Construction of ak systems for 4gd 35. Self-made acoustic systems for a tube amplifier

Let me remind you of his performance characteristics:
Broadband.
Intended for installation in the speaker. It was widespread, now (it turns out!) A rarity appreciated by some amateurs.
Effective operating frequency range 63 - 12500 Hz;
The unevenness of the frequency response is not more than 16 dB (this is a lot, if that);
The level of characteristic sensitivity, not less than 92 dB (from 90 dB / W / m is already suitable for low-power tube amplifiers);
Rated electrical resistance 4 ohms;
The frequency of the main resonance is Fs 65 Hz;
Passport power 8 W;
Full Q factor 1.4 + -0.3;
Overall dimensions 200x200x75.6 mm (8 "speaker);
Weight 0.88 kg.

Now a few words about the sound. 10 point grid 3 points. An ordinary paper sound. Nothing special. In any of the most magical design (OYa, ZYa, labyrinth). 4A28 beats him quite easily. What with a high-quality transistor, what with a good (for example) tube amplifier.


So do not chase such a rarity. And then listen to the recordings and convince yourself of "great sound." One can, of course, deceive oneself, but the nature of the sound cannot be obtained.

Recently, an article about a homemade tube amplifier was published in the Homemade World, today it’s turn to tell how to make acoustics for a tube amplifier with our own hands.

Want great acoustics? Not prestigious, from well-known companies, namely cool, giving out detailed, deep, atmospheric and energetic enough sound? If you do not live in a castle and do not plan to sound the stadium or try to impress your neighbors three floors below, but want to enjoy music, perhaps my acoustic recipe is for you.

There are things that I really want to do, but the lack of funds, materials, equipment, etc. But I really want ... I gave myself a pendel, and I got it ... By the principle: "I want it because I like it!" and because we must finally make acoustics for mine. I wanted to build a speaker that sounds good in both large and small rooms, and at the same time elegant, and not cabinet-like. Still - an element of the interior. In any case, it should be a floor-standing speaker, since "shelf speakers" still need to be placed on stands.

Creation

For the tube one-cycle dynamics, I also chose "tube" speakers: 4GD-35 (4GD-36) and 3GDV-1 (2GD-36), widely known from Soviet radio.

3GDV-1 (passport data):

• Frequency range: 3150-20000 Hz;
• Frequency response: 14 dB;
• Sensitivity: 90dB;
• Working power: 2W;
• Harmonic coefficient in the frequency range 4000 - 10000 Hz: 3%;
• Resistance: 8 ohm;
• Passport power: 3 W;
• Long-term power: 3W;
• Short-term power: 6 W;
• Main resonance frequency: 1200 - 2000 Hz;
• Weight: 0.11 kg.

4GD-35 (4GD-36) (passport data):

• Frequency range: 63 - 12500 Hz;
• Frequency response: 16 dB;
• Sensitivity: 92 dB;
• Working power: 0.8W;
• Harmonic distortion at operating power: 125 Hz: 7%, 200 - 300 Hz: 5%, 1000 - 8000 Hz: 3%;
• Resistance: 4 ohm;
• Passport and long-term power: 8 W;
• Short-term power: 15 W;
• Full Q factor: 1.4 (± 0.3);
• Main resonance frequency: 65 Hz (+20, -15);
• Dimensions: Ø 200 x 75.6 mm .;
• Weight: 0.88 kg.

According to my comparative measurements, it turned out that 4GD-36 and 4GD-35 have the same frequency response, but 4GD-36 sounded more comfortable for hearing.

Four design options were considered:

1. Open box. Semi-open drawer.
2. Closed box. A box with a cat-hole.
3. TQWP - (TaperedQuarter-WavePipe - Expanding Quarter Wave Pipe).
4. Quarter-wave resonator (TwinPipe) in the image of Castle Knight 5, from the British Castle Acoustics.

All variants are noteworthy and have shown certain advantages and disadvantages in sound, although officially these speakers are intended for open design.

After sawing and breaking a bunch of boards, I share the results.

Open box. Classic light sound. Many people may like it, but reflections from the walls do not allow for a clear scene. If the body is closed at the back with a piece of felt (about 1 cm thick), then the situation is somewhat improved. This design (called: aperiodic loading) showed the best linearity, but at the same time more modest dynamic performance and sensitivity. The construction of the stage was also not impressive. Everything is very academically correct, but dry. The room has 10 sq. m you can still listen, but at twenty meters it became boring - a weak attack.

Box with sound port in front. Simple and quite effective. Niza stick out, the rest is in the norms of decency. Sensitivity and dynamics are average. For a room of 20 meters may be suitable, but at 10 meters "rowdy".

Since I wanted something with a twist, I settled on the option "from the knight's castle." In fact, the Twin Pipe has demonstrated the greatest sensitivity and better bass control than the popular TQWP. The sound is light, aggressive and open. The bass is adequate in both small and large rooms. The scene and the detail are pleasing. The linearity, however, of this "mustang" is not at all high-quality: 18db (relative measurement in the resonating room), with a certain resonator setting, you can lower it to 14db. The only significant drawback is that the sound is like a "box". The sound port on the front panel is to blame. You can deal with this by densely damping the case with various materials, but the sensitivity drops and then the meaning of choosing such a design is lost.

In general, all sound ports and bass reflexes serve to output any uterine sounds outside, so if you do not want to listen to how your insides sound, it is better not to use them.

On the album "Bad Dream" Ronnie Dio has such a song - "All the fools swam far away." This is apparently about me, and now I am proud of it.

So what am I inventing? A bike with square wheels ... based on a speaker for an open design ... The problem is that acoustics need a back wall for a good attack. Take 10GDSH-1-4 and such questions would not arise at all. Put it in a box, or in a big box with a fazik, and it will “sing”. But I have already “swam far away”. The case for 4GD-36 is ready, and this is ZYA!

Test and listen ... ???

Regardless of the style of music, the sound is aggressive and harsh, like a heavy metal concert. Drums and guitars punch walls and ears. Having enjoyed this "joy", I wanted to return to musicality.

Where to "sail" now? I return to the aperiodic load, but now with the addition of padding polyester. I lay the fluffed synthetic winterizer in a wave, and I close the back of the speaker with felt 5 mm thick.

Here he is! The musical sound I need is received. Light open sound, not traumatic for the ear, but with sufficient dynamics. Even Le Binks' drum introduction to the song Judas Priest Better by you, better than me sounds convincing in concert. Arranging "deep" listening ... Three months have passed ...

Conclusion

First, about the bad. Acoustics and my one-beat are not at all "friendly" with modern compressed music and with musical styles of increased density and assertiveness, whether it be electronics or thrash metal and the like. At the exit, creaking porridge.

Good. Classic heavy metal plays well (but not stunning). Hard rock is not bad. Led Zeppelin and early rock in general sound very organic, until 1976 (apparently old recording technologies let you know). Electronics: Kraftwerk, Yello, Enigma, Era, B-Tribe.

About the magnificent. Antonio Vivaldi and violin classics in general. Chamber jazz, such as Diana Krol and the like. All vocals of all the singers are perfectly reproduced by the combination of Magnifique Evolution & Magnifique Accustic. With a claim to luxury!

Recipe and ingredients

The body is made of laminated chipboard 16mm thick. Internal fixing: slats, screws, glue. I did not additionally thicken the walls (except for the lower part of the front panel), arguing that the acoustics are low-power.

Internal casing damping: felt, synthetic winterizer.

It is necessary to put a seal under the speakers or "put" them on the plasticine.

Base: Particleboard with a decorative aluminum corner around the perimeter.

Adjustable legs based on crab furniture nuts.

Filter - capacitor MBGSCH-2 4mkf 160v.

The speaker connections are in phase.

Internal wires copper 1.5 mm "Odeskabel".

Connectors for "banana" Soviet instrument (chiseled bronze contact in the case of carbolite).

Amplifier-AC cable: Odeskabel 2X4mm; Banana Profigold connectors.

Total: $ 72.

After assembling the case, securing and connecting the speakers, place the speakers in the position you want in the room. Listen to your favorite music at the volume you are interested in. Add felt and synthetic winterizer "to taste", achieving the desired tone in the sound.

Thank you for your attention and wish you success.

application

Reasoning about

There is a widespread belief in the network that it is possible to design speakers only if you have in your arsenal a professional measuring microphone, or even better, a sound level meter (for example: VShV-003) and a thoroughly damped, muffled room, without resonances. Next, you need to thoroughly study the work of design and simulation programs. And already on their basis, from special audiophile materials to make a truly "correct speaker".

All this is partly true. Partly ... More precisely, not quite, really. That is, in real life, it will not be so.

Even if VShV-003 or its analog is lying around in your closet, the measured frequency response in a city apartment will be replete with resonant distortions that cancel out all the potential accuracy of the measuring equipment. For the same reason, one should not "hunt" for any special microphones, such as the Panasonic WM-61A.

Speaker cabinet design programs, of course, help you navigate, but purely virtually.

The main thing is the concept and the desire to implement it on the basis of common sense and proportionality. A live experiment will always show what is better and what sounds worse in a specific setting, for which the speaker is created. For example: a speaker needs air - this is a fact, so you shouldn't squeeze it into a tight closed box, but you shouldn't put it in a huge wardrobe either. Its installation dimensions already indicate the minimum proportions of the speaker cabinet in width, and statistically twice as much in depth and height. The height of the speaker from the floor affects the transparency and density of the lows and is determined only empirically in a specific setting, and not by sophisticated programs that are closed on themselves.

Even more creativity involves the process of damping the case with different materials (felt, linoleum, cotton wool, synthetic winterizer, etc.). Having done everything right, on the recommendation of dealing with resonances, you can completely kill the music in the sound.

The damping factor and impulse response of your amplifier will make your designed speaker work just as they please, breaking all the rules and canons. The conclusion is to experiment, listen and correlate with relative measurements. These measurements can be easily done using a simple measurement microphone assembled from a cheap Chinese microphone and any line amplifier on a single chip. Then connect it, at least to a sound card, or in the old fashioned way "hum" with a generator and a millivoltmeter.

I made my measuring microphone based on a Chinese with an insignificant name JM901, and a linear amplifier on K157UD2. I collected it about 15 years ago, just for fun, and put it in the closet. Now, suddenly, he came in handy and did his job perfectly.

See also:

• Your article will be here if you send it to us :) [email protected]

For a long time in my * storehouse * 4 pieces of 4GD-35 were gathering dust, in fact they lay idle and were just being shifted from place to place. Having little experience in using them in the SHIELD, I wanted compactness and a more or less normal appearance, otherwise the SHIELD is not a very practical thing.

Features (one speaker):

Frequency range: 40 (-3 dB) - 15000 Hz

Sensitivity: 93dB

Resistance: 8 ohm

Power: 16W

Weight: 25KG

Since the speakers had a HUGE, as in my opinion * quality factor (about 1.2 - 1.5), and even in the passport it is also written about them. It was decided to make an Open Box, this is essentially the same closed box, only without a back wall, or the same shield but rolled back (although there is a difference, but in my version of execution it is not critical, the main thing is to grasp the essence and then it will be clear ).

The speakers compare favorably with modern heads with their very high sensitivity - 93 dB, which means that the sound pressure that they are able to create with a small input power (nominal 4 W, maximum 8 W, maximum short-term 15 W) is very high.

It also could not but rejoice that the speakers reproduce the range from 50 to 14000 Hz very well, however, it was not without a fly in the ointment, while the frequency response curve is really a curve.

Having chosen a couple with a small scatter, I inserted them into the box from the TV, somehow grabbing them with self-tapping screws and gave a signal from ITUNTDA 2050. The sound is very pleasant. But it's not grief to use them in a cardboard box.

I scraped * along the bottom * and blinded, no Kolobok, I blinded a box of chipboard slabs with external dimensions of 410x300x750 mm, I took 16 mm thick, I made the front wall in two layers, the first was from fragments and with the second I reinforced it very well. Pasted over inside and outside with a carpet and these are the columns.

After listening for a couple of days (even before pasting outside and installing grills), I began to find fault with the fact that HF \u200b\u200bis not enough. No sooner said than done, got into the * zagashnik * and took out a pair of auto tweeters nobly picked from car coaxials. The situation improved a little, but this * poking * started * pressure on my brain *. I took out these tsykalki and began to look for something like 2GD-36, and was rewarded for this, I found one right away, and the other was brought by a friend.

All that would be nothing, but the chuyka that the squeakers have less than the broadband ones, began to think and wonder * who to send where *. As a result, I decided to try to connect the buzzer through a 0.47 uF capacitor without any resistor. Connected - turned on and was stunned, in the pleasant sense of the word. The tweeter blended into the overall sound very well and there was no longer a shortage of HF.

Having cut out the seat and finally secured the squeaker, I was very pleased with the work done. But, as they say, - * it was getting dark *, and since I recently became a family man (in all his joys, and well, you understand), the question arose with the protection of the speakers from * playful little hands *. I went to the store and bought nets.

I think everything is great, I'll turn it on and listen. But it was not there, the tweeters without grids played very well and did not get tired, but with the grids they began * to put pressure on the brain *, at first I thought to remove the grills, but still changed my mind, just connected a 4 Ohm resistor with the HF speaker in series 5 W) and changed the polarity to the tweeter, so my main speakers play in phase, and the tweeters in antiphase. With such an uncomplicated * poking method * I got rid of these overtones, and now the buzzer has become easier to live.

I have broadband devices with a nominal impedance of 4 ohms each, I connected them in series, and in parallel they have already connected a tweeter in antiphase.

As a result, we have an AC with a nominal impedance of 8 Ohm, a more linear frequency response (since the two speakers smooth out each other a little, compensating for the dips or surges of one with the other), the total resonant frequency has dropped to almost 40 Hz. The maximum output power was 16 W with a sensitivity of 93 dB, so the sound pressure from these * oldies * is almost the same as from 35 AC-012 with a power input of about 120 W, but in terms of low frequencies, of course, they are inferior unconditionally.

These speakers are most likely best used with a tube amplifier, and many * lampachi * use such speakers in their * projects *, for me the ITUN on a simple microcircuit is quite enoughTDA 2050, which, thanks to this switching scheme, resembles the sound of tube amplifiers, but at the same time introduces much less distortion.

Costs for the entire project:

A can of PVA glue dispersion 1 kg (half the can is left): 17 hryvnia

Karpet 2.1 lm (width 1.4 m): 155 hryvnia

Speaker grills + grill mounts: 84 hryvnia

The way to the store and back: 4 hryvnia

Self-tapping screws 120 pieces: 12 hryvnia

Speakers and filter capacitors + resistors + wires \u003d were in stock

Time spent on making hulls (clean without smoke breaks for lunch and outdoor recreation): 2 hours

Total: 272 hryvnia

I think * it was worth it *, considering that * a Chinese bubnelka * for that kind of money, standing next to it, will cause a smile.

How do they sound? - It is quite normal, MF and HF are very well transmitted, and low frequencies, even if not 20 Hz, BUT 40 Hz, will play out easily.

If you also add a subicom or two (one to each speaker and apply an active filter, while connecting a biamping), then in terms of quality, many commercially available speakers will be plugged into the belt, with a price tag much higher than the cost of manufacturing my speakers.

Below are a few photos, and a video on YouTube with a demonstration of the work, I immediately warn you that I was recording with a camera and there is no point in waiting for something supernatural. I will add that in reality * live * so to speak - the sound is very comfortable and not tiring.

The article was prepared for the siteldsound. ru

Finally, I am laying out the diagram and measurements of my active open system based on 8AC-4 (2 4GD-35 speakers per column, open box). I started doing it in 2009, since then it has been in various stages of readiness and is being used, mainly in the bedroom.
The know-how of the system is an amplifier: LM3886 in inverted connection, in ITUN mode (voltage-controlled current source), shift towards class A, OO compensation, resonance compensation, limitation of KUS on ultrasound (using inductance in a current sensor), frequency filtering below the workers. A feature of the 4GD-35 is that the recoil drop at high frequencies is approximately proportional to the inductance of the voice coil, and the use of ITUN allows this to compensate and expand the sound band upwards, while there is no need to introduce high frequency suppression, as on many other speakers when using ITUN.
The midrange-treble plays excellently (the treble is noticeably better than many speakers with tweeters), the bass is also enough for everyday listening (shelf from 100Hz thanks to the compensator). If you need a disc, then you can sew some kind of low-frequency link with a biamp. There is enough bass for adequate room listening, while they are completely devoid of any smearing, hum or "banging on the head" characteristic of small closed speakers.

·· ··

So, the usual 8AS-4 are taken as a basis, in each column there are 2 pieces of 4GD35 in an open box. Speaker modifications: impregnation with a diffuser with a solution of chitosan in vinegar, impregnation for paper suspensions with softening impregnation AIE, anti-magnets, sticking pieces of rubber from the camera on the speaker baskets. The rework of the boxes is minimal: gluing the seams and an old wadded-felt rug on the walls from the inside. The goal was to maintain an authentic appearance and dimensions.
Now to the diagram (clickable):
The amplifier circuit is suitable for any OY or shield on 4GD35 somewhere 1-4 (maybe more) speakers per column in series connection, the only thing that will need to be adjusted for another shield or box is OB Compensation (the values \u200b\u200bcan be selected in the emulator or in Edge, final hearing correction). The resonance corrector may, in any case, require tuning to suit your speakers. For ZA, PAS this place can be thrown out, the input conductor (C8) can be increased to 0.47..1 mF, reconfigured or removed at resonance.
The diagram does not show the delay in turning on the speakers after connecting the power, a trivial one is used on the transistor and the conductor, for about a second, otherwise, before dialing ~ 8 volts on the power supply and entering the microcircuit into the mode "claps" side of class A "). If Rqc is removed, then there is no need to set the delay, and there will also be a good saving in electricity and radiators. At the cost of a very "subtle" change in sound quality.
R8 / C7 / R9 are not explained in the diagram, this is an optional additional high-pass filter to attenuate frequencies below the operating ones, for ZP, again, you can throw it out.
Well, the power supply unit is also not drawn, everything is traditional there.
The frequency graph of the amplifier, built on the model, and may be at slightly different ratings, in any case it is not accurate in relation to real speakers. I attach, just so that you can understand the idea:
Today I had the opportunity to rush the system in haste. Measurement conditions: living room with wild chaos. Speakers on the floor. Measured from about 1 meter. When changing amplifiers, I had to move everything a little from place to place. In the process of measuring, it turned out that the resonance over the years of operation was smeared and dropped by 10-15 hertz (probably, this is a setup from the side of impregnation), so the Itunovskaya squiggle got out on the bass. (I don’t know if I’ll correct it or not, or just forget it. Once upon a time there was a thought to make a sub from ~ 80-100 Hz to them. To fix it, add 33/16 nF to the capacitors in the Twin-T filter.) And also, although the single speaker on the equalizer has -8 dB at 20 kHz, the two speakers already have a much larger blockage. By the way, above ~ 18 KHz, the charts should be a little overwhelmed, because for technical reasons I had to connect a tablet as a source and use 44 KHz digitizing (it lies around 20 KHz compared to 192 KHz, which I usually use). Put 2 dB there in your mind
1. Conventional amplifier, near field of one of the speakers:

2. ITUN + corrector, near field of one of the speakers:

3. Conventional amplifier, 1 meter, on the floor:

4. ITUN + corrector, 1 meter, on the floor:

The sound differs radically - without correction, just listen to bare vocals or a piano, with correction they play more or less any music and much better. To your taste, you can equalize the second Twin-T hump in the region of 300 Hz and adjust the Twin-T at resonance more correctly, and add an active low-frequency section for very good low frequencies.