Loudspeaker with increased efficiency. High quality three-way loudspeaker Main technical data

Improving the sound quality of modern loudspeakers is achieved mainly through the use of new powerful dynamic drivers, and this most often entails an increase in their dimensions, weight, and cost. Meanwhile, a very good loudspeaker can be built on the basis of inexpensive dynamic heads.

Main technical characteristics.

Rated (nameplate) power, W....................................10 (30)

Nominal range of reproduced frequencies, Hz............30...25,000

Number of lanes................................................... ........................................3

Section frequencies, Hz................................................... ....................500; 5000

Nominal electrical resistance, Ohm............................6.3

Average standard sound pressure, Pa...................................0.35

Dimensions, mm................................................... ................................620x350x310

The electrical circuit of the loudspeaker is shown in Fig. 1. It is built on the basis of three dynamic heads. The low-frequency (LF) functions are performed by the 6GD-2 head, the mid-frequency (MF) head - 3GD-38E, and the high-frequency (HF) head - 6GD-13 (new name 6GDV-4). The second-order filter L1C1 is used in the low-frequency section, the first-order filter L2C2 is used in the midrange, and the third-order filter L3C3C4 is used in the high-frequency section. To equalize the frequency response of the loudspeaker in the region of medium sound frequencies, the midrange head is connected through resistor R1. In order to improve the sound of the system at frequencies above 503 Hz, the 6GDV-4 HF head is connected to a filter using resistors R2 and R3. It is important to note that this head is turned on in antiphase with the bass and midrange heads.

Fig.1. Electrical circuit of a three-way loudspeaker filter

The acoustic design of the loudspeaker is a bass reflex. Its body is made of chipboard 20 mm thick. The front panel and side walls are connected to each other with 20 x 20 mm slats using EDP epoxy glue. The back wall is removable; it is attached to the body through 2 mm thick rubber gaskets.

The view from the front panel is shown in Fig. 2, a, and a section of the body along line A-A- in Fig. 2, b. The bass and midrange speakers are attached to the outside of the front panel. Between it and the head diffusers there are rubber (polyurethane foam) rings 1.5 mm thick are laid.

Fig.2. Three-way loudspeaker drawing

Before placement on the front panel, the 6GD-2 head must be modified in order to reduce its overall quality factor. To do this, acoustic resistance panels (ARPs) should be installed in the windows of its diffuser holder, i.e., sealed with synthetic felt or, in extreme cases, medical gauze folded in several layers. The mid-frequency head must be placed in a sealed box with a volume of about 2 liters, filled with cotton wool. The diameter of the box is equal to the diameter of the hole in the front panel for the midrange head. The place where it connects to the panel must be carefully sealed (for example, with plasticine). The 6GDV-4 RF head is mounted on the inside of the front panel, and the side surfaces of the hole for its installation should, as it were, continue the cone existing on the head and form a radiating horn with it. A sealing rubber ring should be placed between the body of this head and the panel. The bass reflex tunnel is a plastic tube with an outer diameter of 70 and an inner diameter of 65 and a length of 150 mm. It is inserted into the corresponding hole on the front panel from the outside. The gaps between the panel and the tunnel are sealed from the inside with plasticine.

The crossover filter parts are placed on a getinax board measuring 250 x 150 mm, installed on the side wall of the housing in its lower corner, opposite the bass reflex tunnel. To avoid rattling, a sound-absorbing gasket must be laid between the board and the case. The filter uses non-polar MBM capacitors. MBGO for a voltage of 200 V and wirewound resistors with a power of 2 (R3) and at least 7.5 W (others). Capacitor C1 is made up of six 10 micron capacitors connected in parallel. Coils L1-L3 are frameless. The internal diameter and height of the first of them is 40 mm, the other two are 25 and 30 mm, respectively. Coil L1 contains 260 turns of PEL 1.5 wire, L2-170 and L3-90 turns of PEV 1.0 wire. The inner surface of the case is covered with sound-absorbing material (batting, foam rubber) with a thickness of 10...15 mm. The body itself is filled with cotton wool, but in such a way that an air passage is left between the woofer head and the bass reflex. All connections of the housing walls are sealed with epoxy glue.

The sound of the described loudspeaker was compared with the sound of the well-known industrial model 35AC-012 (S-90). During the tests, a stereo AF amplifier with a rated power of 2 x 25 W and a harmonic coefficient of no more than 0.2% was used. The softer sound of the homemade loudspeaker was noted in the region of low and medium sound frequencies, as well as the absence of unpleasant overtones created by the 10GD-35 head installed in the 35AS-012 in the range of 5...10 kHz.

P.S. Replacing the 6GD-2 head. Instead of 6GD-2, you can use a dynamic head 75GDN-1L-4 (formerly designated 30GD-2) or 35GDN-4 (25GD-26B). These heads have more than half the standard sound pressure (0.15 and 0.12 Pa, respectively) compared to the 6GD-2 (0.35 Pa), but their significantly higher rated power compensates for this disadvantage. The rated power of the loudspeaker after such a replacement will increase in the first case to 50, in the second - to 40 W, the nominal electrical resistance will drop to 4 Ohms. The capacity of capacitor C1 when using the 75GDN-1L-4 head is 80 µF. PAS is not required in both cases. The first replacement option is preferable, since the 75GDN-1 L-4 head has the same dimensions as the 6GDN-2, and greater efficiency than the 35GDN-4, especially at frequencies below 100 Hz.

Yu. DLI, Gorky

Radio magazine, No. 3.9 1989

Dynamic heads.

The project under consideration was based on the use of a dome midrange VIFA D75MX-41-08, the main properties of which were determined by technical Project compromises in terms of choosing the remaining dynamic heads are approx.

The essence of the compromise is as follows. On the one hand, the main The advantages of the D75 dynamic head are a high acceleration factor (1420) and a low voice coil inductance (0.13 mH at a frequency of 10 kHz). On the other hand, the linear section of the voice coil stroke is 0.5 mm and the resonant frequency of 300 Hz excludes the possibility of using this head with a crossover frequency below 600 Hz. In this regard, part of the mid-frequency range must be reproduced by the bass head. For detailed reproduction in the frequency band up to 600 Hz, you will need a woofer with an acceleration factor of at least 300. This value of the acceleration factor of the woofer conflicts with the ability to provide a low cutoff frequency and a high sound pressure level at low frequencies. Options for a compromise resolution of this contradiction will be determined by the properties of the LF head.

The woofer head must satisfy one more requirement: it is desirable that its diffuser does not have noticeably pronounced resonance phenomena at operating frequencies, i.e. up to 600 Hz. It is difficult to determine compliance with the latest requirement based on studying the manufacturers' reference materials; you will have to purchase heads and take measurements. Table 1 shows the parameters of four LF heads with a diameter of 200 mm with an acceleration factor exceeding 300. Using reference data, the cutoff frequencies F3 are calculated for a volume Vb = 40 liters. For SEAS H1288 it is assumed to use a closed volume, for the remaining heads - a bass reflex.

Table 1.

Manufacturer, model	BL/m	SENS	Xmax
SEAS H1288		89.5				0.41
PEERLESS 830884		89.3		32.4	68.8	0.38
BEYMA 8woofer/P						0.38
AUDAX HM210Z12		90.7			86.3	0.33

Of the four head models listed in the table, we were able to purchase three: H1288, 8woofer/P and HM210Z12. Figure 1 shows the Z-x characteristics of dynamic heads measured by LMS in current generator mode. The SEAS H1288 cone resonates at 680 Hz (blue curve). The BEYMA 8woofer/p diffuser resonates at a frequency of 500Hz (black curve). Z-characteristic of AUDAX HM210Z12 (yellow curve) does not show obvious resonance phenomena. Of the three available models of dynamic heads, the AUDAX HM210Z12 satisfies the requirements of the dynamics project to the greatest extent. The purchased BEYMA 8woofer/P speakers turned out to be unsuitable for further use in the project - their resonant frequencies and Qts values ​​differed too much from those stated in the reference data.

For further work on the project, SEAS H1288 and AUDAX HM210Z12 remained. The H1288 loudspeaker was examined using a mock-up 40 liter cabinet, as this head is available for purchase by amateurs, in addition, it has some advantages over the HM210Z12 in terms of low-frequency reproduction. Listening to the loudspeaker prototype showed that the H1288, when used together with the D75, gives a satisfactory result, but discerning listeners on vocal parts noticed some coloration in the sound associated with the resonance of the diffuser at a frequency of 600 Hz. The copies of the H1288 heads used in the project had a total quality factor of 0.78 in a closed 40-liter case. For better reproduction of low frequencies, it was necessary to increase the volume of the case to 50 liters.

Figure 2 shows the speaker crossover circuit on the H1288.

    

Figure 3 shows the frequency response of the loudspeaker measured by a microphone located along the axis of the high-frequency head at a distance of 1 m.

The final version of the loudspeaker uses HM210Z12, which has more acceptable characteristics for reproducing vocals, since its diffuser does not have pronounced resonance phenomena.

The choice of a high-frequency driver to work with the D 75 is not determined by specific requirements, and the MOREL MDT 33 seems to be a completely acceptable option for a loudspeaker of this class.

Housing design.

A drawing of the speaker enclosure using the HM210Z12 is shown in Figure 4 4.

    

Preliminary calculations showed that the acoustic design of the HM210Z12 requires a volume of 40 liters with a bass reflex tuned to a frequency of 44 Hz. A pipe with an internal diameter of 75 mm and a length of 30 mm provided the specified tuning frequency. The hole for the pipe is located on the rear wall at the top of the case.

In a housing 1 m high, there is a need to suppress a standing wave between the upper and lower walls at a frequency of approximately 150 Hz. For this purpose, the hole in the jumper located below the woofer head is covered with synthetic padding, the volume of the housing under the jumper is filled with batting. The inner surface of the body above the lintel is covered with thin batting. The measures taken turned out to be sufficient to effectively suppress the standing wave, while having little effect on the efficiency of the bass reflex.

As an acoustic design for the mid-frequency head, a hemispherical chamber VISATON AK 10.13 is used, covered on the outside with guerlain and filled with synthetic padding. The camera and midrange head are installed on opposite sides of the front panel. This solution reduces the transmission of head vibration to the camera, which is essential for high-quality reproduction of mid frequencies, but leads to the need to make the back wall removable. The back wall is attached with ten self-tapping screws to frames glued into the body. Sealing of the rear wall is ensured by a polyethylene foam seal. The complexity of the housing design associated with a removable rear wall can be avoided by fixing and sealing the chamber with wires on the front panel before assembling the housing. For a loudspeaker with an H1288 low-frequency driver, you can use a housing of a similar design, increasing its depth to 300 mm.

TO rossover.

The crossover circuit is shown in Fig. 5

The crossover frequencies of 600 and 3500 Hz are selected in the loudspeaker. In the region of joint radiation of the low-frequency and mid-frequency heads, the second-order Butterworth sound pressure drops in the frequency response are summed up, requiring antiphase switching on of the dynamic heads. The correction chain R1L1 serves to compensate for the rise in frequency response associated with the transition of the low-frequency head's radiation mode from space to half-space. Resistors connected in parallel with the low-frequency head reduce unwanted interaction of the low-frequency head with the filter. (This issue is discussed in detail in the work “Amateur Loudspeakers 3”). Capacitance C2 protects the midrange head from overload with low frequencies and forms a specified decline in the frequency response of the head in the lower region of joint radiation.

In the region of joint radiation of the midrange and high-frequency heads, fourth-order Linkwitz-Riley sound pressure frequency response decays are used, obtained using second-order electrical filters. The transfer characteristics of crossover filters are shown in Fig. 6. The crossover uses MUNDORF, VISATON and SOLEN elements.

Figure 7 shows the frequency response of dynamic heads working with filters. Figure 8 shows the frequency response of the loudspeaker, measured along the axis of the HF head at a distance of 1 m. Figure 9 shows the frequency dependence of the loudspeaker impedance.

Conclusion.

Experience with this project shows the possibility of obtaining sufficiently high-quality reproduction of phonograms of vocal parts through the use of a VIFAD75 dome midrange driver. Considering that it is difficult to reproduce a loudspeaker using HM210Z12 due to the lack of these heads on sale, with some reduction in the requirements for reproduction of mid frequencies, you can use the H1288 .

A three-way dividing line with a crossover frequency of 520–4800 Hz was used (Fig. 1). The presence of attenuators allows you to adjust the frequency response of the loudspeaker in the mid-high frequency region by ±4 dB relative to the average (zero) level. The attenuator resistors are made of Provo-PEMS 0.41 - 0.56. They can be made from iron-tiles.

Separating coils. filters are wound on frames made of wood (birch, ) with. external 0 36 mm, length 24 mm (Fig. 2), and contain: LI, L2 - 260 turns each, L3 - 85 turns, L4 - 170 turns with a tap from the middle of the PEL 1.0 wire.

The loudspeaker body and front panel are made of 16 mm thick chipboard (Fig. 3). The front one (Fig. 4) is deepened by 20 mm. The back cover of the speaker is secured with overlapping screws. Between the back cover and the case for sealing, feathery rubber 5 mm thick is laid. The boxes are fastened with birch bars, pre-coated with EDP-3 or EDP-5 glue. glue seals the loudspeaker.

The dynamic heads are installed on the front side of the front panel. For this purpose, recesses are made in the frame of the dynamic heads. Between the front panel and the bars, and to which it is attached, porous rubber is laid for sealing. Then inside the box, seals are created from cotton wool at an angle so that it becomes spherical. The mid-frequency one is covered with a cap made according to the same technology: a cylindrical blank 0 140 mm, 120 mm high, is machined from foam plastic. Then with one it is given a spherical shape (Fig. 5). A thin (1 - 2 mm) amount of plasticine is carefully applied to the surface of the finished sphere. Then, using the papier-Mrshe method, pieces of fiberglass impregnated with EDP-3, EDP-S glue, 2 - 3 mm thick, are glued onto it. After the glue has dried, the sphere is removed from the foam plastic blank - the cap for the frequency head is ready. The windows of its frame are sealed with mar- , the volume between the head and the cap is evenly filled with cotton wool.

MAIN TECHNICAL DATA:

effectively reproduced frequencies (Hz) with unevenness of 14 dB - 20 - 25000,

with unevenness 8 dB - 20 - 22,000;

dimensions, mm - 460X350X260.

Rice. 1. Schematic diagram of a separating filter.

An air passage is formed between the low-frequency head and the phase inverter using a metal mesh. The remaining volume of the box is evenly filled with cotton wool weighing 0.9 - 1.5 kg. The phase inverter consists of a glass and a pipe insert (Fig. C, made of -16T duralumin. They can also be made using the method from fiberglass and ZDP-3 glue.

Rice. 6. Bass reflex: 1 - glass, 2 - insert.

At the exhibition RosHI-End 2013, together with an amplifier by L. Zuev and a DAC by V. Korsakov, a three-way loudspeaker on speakers with metal diffusers was demonstrated. The reproduction of musical material selected by V. Lukhanin by this system has received many reviews, which can be found on the Vegalab website.

The development was carried out with the goal of constructing a compact floor-standing loudspeaker intended for sounding residential premises with an area of ​​up to 15-20 square meters. meters, focused on playing music programs with a dense spectrum and high-quality vocal reproduction against the backdrop of a dense signal spectrum. Below we will consider a version of this loudspeaker, modified based on comments from visitors and exhibitors, as well as taking into account the possibility of repeating the design at home. The increase in the project budget associated with the modification seems to us justified by the increase in the quality of sound reproduction. Below we will talk in more detail about compromises, including between price and quality.

In residential premises with an area of ​​15 -20 sq. m. It is not always possible to optimally place speakers, which leads to problems in reproducing low frequencies and deterioration in the localization of apparent sound sources. This circumstance was taken into account and was reflected in the choice of the main technical solutions of the project.

A drawing of the loudspeaker enclosure is shown in pic 1.

The front panel has a trapezoidal shape, the variable width of the front panel slightly reduces diffraction effects. The low-frequency closed-type acoustic design has a usable volume of 30 liters, which is powered by the RS225 speaker. Inside the low-frequency compartment there is a piece of sound absorber (sintepon) measuring 0.5 by 0.5 m. The choice of a closed acoustic design is due to the desire to obtain the most compact impulse response of the low-frequency section.

In residential premises, as a rule, there are standing waves between the walls, between the floor and the ceiling. In such a situation, it is advisable to favor a compact impulse response over extending the frequency range downward using a bass reflex.

The midrange speakers operate on a closed volume of 6 liters, tightly filled with sound absorber. The use of two W4-1337SD speakers for the midrange leads to an increase in costs, which is justified by the improvement in overload capacity at mid frequencies and allows the construction of an MTM configuration that provides a narrowing of the radiation pattern in the vertical plane. Narrowing the radiation pattern in the midrange seems to be an additional bonus, since it increases the level of the direct signal at the listening point. A simulation of the radiation pattern in the vertical plane is shown in rice. 2. The W4-1337 speaker has a moving mass of 4.6 grams with a cone area of ​​57 square meters. cm, linear portion of the voice coil stroke is 3 mm. The voice coil inductance value of 0.015 mH indicated in the manufacturer's data sheet is questionable.

According to my estimates, W4-1337 has Levc = 0.4 mH, which is quite acceptable for mid frequencies. The low moving mass and rigid diffuser ensure good transmission of dynamic contrasts. This speaker is manufactured in two versions: W4-1337SD has a neodymium magnet, W4-1337SDF has a ferrite magnet. Both versions are suitable for the loudspeaker. Prior to the publication of this work, it was possible to examine 18 specimens of W4-1337SDF and 24 specimens of W4-1337SD. Based on the results of parameter measurements, it became clear that it was possible not to select speakers in pairs for the MTM configuration.

The increase in the budget associated with replacing the Seas H1499 tweeter with a Mundorf AMT 19CM 2.1 is justified by an increase in the quality of high-frequency reproduction. In addition, as a result of the replacement, it was possible to exclude 4 elements from the filter circuit, including those requiring adjustment, since AMT 19CM are supplied in pairs, with a small spread of characteristics.

The choice of speakers for the loudspeaker assumed the use of crossover frequencies of 500 and 3500 Hz. The specified crossover frequencies with a margin ensure that the speakers operate in piston mode.

At a crossover frequency of 500Hz, the two-polar impulse response, which is inevitably obtained when the speakers are turned on in antiphase, does not spoil the sensation of sound perception. I assume that the waveform distortion lasts less than 2 ms. lie beyond the temporal resolution of hearing at frequencies above 500 Hz. A simulation of the impulse response of LF and MF speakers working with filters is shown in rice. 3. The result of the impulse response simulation raises some doubts; this issue will have to be sorted out. For now, you can focus on the listening results, which indicate fast and dynamic sound delivery in the low-frequency range.

The crossover frequency of 3500 Hz is an acceptable compromise due to non-linear distortion of the midrange and tweeters.

The result of the loudspeaker frequency response simulation is shown in rice. 4. The frequency response has been optimized for a listening distance of 2.5 m. The slight increase at the upper edge of the frequency range takes into account the decrease in acoustic power with increasing frequency, which occurs due to a narrowing of the radiation pattern. On rice. 5 shows the phase response of speakers working with filters.

The crossover circuit is shown in rice. 6. At a cutoff frequency of 500 Hz, the filters formed 2nd order frequency response slopes with a quality factor of about 0.5. The LF and MF speakers are switched on with reverse polarity. A wide co-emission region (Fig. 4) and a compact impulse response (Fig. 3) provide a cohesive and dynamic sound delivery. At the crossover frequency of 3500, 4th-order frequency response slopes according to Linkwitz-Reilly are formed. For the AMT 19CM 2.1 high-frequency speaker, the formation of a given frequency response decline was provided by a 2nd order electric filter; for the midrange speakers, a 3rd order electric filter was required.

The tweeter filter places the most stringent demands on the quality of the elements. The option of parallel connection of film and foil capacitors turned out to be a good compromise between price and quality.

The notch filter R5 L4 C5, which, according to a widespread myth, should kill sound, performs the function of protecting the midrange speakers from overload and corrects the phase response at a frequency near 100 Hz. The value of resistor R5 depends on the ohmic resistance of coil L4. The sum of the ohmic resistance of coil L4 should be 4 ohms ± 10%. When repeating a loudspeaker, it is not at all necessary to use the types of components that are indicated in the tables. Crossover filters have a low quality factor and allow deviations of values ​​from those indicated in the diagram of at least 5%, and 10% in the ohmic resistance of the coils. The crossover uses 10 W MOX resistors.

Inductors

L1	Mundorf Aire Core M Coil	0.47 mHn 0.58 Ohm
L2	Mundorf Aire Core M Coil	0.82 mHn 0.44 Ohm
L3	Mundorf Aire Core M Coil	0.22 mHn 0.21 Ohm
L4	ERSE Air Coil ALg 20ga	3.3 mHn 1.37 Ohm
L5	Mundorf Ferrite M Coil BH Drum coil	5.6 mHn 0.62 Ohm

Capacitors

C1-2	Dayton Audio PPF	0.47 mkF 400V
C1	MKP Mundorf M Cap	3.3 mkF 250V
C2	MKP Mundorf M Cap	22 mkF 400V
C3	MKP Mundorf M Cap	10 mkF 400V
C4	MKP Mundorf M Cap	8.2 mkf 250V
C5	Erse Non-Polarized	470 mkF 100 V
C6	MKP Mundorf M Cap	47 mkF 400V

On rice. 8 shows the frequency dependence of the input impedance of the loudspeaker. The minimum input resistance is 6 ohms, the maximum is 13.5 ohms. The phase angle, which characterizes the reactive component of the input resistance, does not go beyond plus - minus 30 degrees in the frequency band 20 - 20000 Hz. The parameters of the input impedance of the loudspeaker allow us to consider it a quite comfortable load for the amplifier.

The transfer characteristics of the filters are shown in rice. 7. Resistor R6 with a value of 22 Ohms was sufficient to eliminate unwanted interaction between the filter and the speaker. This can be judged by the transfer characteristic of the low-pass filter. “Pumping” does not exceed 1.5 dB with a maximum at 70 Hz.

On rice. 9 shows the loudspeaker's frequency response, measured in a room at a distance of 1 m at an input voltage of 2.83 V. The measured frequency response is not smoothed, but is the result of averaging three measurements: along the axis of the tweeter and when the microphone is shifted 5 cm down and up from the axis. This measurement technique allows you to get a clearer idea of ​​the tonal balance of a loudspeaker in a room than a smoothed frequency response along the axis of the tweeter.

In conclusion, I consider it necessary to express gratitude to V. Lukhanin, who resolved all organizational issues and carried out the bulk of the work on modernizing the loudspeaker, to the Difton company, which quickly and efficiently manufactured the enclosures, as well as to all sound lovers for their comments and suggestions on the project.

It is known that the degree of fidelity of sound reproduction equally depends on the quality of the bass amplifier and loudspeaker. A high-quality three-way loudspeaker is offered to radio amateurs. Oya is designed to work with a low-frequency amplifier with a channel power of 10...25 W and contains dynamic direct radiation heads - low-frequency 10GD-30, mid-frequency 4GD-8E, high-frequency ZGD-31 and a separating filter. The acoustic design of the low-frequency head is made on the principle of a bass reflex, which made it possible to expand the frequency band of the loudspeaker towards lower frequencies and reduce distortion at these frequencies.

Main technical characteristics

Power, W:

• nominal……………12.
• maximum………… 25
• Nominal total electrical resistance, Ohm…………….. 8
• Nominal operating frequency range, Hz, with uneven frequency response in sound pressure no more than 12 dB......35...18,000
• Average standard sound pressure, Pa…..0.15

Filter crossover frequencies, Hz:

• first …………… 500
• second…………….. 5000
• The slope of the filter characteristic at the crossover frequencies, dB/octave……….. 12
• Loudspeaker dimensions, mm……. 440X280X263

The schematic diagram of the loudspeaker is shown in Fig. 1. The filter coils are wound on frames made of insulating material. The frames of the coils L1, L2 are made of 36 mm long sections of polyethylene pipe with a diameter of 66 mm, to which cheeks made of 4 mm thick plywood are attached with three MZ screws. Coils L3, L4 are wound on cardboard sleeves from elements 373. Coils L1 and L2 each contain 230 turns of PEV-1 1.12 wire, wound between the cheeks. The inductance of the coils is 3.1 mH. Coils L3 and L4 are wound in several layers with PEV-1 0.86 wire. Number of turns - 145, winding length 42 mm, inductance - 0.4 mH. The design of the coil frames is shown in Fig. 2.
The filter uses MBGP capacitors with a rated voltage of 160 V and PEV-5 resistors.

Rice. 1. Loudspeaker circuit

The box is made of dense plywood 10 mm thick. The dimensions of the side walls are 440×263 mm, and the bottom and top walls are 280×263 mm. Cutting blanks of plywood parts should be done with a saw with fine teeth to avoid chips and cracks at the ends. It is convenient to use a hacksaw for this purpose.
Having cut out the blanks, they cover their outer sides with decorative film or veneer of valuable wood species. The decorative film is glued with 88H glue. Wooden blocks with a cross section of 25X20 mm are glued to the inner sides of the workpieces with epoxy glue, the location of which is shown in Fig. 3. The front panel is glued together with epoxy glue from two pieces of plywood, each 10 mm thick, after cutting out the holes for the heads and the bass reflex tunnel with a jigsaw. The shape and dimensions of the blanks and the assembled panel itself are shown in Fig. 4.
The parts of the box are glued together with epoxy glue, tied together with ropes, a weight is placed on the top panel and left for 1.5...2 days for the glue to completely cure. After that, the ropes are removed, the box is inspected and, if there are gaps in the joints, they are filled with epoxy glue.
The bass reflex tunnel with an internal diameter of 40 mm is glued together from thick hard cardboard or several layers of Whatman paper with PVA glue. Wall thickness 3 mm. The tunnel is glued to the front panel with epoxy glue after adjusting the bass reflex, and is secured with plasticine during adjustment.

Rice. 2. Design of coil frames

Rice. 3. Loudspeaker box design

The 10GD-30 head is installed on the front panel of the box from the inside, and the 4GD-8E and ZGD-31 heads are installed on the outside. The 4GD-8E head is covered with a cap made of plywood or duralumin. The internal volume of the cap is filled with cotton wool (but so that it does not touch the oscillating membrane of the head). This is necessary so that air vibrations created by the woofer head do not affect the operation of the midrange head.
The separation filter parts are mounted on a board, which is then attached to the bottom of the box. The back wall is attached to the box with screws. The wire for lining and drilling the heads is threaded into the hole in the back wall and filled with glue. To ensure the tightness of the rear wall installation, use sealing mastic or a sponge rubber gasket. The inner surface of the box is covered with foam rubber 30...40 mm thick.
The bass reflex is adjusted to the resonant frequency of the 10GD-30 head in open air. The resonant frequency is measured by impedance (curve 1 in Fig. 5). Then, having installed the head in the box, they remove the dependence of the impedance on the frequency and, by changing the length of the tunnel, ensure that at the resonant frequency of the head there is a minimum impedance (curve 2). If the minimum of curve 2 is located to the left of fpe3, then it is necessary to reduce the tunnel length, and vice versa. To do this, make a tunnel of obviously greater length and, by shortening it, adjust the bass reflex. In the described loudspeaker, the length of the tunnel is 190 mm. It should be noted that if the loudspeaker is manufactured in exact accordance with the description, adjusting the bass reflex will most likely not be required. It will be necessary when the internal diameter of the tunnel changes by more than 7...10% and the volume of the box by 10...20%.

It is best to make a decorative frame as stated in O. Saltykov’s article “Small-sized loudspeaker” (see “Radio”, 1977, No. 11, pp. 56, 57).
When listening to a variety of music programs, a noticeable advantage of this loudspeaker was noted compared to factory ones with a power of up to 20 W (10MAS-1, 20AC-1), especially at lower frequencies.

Literature

To help the radio amateur: Collection. Vol. 79/B80
F. Budankov