The output power is quite high, since the main amplifier is built according to the famous LANZAR circuit, which provides a maximum power of 390 watts, but of course the amplifier does not operate at full power. This amplifier is designed to power the SONY XPLOD XS-GTX120L subwoofer head, head parameters are shown below.
>>
Rated power - 300 W
>>
Peak power - 1000 W
>>
Frequency range 30 - 1000 Hz
>>
Sensitivity - 86 dB
>>
Output impedance - 4 ohms
>>
Diffuser material - polypropylene
.
In addition to the subwoofer amplifier, there are also 4 separate amplifiers in the complex, two of which are made on a well-known microcircuit TDA7384, as a result, 8 channels of 40 watts each are designed to power the interior acoustics. The remaining two amplifiers are made on a chip TDA2005, I used these particular microcircuits for one reason - they are cheap and have good sound quality and output power. The total power of the installation (nominal) is 650 watts, the peak power reaches 750 watts, but it is difficult to overclock to peak power, since the power supply does not allow this. Of course, 12 volts of a car is not enough to power a subwoofer amplifier, so a voltage converter is used.
Voltage transformer- perhaps the most difficult part of the whole structure, so let's consider it in a little more detail. Of particular difficulty is the winding of the transformer. The ferrite ring is almost never found on sale, so it was decided to use a transformer from a computer power supply, but since the frame of one transformer is clearly too small for winding, two identical transformers were used. First you need to find two identical ATX PSUs, solder large transformers, disassemble them and remove all factory windings. Ferrite halves are glued to each other, so they should be heated with a lighter for a minute, then the halves can be easily removed from the frame. After removing all the factory windings, you need to cut off one of the side walls of the frame, it is advisable to cut off the wall free from contacts. We do this with both frames. At the last stage, you need to attach the frames to each other as shown in the photographs. To do this, I used ordinary tape and electrical tape. Now you need to start winding.
The primary winding consists of 10 turns with a tap from the middle. The winding is wound immediately with 6 wires of 0.8 mm wire. First, we wind 5 turns along the entire length of the frame, then we isolate the winding with insulating tape and wind the remaining 5.
Now we wind the secondary winding. It is wound according to the same principle as the primary one, only it contains 40 turns with a tap from the middle. The winding is wound immediately with 3 cores of wire 0.6-0.8 mm, first one shoulder (along the entire length of the frame), then the other. After winding the first winding, we put insulation on top and wind the second half identically to the first. At the end, the wires are stripped of varnish and coated with tin. The last stage is to insert the halves of the core and fix it.
IMPORTANT! Do not allow a gap between the halves of the core, this will lead to an increase in the quiescent current and to abnormal operation of the transformer and the converter as a whole. You can fix the halves with tape, then fix with glue or epoxy. While the transformer is left alone and proceed to the assembly of the circuit. Such a transformer is capable of providing a bipolar voltage of 60-65 volts at the output, a rated power of 350 watts, a maximum of 500 watts, and a peak of 600-650 watts.
master oscillator rectangular pulses is made on a two-channel PWM controller TL494 tuned to a frequency of 50 kHz. The output signal of the microcircuit is amplified by a driver on low-power transistors, then it goes to the gates of the field switches. Driver transistors can be replaced with BC557 or domestic ones - KT3107 and other similar ones. The field effect transistors used are the IRF3205 series - this is an N - channel power transistor with a maximum power of 200 watts. 2 such transistors are used for each arm. In the rectifier part of the power supply, diodes of the KD213 series are used, although any diodes with a current of 10-20 amperes that can operate at frequencies of 100 kHz or more are suitable. You can use Schottky diodes from computer power supplies. To filter high-frequency interference, two identical chokes were used, they are wound on rings from computer power supplies and contain 8 turns of 3-wire wires 0.8 mm.
The main inductor is powered, wound on a ring from a computer power supply unit (the largest ring in diameter), it is wound with 4 strands of wire with a diameter of 0.8 mm, the number of turns is 13. The converter is powered when the remote control output is supplied stable plus, then the relay closes and the converter starts working. The relay must be used with a current of 40 amperes or more. Field keys are installed on small heat sinks from a computer PSU, they are screwed to the radiators through heat-conducting pads. The snubber resistor - 22 ohms should overheat a little, this is quite normal, so you need to use a resistor with a power of 2 watts. Now back to the transformer. It is necessary to phase the windings and solder it to the converter board. We first phase the primary winding. To do this, you need to solder the beginning of the first half of the winding (shoulder) to the end of the second or vice versa - the end of the first to the beginning of the second.
If the phasing is incorrect, the converter will either not work at all, or the field workers will fly off, so it is desirable to mark the beginning and end of the halves when winding. The secondary winding is phased exactly according to the same principle. Printed circuit board - in .
The finished converter should work without whistles and noises, at idle the heat sinks of transistors may overheat slightly, the quiescent current should not exceed 200 mA. After the completion of the PM, you can consider that the main work is done. You can already start assembling the LANZAR circuit, but more on that in the next article.
Discuss the article AMPLIFIER WITH YOUR HANDS - POWER SUPPLY
It would seem that it could be simpler - I took the power supply, connected it with two or three wires to the amplifier and everything ... should sing? It turns out not always. As we have already found out in this series of articles, there are many pitfalls here.
Let's continue to understand the intricacies of the wires supplying the amplifier. And oddly enough, the common (earth) conductor can deliver the most problems.
Let's fix one mistake first. The article published a diagram of a bipolar amplifier power supply, but its wiring diagram was missing.
Here's both for you:
Bipolar power amplifier power supply.
Bipolar Power Amplifier Power Supply Wiring Diagram
In fact, there are two “mirrored” unipolar blocs here.
Speaker reverse current
As you know, the acoustic system is a reactive load. So, it can return current to the amplifier. This current, flowing through the conductors, creates a potential difference, which can lead to positive feedback and, as a result, instability of the amplifier.
To avoid this, the loudspeaker ground terminal should be connected to common terminal of the filter capacitors nutrition. Often the output of the loudspeaker is connected to the common output of the microcircuit, as shown in the figure:
This connection closes the negative half-wave of the signal in the local loop, eliminating the filter capacitor, which could reduce radiated noise and improve system stability.
The figure shows how the leakage current to ground of one half-wave of a signal can induce unpleasant noise and distortion if the common wire of the loudspeaker is connected to the output stage of the microcircuit:
Similarly, if there are bypass capacitors on the amplifier board in the power circuits (and they usually are) of a rather large capacity of several hundred microfarads, then charging current pulses will also create a potential difference on the common conductor. Therefore, we repeat once again, the best point for connecting the common wire of the speaker system is the common terminal of the power filter capacitors.
The more power, the worse...
Often, radio amateurs try to make their amplifier as powerful as possible (like, so cool), and audiophiles often equip their systems with amplifiers with a power many times higher than necessary to sound a normal room to a normal volume level, motivating them to get a greater dynamic range. Such amplifiers (high power) sometimes solve some problems, but create others.
The inductance of the power conductors is the main "weak link" of class AB power amplifiers. In such amplifiers, the output transistors turn on and off alternately, respectively, half-waves of charging currents flow through the positive and negative power buses.
If these pulses get into the audio path through capacitive and inductive couplings, this leads to a terrible blurry sound.
This happens if some sensitive track (conductor) passes next to the power bus. The bifilar lay of power wires effectively suppresses radiated interference due to mutual compensation of positive and negative half-waves.
On a printed circuit board, this method can be implemented by placing the power rails on top of each other on both sides of the board (requires a double-sided printed circuit board)
A worthy example of PCB design for a power amplifier is the Ultra-LD 200W design, presented in one of the Practical Electronics Every Day magazine. On the printed circuit board of this amplifier, all the installation recommendations presented in this series of articles are implemented. And largely due to this, it was possible to obtain a noise level of -122 dB and a level of non-linear distortion below 0.001%.
Note from the editors of RadioGazeta: if our readers are interested, write in the comments and we will publish a description of this amplifier.
Grounding one side of the PCB works well in high frequency and low current designs. This is not suitable for power amplifiers, because it is difficult to predict the flow of currents depending on the choice of ground points.
In modern tube amplifiers, the common bus is often made in the form of a piece of thick tinned wire. Many gurus preach star wiring with a single connection point. There are cases when amplifiers do not work well with this approach. Says a large number of long wires, which reduce the stability of the structure.
As a rule, in a good amplifier there are several ground points.
denouement
When using two filter capacitors with a bipolar supply, it is necessary to ensure that the two half-waves of the signal are summed at one point, as it shown on the picture:
Often the use of a single capacitor connected between the plus and minus of the power supply solves this problem. This method works well with 5532 type op amps, and LM3886 type power amplifiers.
When the driver stage and output stage are powered by separate wires, this can cause some instability of the amplifier at high frequencies. The problem is solved by connecting a small ceramic capacitor between the power pins of the microcircuit:
zoom on click
If the bypass (blocking) capacitors are larger than 100uF, their common wire must be connected to a "dirty" ground, since large charging currents can create noticeable interference if the capacitors are connected to signal ground.
Zobel chain
A Zobel circuit at the output of the amplifier prevents it from being excited at high frequencies. The current pulses in this circuit can cause problems, so they must be shorted to a "dirty" ground, that is, to the common terminal of the filter capacitors or bypass capacitors.
For some power amplifier ICs, long wires in Zobel circuits cause instability on the negative half-waves of the signal.
Example of mounting a mono amplifier
Typically, the "star" in a single-supply amplifier is three-path: signal ground, power filter capacitor ground and "dirty" ground. An example is shown in the figure:
zoom on click
Here, an amplifier should be understood as an integrated design, as well as amplifiers based on discrete elements.
As you can see, a signal ground is connected to one beam - here the currents are very small, so there is no need to connect all the elements with separate conductors. To the second beam individual conductors the outputs of high-current circuits are connected: the output stage, the Zobel circuit, the common output of the speaker system and bypass capacitors. The common output of the filter capacitor of the power supply is connected to the third beam.
The correct connection of the common wire to the pins of the microcircuits is shown in the figure:
Option "c" is the wrong option. Due to the track resistance, a large current will raise the potential of the low-current common wire relative to the output of the microcircuit, which will lead to an increase in distortion.
To be continued...
The article was prepared based on the materials of the journal "Practical Electronics Every Day"
Free translation: Chief editor « »
It would seem that it could be easier to connect the amplifier to power supply and enjoy your favorite music?
However, if we recall that the amplifier essentially modulates the voltage of the power supply according to the law of the input signal, it becomes clear that the design and installation issues power supply should be approached very responsibly.
Otherwise, mistakes and miscalculations made at the same time can spoil (in terms of sound) any, even the most high-quality and expensive amplifier.
Stabilizer or filter?
Surprisingly, most power amplifiers are powered by simple circuits with a transformer, a rectifier, and a smoothing capacitor. Although most electronic devices today use stabilized power supplies. The reason for this is that it is cheaper and easier to design an amplifier that has a high ripple rejection ratio than it is to build a relatively powerful regulator. Today, the level of ripple suppression of a typical amplifier is about 60dB for a frequency of 100Hz, which practically corresponds to the parameters of a voltage regulator. The use of direct current sources, differential stages, separate filters in the power supply circuits of the stages and other circuitry techniques in the amplifying stages makes it possible to achieve even greater values.
Nutrition output stages most often made unstabilized. Due to the presence in them of 100% negative feedback, unity gain, the presence of LLCOS, the penetration of the background and ripple of the supply voltage to the output is prevented.
The output stage of the amplifier is essentially a voltage (power) regulator until it enters clipping (limiting) mode. Then the ripple of the supply voltage (frequency 100 Hz) modulates the output signal, which sounds just awful:
If for amplifiers with a unipolar supply only the upper half-wave of the signal is modulated, then for amplifiers with a bipolar supply, both half-waves of the signal are modulated. Most amplifiers have this effect at large signals (powers), but it is not reflected in any way in the technical characteristics. In a well-designed amplifier, clipping should not occur.
To test your amplifier (more precisely, the power supply of your amplifier), you can conduct an experiment. Apply a signal to the input of the amplifier with a frequency slightly higher than you can hear. In my case, 15 kHz is enough :(. Increase the amplitude of the input signal until the amplifier enters clipping. In this case, you will hear a hum (100 Hz) in the speakers. By its level, you can evaluate the quality of the power supply of the amplifier.
Warning! Be sure to turn off the tweeter of your speaker system before this experiment, otherwise it may fail.
A stabilized power supply avoids this effect and results in less distortion during prolonged overloads. However, taking into account the instability of the mains voltage, the power loss on the stabilizer itself is approximately 20%.
Another way to reduce the clipping effect is to feed the stages through separate RC filters, which also reduces power somewhat.
In serial technology, this is rarely used, since in addition to reducing power, the cost of the product also increases. In addition, the use of a stabilizer in class AB amplifiers can lead to excitation of the amplifier due to the resonance of the feedback loops of the amplifier and regulator.
Power losses can be significantly reduced if modern switching power supplies are used. Nevertheless, other problems emerge here: low reliability (the number of elements in such a power supply is much larger), high cost (for single and small-scale production), high level of RF interference.
A typical power supply circuit for an amplifier with an output power of 50W is shown in the figure:
The output voltage due to smoothing capacitors is approximately 1.4 times greater than the output voltage of the transformer.
Peak power
Despite these shortcomings, when the amplifier is powered from unstabilized source, you can get some bonus - short-term (peak) power is higher than the power of the power supply, due to the large capacity of the filter capacitors. Experience shows that a minimum of 2000µF is required for every 10W of output power. Due to this effect, you can save on the power transformer - you can use a less powerful and, accordingly, cheap transformer. Keep in mind that measurements on a stationary signal will not reveal this effect, it appears only with short-term peaks, that is, when listening to music.
A stabilized power supply does not give such an effect.
Parallel or series stabilizer?
There is an opinion that parallel regulators are better in audio devices, since the current loop is closed in a local load-stabilizer loop (power supply is excluded), as shown in the figure:
The same effect is obtained by installing a decoupling capacitor at the output. But in this case, the lower frequency of the amplified signal limits.
Protective resistors
Every radio amateur is probably familiar with the smell of a burnt resistor. It's the smell of burning varnish, epoxy and... money. Meanwhile, a cheap resistor can save your amp!
When the author first turns on the amplifier in the power circuits, instead of fuses, he installs low-resistance (47-100 Ohm) resistors, which are several times cheaper than fuses. This has repeatedly saved expensive amplifier elements from installation errors, incorrectly set quiescent current (the regulator was set to maximum instead of minimum), reversed power polarity, and so on.
The photo shows an amplifier where the installer mixed up TIP3055 transistors with TIP2955.
The transistors were not damaged in the end. Everything ended well, but not for the resistors, and the room had to be ventilated.
The key is voltage drop.
When designing printed circuit boards for power supplies and not only, one should not forget that copper is not a superconductor. This is especially important for "ground" (common) conductors. If they are thin and form closed circuits or long circuits, then due to the current flowing through them, a voltage drop occurs and the potential at different points turns out to be different.
To minimize the potential difference, it is customary to wire the common wire (ground) in the form of a star - when each consumer has its own conductor. The term "star" should not be taken literally. The photo shows an example of such a correct wiring of a common wire:
In tube amplifiers, the resistance of the anode load of the cascades is quite high, of the order of 4 kOhm and higher, and the currents are not very large, so the resistance of the conductors does not play a significant role. In transistor amplifiers, the resistance of the cascades is significantly lower (the load generally has a resistance of 4 ohms), and the currents are much higher than in tube amplifiers. Therefore, the influence of conductors here can be very significant.
The resistance of a track on a printed circuit board is six times higher than the resistance of a piece of copper wire of the same length. The diameter is taken 0.71mm, this is a typical wire that is used when mounting tube amplifiers.
0.036 Ohm as opposed to 0.0064 Ohm! Considering that the currents in the output stages of transistor amplifiers can be a thousand times higher than the current in a tube amplifier, we find that the voltage drop across the conductors can be 6000! times more. Perhaps this is one of the reasons why transistor amps sound worse than tube amps. This also explains why PCB-assembled tube amps often sound worse than surface-mounted prototypes.
Don't forget Ohm's law! Various techniques can be used to reduce the resistance of printed conductors. For example, cover the track with a thick layer of tin or solder a tinned thick wire along the track. The options are shown in the photo:
charge impulses
To prevent the penetration of the mains background into the amplifier, measures must be taken to prevent the penetration of charge pulses of the filter capacitors into the amplifier. To do this, the tracks from the rectifier must go directly to the filter capacitors. Powerful pulses of charging current circulate through them, so nothing else can be connected to them. the power supply circuits of the amplifier must be connected to the terminals of the filter capacitors.
The correct connection (mounting) of the power supply for an amplifier with unipolar power supply is shown in the figure:
Zoom on click
The figure shows a PCB variant:
Ripple
Most unregulated power supplies have only one smoothing capacitor after the rectifier (or several connected in parallel). To improve the quality of power, you can use a simple trick: split one container into two, and connect a small resistor of 0.2-1 ohm between them. At the same time, even two containers of a smaller denomination can be cheaper than one large one.
This gives a smoother output voltage ripple with less harmonics:
At high currents, the voltage drop across the resistor can become significant. To limit it to 0.7V, a powerful diode can be connected in parallel with the resistor. In this case, however, at the peaks of the signal, when the diode opens, the output voltage ripples will again become “hard”.
To be continued...
The article was prepared based on the materials of the journal "Practical Electronics Every Day"
Free translation: Editor-in-Chief of Radio Gazeta
A switching power supply that provides a bipolar voltage of +/-50V with a power of up to 300 W is designed for use or high-power laboratory PSUs (). This relatively simple switching power supply circuit is assembled mainly from radio elements taken from old AT / ATX power supplies.
Schematic diagram of the converter 220 / 2x50V
Scheme of a home-made pulsed PSU for UMZCH The inverter transformer was wound on an ETD39 ferrite core. The winding data are practically the same, only the output windings are slightly wound up for an increase in voltage. Key transistors - powerful IRFP450. The driver is the popular TL494 chip. Power is supplied through a special stabilizer. In it, a starting resistor with a rectified mains voltage charges the power capacitor, on which, when the voltage reaches the threshold, the stabilizer will turn on, starting the driver. It will be powered only at the moments of energy accumulation on the capacitor, and after starting the converter, the driver will be powered by an additional transformer winding. The principle of operation of this launch option has been known for a long time and is used in the popular m / s UC384x.
Printed circuit board power stage
Another feature of the PSU circuit design is the control of field-effect transistors. Here, according to the IRFP450 scheme, the lower one is controlled directly from the driver output, and the upper one with the help of a small transformer.
In addition, the system was equipped with current protection, monitoring the current of the lower field using its resistance Rdson.
PSU test results
Ready power supply - board with parts In practice, it was possible to get about 100-150 output power into 4 ohm speakers. The voltage +/-50V is set by the resistor P1 10k. Of course, it can take on any values, depending on the ULF scheme used. The system is currently in operation.
Other articles on the construction of this ULF.
Schematic diagram of the power supply.
The power supply is assembled according to one of the standard schemes. A bipolar power supply is selected to power the final amplifiers. This allows the use of low cost, high quality integrated amplifiers and eliminates a number of problems associated with supply voltage ripple and turn-on transients. https://website/
The power supply must provide power to three microcircuits and one LED. Two TDA2030 microcircuits are used as final power amplifiers, and one TDA1524A microcircuit is used as a volume control, stereo base and tone control.
The electrical circuit of the power supply.
|
VD3... VD6 - KD226 |
C1-680mkFx25V C3... C6 - 1000mkFx25V |
On diodes VD3 ... VD6, a bipolar full-wave rectifier with a midpoint is assembled. This switching circuit reduces the voltage drop across the rectifier diodes by half compared to a conventional bridge rectifier, since current flows through only one diode in each half-cycle.
Electrolytic capacitors C3 ... C6 are used as a rectified voltage filter.
On the IC1 chip, a voltage regulator is assembled to power the electronic volume control circuit, stereo base and tone. The stabilizer is assembled according to a standard scheme.
The use of the LM317 chip is due only to the fact that it was available. Here you can apply any integral stabilizer.
The protective diode VD2, indicated by a dotted line, is not necessary when the output voltage on the LM317 chip is below 25 volts. But, if the input voltage of the microcircuit is 25 Volts and higher, and the resistor R3 is trimmer, then it is better to install the diode.
The value of the resistor R3 determines the output voltage of the stabilizer. During prototyping, I soldered a trimmer instead, set the voltage to about 9 volts at the output of the stabilizer with it, and then measured the resistance of this trimmer so that I could install a constant resistor instead.
The rectifier that feeds the stabilizer is made according to a simplified half-wave circuit, which is dictated by purely economic considerations. Four diodes and one capacitor cost more than one diode and one slightly larger capacitor.
The current consumed by the TDA1524A chip is only 35mA, so this scheme is fully justified.
LED HL1 - power-on indicator of the amplifier. A ballast resistor of this indicator is installed on the power supply board - R1 with a nominal resistance of 500 Ohms. The current of the LED depends on the resistance of this resistor. I used a green LED rated at 20mA. When using a red LED type AL307 for a current of 5mA, the resistance of the resistor can be increased by 3-4 times.
Printed circuit board.
The printed circuit board (PCB) is designed based on the design of a particular amplifier and the available electrical components. The board has only one mounting hole, located in the very center of the PCB, which is due to an unusual design.
To increase the cross-section of copper tracks and save ferric chloride, the places free from tracks on the PCB were filled using the "Polygon" tool.
Increasing the width of the tracks also prevents peeling of the foil from the fiberglass in case of violation of the thermal regime or during repeated soldering of radio components.
According to the drawing given above, a printed circuit board was made of foil fiberglass with a cross section of 1 mm.
To connect the wires to the printed circuit board, copper pins (soldiers) were riveted in the holes of the board.
This movie requires Flash Player 9 |
||
And this is the already assembled printed circuit board of the power supply.
To see all six views, drag the picture with the cursor or use the arrow buttons located at the bottom of the picture.
The mesh on the PP copper tracks is the result of using this technology. 
When the board is assembled, it is desirable to test it even before connecting the final amplifiers and the regulator unit. To test the power supply, you need to connect a load equivalent to its outputs, as in the above diagram.
As a load of +12.8 and -12.8 Volt rectifiers, resistors of the PEV-10 type for 10-15 Ohms are suitable.
The voltage at the output of the stabilizer, loaded on a resistor with a resistance of 100-150 ohms, is a good idea to look with an oscilloscope for the absence of ripples when the AC input voltage is reduced from 14.3 to 10 volts.
P.S. Finalization of the printed circuit board.
During commissioning, the printed circuit board of the power supply came in.
When finalizing, I had to cut one track pos.1 and add one contact pos.2 to connect the transformer winding that feeds the voltage stabilizer.
