The other day, at the “Domestic Radio Engineering of the 20th Century” forum, a conversation began about RIAA tube correctors. I also got involved in this conversation and as the conversation progressed I remembered another old, forgotten construction of mine. This is a tube preamp with an RIAA equalizer for an MM head, which I made back in 1999. It was assembled according to the scheme of Yu. Makarov “Neophyte” and was described in the magazine "Hi-Fi & Music" No. 11 - 1997.

Schematic diagram of a corrector-preamplifier.

I had to spend a lot of time to find this structure in the “deposits” in the pantry. I found it, but it turned out that over the years I had thoroughly “gutted” it. And although the remaining blocks were found (except for the power transformer and inductor), the design is already a “pathetic sight”:

In the photo: the remains of a once finished structure.

I don’t remember when and why I took it apart. But I remember that for quite a long time I listened to gramophone records through this corrector (I then had a Vega-106 player) and Arkam. And with the help of a preamplifier, I conducted experiments: I tried to “ennoble” the sound of a CD player with harmonics.
The preamp board was found in another box. I suspect that it is still working :) Once upon a time it stood next to the RIAA board. Well, the anode power supply has also been preserved. At the input there was a kenotron, then an LC filter, then a stabilizer on the KT805 at +300 V.

In the photo: preamplifier and anode stabilizer boards.

Actually, I wanted to check the performance of the corrector and, if it is still working, listen to it and compare it with the one I am currently “listening to”. To do this, I dismantled the board from the case, inspected the installation, checked for the absence of a short circuit, etc. - after all, the board has not been turned on for at least 8-9 years:

In the photo: top and bottom views of the corrector board.

The date of its manufacture is written on the board: January 26, 1999. Naturally, I didn’t have a PC then (well, except for the home-made Sinclair, of course :)), but I learned about LUT, Sprint Layout and other amateur radio “useful things” much later :) Therefore, the board is drawn in the old fashioned way, with a glass drawing pen and nail polish.

Drawing of the printed circuit board of the corrector and the date of manufacture of the board.

I was pleased with the inspection, so I connected it to my “copper” power supply (I had to make a small modification to the power supply - bring the voltage to the block after the kenotron and filter, since the stabilizer produces a maximum of +220 V). After turning it on, nothing smoked or exploded, which is already good :) Under load, the anode voltage turned out to be +291 V, which is quite normal (normally +300 V). I checked and slightly adjusted the constant voltages on the electrodes of both 6Zh32P lamps. There are small deviations from those indicated in the diagram, but everything is within normal limits.
After that, I connected it to the Denon receiver and listened to music for a while. I honestly didn't like it. The sound is completely flat, like coming from a bucket. I ran it in “background” mode for an hour and a half or two, after which I decided to listen to music one more time.
It was as if the device had been replaced! The sound became rich, rich, the way you expect from a record :) Just for fun, I connected the player to my “standard” corrector. In principle, there are differences, but at the level of “nuances”. But again, if you mount the Neophyte in a normal case, make a good power supply, dilute the ground, carefully set all the modes, and even replace the feed-through capacitors (and there are not very high-quality capacitors there - I installed the ones that I was able to find then) - I think it will "sound" very good.

In the photo: a corrector with a power supply and a general view of the “test stand”

The next stage was an experiment with replacing lamps. There were 3 Tesla EF86 lamps on the farm. Moreover, one lamp does not have 2 and 7 legs (screen). I thought someone had cut them off, but when I looked closer I saw that they didn't seem to come from the factory.

In the photo: EF86 lamps; The missing legs are circled in red.

After I installed them and turned on the corrector, a real “firing” began in the speakers, a crackling sound, so much so that the Denon protection quickly worked. In general, I let them warm up for half an hour, after which I carefully turned on the Denon again. The shooting ended and I was able to listen to the corrector with these lamps. My hearing seems to be fine, but to be honest, I didn’t hear any difference. Well, none at all. The only difference is that when I pounded the working 6Zh32P with the handle of a screwdriver, the sound was very clear and loud, while with the Tesla lamp it was “dull”. In this sense, of course, the EF86 looks better.
In a word, I checked the old proofreader and now with a clear conscience I will send it to my colleague. If he puts in a little effort, he will get a very good equalizer for a pleasant listening experience. :)

Finally, a couple of nice photos.

In the photo: corrector lamps in operation and a Yamaha TT-400 player.

RIAA correction is the reduction of the signal spectrum to the amplitude-frequency characteristic of the human ear. RIAA correction is also called “weighting” (weighted RIAA filter), which is used in measuring equipment.

When recording a vinyl record, the level of high-frequency components increases, and low-frequency components decrease. The fact is that the power of high-frequency components in a musical recording is, as a rule, less than that of low-frequency ones. Therefore, disk noise is more pronounced at high frequencies. To make noise less noticeable, high-frequency components are raised during recording and lowered during playback. As for the low-frequency components, they are reduced so that the needle does not “fly out” from the track. Accordingly, during playback, their level is increased.

Frequency response for recording and playback of records was first standardized in 1953 by the Recording Industry Association of America (RIAA). Therefore, the amplitude-frequency response during playback is called the RIAA response. This curve describes the amplitude-frequency response for the frequency range from 30 Hz to 15 kHz. The RIAA standard has been adopted worldwide. With the development of technology, it became possible to record sounds at lower frequencies. Therefore, in 1978, the RIAA-78 standard was adopted, which describes the amplitude-frequency response at frequencies over a wider range. In some publications it is called the IEC characteristic, since the amplitude-frequency characteristic for vinyl records has also been standardized by the International Electrotechnical Commission.

To ensure compatibility with both old and new records, many phono preamp models have a frequency response that is somewhere between RIAA and RIAA-78. Music lovers' collections also include records released before the introduction of the RIAA standard. To play them, some phono stages have a special operating mode called Old Columbia LPs. The phono stage may also have a 78 rpm record playback mode. In this mode, the phono stage simply amplifies the signal.

To suppress low-frequency beats associated with the transmission of vibration from the engine or warping of the record, some phono stages have a special filter.

All corrections, both during recording and during playback, were always made using minimal phase circuits, for which there is a natural unambiguous relationship between the amplitude-frequency response and the phase-frequency response, and which have pre-emphasis on both the amplitude-frequency response and the phase-frequency response when recording completely are compensated by pre-emphasis in a corrector with an inverse transfer function.

This is my third approximation to the optimal corrector configuration on 600 LCR modules. This time I decided to test the classic version, with impedance matching using interstage transformers. So, here is a diagram of one of the two options I tried:

As you can see, there are four stages, two interstage (one of them serves as an output) transformer, two interstage capacitors. Completely ignoring the idea of ​​a “short path”, and taking into account the fact that the corrector is connected to the preamplifier - cynical ignoring. 🙂 It’s all the more surprising that today this corrector (in my system) is the most transparent, dynamic and “stable” in sound that I’ve heard. I was very puzzled by this sonic result - since it is generally contradicts technical common sense. Apparently, even taking into account twice the number of amplification stages (than is usually required), the positive contribution that low-impedance LCR correction brings to the sound is significant.” outweighs ” those (previously unnoticeable!!!) sound artifacts associated with the use of classic high-impedance RC circuits.

According to the scheme.

The first stage is assembled on a 7F7 double triode (6113, 6SL7, 5751,12AХ7, etc. can be used) and has a gain = ~ 30, the second stage is assembled on a 7C5 tetrode (6V6GT, 6F6GT can be used) in a triode connection, its coefficient gain = ~1.8, the LCR module attenuates the signal by about another ~14 dB, so with an input voltage of ~5mV (RMS)@1000Hz at the output of the LCR module we get ~55 mV. Next, the signal is amplified by the third stage (gain = ~12) and, through an interstage capacitor and a level regulator, is fed to the fourth stage - with a transformer load. Depending on what maximum output signal level is required and how low the output impedance is required, the output transformer can be connected with a transfer ratio of 1:1 or 1:0.5, the stage gain will be 8 or 4, and the output voltage will be ~ 5.4 or 2.7V (RMS), the output resistance of the corrector in the second case will be ~ 1 kOhm. In practice, if the output voltage within ~ 1...2V (RMS) is sufficient, then the output transformer can be the same as in the second stage and the output resistance of the corrector in this case will be ~ 600 Ohms. Moreover, if you use transformers with a reduced primary winding resistance of ~ 20K - for example, Hashimoto HL-20K-6 or Silk L-941S, then it is quite possible to use a “classical” double triode with Ri ~ 7K (VT231, 6SN7, 7N7,12AU7, etc.). This will slightly reduce the dimensions of the structure and ease the requirements for the power supply. In my opinion this is very promising version of the corrector - the circuit remains approximately the same, only the lamps are different. 🙂

The power supply is made according to the classic (for my designs) scheme, the anode and filament voltages are stabilized. In principle, if you use high-quality Hashimoto power transformers, then with carefully thought-out installation, it is quite possible to supply the filament with AC voltage, and the anode voltage may not be stabilized by using RCLC filters.

The design is assembled on a standard “classic” Hammond chassis, consisting of a wooden frame and two aluminum (top and bottom) panels. I can’t say that this is the optimal chassis option for a corrector, however, the level of noise, interference and interference at the output is very low. Probably, the power supply voltage is stable and well filtered, and the installation is more or less optimal. 🙂

The corrector has outstanding resistance to overloads, “clicks” and infra-low-frequency interference - interstage transformers help very well with this. In my opinion, although the cost of the structure is quite high, it is reasonably justified, since the price/quality ratio is very good. In this particular case, the use of expensive high-quality transformers and LCR modules gives an obvious, audible and impressive sound result.

A few photos.

May 2018 Vladivostok

Last week, a very peculiar proofreader came to see me for an “examination.” The design from a famous Moscow region master was purchased by the happy owner several years ago, and during all this time it was not possible to “extract” any interesting sound from it. The system in which this corrector was installed is quite good - Audio Note acoustics, a single-ended amplifier with 45 (or 2A3) triodes, a Nottingham table with an excellent set of tonearms and cartridges. However, the system did not “sound”; the sound was flat, compressed and enriched with sibilants. At the same time, the sound from the CD player was significantly better than from vinyl - which, of course, in my opinion is already very strange and suspicious. 🙂 The situation needed to be sorted out.

So, here is this design - a few photos -

At first glance at the printed circuit board, I felt somehow unwell and the reason for this was not the printed circuit board at all. 🙂 And after I drew the diagram, I felt really bad. NOT EXPECTED .

Scheme -

So, the design is based on the well-known classic corrector Marantz-7, built on the principle of active correction, that is, as an amplifier with a high gain, covered by a deep loop of frequency-dependent general feedback. In the case of Marantz, such a circuit design solution was completely justified - firstly, it was “fashionable” then, and secondly, deep OOS makes it possible to obtain and stabilize the specified characteristics of the corrector even when the lamp parameters vary, as well as during their aging, which is very important for serial manufactured product. Nobody paid attention to the “harmful” impact of environmental feedback on sound during the development of the Marantz-7. 🙂

But the “Moscow region” version was more than original - the original amplifier with a high gain remained practically unchanged, and the RC correction circuits were made passive and included at the output of the amplifier, before the output stage - the cathode follower. The first question I had practically At once- but what about overload capacity? Unfortunately, the measurements confirmed my worst expectations.

Shape and level of signals at the outputs of the corrector, input signal 5mV@1000Hz. So far everything looks pretty good.

And here are the waveforms of signals at various points in the circuit at various levels of input voltage. Read the details in the comments to the photo.

First stage lamp anode (Yellow), Corrector output (Blue), input voltage 5 mV@1000Hz

First stage lamp anode (Blue), Second stage lamp anode (Yellow), input voltage 15 mV@1000Hz Noticeable second stage INPUT overload

First stage lamp anode (Blue), Second stage lamp anode (Yellow), input voltage 20 mV@1000Hz Second stage INPUT overload is quite obvious

Second stage lamp anode (Yellow), Corrector output (Blue), input voltage 5 mV@1000Hz

Second stage lamp anode (Yellow), Corrector output (Blue), input voltage 10 mV@1000Hz

Second stage lamp anode (Yellow), Corrector output (Blue), input voltage 15 mV@1000Hz, second stage overload BY INPUT, correction circuit slightly shifts the phase and smoothes the waveform AT THE OUTPUT.

The measurement results are quite obvious - the entire structure as a whole and the second stage in particular starts to be overloaded already at the voltage at the input of the corrector = 15mV, which is completely insufficient.

Based on the average reference data of the most common MM cartridge models, the nominal input signal level for measurements and characterization can be considered as a voltage of 5mV @1000Hz. In this case, if we assume that the HF level on the record is recorded at 0dB, then at a frequency of 20 kHz the nominal input signal level will be ~ 50mV, that is, the corrector must provide a margin of overload capacity at the input of at least +20dB.

According to Shure research, the absolute maximum music signal ever recorded on a long-playing record is 38 cm/s at 2 kHz; at low and high frequencies, record levels drop to 26 cm/s at 400 Hz and 10 cm/s at 20 kHz. In addition, for example, in the famous article - Douglas Self. Design of moving-coil head amplifiers // Electronics & Wireless World 1987 No. 12 - the author’s reasoning leads to the conclusion that the maximum root-mean-square level of the input signal voltage, which you need to focus on when designing a vinyl corrector, should be no less than 64 mV (40 cm /c at sensitivity 8 mV@1000Hz)

Thus, the corrector does not have any significant reserve in overload capacity, which, in fact, is manifested in its characteristic sound - pinched, limited and dull. In addition to the fundamentally incorrect circuit design, the circuit still contains a number of “atavisms” from Marantz - an unshunted resistor in the cathode of the first stage lamp (in the original circuit, the OOS loop was connected to it) and a somewhat strangely chosen value of the grid resistor of the first stage, which determines the input resistance of the corrector. For some reason, instead of the generally accepted standard of 47 kOhm, a 100 kOhm resistor was installed. The ratings of the correction circuits also raise some questions, since measurements revealed a discrepancy (up to +- 2 dB) in the frequency response of the RIAA curve corrector both in the low (20….100 Hz) and in the high (10….20 kHz) frequencies .

The corrector's power supply is built according to a linear-standard circuit - a rectifier with a midpoint, a multi-link RCRCRCRC power filter. The lamp filaments are powered by rectified and stabilized DC voltage.

Power Supply Diagram -

Well, well, this means that the design clearly needs to be improved and, fortunately, if you modify the power supply, reconnect several tracks on the corrector printed circuit board and swap several resistors, you can get a fundamentally better result even without significantly changing the ratings of the parts. *** designations B1 and B2 need to be swapped ***

Here is the new, improved corrector circuit -

As you can see, I have assembled a completely “classical” version of a tube corrector based on triodes with concentrated passive correction connected between the first and second stages. A cathode follower is used as an output stage - a “buffer”. I more accurately recalculated the values ​​of the correction circuits, used capacitors of a different type in the correction circuits and at the output, and also reduced the value of the output capacitor. Taking into account the fact that, as a rule, the input impedance of a power amplifier is about 50 kOhm, it is quite reasonable to limit the capacitance of the output capacitor to a nominal value of 2.7...4.7uF. In addition to reducing turn-on transients, choosing a relatively small capacitance allows you to limit the level of infra-low-frequency interference that penetrates the input of the power amplifier.

Power unit -

In the power supply, I changed the values ​​of several filter resistors, which made it possible to more efficiently distribute the supply voltage between the stages. In order to reduce the likelihood of a breakdown between the filament and the cathode of the output stage lamp, I added a circuit to “raise” the potential of the filament circuit above the general one.

Several photos and oscillograms of signals -

As can be seen from the measurement results, the overload capacity of the corrector has significantly (10 times) 🙂 improved (see last photo - 150 mV at the input instead of the original 15 mV), which is approximately 2.5 times more than recommended by Douglas Self : the corrector will be clean, free, open, dynamic, voluminous and airy. The level of distortion is very low, the resistance to “clicks” is extremely high. The deviation of the frequency response from the RIAA curve in the low frequency region is no more than 0.3 dB, in the high frequency region (12...20 kHz) no more than 0.7 dB.

To date, the design has been tested in three very high-quality setups and has shown itself to be very worthy. Of course, it clearly doesn’t match the sound of an LCR corrector, but among the usual classic RC triode correctors, this design can rightfully be considered one of the best.

January 2018, Vladivostok.

Posted in |

Somehow, on one of the forums, a topic flashed - "The right amplifier - the right acoustics." And I will say this - “ The right amplifier needs the right preamplifier.”

Why is your preamplifier so “correct?” – you ask me. And you will be right in your own way. 🙂

Functionally, the preamplifier consists of three blocks - a power supply, an RIAA corrector block and, in fact, a preamplifier stage with level controls and an input switch. To reduce interference and for greater convenience in a rack with audio equipment, the power supply is made in a separate case.

The power supply circuit is quite traditional for my designs and does not have any features. All power supplies are stabilized, the rectifier is based on semiconductor diodes - a bipolar transistor is used as a regulating element. The voltage for feeding the glow is rectified and stabilized.

The preamplifier block is circuit-technically equivalent to the “Zen Guru” amplifier and today I consider this solution to be the best for a preamplifier stage. This option provides only RCA inputs and outputs, without an isolating balanced transformer at the input. Output transformers - Hashimoto, tubes - Zenith 6J5GT from the 50s. The level control - Gold Point, based on ELMA switches and KOA Speer resistors - in my opinion, is the optimally best solution both in terms of reliability and sound characteristics.

A few words about the RIAA proofreader. During the discussion of the design, it was decided that the corrector, firstly, should introduce as little color as possible into the overall sound signature, have excellent resolution, clarity over the entire frequency band and stable sound characteristics - the “scene” should not “float” depending on on the spectral composition and playback volume. I think that to the best of my ability and within the allocated budget 🙂 I completely coped with the task. The proofreader attended several auditions in very high-class audio systems and always noted both the exceptionally clear elaboration of the subtleties of the rhythmic component of the music, as well as the clarity, stability of the stage, echeloning of musical instruments and voices of the performers. Perhaps, for recordings of “old” jazz, the “separation” is even too good, for example, it is quite obvious that the drum solo in “Take Five” by Dave Brubeck (about the 3rd minute) on the recording is “closer” by the sound engineer, and in “Our” Love Is Here To Stay” it is heard that Ella and Louis were located in the studio at some distance...

According to the scheme:

Two stages using 6AC7 tubes in triode connection. I used integrated current sources as the anode load; this solution made it possible to obtain maximum gain at a very low level of harmonic distortion, which increases very slightly with increasing amplitude of the output signal before it begins to limit. The first stage is with a current source as an anode load, the second stage is the so-called “hybrid” SRPP. In particular, the cascade shown in the diagram has a gain of 42, an output resistance of ~ 800 Ohms, a maximum output voltage swing across a load of 10 kOhm ~ 36V rms, while the harmonic coefficient is no more than 0.3%. The correction circuit is connected between the cascades; as correction elements I used rolled polystyrene capacitors and carbon film resistors, the interstage capacitor is metal-paper, the output capacitor is a composite of MKP film and metal-paper capacitors connected in parallel. Naturally, the lamps for the corrector had to be carefully selected both for the microphone effect and for the required gain and distortion. I was able to pick out two matching pairs out of about 30 pieces. Structurally, the sockets of the lamps of the first stage are placed on mounting panels with vibration isolation, the remaining sockets are on the upper side of the chassis. In this design, I moved away from the typical “common bus from input to output” wiring diagram of correctors. To minimize interference, it turned out to be more correct not to combine the common one with the housing at the ground terminal near the input connectors, but to stretch a separate wire from the terminal and connect the common one to the housing near the first stage.

In general, the construction of a corrector based on the circuit of two successive cascades loaded on current sources seems to me to be a promising idea in terms of “sound”, which, in my opinion, for example, it definitely makes sense to try it on “our” 6S45P lamps, reviews about the sound characteristics of which are very contradictory . It seems to me that in this case, 6S45P can reveal itself in a very unexpected way.

PS (2019) Over the past two years, the Corrector has been repeated several times and acquired its name - “CODA!” (Coda), just like the name of the wonderful Led Zeppelin vinyl. “...Coda in music is an additional section at the end of a piece of music.The content of the code can be an “afterword”, a conclusion, a denouement and a generalization of the themes developed in the development...”

October 2017 Vladivostok

Posted in |

While looking for interesting circuit solutions for RIAA phono preamplifiers on forums, I often came across questions like “...but has no one tried correction on LCR modules and, if so, what is the difference in sound?” Frankly, until recently I also did not “fully” try this type of correction, assuming that the traditional RC option was more than sufficient. However, about a year ago, in the process of a complex and multi-stage exchange of components 🙂, I happened to receive a pair of LCR RIAA 600 Ohm modules from Silk Audio. Around the same time, I listened to them on a breadboard, noting the smooth and dense sound - but with too high sensitivity to interference. At this point the tests ended and the modules went into the “bedside table” until better times. The best times came this summer after I tested various vintage cartridges and the Opera Consonance T1288 single post tonearm on my turntable. Since everything is more or less clear to me with cartridges and tonearms, I decided to further study the types of correction and bring the layout with LCR modules to the finished result.

1. What is the point of using LCR correction?

Firstly, this is a ready-made correction module with a strictly standardized frequency response. Secondly, since the input impedance of the LCR module = 600 Ohms, it is circuit-technically possible to build a corrector without capacitive coupling between stages, using “standard” 600-ohm transformers, which were previously widely used in studio equipment. Moreover, the signal currents passing through the correction circuits have significantly larger amplitudes compared to traditional RC circuits. Thirdly, the DC resistance of the LCR module is low and the output impedance = 600 Ohms, which makes it possible to further amplify the signal using a cascade with a relatively low input impedance, which, in turn, significantly reduces the level of interference at its input. However, in practice this does not eliminate the need for careful shielding of the module. Fourthly, experts I respect claim that LCR, LR and especially Rx correctors sound “more reliable, crisper, clearer and more musical” than RC. I should have heard this too :)

2. Difficulties with the first cascade.

Apparently, modules from Silk Audio are assembled according to the following scheme:

Capacitors, according to Silk Audio, are designed for an operating voltage of no more than 100V DC. As one of the possible options, a “classic-vintage” corrector scheme could look like this:

I could, of course, use other lamps, a galvanic connection between the stages, a rectifier on pp diodes, a stabilizer-filter on a transistor, etc. – but even in this case the device would have turned out ( In my opinion) excessively large and heavy.

The main problem lies in the first stage and its matching to the low load resistance. Firstly, it must amplify the signal by at least 30...50 times, secondly, its output impedance must be below 600 Ohms and thirdly, the constant potential at its output should not exceed 100 Volts. That is, if we consider a simple cascade with a resistor as an anode load, we need a lamp with an internal resistance of no more than 600 Ohms, u = 50....70, with a decent range of characteristics and good linearity at the operating point with +70...+90V at the anode and -1…-2V – on the grid. For example, I don’t know such lamps. 🙂 If we consider “ composite” cascade, then, in principle, SRPP on 6S45P-EB may well be suitable, you just need to check the grid current in the selected mode. In addition to lamps, I also considered options for the input stage using low-noise field-effect transistors. Something like these configurations may well work, although, of course, transistors are not our method:

3. Layout and final diagram.

During the prototyping process, I decided to try the so-called “hybrid SRPP”:

4. Corrector circuit:

Just in case, for greater clarity, I provide an approximate calculation by overload capacity first cascade.

Output voltage of a “typical” MM cartridge at 1000Hz at a linear recording speed of 5cm/sec it is ~ 5mV. The maximum linear recording speed on an LP disc is limited by the width of the audio track and cannot be more than ~ 12cm/sec, the voltage at the corrector input will be = 12mV. Let the first stage have a gain = 50, then the voltage at its output will be ~ 0.6V. Based on the selected mode, the maximum output voltage at a load of 600 Ohms = 6...7V, which, in general, provides a good margin for overload capacity. However, it is worth noting that if your collection contains a lot of EP discs of “45”, the maximum linear recording speed of which can be up to 33cm/sec, then it is advisable to slightly modify the input stage of the corrector. In particular, the option on field-effect transistors with a bias voltage of 200mV and a power supply voltage of less than 40V in this case does not look at all attractive.

So - First stage - 6AC7 (6Zh4) in triode connection, operating point 90V@15mA offset = - 0.7...1V. An IXYS IXCP10M45S integrated current source is used as an anode load; the signal is taken from its cathode. In this configuration, the cascade has a gain of ~ 40...50, an output resistance of ~ 50 Ohms, with a maximum load current of about 10... 12 mA, which at a load of 600 Ohms allows you to obtain a signal amplitude of up to ~ 6... 7V.

The second cascade has no special features; a 1:1 transformer with Ra = 5K is used as a load. It is quite possible to build a second cascade according to the same scheme as the first.

The power supply is typical for my designs - the anode voltage is stabilized, the stabilizer is a simple parametric one on a field-effect transistor. The filament is powered by rectified and stabilized DC voltage.

Main characteristics:

• Input impedance = 47 kOhm (can be changed by installing additional resistors)
• Output impedance =< 2 кОм (в варианте коммутации выходного трансфоматора 1:1)
• Nominal output voltage ~ 1V RMS
• Maximum output voltage at 10 kOhm load = 60V RMS
• Gain ~ 180
• <150uV (“взвешено” по кривой “A”)
• <= 0.2%, в основном 2-я и 3-я гармоники. Уровень третьей гармоники относительно уровня второй <= -20 dB.

A few photos -

Note the “cornered” Nagra PL-P preamplifier.

The corrector is assembled in one housing with a preamplifier, which is also the previously mentioned amplifier for high-impedance headphones - Zen Guru . I will publish the diagram a little bit later.

August 2016 Vladivostok

P.S. About sound. In the same case, in the same place before the LCR corrector, there was an RC corrector for 6SF5 + 6AC7. The power supply and internal wiring remained almost the same as before the modification.

Therefore, I believe that I fully grasped the characteristic differences in the presentation of “sound” from changing the type of correction. Firstly, this is the low-frequency region - with LCR they are more full-bodied, the resolution is higher, the transition from low-frequency to mid-range frequencies has become, as it were, “smoother and clearer” :) Secondly, this is a more stable “scene” when changing the volume and slightly better volume , fullness of sound. Thirdly, the transition from midrange to high frequency has also become “smoother and clearer”. We can say that the sound from LCR is while maintaining musicality and plasticity – made it possible to more clearly hear some previously eluding subtle features of the recording. In general, the use of LCR modules in correction is quite justified 🙂 and I’ll probably continue experiments with them.

Posted in |

Novice “vinyl lovers” often ask me about a corrector circuit that is easy to assemble and does not require special adjustments, based on inexpensive and accessible Soviet-made lamps. Well, I have such a scheme :)

Comments on the corrector diagram.

In my opinion, this is the most optimal and high-quality circuit based on 6N2P-EV, 12AX7 lamps. The first cascade - the lamps of one cylinder are connected in parallel, this reduces the internal resistance, which, in turn, reduces noise and reduces the output impedance of the cascade. Thus, the correction circuits load the first stage less and the signal losses on them are less. The second stage is with a cathode follower at the output, which provides low output resistance and makes it possible to work with a long cable and a load resistance of 10 kOhm.
As for the capacitors in the correction circuit, there is no high voltage on them, so you can use high-quality foil low-voltage polystyrene capacitors. The interstage and output capacitors must have an operating voltage of at least Ua. Cathode
capacitors – Panasonic FK, FC series. It is better to use lamp panels
with “glasses”. Voltage power source can be within +220...+300V (maybehigher, but correction of the values ​​of resistors R9, R10 will be required). Setting up the circuit comes down to monitoring the operating modes of the lamps and selecting lamps based on the same final gain of the left and right channels. The voltage at the anodes of the lamps of the first and second stages, depending on the voltage of the power source, should be within 100...150 Volts. I recommend stocking up sufficient number of lamps, 10 pcs 6N2P-EV - this is the minimum for selecting an identical set. And also - 6N2P lamps Necessarily must have an index EV . Regular“simple” 6N2P will not work, don’t waste your time on them.

Power unit.

Since novice vinyl drivers do not use transformers “as they should be,” but “as they are available” 🙂, to eliminate various difficult “surprises,” I recommend making the power supply in a separate case. The circuit is quite standard - a rectifier, a field-effect transistor filter. If the secondary winding of the existing transformer is one without a tap from the middle and for a voltage of 200...250V, then a bridge rectifier can be used.
The filter transistor and stabilizer are on radiators, can be mounted on
metal body through insulating gaskets. The filter transistor practically does not heat up, and the filament voltage stabilizer will
quite hot.

Good Sound!

January 2015 Vladivostok

Posted in |

The other day I had a “medical examination” 🙂 quite an interesting preamplifier from YBA – model 2 “Alpha”. The signal level when connecting the player to the “Phono” input was low and there was some level imbalance across the channels. But this not the most important. 🙂 It’s interesting how this design solves the “problem” (*** But in general, how significant is it for transistors?) reducing the influence of external vibrations on the signal. I simply have no words, only photographs.

The amplifier circuit is mounted on the back side of the board, surface mounting. Almost classic transistor circuit design, nothing interesting.

October 2014 Vladivostok

Posted in |

One long winter evening, while raking out the bins, I suddenly found a wonderful pair of lamps -

And it just so happened that at the same time, my very good friend Vladimir asked me to make a vinyl corrector for him. This is definitely fate :)

The development and calculation of the scheme took several days. The basic operating conditions were as follows: MM or MI cartridge, relatively short connecting wires, input impedance of the power amplifier (also, by the way, made by me) = 20 kOhm, sensitivity 300 mV. I decided to use the classic solution - three cascades + passive concentrated correction. The lamp triodes of the first stage are connected in parallel - this allows, firstly, to reduce the noise level and, secondly, to reduce the internal resistance - which, in turn, allows the use of resistors with a nominal value of no more than 200...250 kOhm in the correction circuit. I can’t say that I wasn’t at all worried about the increase in the input dynamic capacitance of the parallel triode, but preliminary calculations and subsequent measurements showed that my worries were excessive. The calculation of the correction circuits was performed in an Excel table (see section Literature).

“Through” frequency response taken with an inverse RIAA filter - (Note the scale along the “Y” axis)

Briefly about the scheme.

The first stage is with a common cathode, gain = 48, output resistance ~18 kOhm. The correction circuit uses polystyrene foil capacitors and Dale resistors with an accuracy of 1%. The interstage capacitor is “our” K40-U9, Jensen PIO is also quite suitable. The signal attenuation in the correction circuit is approximately -18dB. The output stage is composite, with galvanic coupling, according to the circuit of a cascade with a common cathode + cathode follower. The second stage gain = 16, the cathode follower provides the necessary matching with the interconnect cable and the power amplifier input. There is a certain “audiophile prejudice” about the use of cathode followers in audio circuits. In my opinion and hearing, everything is fine with repeaters, you just don’t need to demand the impossible from them, for example, linear operation with a load exceeding the calculated output impedance by only 10 times. Exceed by 20 times– and everything will be fine with the music :)

The power supply is made in a separate housing. The transformer is toroidal, with a power of 50VA, covered with a thick steel casing. The anode voltage rectifier is a bridge, on FR157 diodes, the voltage is filtered by an electronic filter on the VT1 transistor, which also ensures its smooth supply. Lamp incandescence relevant cascades connected in series and powered by rectified and stabilized DC voltage. Since the maximum allowable voltage between the cathode and the filament for 7N7 lamps is 90 Volts, the filament is “raised above ground” by approximately 50 Volts by the R4R5 divider.

Main technical characteristics.

• Output impedance =< 1 кОм
• Nominal output voltage = 0.32V RMS
• Nominal input voltage = 4mV RMS.
• Maximum output voltage at load 20 kOhm ~ 35V RMS
• Gain at 1 kHz ~ 80
• The level of intrinsic noise and interference at the output with a “closed” input =<190uV (“взвешено” по кривой “A”)
• Deviation of the total frequency response from the RIAA standard in the frequency range 20Hz…20kHz = no more than 0.5dB.
• Harmonic distortion at 1 kHz into a 20 kΩ load at rated output voltage<= 0.3%, в основном 2-я и 3-я гармоники. Уровень третьей гармоники относительно уровня второй <= -20 dB.

Complete with the MI cartridge Grado Prestige Gold, the sound of the corrector is very free, spacious, with excellent musical resolution and excellent tonal balance. To be fair, it should be noted that the corrector on C3g pentodes is somewhat more “fast and dynamic”. But for the musical genres that Vladimir prefers, this is completely unimportant. 🙂

A few photos -

May 2014 Vladivostok

Update dated September 15, 2014- 7F7 lamps are also used in the output stage. In this case, resistors R10 and R11 = 100 kOhm. Output stage gain = 39...42, the final corrector gain increased to 190..193. Thus, with a “standard” output voltage for most MM/MI cartridges of ~ 4mV (@1000Hz, 5 cm/sec), the signal level at the output of the corrector is ~ 0.77 V RMS (0 dbU). The output impedance at this output level is approximately 600 ohms. The minimum load resistance at the output of the corrector must be >= 10 kOhm.

Posted in |

Preamplifier-corrector based on field-effect transistors

This circuit was created by me back in 1988 - for the Aria -102 player. I remember that initially I assembled a version on the K157UD2 microcircuit, but upon direct comparison, the design on the operational amplifier seemed to me to be significantly poorer in sound than on field-effect transistors. Therefore, after the recent revival of vinyl in my collection - the first proofreader I decided to collect - was this very diagram. I really wanted to check its sound properties - was it really as good as it seemed to me then :) Moreover, to my surprise, on the Internet I found a kit for assembling a corrector with approximately the same circuit as the “drawn” one. me 25 years ago. The set was immediately purchased, the ratings of the correction circuits and transistor modes were recalculated. As a result, the scheme took on the following form:

“Through” frequency response recorded with a reverse RIAA filter –(Note the scale on the “Y” axis)

The corrector is extremely simple - in the “basic” version there are only two amplification stages, the first on a low-noise field switch 2SK170GR (Idss = 2.6...6.5 mA), the second simply on a suitable field switch 2SK246GR (Idss = 2.6...6.5 mA). Operating mode of the first stage: quiescent current = 1.5mA. offset voltage = -0.27V, gain = 125 (with source bypass capacitor). A passive RC correction circuit is included between the stages. With a good degree of accuracy, the output resistance of the first stage can be considered = R3, and the values ​​of the elements of the correcting circuit can be easily calculated using the Excel table given in section Literature. The signal loss in the correction chain at a frequency of 1 kHz is approximately 20 dB. Mode of operation of the second stage: quiescent current = 2mA, bias voltage = -0.47V, gain = 15, output impedance approximately 10 kOhm. To work on a long (more than 1.5m) cable, it is advisable to supplement the output stage circuit with a source or emitter follower on another transistor. Thus, the total gain of the circuit at 1 kHz = 188, the input overload capacity is approximately 20dB at 100Hz, the nominal output voltage = 1V rms, the maximum output voltage = 12V rms. In general, very good parameters for such a simple design.

The power supply is assembled according to a voltage multiplication circuit, this can significantly reduce the switching interference of rectifier diodes, the rectified voltage is filtered by a filter on a bipolar transistor T1.

Main technical characteristics –

• Input impedance = 47 kΩ (can be reduced by adding additional resistors)
• Output impedance =< 10 кОм (в “базовом” варианте)
• Nominal output voltage = 1V RMS
• Maximum output voltage at 100 kOhm load = 12V RMS
• Gain ~ 188
• The level of intrinsic noise and interference at the output with a “closed” input =<190uV (“A-weighted”)
• The deviation of the total frequency response from the RIAA standard in the frequency range 20Hz ... 20kHz = no more than 0.8dB.
• Harmonic distortion at a frequency of 1 kHz into a load of 100 kΩ at rated output voltage<= 0.3%, в основном 2-я и 3-я гармоники. Уровень третьей гармоники относительно уровня второй <= -15 dB.

Some time ago, to my good friend, music lover and esotericist Nicholas a Dual vinyl turntable came into use with a very promising MC cartridge Audiotechnica AT-33EV. Naturally, I urgently needed a proofreader and he turned to me. 🙂 The requirements were the following - crisp, clear and dynamic sound, without any “vintage” touch. Power supply - without electrolytic capacitors. Transformer output stage, signal and output transformers – Sowter. One block. Dimensions don't matter. Well then - they don't have it, they don't have it🙂 This is how this design arose - in a large aluminum case of natural color, dimensions 45x25x35cm. Der Frankenstein.

The corrector is two-stage, with passive correction; wonderful C3g lamps from Siemens are used in the first and second stages. The requirements for the first stage with this corrector configuration are quite stringent - it must have a relatively high gain with a minimum noise level, good overload capacity, stable output impedance and low dynamic input capacitance. Based on these requirements, it is quite logical to use C3g in the “native” pentode connection. The second stage should have low output impedance and excellent overload capability with moderate gain. C3g in triode connection is an option close to ideal :) The operating mode of the first stage is voltage at the anode = +175...180V, voltage on the second grid = +110...115V, bias voltage = +1.5...1.7V. gain = 95…100. I should note that C3g in a pentode connection “sounds” well over a fairly wide range of anode loads. To match the cartridge, a specialized step-up MC transformer Sowter 1990 (1:10) was used. Low-voltage “rolled” polystyrene capacitors, known for their excellent sound properties, are used in the correction circuits. Due to their low operating voltage, the correction circuit is connected “traditionally”, between the amplification stages. The signal loss in the correction circuits is approximately 20dB. Modes of the second stage - voltage at the anode = + 155 ... 160V, bias voltage = + 2.6 ... 2.8V, gain = 45 ... 50 output resistance = 2.3K. Output transformers Sowter 9525. Taking into account the transfer coefficient of the input transformers, the final gain of the corrector at the MC input is about 5000, when playing the “0 dB@1000Hz” track of the test disc with an AT-33EV cartridge, the voltage at the corrector output is 1.5V RMS. The secondary winding of the output transformer has several taps, which allows you to adjust the output voltage level and, if necessary, lower the output impedance of the corrector. Takman resistors of the REX series are used in the correction, all other resistors are Kiwame. Capacitors shunting cathode resistors - Panasonic, interstage capacitor - Jensen (copper foil paper oil). Capacitors in power circuits - ASC. (Teflon + oil). The installation was done using Siltech silver-gold wire.

Corrector power supply circuit –

The anode voltage rectifier is assembled according to a midpoint circuit; the rectified voltage is filtered by an electronic transistor filter, which also ensures a smooth increase in the anode voltage when the device is turned on. From the filter, power is supplied to each of the corrector channels through additional decoupling esoteric LC chains. The filament of the lamps is powered by a rectified and stabilized voltage of 12.6V, the filaments of the lamps of each channel are connected in series. As I mentioned earlier, the corrector and power supply are mounted in one large aluminum case. The bottom of the case is made of two aluminum plates, fastened together with a vibration-damping adhesive. The lamps and circuit parts are mounted on a separate thick (12mm) aluminum plate, attached to the bottom of the case through four vibration-damping posts.

1. Corrector

To control the frequency response of the corrector, it is convenient to use the so-called Anti-RIAA chain, for example the one in the article “On Reference RIAA Networks” by Jim Hagerman. (see section ) Scheme -

To measure the final frequency response, the circuit is connected between the generator and the corrector being tested. When using capacitors with an accuracy of 5% and resistors of 1%, when measuring the final frequency response, compliance with the RIAA standard is ensured with an accuracy of 0.5dB - which is more than enough. It is convenient to use a computer with a professional-quality sound card and relevant a set of connecting cables. To take measurements I I recommend use the program True RTA (Level 4).

Anti-RIAA chain o It’s very convenient to implement it as a separate module -







2. Cartridge + cable + corrector

After bringing the frequency response of the corrector to the standard, it is advisable to remove the frequency response of the “cartridge + connecting cable + corrector” system in the HF region, this is especially true for MM cartridges and correctors, the input stage of which is made on a triode with a high gain. The purpose of these measurements is to check the absence of frequency response deviations in the HF region caused by the joint interaction of the cartridge inductance, the capacitance of the connecting cable and the input capacitance of the first stage of the corrector. To do this, use the simplest scheme -

Deviations in the frequency response are compensated by selecting the value of the load resistor at the corrector input. “ The recommended value of 47...51K by most manufacturers is only a “starting point”. A corrector, the first stage of which has a small input capacitance, complete with an MC cartridge loaded on a matching transformer, will have smoother frequency response in the HF region, compared to most MM and MI cartridges connected to the input of the same corrector. The combination of a high-gain triode front end, a long connecting cable, and an MM (MI) cartridge is the most problematic in terms of the “behavior” of the resulting frequency response at high frequencies.

3. Table + tonearm + cartridge + cable + corrector

The next stage is the removal of the final frequency response of the entire system - a player + cartridge + connecting cable + corrector. After checking, using the appropriate templates, the correct installation of the tonearm, the cartridge on the tonearm and setting the optimal downforce, a measuring plate is installed on the player. For example, these would be suitable:



Before starting work - on the corresponding track, by controlling the balance of the channels, it is necessary to check the correct installation of the cartridge in the horizontal plane. Then the frequency response is taken, special attention should be paid to the low-frequency region, any deviations (constant or periodic) in the frequency response in this area may be due to the mechanical resonance of the tonearm, the penetration of noise and background noise from the engine control circuits to the input of the amplifier, uneven rotation or violation of disk geometry. As a rule, if the mechanics of the player are in good order, the cartridge is installed accurately and the final frequency response of the “cartridge + cable + corrector” system was previously set correctly, the measuring plate will not show any significant deviations in the frequency response. In this case, your kit can be considered More or less tuned.

If you want the sound of your system to always be reference– carry out the setup procedure every time you replace the cartridge :)

Vladivostok, 2013

Posted in |

If you look at the tracks of a record with a magnifying glass, you will see that the tracks are by no means perfectly parallel to each other. Their edges waver and twist from side to side, sometimes ending up dangerously close to neighboring paths. These throws are determined by the amplitude of the low-frequency components of the signal and it is they that limit the recording density, and therefore the playing time of the record.

Recording high-frequency signals involves a different kind of nuance. If the amplitude of the high-frequency details of the recording is small, then the level of these details will be comparable to the level of the record's own noise. In addition, high-frequency oscillations are troublesome to read - the mechanical elements of the reading system have mass, that is, they are inert, which imposes restrictions on the frequency of oscillations that can be read and converted into an electrical signal, and they are not absolutely elastic bodies, that is, part of the read high-frequency information does not reach the plate surface to its destination - the sensor, but is damped in mechanics - therefore, high-quality needle holders tend to be made from the lightest and hardest materials, such as beryllium. Among other things, the lighter the element, the higher its own resonant frequencies, and the shift of the resonance frequencies of the mechanical elements of the sound-producing path further beyond the audible region is a problem that has long been familiar to developers.

It seems obvious that in order to restore the output signal in a form as close as possible to its original state, the transformation curves carried out during recording and playback must a) correspond to each other, be mirror images of each other, and b) be regulated by the appropriate standard so that any record can be play on any player. This was not obvious, however, for about a quarter of a century - until the 1950s, record manufacturers implemented similar frequency correction "who cares what", which now results in a headache for those who want to hear an old record in the "correct" quality.

Strictly speaking, attention was paid to the nonlinearity of the frequency response of records back in 1926 - almost immediately after the advent of electric recording, in 1930 the question arose about what to do with the noticeable increase in the mid-frequency region introduced by condenser microphones, and by the mid-1930s, correction of the reproduced signal was already in full swing practiced - for example, on the radio. Accordingly, correction began to be used in the production of records. But only in the 1940s did a premonition of the need for a single standard arise, which transferred from a premonition to a requirement of the time at the border of the 1940s/1950s - when the marketing battles Columbia vs RCA with media formats and recording speeds spread to correction circuits, darkening the cloudless future of the recording industry with an anarchic multiplication of entropy.

Since 1942, NAB (National Association of Broadcasters) began work on the standard and in 1949, NAB recommendations began to be used in the production of records; after the presentation in 1948, Columbia made their correction scheme public; in 1949 RCA responded with its "New Orthophonic" equalization scheme, the details of which were published in 1953. As a result, the RIAA (Recording Industry Association of America) was created in 1952 to develop a single standard. Through her efforts, by 1955-1956, a standard was formed, which, with minor additions, is used to this day. It's funny, but now on the RIAA website technical standardization is last on the list of tasks, and the first place is - that's right - the fight against piracy. Standards are standards, but the most sensitive place in the body is still the wallet.

But it was a saying: so to speak, the generally accepted version of events, and now -.

Article published 2011-09-21
The author of the articles or the translator is Dmitry Shumakov, unless otherwise indicated. When quoting, please put a link to the record store siteBe the first to comment!

This article is for those who still love and appreciate vinyl sound, despite all the digital modern things :)

The corrector is used to amplify and correct the signal that comes from the electric playing head of the EPU with a diamond or corundum needle. The work of the corrector is based on the RIAA standard; it regulates the basic requirements for recording and playback of recordings from vinyl discs. According to the RIAA standard, the frequency response has the form shown in Fig. 2. For this reason, in order to achieve linearity of the frequency response of the playback track, you need to use a phono stage; its frequency response is shown in Fig. 3.

Rice. 2

Rice. 3

A diagram of a practical amplifier - phono stage is shown in fig. 4, and the diagram of the power supply unit is shown in fig. 5.

Rice. 4

Rice. 5

The basis of the circuit consists of a two-stage amplifier, which is built according to the classical circuit of a voltage amplifier with a resistive load. The frequency correction of the signal is created by a passive frequency correction circuit. To ensure reliable operation of the filter, it is placed in the cut between two amplification stages.

The graph of the actual frequency response of the phono stage is shown in Fig. 6. As you can see, the type of practical characteristic is almost no different from the theoretical one.

Rice. 6

Elements, design and setup

For correct and reliable operation of the corrector, all elements that are used in its assembly must be of the best quality and must have a minimum nominal error tolerance. Maximum rating tolerance for frequency correction circuits is ±1%. For the rest of the circuit ±5%. It is possible to use elements with a large tolerance, but then you need to individually select the elements according to their nominal value. It is also recommended to use radio tubes with military approval and EB marking (that is, with increased durability and mechanical strength).

The body of this device can be made with closed or open radio tubes. The body can be made of metal (steel, copper, brass, etc.), plastic and wood. In the last two cases, additional shielding of the internal circuit with copper or brass foil is required. Figures 1 and 7 show one of the possible design options for a phono preamplifier.

Rice. 7

Particular attention should be paid to the phono stage's power supply, since the main problem with preamplifiers is considered to be a high background level. To minimize the background level when assembling the power supply, you need to take several measures. First of all, the power supply must be made in its own separate housing (to prevent the influence of electromagnetic fields from the network transformer). It is better to place the network transformer in the screen, or at least wind an additional screen winding on it. The diagram shows the minimum values ​​of all electrolytic capacitors. To reliably eliminate the background of their capacity, it is better to increase it by 1.5 - 2 times. The value of capacitor C1 is especially important, because the filament voltage of the device (unlike the anode voltage) is not stabilized. Stabilization of the anode voltage is achieved using an “Electronic choke”. Separating the power supply of stereo channels is not necessary, since the separation of channels during recording is very small.

This is all. Goodbye.