal that is always alternately saturated. Let's take a closer look at the principle of positive feedback . Due to their general mode of operation, the signal at the plates of a tube is phase-shifted by 180 degrees compared to the signal at the grid. This means that if we have a negative voltage drop at the grid, then this will be noticeable at the plate in a 180 degree rotated increase in the plate voltage. So here we have the input signal, there the output signal. And they are rotated 180 degrees to each other. This is the working principle of tubes; the phasing between grid and anode is rotated by 180 degrees. Yes, and the signal that we see here, for example, on grid G2, is rotated once through tube 2, 180 degrees, and then it is transmitted here through capacitor C2 to the grid of tube 1 and is also inverted again by this tube 180 degrees out of phase. Then transmitted back to the original tube via C1, where it arrives again phase-shifted by a total of 360 degrees. That means: the original signal was rotated 180 degrees, then rotated 180 degrees again and there it comes back, so it's rotated 360 degrees overall. So it arrives in phase, but as we can see here, it no longer arrives in exactly the same way as it started there, but rather it arrives a little amplified overall and so it keeps building up. In-phase signals are amplified more and more and the system begins to oscillate, it oscillates. And with each pass, the swaying gets stronger until the tube runs into saturation, so that at some point it can no longer be amplified at all and is therefore heavily distorted. Yes, and the voltage amplification of each tube system here in the Echolette oscillator is about 1.5. Means: If I have 20 volts peak to peak at the grid, then I have 30 volts peak to peak at the plate. So we also meet the Barkhausen criterion with this, we have a loop gain that is greater than 1. And the astable thing about the astable multivibrator is that the swinging up of a tube does not remain stable up to the maximum state, but tilts by itself in such a way that the other tube also swings up to its maximum state in turn. That's exactly what I just said, it's always a tube on, tube off, tube on, tube off. And the whole thing swings back and forth like a swing and one tube is at its maximum and then it swings to its minimum. And the other goes to the maximum state. And in the end there is this square-wave signal that we have at the output. Another little thing that is interesting to know: well, we have seen that these capacitors C1 and C2 are a very important component of this astable multivibrator. These charging and discharging circuits have so-called time constants. Yes, and these time constants also result from the resistances and capacitance of the capacitor. The resistances in the circuit determine the current and the capacitance of each capacitor determines its ability to store charges. This means that a current must flow in order to charge the capacitor. And how much current flows is ultimately determined by the resistances in this charging circuit. And how long the current can flow is determined by the capacity of the capacitor. So how many more charges can it store? The time constant thus varies with the dimensioning of the resistors and the capacitors and this ultimately determines the frequency of the multivibrator. So frequency: the frequency of the transitions of charge and discharge just described. It goes back and forth - they charge, they discharge. And how fast the switching back and forth between charging and discharging is ultimately determined by the resistances and the capacitance in these charging and discharging circuits. A small example: If C1 has a larger capacity than C2, then tube 2 would drive longer than tube 1. The red marked spots here on the left would then no longer be of the same length. I did it on the breadboard and measured it. So we're looking at two separate examples here, measured against one of the grids. And here the capacitors C1 and C2 each had very different capacitances and you notice that the result is no longer such a nice, even rectangle. But this is much, much longer than this section and here even greater capacity is introduced in one of the capacitors, the whole thing gets completely out of hand. One thing I would like to point out: We left out the common cathode resistor in the thought experiment, but of course it also has an impact on the circuit. The cathode resistor, which I have labeled RC here, is of course used for the automatic adjustment of the grid bias. How's that going now? When plate current flows through RC, a positive voltage drops across the cathodes K1 and K2. However, since the grids are at ground potential, i.e. 0 volts, via RG1 and RG2, the grids are automatically more negative than the cathodes. This is called automatic grid bias adjustment. This means that we don't have to introduce a negative grid bias here using an external power supply, for example, and apply it to the grid, but the grid is simply automatically negative with respect to the cathode because we made that positive with respect to ground. Yes, and the values ​​of the resistors ultimately result from the desired biasing point, i.e. the degree of negative bias, as can be seen from the data sheet for the tubes. But now consider the following situation: Tube1 drives fully, the plate current through RC increases to maximum. At the point I have marked with red A there is an even more positive voltage drop than before. However, this voltage drop also occurs at K2, i.e. at A dash, since both points are at the same potential. This ensures that K2 becomes even more positive with respect to ground and G2 becomes even more negative in relation to K2 and tube 2 blocks even more. The coupled cathodes make this on/off effect of the tubes even more stable. Alright, now we come to the second component that belongs to this Echolette oscillator, the LC resonant circuit. Here again briefly some theory based on a film example. As in the example circuit below, a parallel oscillating circuit consisting of a coil and a capacitor is supplied with a DC voltage via a switch. The switch is closed, allowing the capacitor to fully charge. The switch is then opened again and the capacitor is thus separated from the voltage source. A free oscillation now occurs at the natural frequency of the oscillating circuit, but this oscillation is damped and decays very quickly. This type of excitation of the resonant circuit is not the direct explanation for its use in the Echolette oscillator, but it opens up an understanding of the topic of resonance, which will be very useful for further understanding. So this LC resonant circuit consists of two types of energy storage, a capacitor, that's the C in the name, and an inductor, which is usually abbreviated to L. So LC resonant circuit: capacitor and inductor. And both components have different properties in a circuit. We want to take a quick look at two of these properties here, namely the phase relationship between current and voltage. With a capacitor, the current leads the voltage by about 90 degrees, and with an inductor, the current lags the voltage by about 90 degrees. So that means the other way around. The second property is the reactance, there is a capacitive and an inductive reactance. The capacitive reactance XC decreases with increasing frequency and with increasing capacity of the capacitor. The inductive reactance XL increases with increasing frequency and with greater inductance of the coil. Now let's look at this LC resonant circuit from the example again and break it down into its individual parts, from T0 to T4, for a better understanding, and look at what exactly happens that this resonant circuit oscillates at all. Yes, the capacitor is now discharging through the coil at time T0. We see this here: the voltage slowly drops because the capacitor is discharging. And here the current flowing through the coil slowly increases, the current flow through the coil now creates a magnetic field around the coil and the electrical energy in the capacitor is now practically converted into magnetic energy and also stored in this magnetic field. However, inductivities resist abrupt changes in the current flow, so that it does not increase abruptly, but - as shown here in the green waveform - increases slowly sinusoidally. And when the current flow is at its maximum up here, the voltage on the capacitor is at zero. So it's completely discharged. And at that moment the magnetic field around the coil collapses and induces a voltage on itself and the current through the coil thus continues to flow in the other direction and slowly recharges the capacitor in the other direction until it is at its maximum value , i.e. fully charged, only this time the other way around, here plus there minus. And then the flow of current subsides again. Yes, and so it goes from the other direction just as before. The capacitor discharges again through the coil, the current through the coil slowly increases until it reaches a maximum. The magnetic field slowly collapses again at T3: self-induction - the current flows in the other direction and in turn fully charges the capacitor in the other direction. And so on. So theoretically we have a constant back and forth of capacitor discharge, magnetic field build-up and collapse around the coil. And it recharges the capacitor in the other direction and it goes back and forth in theory all the time. Yes, the values ​​of the inductance L and the capacitor C determine the frequency of this polarity reversal or, to put it another way, the duration of the individual processes shown, i.e. the frequency of the oscillation. This specific frequency is also called the resonance frequency, or F0. Lower inductance and lower capacitance increase F0, which follows from the following equation, in which L must be written in Henry and C in Farads. Yes, if you don't want to do it yourself, you can also use online calculators like the ones shown on the left, which also make it a bit easier to test and try out values. I did a small test here: In the upper left row, only the capacitance value of the capacitor is changed and in the lower left row we only change the inductance value of the coil in the same way. And now note the respective frequency changes as they are always shown here in the individual images at the bottom right. Yes, and of course this can also be measured in our test setup by replacing the capacitor of the oscillating circuit in this oscillator with capacitors of other values. Yes, and here on the right side you can always see my frequency counter, which counts what frequency we have at the moment. Yes, and in the picture above on the left is the 1000 picofarad capacitor, the original, so to speak, as intended in the circuit. And I measured a frequency of about 41 kilohertz. Yes, the test setup is actually not oscillating at 55 kilohertz, as it should be, because as I said, both on the breadboard and in the cables that I used to connect the coil, so these are the stray capacitances that are in there that cause the simply change the capacitive part of the LC resonant circuit. But I've already managed to get it to 50 kilohertz with other cable combinations. That means for me it is absolutely understandable that it is due to these stray capacitances. So I measured around 41 kilohertz. In the picture above on the right I inserted a capacitor of 220 picofarads and the frequency of the oscillator increased to 58 kilohertz. In the picture below on the right, I even went up to 3.8 nanofarads, where the frequency now drops to around 25 kilohertz. And on the bottom left I went up to 95 nanofarads and there the frequency drops to 5.59 kilohertz. That means you can see that if the oscillator does not oscillate at the right frequency, you can definitely change the value of this capacitor in the oscillating circuit so that the frequency is the way we want it to be again. Yes, and here is an example measurement on a real NG-51S, namely a test measurement on L1 of the coil and this results in 56.31 kHz. We had already seen in the theoretical considerations that the HF for the desired frequency range of the NG-51S should be somewhere around 60 kilohertz. So yes, 55/56 kHz is what you will usually find in real devices. Next, let's take a look at the oscillator coil itself. Yes, the design, the design of this coil is called "pot core" or "P-Core". It is a pot core made of ferrite. Yes, as we see here, on the left is the Siemens symbol. So it was once made by Siemens. Ferrite has a high magnetic permeability, which means the magnetic fields do not penetrate into the air space around the coil. That is, the magnetic fileds flow through the pot core and are deflected back inside. But at the same time, this ferrite also has low electrical conductivity, which means that the eddy current losses that we would otherwise have to contend with are also minimized as a result. The core encloses the coil almost completely, which means we have high magnetic shielding, very low leakage inductance, very low electromagnetic radiation - because we don't want that. In the NG-51 this oscillator is in the middle of the device and if it spread into everything else, that would certainly lead to problems and that is why the pot core is very suitable here. Pot core coils are also usually used where a large amount of magnetic energy has to be stored in a small space at high frequencies. A small example also shows how important this pot core is for the inductance. In the first example we have an inductance measurement of only one side of the coil without the core material. That means the pot core is completely removed and the meter reads 0.364 millihenry. And now in the next example we have used the lower pot core again and then the inductance already increases to 0.942 millihenry. And if we also add the second pot core, then the inductance makes a huge leap upwards, because we are now at 10.6 millihenry. This means that the core has a significant share in achieving the required inductance and pot-core coils are particularly effective here in a relatively small space. Yes, about the material of this pot core: It says "N22", that's "Siferrit". And what else is written on it is almost even more interesting, namely the value AL400. "AL" is the so-called inductance constant and is given here as 400 nanohenry. That is, the value is always nanohenrys - per square turn, to be exact. Yes, what exactly is this AL value supposed to express? So anyone who manufactures such coils must of course know how many turns of a wire are required to ultimately achieve a certain inductance. And since the core material, as just described, has a significant influence on the inductance, the core manufacturers specify an inductance constant for each of their products. The coil manufacturer can then work more easily with this because he only has to worry about the number of turns. Yes, and these relationships between the inductance constant and the inductance and the turns of wire are shown in the formulas on the left. And if you put in the measured values, i.e. these ten millihenrys that we measured, we know that it is an inductance constant of 400 and then you can practically calculate that there must be about 50 turns of wire on each side of this coil. And I measured the wire diameter at 0.33 mm. Yes, we have these three formulas here above: L is the inductance, N is the number of windings and AL, the inductance constant, results from the ratio of inductance per square turns, i.e. nanohenrys per square turns. Yes, and if you want to play with values ​​a bit, you can use these three simple formulas. Alright, on with the theory of the LC resonant circuit. We will now reassemble both assemblies of the oscillator and take a look at how the parallel LC resonant circuit works specifically with the astable multivibrator. We have found that the astable multivibrator generates an approximate square-wave signal within the limits of its structure, and if you look more closely at the frequency spectrum of this square-wave signal, it becomes clear that it is a relatively broadband signal with many overtones. Of course, the signal I measured contains a lot of noise and the signal-to-noise ratio is significantly lower than normal. On the one hand, this is of course related to the very unideal test setup on the breadboard and the reduced operating voltage. On the other hand, also with my limited measuring equipment for frequency spectra. So I still have to step this game up significantly, but it serves to illustrate what is said below and not as an absolute reference that can be used for comparative measurements. Yes, and if we now look at the peaks of this broadband signal, then F0 of the astable multivibrator appears to be around 27 kilohertz. This means that the first overtone of these 27 kilohertz is around 55 kilohertz, which in theory should be the frequency at which the HF for the oscillator should be. That means we don't even need the base frequency of this oscillator or this rectangle, but it's harmonics. Yes, and how we get to that now, let's take a closer look. In the following, we now want to consider how the parallel LC resonant circuit behaves when it is excited by such a signal as we have just seen and how it happens that the HF oscillator ends up only oscillating at a single frequency and no longer on this frequency mixture. Yes, as we have already seen: A combination of coil and capacitor in the oscillating circuit has a natural frequency, the resonant frequency. The LC combination adapts to this in free oscillation. However, it would now be interesting if one were to excite such an oscillating circuit with the broadband spectrum of the astable multivibrator. The resonant circuit will oscillate here as well, but not freely, but in the form of forced oscillations. At the resonant frequency, the LC resonant circuit now has some interesting properties in the case of forced oscillations. These would be: phase angle. The phase angle between current and voltage is 0 at resonance. The reactances of the coil and the capacitor are the same at resonance. The reactance of the oscillating circuit as a whole is at its maximum at resonance, and the highest signal amplitude can be measured at the oscillating circuit at its resonance frequency. And if the forced oscillation deviates up or down from the resonant frequency, then the most noteworthy property of the oscillating circuit in the focus of this consideration is that the impedance decreases for higher or lower frequencies to the point of close to a short circuit. Yes, if you now plot the voltage that can be measured on the resonant circuit against the frequency in a graph, then something like the following results: Above and below a more or less narrow range, so this is our narrow range, and above and below F0 around the signal amplitudes decrease rapidly. The parallel LC resonant circuit is therefore suitable as a filter to filter out the frequency from the mixture of frequencies that corresponds to the resonant frequency of the resonant circuit. So how accurate is this filter now? This is where the concept of "Quality" comes into play. Starting from the maximum amplitude at resonance, the two points on the curve are now drawn on the left and right at which the amplitude has dropped to 70% of the maximum value, or to put it another way, by 3 dB. And the difference between these two frequencies is the bandwidth "b" of this LC filter. And the smaller this bandwidth is, the higher the quality. Of course, a high quality should be aimed for, because ideally we only want to use a single frequency in the oscillator. Everything else should be filtered out. And the smaller the bandwidth b of the filter, the higher its quality and the higher the quality, the more selective the resonance. That means we want selectivity here, we want to filter out a single frequency from this mix of frequencies and use it as our HF frequency. Yes, and how is the resonant frequency filtered out? We can read this curve a little differently, then this connection will also become clearer. If we do not plot the voltage V on the Y-axis, but the impedance Z, then it becomes clear that the parallel LC resonant circuit has the greatest impedance at the resonant frequency. All other frequencies apart from the resonant frequency are therefore short-circuited via the coil ends. They disappear. And the steeper the slopes of the impedance frequency curve rise and fall, the smaller the bandwidth and the higher the quality - the following applies: ideally, all frequencies next to the resonance are either completely short-circuited or at least sufficiently damped so that they no longer matter. Yes, and the quality of this oscillating circuit now depends mainly on the quality of the coil, which is why it is so important and why it is particularly important here if the coil is replaced . And different factors are relevant for the quality of the coil. A look at the equivalent circuit diagram here should reveal that. A coil of course has an inductance. But it also has an ohmic resistance that depends on the wire length and the number of turns. And these wires lying next to each other on the coil angle also have a capacitance. This means that these windings next to each other have a capacitance that increases with the number of windings. This means that the smaller the ohmic resistance in the equivalent circuit diagram, the higher the quality. So you have to pay attention to it, the smallest possible resistance. And to reduce R, among other things, special stranded wire can be used to wind it onto the bobbin, so-called HF stranded wire. The skin effect is reduced, which means that the skin effect is that at high frequencies, the currents no longer flow through the entire wire diameter, but mostly on its surface. And by not using the whole wire, the resistance also increases. Yes, low-loss core material also improves the quality, so we have Siferrite here in the Echolette oscillator or in the Echolette oscillator coil - that is low-loss core material. And there are also special winding techniques, so-called cross-winding, which can also be used to minimize the capacitance between the wire turns. This means that high-quality coils have a very high Q-Factor and we need something like that here, because this parallel capacitance of the winding has the effect, in the worst case, that the coil has a natural resonance even without the actual oscillating circuit capacitor and above this own resonance frequency the coil can no longer be used at all. That means that would be crap, you can't use that here. Now that we've covered the basics of the LC filter, let's look at what it does with the multivibrator's square wave. I did a few measurements for this. I have always pointed out where I attached my two probe ends so that you can understand. And now, in the first example, I measured from one winding end of the coil to ground. And that's where this sine comes out. Here we have a plate voltage in the test setup of approx. 30 volts RMS and the display area, you can still see it here in the background of the picture here, these boxes are 10 volts per division, which means I measure at the points shown here an AC voltage peak to peak of 28 to approx. 30 volts. Yes, of course I measure the same thing on the other side, also 28 to 30 volts. Yes, and here we used both channels of the oscilloscope and measured simultaneously and you can see that the values ​​of this AC voltage at the ends of the probe are the same, but phase-shifted by 180 degrees. Now let's measure the two ends of the coil, but there is one thing to note here if you want to copy it: measurement over the winding ends. We have measured flying voltages here, which means that a voltage without ground potential and ground reference is measured. And as we have seen, we have an AC voltage of 30 volts peak to peak at each of these winding ends. And the ground terminal of a normal probe is of course connected to the circuit ground of the oscilloscope in the normal case and if we were to simply measure that with our oscilloscope now, it would mean that we have 30 volts on the oscilloscope chassis. Which can be very, very dangerous at other voltages than used in this test setup. Yes, and that's why we don't do it that way! In order to measure something like the picture here, we need a special device, namely a "differential probe". This is an active probe that only displays the difference between the two probe tips and provides a ground reference. Yes, and it looks like this: I suddenly measure 60 volts peak to peak here, i.e. a doubling. That means we had previously measured something like 28, 30 volts at each end of the coil to ground and now we suddenly measure 60 volts peak to peak, which is a doubling. How does that come about now? The voltage doubling is relatively easy to explain, it simply has to do with the reference potential of the voltage. First, let's look at a very simple experiment. I have a 9 volt battery that I'm going to measure with my multimeter. If we measure the voltage of this 9 volt battery with a multimeter, then the voltage displayed on the multimeter is related to the reference potential of the ground point. So it has something to do with where I put the black measuring tip on it. By convention, the black tip is to be placed on the negative pole and the red tip on the positive pole. The multimeter then shows in the example that the battery has a voltage of 8.53 volts. But if we swap the tips now, the black probe tip is now on the higher potential, on the positive pole, and the red on the lower, on the negative pole of the battery. And then the voltage is no longer +8.53 volts, but suddenly it is -8.53 volts. Although that's the same battery. That means whether I measure plus or minus always depends on the reference potential. And that's important to understand next. So we now have voltage at the ends of the winding rotated 180 degrees to each other and we start with the measurement at the zero point. The difference between the two probes is 0 volts here, because they are both at 0 volts. Now let's go to the next point, a quarter of a full wave further. The voltage difference between -2 volts and +2 volts is now +4 volts. Positive, +4 volts because the black tip is at the lower potential. At the third measuring point, both curves are again at 0 volts, so we have a voltage difference of 0 volts. And at the last measuring point, red is now on the lower potential and black on the higher. We have a voltage difference of -4 volts between +2 volts and -2 volts. Yes, so we measured; 0 volts, +4 volts, 0 volts, -4 volts. This results in this composite voltage across the coil, across the transformer. This results in the coil ends - with their constantly changing potential, because we don't have a DC voltage there, we have an AC voltage there. And depending on the period of time where I look at the whole thing, the voltage is at a different potential and our reference potential of the black probe is also at a different potential depending on the period of time - this composite voltage results across the coil ends and we therefore measure a doubling of the voltage across the entire coil. Yes, this has not yet been explicitly mentioned, but the operating voltage B+ of the tubes, i.e. approx. 290 volts DC in the Echolette, is fed to the oscillator coil via the center tap on the L2 side. This is purely practical because it guarantees a completely symmetrical feeding of both triodes. And the high frequency is ultimately not tapped directly at the LC resonant circuit for use on the tape heads, but on the L1 side of the transformer coil. The transmission ratio is one to one between L2 and L1, that is, it is not a matter of up-transforming or down-transforming. The transformer is used here primarily for decoupling from the plate voltage, because DC voltage on the tape heads would be absolutely not beneficial for the reasons mentioned at the beginning - we remember: erasing with DC voltage - and at the given level of 290 volts that would even be devastating, it would burn out the tape heads. In addition, the transformer is also necessary here because we have to pick up two different voltages for the erase head and the recording heads on the L1 side, i.e. on the side where the tape heads and the erase head are. And that's why we need a transformer, because we can't change the winding ratio on the L2 side. The two coil partial windings must be exactly identical so that both triodes get exactly the same plate voltage. That 's why we have to pick up the partial voltage that we need on the other side of the transformer. Yes, what's next? I already said it at the beginning, next time we'll take a look at what actually belongs to the oscillator and HF topic, but for the sake of clarity I wanted to leave it out here: namely recording and playback. The NG-51S or successor E51, they have a very special construction of the recording and playback heads, as well as the use of the HF here. This special feature is absolutely decisive for the entire special sound of the device and that is why we will take a closer look at what I have marked in red here in the next video, if we also take a closer look at the recording and playback amplifier next to it. Thank you very much for sticking with me until the end. As I said, I would be happy if you subscribe to this channel. It's free and of course gives me at least some moral support because I know you guys love watching my videos and want to see more. Thank you, see you next time. Take care, bye!