Here are located scheme,detailed description, options Acoustic speakers speakers Radiotehnika class S90 (S90, S90B, S90D, S90F)

High-quality acoustics from Soviet times, after minor modifications and restoration, we can confidently say that they will give a head start to many modern acoustic systems. If you have similar ones lying around or bought them somewhere cheap, then put them in order and they will delight you for a long time with powerful bass, rich mids and high frequencies in musical works of any style and direction.

S-90 first model

In the speaker system
S-90 There are two step playback level controls separately for mid and high frequencies in the ranges from 500 to 5000 Hz and from 5 to 20 kHz, respectively. Both regulators have three fixed positions: “0”, “-3dB” and “-6 dB”. In position "0", the signal from the crossover filter is supplied to the corresponding head directly. In the “-3 dB” and “-6 dB” positions, the signal is weakened relative to the “0” position by 1.4 and 2 times, respectively.
With the appropriate spectral composition of the program, switching the regulator changes the timbre coloring of the sound.

S-90

Nameplate power 90 W
Rated power 35 W
Nominal electrical resistance 4 ohms
Frequency range 31.5-20000 Hz
Nominal sound pressure 1.2 Pa
Overall dimensions of speakers 360x710x285 mm
Speaker weight no more than 30 kg

Schematic diagram of S90

IN AC There is an indication of speaker head overload. Regulators located on the front panel AC, make it possible to smoothly adjust the sound pressure level of the high-frequency and mid-frequency loudspeaker heads in the range from 0 to minus 6 dB.
There is also a model of the speaker system" S-100D", it uses a mid-frequency head 30 GDS-3 with MAHID magnetic fluid, which allows you to increase the rated power of the speaker system to 100 W. The rest of the design" S-90D" And " S-100D" are similar.
To operate, the speakers must be connected to an amplifier that has the highest (maximum) power at the output of each channel, ranging from 50 to 150 W.
If the OVERLOAD indicators start to light up when the speaker is operating, then you should reduce the level of the input signal supplied to it (by using the volume control in the amplifier to which the speaker is connected).

Passport specifications S-90D

Nameplate electrical power " S-90D"/"S100-D" not less than 90 W 100 W
Rated electrical power 35 W
Nominal electrical resistance 8 ohms
The range of reproduced frequencies is no longer 25-25000 Hz
Characteristic sensitivity in the frequency range 100-8000 Hz, at a power of 1 W, not less than 89 dB
Overall dimensions of speakers 360x710x286 mm
Speaker weight no more than 23 kg

The figure below shows the principle scheme speakers S90D.

Schematic diagram of S90D

Speakers S90 diagram, description.

Related materials:

Reworking the S-90 filter

Having been an opponent of audiophilia as a simplification, after experiments I changed my point of view and now I’m even ready to sacrifice something for the sake of a small number of obstacles in the path of sound :). This is actually very important, even on the speakers discussed below. But this also forces you to sacrifice some things: high power and congestion of frequency bands.

I used the crossover below, broken down into pieces, for my s-90de with speakers: 30GD-2, 6GDSH-5-5, 3GD-2, where it plays simply wonderful with any genre of music. 3GD-2 (its worse analogue 6GDV-1-16) is a very old high-frequency speaker (my copy is from 1977) with a resonance frequency of as much as 4500 Hz (but there is an opinion that in this place it is quite calm), so the high frequency of the midrange section is HF is due precisely to this fact. However, most domestic tweeters have not gone far, so I consider this cut to be very good for them.

This filter will work great on good foreign mid-high-frequency speakers, which I tried myself :). But, of course, it needs to be changed taking into account everything new (including the frequency of the section) - taking the principle itself as a basis.

p.s. Still, we should not forget that everything in the world is not only relative, but also subjective :). In addition, at the moment I have absolutely no means of measuring the frequency response of my system - everything is adjusted by ear in the same room...

speakers

NC: Let's look at the generally good bass driver used in the s-90. 30GD-2 (75GDN-1-4) with a nominal resistance Z=4Ohm, sensitivity S=86dB (or dB/W*m) and frequencies F=30-1000Hz does not provide the best IFC (impedance-frequency characteristic:)) in the compartment with poor sound at frequencies above 500Hz.

Our cutoff will be at 500Hz. Ideally, to make this speaker work really well, you would cut off everything above 200Hz. After all, the main drawback of the 30GD-2 is that at these frequencies it mumbles (“sound from under the diffuser hood”) and plays very poorly. But to make such a low crossover frequency you need an excellent midrange speaker with a resonance frequency of no more than 70Hz.

MF: The standard mid-range driver 15GD-11 (20GDS-4-8), with parameters Z=8Ohm, S=89dB, F=200-5000Hz, does not stand up to any criticism either in terms of sound or the characteristics we need. Therefore, it needs to be replaced with the nice baby 6GDSh-5-4 (Z=4Ohm, S=92dB, F=150-12000Hz) which looks completely frivolous, but in fact turns out to be very good. In addition, it has the sizes we need, price (no more than $4!) and availability in Russia.
It should be noted the low power of the 6GDSH-5 (as a result, the inability to work at discos/parties) and bursts in some parts of the frequency range (“loudness”).

There were opinions that the 6GDSH-5 has poor directivity at high frequencies, which is why the stereo panorama is “unstable” at a relatively high section. It seemed to me that this was not so, so if there are problems, act according to the circumstances :).

HF: Any tweeter with parameters S=89-92dB and Z=16Ohm will do. It is important to note F (actually, the minimum operating frequency of the speaker) - it should not be more than 4500Hz, and the lower the better.
Structural dimensions and fastenings are selected on site using available means.

sensitivity

MF: To cut off the extra 7 dB (92-85 = 6), I suggest using the option of one resistor, which will avoid unnecessary elements in the circuit and at the same time reduce the ratings of the filter elements due to an increase in the speaker resistance. Resistor R2=4.3Ohm will give us a reduction of 6dB. The sensitivity is reduced by a resistor in the approximate ratio of 1 dB/0.7 Ohm. Coil L1 has its own resistance of 0.75 Ohm and will help us remove another 1 dB. Voila! :)

However, the disadvantage here is that there are no exact formulas and dependencies, and the values ​​I gave appeared as a result of my personal feelings.

VC: We use the same method, selecting the desired resistor until the desired result is achieved. However, in this circuit there are no filter elements with a high intrinsic resistance, so resistor R1 must be taken with a margin of 1 dB. We also note that the volume of the high-frequency speakers relative to others in the system strongly characterizes its “inclinations” - for example, most listeners like a slightly muffled high-frequency sound (by about 1-2 dB), the system seems to be “softer”. What is important for domestic high-frequency speakers is not of the best quality :)). For heavy music, emphasizing the high frequencies may be more important.

It's nice to know that changing the sensitivity resistors within one unit (1 Ohm) has virtually no effect on the filter itself and the cutoff frequencies, which makes it possible to experiment.

But you shouldn't cross the 0.7 Ohm difference when experimenting with R2 - the L1 coil is much more sensitive to this change.

inductors

The most difficult thing. We urgently need to find ways to measure inductance, otherwise precise tuning will not work.

In the absence of a way to measure, I suggest the following: compare the coils by their own resistance, taking into account all design parameters. Theoretically, if all the factors influencing the inductance rating coincide (there are some very interesting ones - the density of turns, the content of iron impurities in the frame :)), then you can obtain the necessary inductance, as if “following a model”.

Despite everything, this method, it must be said, is very inaccurate. There is no difference between the inductance L2, for example, 1.5 mH and 1.27 mH in terms of resistance.

LF: I’ll give you my parameters for a large coil (it also has “ears” on the sides): inner diameter of the ring: 35mm, outer: 70mm, coil height: 37mm, width of the winding area (height without sides): 30mm, wire thickness (copper, enameled ): 1mm. With these parameters, the DC coil resistance (measured with a digital tester): 0.8 Ohm.
If these parameters are observed, you should get an inductance in the region of 1.0-1.6 mH, congratulations :).

You can wind the coil the “old fashioned” way, knowing how many turns you need to make. This has recently become known: for 1.27 mH, 210 turns of “manual” (not very neat) winding are required. In this case, for every 0.05 mH there are approximately 5 turns.

SC: The small coils should all be the same in frame; I took the one with the smallest inductance. Inner ring diameter: 12mm, outer: 32mm, coil height: 23mm, winding area width (height without edges): 18mm, wire thickness (copper, enameled): 0.5mm. Resistance: 0.7 Ohm, inductance 0.18-0.21 mH.

At 0.18 mH the number of turns is 127 pieces. At 0.21 mH - 136.

By the way, do not repeat the mistakes of the USSR assemblers, do not fasten small coils with screws inside - the inductance will change and nonlinearity will be added; attach with glue.

For those who measure themselves: it is useless to try to rewind a small coil with a thick wire from a large one, and you probably want to do this :). Even having completely wound the entire frame, I did not get an inductance of more than 0.1 mH.

At the same time, if you build a new optimal frame (see links, “Cec”), which is not as simple as it seems, then the coil’s own resistance will allow you to gain 1 dB to the speaker sensitivity - you will need to slightly calibrate the sensitivity resistors in front of the speakers.

If you try to find the same large frames somewhere else and wind the L1 coils with thick wire, then their resistance will be approximately 0.4 Ohm - also better.

p.s. I kindly ask you not to write me letters asking for help in calculating inductance on other frames and other values ​​using this method. Assemble the “box” (see links), it is very easy and will solve all your problems with precise winding of coils.

capacitors

Everything is extremely simple. You need to find the same values ​​for decent quality capacitors, you can read about the types here, and about resistors there, by the way. Capacitors can be combined (summed) in parallel (as well as reduced according to the resistance rule by connecting in series). If you have disassembled the s-90 filters, then you should already have a good set of necessary containers :).

Among the domestic ones, instead of the film K73-xx that you probably came across, I recommend trying the metal-paper MBxx - a “softer” sound. If you have the funds and availability, it is advisable to use foreign MKP (1uF ~ $1.1, domestic equivalent - k78).

The capacitors, of course, are non-polar and for a voltage of at least 40V. The quality of the elements in Zobel circuits is just as important.

Here you can experiment with changing the “color” of the system that the capacitors give. I recommend trying to bypass all capacitors (except those in the Zobel circuit) with small (around 0.1 µF) capacitors of other, usually higher quality, types. For example, polystyrene (k71-7) or mica (SGM) - the result is more detailed sound at mid-high frequencies and increases the transparency of the system. In addition, metal-paper (MBxx) capacitors give a slightly “muddy” sound. To bypass means to combine together in parallel :).

resistors

With a power of at least 2W, with less, overheating and a change in rating are possible. Among the domestic ones, MLT-2 can be used. PEV-10 from the s-90 kit are not the best, but they will reluctantly go... I recommend Chinese ceramics - they look like white teeth, they are large, they are inexpensively sold everywhere in radio stores (power up to 15 W), but the range of ratings is fully present.

In other matters, low-power MLT resistors also work well at non-disco power levels, at least in place of R1.

Please note that the value written on the resistor is not necessarily the same as what it actually is. I strongly recommend selecting resistors by measuring them with an ohmmeter/tester. The diagram shows clearly measured resistors.

When finalizing the speakers, it is highly recommended to place resistors R1 and R2 as close to the speakers as possible - directly on the terminals. This will greatly reduce the influence of the cable (which is after these resistors, but not before them) on the sound.
Zobel chains

The reason is that the impedance of the speaker is not constant and increases as the frequency response decreases. This effect occurs in all dynamic type heads without exception, regardless of the country and year of production. More precisely, the Zobel circuit (in my filter only a simplified version of it is used; full ones allow you to adjust the impedance at low frequencies, which is not always necessary) is necessary for the normal operation of the filter inductors, with a sufficiently large self-inductance of the speaker coil. Without a Zobel circuit, the operation of the inductor as a low-pass filter is grossly disrupted and filtering is practically not carried out at all (!).

LF: Elements R4 and C4. It is advisable to set C3 to more than 60 µF, but even this is sufficient for a crossover frequency of 500 Hz. R4 is equal to 4.3 Ohm.

Compare the ICHH of 30GD-2 without and with Zobel. The graphs are approximate, but there you can see the tuning frequency of the s-90 bass reflex - the second huge rock on the left, before 100Hz :).

SC: ICHH 6gdsh-5. You can try smoothing above 3 kHz with Zobel R3, C3. For this, 10-20 µF and an 8.0 Ohm resistor are enough.

Important: a Zobel circuit on the midrange is required for the normal operation of this crossover. Without it, the “new lightweight filter” showed its complete inconsistency in the midrange and high frequencies.

HF: Due to the low inductance of the speaker's own coil and the cutoff at low frequencies, the circuit is irrelevant.

filter

In all frequency sections, a passive all-pass filter of the first order is used with an attenuation of 6 dB per octave (frequency change by a factor of two), Butterworth approximation. Actually, the filter itself was calculated by the JBL Speaker Shop program and slightly adjusted manually :)).

LF: Low pass filter. As you can already understand, the cutoff frequency is 500 Hz (for 30GD-2/75GDN-1-4, lower is desirable, but was chosen as a compromise to 6GDSH-5). Provided by element L2, speaker load, coupled with a simplified Zobel correction circuit.

Midrange: Bandpass filter. The lower part (C2) is matched with the low-pass filter and is tuned to a cutoff frequency of 500 Hz based on considerations of the resonant frequency of the 6GDSh-5. The upper part (L1) is matched with the high-pass filter and is tuned to 7500Hz, which allows for a broadband speaker structure, coupled with Zobel.
Both parts are loaded at 8 ohms (4 ohms from 6GDSH-5-4 + 4 ohms from R2).

HF: High pass filter. The frequency is matched to the upper part of the midrange filter and operates at 7500Hz, which avoids problems associated with the high frequency of the main resonance of domestic high-frequency speakers. Load 21 Ohm (16 Ohm speaker + 5 Ohm from R1).

All speakers are switched on in phase, which has a lesser effect on the phase characteristics of the system.

scheme

Electrical circuit diagram. Click to enlarge :).

The arrow on the right shows the “sound input” from the amplifier. The dotted lines are bi-wiring (the low-pass and mid-high-frequency filter sections are connected to each other in parallel at the amplifier - plus the low-pass with the plus of the mid-range and high-frequency to the plus of the amplifier, the minuses are the same).

The gray numbers in brackets above the filter elements indicate their load. The gray numbers with an "r" in front of them are the element's own resistance. Gray marks -1dB - loss of speaker sensitivity on the elements.

Next to the speakers, their important characteristics are briefly written out; below are the frequencies at which the bands/links intersect.

Inductance in mH, capacitance in µF, resistance in Ohm. After assembling the filter, the nominal speaker impedance for the amplifier remains equal to 4 Ohms.

A version of the “new lightweight” filter for s-90 clones, more precisely for Orbit 35AC-016. Speakers: 10gdv-2-16, 6gdsh-5-4, 75gdn-1-4 - a fairly common set.

They are distinguished by an expanded range of reproduced frequencies, the introduction of an indication of electrical overload of loudspeakers and a new appearance. The recommended power of a high-quality household amplifier is 20-90 W. The preferred installation option is floor-mounted.

A distinctive feature of the acoustic system is the use of a loudspeaker with magnetic fluid “MANGO” as a mid-frequency link, which made it possible to increase the rated power of the loudspeaker and the speaker as a whole.

Specifications:Сг3.843.050 TU.

Shape of the frequency response of sound pressure measured along the acoustic axis: a) , b) ,

Specifications:

Specifications	Meaning
Reproducible frequency range in free field conditions, Hz:
	25...25000
,	25...25000
Uneven frequency response of sound pressure, dB, at the lower limit frequency of the reproduced frequency range relative to the average sound pressure level:
	-15
,	-14
Uneven frequency response of sound pressure, dB, in the frequency range 100...8000 Hz relative to the average sound pressure level:
	±4
,	±4
Level of characteristic sensitivity (characteristic sensitivity), dB, not less:
	85 (0,338)
,	89 (0,56)
Directional characteristics of the speaker, dB, determined by the deviation of the frequency response of sound pressure, measured at angles of 25 ± 5° in the horizontal plane and vertical plane, from the frequency response measured along the acoustic axis of the speaker (0°):
:
in the vertical plane	±8
in the horizontal plane	±6
, :
in the vertical plane	+ 2
	-4
in the horizontal plane	+4
	-3
Harmonic distortion of speakers, %, determined by the total characteristic harmonic coefficient at an average sound pressure level of 90 dB at frequencies, Hz, no more than:
250…1000	2
1000...2000	1,5
2000...6300	1
Nominal electrical resistance (nominal value of total electrical resistance), Ohm:
	4/8
,	8
	8
Minimum value of total electrical resistance, Ohm:
	3,2/7,6
,	7,6
	7,0
Maximum noise (nameplate) power, W:
,	90
	100
Maximum short-term power, W:
,	600
Type of low-frequency acoustic design	bass reflex
Weight, kg	23
Dimensions, mm	360x710x285

Design Features:

The enclosures of all modifications of the speakers are made in the form of a rectangular non-dismountable box made of chipboard veneered with valuable wood veneer. The thickness of the case walls is 16 mm, the front panel is plywood 22 mm thick. At the joints of the housing walls, elements are installed on the inside that increase the strength and rigidity of the housing.

The following sets of heads are used in acoustic systems:

• : ; ; ;
• ; : ; ; ;
• : ; ; .

The heads are produced by Radiotekhnika software.

The heads that make up the speakers are each framed with decorative blackened overlays, made by stamping from aluminum sheet, with four mounting holes. The midrange head is isolated on the inside from the total volume of the housing by a special plastic casing in the shape of a truncated cone. The LF head is located on the front panel along a vertical axis, and the MF and HF heads are shifted relative to this axis to the left or right. On the front panel there are also knobs for the midrange and treble level controls, and in the lower part there is a plastic overlay panel with a nameplate and a rectangular hole (100x80 mm), which is the bass reflex output. The nameplate shows frequency response curves corresponding to various positions of the level controls, as well as the name of the speaker and the manufacturer's logo. In addition, the front panel has special plastic bushings for attaching a decorative frame with fabric.

Crossover frequencies provided by filters: between low- and mid-frequency heads - 750±50 Hz, between mid- and high-frequency heads - 5000±500 Hz.

The design of the filters and the overload indication unit uses resistors such as BC, MLT, SPZ-38B, S5-35V, PPB, capacitors such as MBGO-2, K50-12, K75-I and inductors on plastic cast frames.

The removable decorative frame included in the kit is covered with knitted fabric with high acoustic transparency.

The delivery set includes four plastic feet, which can be attached to the base of the case if necessary.

Modification of 35AC-212 (S-90) with original speakers and switches.

According to the prospectus of the 90s, the Riga Radio Plant mass-produced two models of acoustic systems: 35AC-212 or “S-90” and 35AC-012 in the modifications “S-90B”, “S-90D”, “S-100B”. The time has come to modify the older model 35AC-212, as well as its predecessor 35AC-1, which have a similar set of speakers.
These models have switches for the level of attenuation of the energy supplied to the midrange and tweeters, allowing you to adjust them to the level of the woofer and adjust the system to specific listening conditions. All this is good, of course, but it “thumps and clicks” no matter how you turn the switches. I want it to be musical. Somehow I was talking about alternative thoughts about finalizing the S-90. These thoughts happily dissipated without being realized. They were replaced by others, more interesting ones. It seemed the most promising to use the “Nivaga 9” filter from the previous article and convert it to another set of speakers, and leave the midrange and high-frequency switches in their original factory form. The resulting filter diagram for the S-90 is shown in the figure. I propose to call it “Nivaga 10”.

A distinctive feature of the filter is the presence of resistors R1, R2, R3, R4, which provide direct potential contact of all speakers with the PA output and do not allow the phase response to deviate far from a linear frequency dependence. If you look closely at the diagram, you will notice that the resistance of these resistors is close to the active resistance of the corresponding speakers. Meticulous comrades can of course add the inductive equivalent of these speakers. I was lazy, because even in this form the sound quality completely satisfied me, but I didn’t have the opportunity to experiment in a sound chamber. Well, if you look even more closely at the circuit of the bandpass filter leading to the midrange speaker, you can see that it was created from a previously developed filter like “Nivaga 6 or 8” by replacing the speakers with equivalent resistors. Similarly in the lowpass and highpass filters resistors R1 and R3 are equivalent to the corresponding speakers. Therefore, this circuit with parallel connection of speakers is a logical development of the previous one with series connection of speakers, which means it retains all its advantages, which were written about in earlier articles. And at the same time, it creates new opportunities to move the cutoff frequency of all four filters included in the circuit independently of each other, controlling the peaks and dips in the frequency response of the speaker, which was not the case in the previous circuit.In the specific case of this circuit, I sought to expand the cutoff frequencies of the low and midrange speakers, as well as the cutoff frequencies of the midrange and high frequencies speakers by half an octave. The results are brilliant. Elastic bass, stereo panorama, volume, clear mids - everything that the ear of a music lover craves is present in the modified S-90 speakers.
The fear that the introduced resistors would heat up was not justified. Their power is theoretically based. In practice, it can be reduced by 2 - 3 times, but the resistors must be wirewound.
Practice shows that not everything that I like suits others. Well, the proposed scheme is open to reasonable modifications, and I am ready for serious discussions.
This treatise was compiled on February 20, 2012.

Video

Description

Speaker system Radiotehnika S-90F

Radiotehnika S-90F - 3-way speaker system with bass reflex.
Production since 1980 at the Riga plant "Radiotehnika"

Designed for high-quality sound playback in standard living conditions.
"S-90F" is a speaker that meets the requirements of international documents for Hi-Fi category equipment. The difference between this modification of the speaker and the S-90 is: an expanded frequency range, the introduction of an indication of electrical overload of the loudspeakers and a new appearance. The recommended power of the amplifier used is 20-90 W.

The enclosures of all modifications of the speakers are made in the form of a rectangular non-dismountable box made of chipboard veneered with valuable wood veneer. The wall thickness of the case is 16 mm, the front panel is plywood with a thickness of 22 mm. At the joints of the housing walls, elements are installed on the inside that increase the strength and rigidity of the housing.

The speaker uses the following set of heads produced by "Radiotehnika" software:

The heads included in the "S-90F" are each framed with decorative overlays and protected by a mesh; The arrangement of the heads is similar to that in the "S-90". On the front panel, in addition, there are: indicators of overload of the loudspeaker heads, regulators of the sound pressure level of the MF and HF loudspeaker heads, and a nameplate with the name of the speaker. At the bottom of the front panel there is a rectangular outlet 95x75 mm. bass reflex, the tuning frequency of which is 31 Hz.
The internal volume of all modifications of the speaker is 45 dm3. To reduce the influence of the frequency response of sound pressure and the sound quality of the AC resonances of the internal volume of the housing, it is filled with a sound absorber, which is mats made of technical wool, covered with gauze. The mats are located and fixed on the inner surfaces of the housing walls.
Electrical filters are mounted inside the housing walls on one board, providing electrical separation of the bass, midrange, and high-frequency speaker bands. Electrical filters of all modifications of speakers have the same design and electrical circuit diagrams.
Crossover frequencies provided by filters: between LF and MF - 750±50 Hz, between MF and HF - 5000±500 Hz.
Resistors of the BC, MLT, SP3-38B, S5-35V, PPB types are used in the design of the filters and the overload indication unit; capacitors such as MBG0-2, K50-12, K-75-11 and inductors on plastic cast frames.
The removable decorative frame included in the “S-90” kit, “S-90D” is covered with knitted fabric with high acoustic transparency.
The delivery set includes 4 plastic feet, which can be attached to the base of the case if necessary.

Technical data:

Frequency range - 25 (-14 dB) – 25000 Hz;
- Sensitivity - 89 dB (0.56 Pa/? W);
- Frequency response unevenness in the frequency range 100 – 8000 Hz: ±4 dB;
- Directional characteristics of the speakers, measured by the deviation of the frequency response of sound pressure at angles to the acoustic axis:
vertical +2°: -4 dB
horizontal +4°: -3 dB
- Harmonic distortion at 90 dB sound pressure in the frequency range:
250 – 1000 Hz: 2%
1000 – 2000 Hz: 1.5%
2000 – 6300 Hz: 1%
- Resistance - 8 Ohms;
- Minimum impedance value -7 Ohm;
- Rated power - 90 W;