Distinctive feature and the disadvantage of conventional linear voltage stabilizers operating in the mode of strong input level deviations is their low efficiency. This situation, as a rule, is explained by significant heat losses in the circuit elements. In addition, such devices at high load currents (up to tens of amperes) look very bulky and have significant weight. It is possible to significantly improve all the indicated parameters of the converter device in the case of using the pulsed stabilization method.

Switching stabilizer voltage is a device of a special class that allows you to maintain the output voltage within the specified limits due to the key mode of operation of the main elements of the circuit. Consider the principle of operation of this device in more detail.

Fundamentals of Pulse Conversion

First of all, you should know that pulse devices for obtaining a stabilized voltage, like their linear counterparts, can be performed in parallel and in series. In both cases, the function of a key element is traditionally performed by a powerful field-effect transistor. Since in key mode it operating point is instantly shifted from the saturation region to the cutoff zone (quickly "overshooting" the active section), such a scheme has minimal heat losses. And this indicates that the switching voltage regulator has a high efficiency.

The stabilization of the output signal is carried out by controlling the duration or repetition rate generated by a special pulse generator, which in electronics is called latitude (SHI) or frequency (FH) pulse regulation.

Note! In some models of such devices, a combined width-frequency control method (CHSHI) is used.

In stabilizers of the first type (SHI), the pulse repetition frequency remains constant, and only their duration changes. In the second case, the frequency is subject to change, and the length (duty cycle) of the pulse signal does not change with time.

At the output of the regulating converter (inverter), there is a square wave signal, which is not suitable for supplying to the working load. Therefore, it must first be straightened or smoothed to a form suitable for use. This explains the presence of a special filtering module at the output of the device, consisting of elements smoothing pulsations. Their function is traditionally performed by capacitive-inductive chains of the P- or L-shaped type.

Depending on the parameters of these circuits (on the inductance of the inductor, in particular), the current through the LC filter element may be intermittent or constant. Everything is determined by whether, by the time the next pulse arrives, the previously charged capacitor has time to discharge through the inductance. When presenting special requirements for the level of ripples, preference is given to the inseparable principle of generating the output current.

Additional Information. A kind of "retribution" for this is a significant consumption of copper material used to manufacture the throttle coil.

In cases where the value of the ripple coefficient is not standardized, it is allowed for the circuit to operate in the intermittent current mode.

block diagram

The classic switching voltage regulator contains the following mandatory modules:

• master oscillator;
• Directly converter (inverter);
• Comparing device;
• filter element.

The master oscillator (MG) provides the formation of pulses with a shape close to a rectangular standard. The latter enter the converter device, where they are processed according to the selected control parameter (frequency, duration, or both at once). Then the processed pulses are fed to the filter element, and after it - to the output and to the feedback (control) link chain.

The block diagram below will help you get acquainted with the operation of the device.

Important! The key link in this scheme is the chain feedback(comparison device), the presence of which allows, by the state of the output signal, to determine the need for additional actions (adjustments).

That is, when the output signal has ideal parameters, the device compares it with exemplary voltages and perceives this as a command to interrupt the control operation. If the shape or other characteristics of the output signal begin to differ from the parameters specified in the technical specifications, the comparing module (CS) generates a signal for additional correction of the pulses generated by the generator.

Benefits of OS regulation

A differential signal proportional to the deviation of the output voltage parameters from the norm is supplied to the master oscillator, so that the whole circuit works on the principle of a differential amplifier. Such a circuit design makes it possible to multiply the sensitivity of the feedback loop (FB) and increase the efficiency of the control process.

In this mode, the generated control pulses are sent to the key elements of the converter device, where they are processed with simultaneous preparation for subsequent filtering. When the frequency or pulse width is changed by a signal from the control system, it is possible to achieve the required quality of the output voltage.

Additional Information. There may be situations where the need for adjustment is completely eliminated. This usually happens when the output voltage is within specification.

Control device diagrams

Increasing

Boosting pulse stabilization circuits are in demand when it is necessary to connect a load, the voltage on which must exceed the input parameter by some amount. In this case, galvanic isolation between the consumer and the 220 Volt power supply network is not provided. Abroad, this conversion principle is called a “boost converter”, and its diagram is shown in the figure below.

When a control voltage is applied between the gate and the source of the transistor VT1, it enters a state of saturation, ensuring the unhindered flow of current through the storage inductor L1. In this case, the output current component is created by charging the capacitor C1.

After removing the potential from the transistor VT1, it goes into a cutoff state; in this case, an EMF of self-induction appears on the inductor L1, which is transmitted through the diode VD1 to the load with the same polarity. At the end of the flow of current through the inductor L1, the coil fully transfers energy to the circuit. It is taken by the capacitor C1, which is charged until the transistor VT1 is again in saturation.

Step Down Stabilizer

The step-down stabilizer works on the same principle, but only the inductor in this case is turned on after the controlled field-effect transistor (see figure below).

The foreign name for this transformation principle is "chopper", and its characteristic feature is the reduced output voltage .

After the control pulse is applied to VT1, the transistor saturates, as a result of which a current begins to flow through it, which flows through the smoothing inductor L1 directly to the load (the diode VD1 is closed by reverse voltage).

After the input signal is removed, the key transistor will go into cutoff mode, which will lead to a sharp decrease in current. The self-inductance emf of the inductor L1 will strongly prevent its decrease, maintaining the process under load. However, due to the voltage drop across the L1 coil, its value at the output of the device will always be less than the input value (due to the opposite sign of the EMF).

Inverter

This type of stabilizer is used when working with loads that have a fixed voltage output voltage, phase-shifted relative to the input. Moreover, its value itself can be either greater or less than the input (it all depends on how the inverting device was rebuilt).

Similarly to both previous schemes, here the galvanic isolation of the supply and output circuits is completely absent. In a foreign lexicon, such stabilizers are referred to as "buck-boost converter". The main circuit difference from the buck converter is that the inductor and diode in this case are reversed. Moreover, the semiconductor element is switched on in the reverse (closed for direct current) direction.

Such a replacement leads to a phase shift between the input and output signals by 90 degrees (in other words, to its inversion).

In the final part of this review, let's pay attention to one more detail that is characteristic of all the considered types of converting devices. As a switching key in all circuits, a special semiconductor element with a field structure is used, controlled not by voltage, but by potential. Due to this, it is possible to repeatedly reduce the input control currents, as well as further increase the efficiency of the entire device as a whole.

Video

Attachment to the power supply

This converter was conceived as a prefix that allows you to expand the voltage range of a laboratory power supply, designed for an output voltage of 12 volts and a current of 5 amperes. The schematic diagram of the converter is shown in Figure 1.

The basis of the device is a microcircuit of a single-cycle pulse-width controller UC3843N, included according to a typical scheme. This scheme of the ball was directly borrowed from the German radio amateur Georg Tief (Tief G. Dreifacher Step-Up-Wandler. Stabile Spennunger fϋr den FieldDay). Data in Russian for this microcircuit can be viewed in the reference book “Microcircuits for impulse sources power supply and their application ”publishing house“ Dodeka ”on page 103. The circuit is not complicated, and with serviceable parts and proper installation, it starts working immediately. The output voltage of the converter is adjusted using the trimmer resistor R8. But if desired, it can be changed to a variable resistor. The output voltage can be changed from 15 to 40 volts, with the values ​​​​of the resistors R8, R9, R10 indicated in the diagram. This converter has been tested with a 24 volt, 40 watt soldering iron.
So:

Output voltage ……………… 24 V
The load current was ……....... 1.68 A
Load power ………………. 40.488 W
Input voltage ………………... 10.2 V
Total current consumption ………. 4.65 A
general power…………………... 47.43 W
The resulting efficiency ………………... 85%
The temperature of the active components of the circuit was around 50 degrees.

In this case, the key transistor and the diode with a Schottky barrier have small radiators. As a key transistor, an IRFZ34 transistor is used, which has an open channel resistance of 0.044 Ohm, and one of the diodes of the S20C40C diode assembly, soldered from the power supply of an old computer, is used as a diode. The printed circuit board provides switching of diodes using a jumper. Other Schottky barrier diodes with a forward current of at least twice the load current can be used. The inductor is wound on a yellow and white sprayed iron ring, also taken from a PC power supply. You can read about such cores in Jim Cox's brochure. You can download it from the web. In general, I advise you to download this article and read it in full. A lot of useful material on chokes.

The magnetic permeability of such a ring is 75, and its dimensions are D = 26.9 mm; d = 14.5 mm; h = 11.1 mm. The inductor winding has 24 turns of any winding wire with a diameter of 1.5 mm.

All parts of the stabilizer are installed on a printed circuit board, and on the one hand, all the “high” parts are installed, and on the other, all, so to speak, “undersized”. The circuit board drawing is shown in Figure 2.

The first switching on of the assembled device can be done without a key transistor and make sure that the PWM controller is working. At the same time, there should be a voltage of 5 volts at pin 8 of the microcircuit, this is the voltage of the internal reference voltage source of the ION. It must be stable when the supply voltage of the microcircuit changes. Both the frequency and the amplitude of the sawtooth voltage at output 4 DA1 must be stable. After making sure that the controller is working, you can also solder a powerful transistor. Everything should work.

Do not forget that the load current of the stabilizer must be less than the current that your power supply is designed for and its value depends on the output voltage of the stabilizer. Without load at the output, the stabilizer consumes a current of approximately 0.08 A. The frequency of the pulse sequence of control pulses without load is in the region of 38 kHz. And a little more, if you draw a printed circuit board yourself, read the rules for mounting a microcircuit according to its documentation. Stable and trouble-free operation of impulse devices depends not only on high-quality parts, but also on the correct wiring of the printed circuit board conductors. Good luck. K.V.Yu.

Making a power supply with your own hands makes sense not only for an enthusiastic radio amateur. homemade block power supply (PSU) will create convenience and save a considerable amount also in the following cases:

• To power a low-voltage power tool, in order to save the resource of an expensive battery (battery);
• For the electrification of premises that are especially dangerous in terms of the degree of electric shock: basements, garages, sheds, etc. When powered by alternating current, its large value in low-voltage wiring can interfere with household appliances and electronics;
• In design and creativity for precise, safe and waste-free cutting of foam plastic, foam rubber, low-melting plastics with heated nichrome;
• In lighting design - the use of special power supplies will extend the life led strip and get stable lighting effects. Power supply of underwater illuminators, etc. from a household power supply is generally unacceptable;
• For charging phones, smartphones, tablets, laptops away from stable power sources;
• For electroacupuncture;
• And many others who do not have direct relationship to electronics, purposes.

Permissible simplifications

Professional PSUs are designed to power loads of any kind, incl. reactive. Among the possible consumers - precision equipment. The pro-PSU must maintain the specified voltage with the highest accuracy indefinitely for a long time, and its design, protection and automation must allow operation by unskilled personnel in harsh conditions, for example. biologists to power their instruments in a greenhouse or on an expedition.

An amateur laboratory power supply is free from these restrictions and therefore can be significantly simplified while maintaining quality indicators sufficient for its own use. Further, through also simple improvements, it is possible to obtain a special-purpose power supply unit from it. What are we going to do now.

Abbreviations

1. Short circuit - short circuit.
2. XX - idling, i.e. sudden disconnection of the load (consumer) or a break in its circuit.
3. KSN - voltage stabilization coefficient. It is equal to the ratio of the change in the input voltage (in% or times) to the same output voltage at a constant current consumption. Eg. the mains voltage dropped "in full", from 245 to 185V. Relative to the norm at 220V, this will be 27%. If the PSV of the PSU is 100, the output voltage will change by 0.27%, which at its value of 12V will give a drift of 0.033V. More than acceptable for amateur practice.
4. PPN is a source of unstabilized primary voltage. This can be a transformer on iron with a rectifier or a pulsed mains voltage inverter (IIN).
5. IIN - operate at an increased (8-100 kHz) frequency, which allows the use of lightweight compact transformers on ferrite with windings of several to several tens of turns, but are not without drawbacks, see below.
6. RE - the regulating element of the voltage stabilizer (SN). Maintains the specified output value.
7. ION is a reference voltage source. Sets its reference value, according to which, together with the feedback signals of the OS, the control device of the control unit affects the RE.
8. CNN - continuous voltage stabilizer; simply "analogue".
9. ISN - switching voltage stabilizer.
10. UPS - switching power supply.

Note: both CNN and ISN can work both from power frequency PSU with a transformer on iron, and from IIN.

About computer power supplies

UPSs are compact and economical. And in the pantry, many have a power supply from an old computer lying around, obsolete, but quite serviceable. So is it possible to adapt a switching power supply from a computer for amateur / work purposes? Unfortunately, a computer UPS is a fairly highly specialized device and the possibilities of its use in everyday life / at work are very limited:

It is advisable for an ordinary amateur to use a UPS converted from a computer one, perhaps, only to power a power tool; see below for more on this. The second case is if an amateur is engaged in repairing a PC and / or creating logic circuits. But then he already knows how to adapt the PSU from the computer for this:

1. Load the main channels + 5V and + 12V (red and yellow wires) with nichrome spirals for 10-15% of the rated load;
2. Green soft start wire (with a low-voltage button on the front panel of the system unit) pc on short to common, i.e. on any of the black wires;
3. On / off to produce mechanically, a toggle switch on the rear panel of the PSU;
4. With a mechanical (iron) I / O "duty room", i.e. the independent +5V USB power supply will also be turned off.

For business!

Due to the shortcomings of the UPS, plus their fundamental and circuitry complexity, we will only at the end consider a couple of these, but simple and useful, and talk about the method of repairing IIN. The main part of the material is devoted to SNN and PSN with industrial frequency transformers. They allow a person who has just picked up a soldering iron to build a very high quality PSU. And having it on the farm, it will be easier to master the “thinner” technique.

IPN

Let's look at the PPI first. We will leave the impulse ones in more detail until the section on repair, but they have something in common with the “iron” ones: a power transformer, a rectifier and a ripple suppression filter. Together, they can be implemented in various ways according to the purpose of the PSU.

Pos. 1 in Fig. 1 - half-wave (1P) rectifier. The voltage drop across the diode is the smallest, approx. 2B. But the ripple of the rectified voltage is with a frequency of 50 Hz and is “torn”, i.e. with gaps between pulses, so the ripple filter capacitor Cf must be 4-6 times larger than in other circuits. The use of a power transformer Tr in terms of power is 50%, because only 1 half-wave is straightened. For the same reason, a magnetic flux distortion occurs in the Tr magnetic circuit and the network “sees” it not as an active load, but as an inductance. Therefore, 1P rectifiers are used only for low power and where it is impossible to do otherwise, for example. in IIN on blocking generators and with a damper diode, see below.

Note: why 2V, and not 0.7V, at which the p-n junction opens in silicon? The reason is through current, which is discussed below.

Pos. 2 - 2-half-wave with a midpoint (2PS). Diode losses are the same as before. case. The ripple is 100 Hz continuous, so SF is the smallest possible. Use Tr - 100% Disadvantage - double the consumption of copper in the secondary winding. At a time when rectifiers were made on kenotron lamps, this did not matter, but now it is decisive. Therefore, 2PS is used in low-voltage rectifiers, mainly at increased frequency with Schottky diodes in UPS, but 2PS have no fundamental power limitations.

Pos. 3 - 2-half-wave bridge, 2PM. Losses on diodes - doubled compared to pos. 1 and 2. The rest is the same as for 2PS, but almost half as much copper is needed for the secondary. Almost - because several turns have to be winded to compensate for the losses on a pair of "extra" diodes. The most common circuit for voltage from 12V.

Pos. 3 - bipolar. The "bridge" is depicted conditionally, as is customary in circuit diagrams(get used to it!), and rotated 90 degrees counterclockwise, but in fact this is a pair of 2PS switched on in opposite directions, as can be clearly seen further in fig. 6. Copper consumption as in 2PS, diode losses as in 2PM, the rest as in both. It is built mainly to power analog devices that require voltage symmetry: Hi-Fi UMZCH, DAC / ADC, etc.

Pos. 4 - bipolar according to the scheme of parallel doubling. Gives without additional measures increased voltage symmetry, because the asymmetry of the secondary winding is excluded. Using Tr 100%, ripple 100 Hz, but torn, so SF needs double the capacity. The losses on the diodes are approximately 2.7 V due to the mutual exchange of through currents, see below, and at a power of more than 15-20 W they increase sharply. They are built mainly as low-power auxiliary for independent power supply. operational amplifiers(OU) and other low-power, but demanding on the quality of the power supply of analog nodes.

How to choose a transformer?

In a UPS, the entire circuit is most often clearly tied to the size (more precisely, to the volume and cross-sectional area Sc) of the transformer / transformers, since the use of fine processes in ferrite makes it possible to simplify the circuit with greater reliability. Here, "somehow in your own way" comes down to strict adherence to the recommendations of the developer.

The iron-based transformer is selected taking into account the characteristics of the CNN, or is consistent with them when calculating it. The voltage drop across the RE Ure should not be taken less than 3V, otherwise the KSN will drop sharply. With an increase in Ure, the KSN increases somewhat, but the dissipated RE power grows much faster. Therefore, Ure take 4-6 V. To it we add 2 (4) V losses on the diodes and the voltage drop on the secondary winding Tr U2; for a power range of 30-100 W and voltages of 12-60 V, we take it 2.5V. U2 occurs mainly not on the ohmic resistance of the winding (it is generally negligible for powerful transformers), but due to losses due to remagnetization of the core and the creation of a stray field. Simply, part of the energy of the network, "pumped" by the primary winding into the magnetic circuit, escapes into the world space, which takes into account the value of U2.

So, we counted, for example, for a bridge rectifier, 4 + 4 + 2.5 \u003d 10.5V in excess. We add it to the required output voltage of the PSU; let it be 12V, and divide by 1.414, we get 22.5 / 1.414 \u003d 15.9 or 16V, this will be the smallest allowable voltage of the secondary winding. If Tr is factory, from type series take 18v.

Now the secondary current comes into play, which, of course, is equal to the maximum load current. Let us need 3A; multiply by 18V, it will be 54W. We got the overall power Tr, Pg, and we will find the passport P by dividing Pg by the efficiency Tr η, depending on Pg:

• up to 10W, η = 0.6.
• 10-20 W, η = 0.7.
• 20-40 W, η = 0.75.
• 40-60 W, η = 0.8.
• 60-80 W, η = 0.85.
• 80-120 W, η = 0.9.
• from 120 W, η = 0.95.

In our case, it will be P \u003d 54 / 0.8 \u003d 67.5W, but there is no such typical value, so we have to take 80W. In order to get 12Vx3A = 36W at the output. Steam locomotive, and only. It's time to learn how to count and wind "trances" yourself. Moreover, in the USSR methods for calculating transformers on iron were developed, which made it possible to squeeze 600W out of the core without loss of reliability, which, when calculated according to amateur radio reference books, is capable of producing only 250W. "Iron Trance" is not at all as stupid as it seems.

SNN

The rectified voltage needs to be stabilized and, most often, regulated. If the load is more powerful than 30-40 W, protection against short circuit is also necessary, otherwise a PSU malfunction can cause a network failure. All this together makes SNN.

simple support

It’s better for a beginner not to go into high powers right away, but to make a simple highly stable CNN for 12V for testing according to the circuit in Fig. 2. It can then be used as a source of reference voltage (its exact value is set to R5), for checking instruments or as a high-quality CNN ION. The maximum load current of this circuit is only 40mA, but the KSN on the antediluvian GT403 and the same ancient K140UD1 is more than 1000, and when replacing VT1 with medium-power silicon and DA1 on any of the modern op-amps, it will exceed 2000 and even 2500. The load current will also increase to 150 -200 mA, which is already good for business.

0-30

The next step is a voltage regulated power supply. The previous one was made according to the so-called. compensatory comparison circuit, but it is difficult to convert this to a large current. We will make a new CNN based on an emitter follower (EF), in which RE and CU are combined in just 1 transistor. KSN will be released somewhere around 80-150, but this is enough for an amateur. But the CNN on the EP allows you to get an output current of up to 10A or more without any special tricks, how much Tr will give and withstand the RE.

A diagram of a simple power supply unit for 0-30V is shown in pos. 1 Fig. 3. PPN for it is a ready-made transformer of the TPP or TS type for 40-60 W with a secondary winding for 2x24V. Rectifier type 2PS on diodes of 3-5A or more (KD202, KD213, D242, etc.). VT1 is installed on a radiator with an area of ​​50 sq. cm; the old one from the PC processor is very well suited. Under such conditions, this CNN is not afraid of a short circuit, only VT1 and Tr will heat up, so a 0.5A fuse in the Tr primary winding circuit is enough for protection.

Pos. 2 shows how convenient it is for an amateur CNN on an electric power supply: there is a power supply circuit for 5A with adjustment from 12 to 36 V. This power supply unit can deliver 10A to the load if there is Tr at 400W 36V. Its first feature - the integrated CNN K142EN8 (preferably with the index B) acts in an unusual role of UU: to its own 12V at the output, all 24V is added, partially or completely, the voltage from the ION to R1, R2, VD5, VD6. Capacitances C2 and C3 prevent excitation on the RF DA1, operating in an unusual mode.

The next point is the protection device (UZ) against short circuit on R3, VT2, R4. If the voltage drop across R4 exceeds approximately 0.7V, VT2 will open, close the base circuit VT1 to a common wire, it will close and disconnect the load from the voltage. R3 is needed so that the extra current does not disable DA1 when the ultrasound is triggered. It is not necessary to increase its face value, because. when the ultrasound is triggered, VT1 must be securely locked.

And the last - the apparent excess capacitance of the output filter capacitor C4. In this case, it is safe, because. the maximum collector current VT1 of 25A ensures its charge when turned on. But on the other hand, this CNN can deliver current up to 30A to the load within 50-70 ms, so this simple power supply is suitable for powering low-voltage power tools: its starting current does not exceed this value. You just need to make (at least from plexiglass) a contact shoe with a cable, put on the heel of the handle, and let the "akumych" rest and save the resource before leaving.

About cooling

Let's say in this circuit the output is 12V with a maximum of 5A. This is just the average power of a jigsaw, but, unlike a drill or screwdriver, it takes it all the time. About 45V is kept on C1, i.e. on RE VT1 remains somewhere 33V at a current of 5A. The dissipated power is more than 150W, even more than 160W, given that VD1-VD4 also needs to be cooled. From this it is clear that any powerful regulated PSU must be equipped with a very efficient cooling system.

A ribbed/needle radiator on natural convection does not solve the problem: the calculation shows that a scatter surface of 2000 sq. see also the thickness of the radiator body (the plate from which the ribs or needles extend) from 16 mm. To get so much aluminum in a shaped product as a property for an amateur was and remains a dream in a crystal castle. A blown CPU cooler is also not suitable, it is designed for less power.

One of the options for a home master is an aluminum plate with a thickness of 6 mm or more and dimensions of 150x250 mm with holes of increasing diameter drilled along the radii from the installation site of the cooled element in a checkerboard pattern. It will also serve as the rear wall of the PSU case, as in Fig. 4.

An indispensable condition for the effectiveness of such a cooler is, albeit a weak, but continuous flow of air through the perforation from the outside to the inside. To do this, a low-power exhaust fan is installed in the case (preferably at the top). A computer with a diameter of 76 mm or more is suitable, for example. add. cooler HDD or video card. It is connected to pins 2 and 8 of DA1, there is always 12V.

Note: in fact, a radical way to overcome this problem is the secondary winding Tr with taps for 18, 27 and 36V. The primary voltage is switched depending on which tool is in operation.

And yet UPS

The described PSU for the workshop is good and very reliable, but it’s hard to carry it with you to the exit. This is where a computer PSU will come in handy: the power tool is insensitive to most of its shortcomings. Some refinement comes down most often to installing an output (closest to the load) high-capacity electrolytic capacitor for the purpose described above. There are many recipes for converting computer power supplies to power tools (mainly screwdrivers, as they are not very powerful, but very useful) in Runet, one of the methods is shown in the video below, for a 12V tool.

Video: PSU 12V from a computer

With 18V tools it's even easier: with the same power, they consume less current. Here, a much more affordable ignition device (ballast) from an economy lamp of 40 or more W can come in handy; it can be completely placed in the case from the unusable battery, and only the cable with the power plug will remain outside. How to make a power supply for an 18V screwdriver from ballast from a burnt housekeeper, see the following video.

Video: PSU 18V for a screwdriver

high class

But let's get back to the SNN on the EP, their possibilities are far from being exhausted. On Fig. 5 - bipolar powerful power supply with 0-30 V regulation, suitable for Hi-Fi audio equipment and other fastidious consumers. Setting the output voltage is done with one knob (R8), and the symmetry of the channels is maintained automatically at any value and any load current. A pedant-formalist at the sight of this scheme may turn gray before his eyes, but such a BP has been working properly for the author for about 30 years.

The main stumbling block in its creation was δr = δu/δi, where δu and δi are small instantaneous voltage and current increments, respectively. For the development and adjustment of high-end equipment, it is necessary that δr does not exceed 0.05-0.07 Ohm. Simply put, δr determines the ability of the PSU to instantly respond to surges in current consumption.

For SNN on the EP, δr is equal to that of the ION, i.e. zener diode divided by the current transfer coefficient β RE. But for powerful transistors β on a large collector current drops strongly, and δr of the zener diode ranges from units to tens of ohms. Here, in order to compensate for the voltage drop across the RE and to reduce the temperature drift of the output voltage, I had to dial their whole chain in half with diodes: VD8-VD10. Therefore, the reference voltage from the ION is removed through an additional EP on VT1, its β is multiplied by β RE.

The next feature of this design is short circuit protection. The simplest one described above does not fit into the bipolar scheme in any way, therefore the protection problem is solved according to the principle “no reception against scrap”: there is no protective module as such, but there is a redundancy in the parameters of powerful elements - KT825 and KT827 for 25A and KD2997A for 30A. T2 is not able to give such a current, but while it warms up, FU1 and / or FU2 will have time to burn out.

Note: it is not necessary to make a blown fuse indication on miniature incandescent lamps. It's just that then the LEDs were still quite scarce, and there were several handfuls of SMok in the stash.

It remains to protect the RE from the extra currents of the discharge of the ripple filter C3, C4 during short circuit. To do this, they are connected through limiting resistors of low resistance. In this case, pulsations with a period equal to the time constant R(3,4)C(3,4) may occur in the circuit. They are prevented by C5, C6 of smaller capacity. Their extra currents are no longer dangerous for RE: the charge will drain faster than the crystals of powerful KT825/827 will warm up.

Output symmetry provides op amp DA1. The RE of the negative channel VT2 opens with a current through R6. As soon as the minus of the output exceeds the plus in modulo, it will slightly open VT3, and it will close VT2 and the absolute values ​​of the output voltages will be equal. Operational control over the output symmetry is carried out using a pointer device with zero in the middle of the P1 scale (in the inset - its appearance), and adjustment if necessary - R11.

The last highlight is the output filter C9-C12, L1, L2. Such its construction is necessary to absorb possible RF pickups from the load, so as not to rack your brains: the prototype is buggy or the power supply unit is “bogged down”. With some electrolytic capacitors shunted with ceramics, there is no complete certainty here, the large intrinsic inductance of the “electrolytes” interferes. And the chokes L1, L2 share the "return" of the load over the spectrum, and - to each his own.

This PSU, unlike the previous ones, requires some adjustment:

1. Connect the load to 1-2 A at 30V;
2. R8 is set to the maximum, to the highest position according to the scheme;
3. Using a reference voltmeter (any digital multimeter will do now) and R11, the channel voltages are set equal in absolute value. Maybe, if the op-amp is without the possibility of balancing, you will have to choose R10 or R12;
4. Trimmer R14 set P1 exactly to zero.

About PSU repair

PSUs fail more often than other electronic devices: they take the first hit of network surges, they get a lot of things from the load. Even if you do not intend to make your own PSU, there is a UPS, except for a computer, in a microwave, washing machine, and other household appliances. The ability to diagnose a power supply unit and knowledge of the basics of electrical safety will make it possible, if not to fix the malfunction yourself, then with knowledge of the matter to bargain for a price with repairmen. Therefore, let's see how the PSU is diagnosed and repaired, especially with IIN, because over 80% of failures are accounted for by them.

Saturation and draft

First of all, about some effects, without understanding which it is impossible to work with the UPS. The first of these is the saturation of ferromagnets. They are not able to accept energies of more than a certain value, depending on the properties of the material. On iron, amateurs rarely encounter saturation, it can be magnetized up to several T (Tesla, a unit of measurement of magnetic induction). When calculating iron transformers, induction is taken 0.7-1.7 T. Ferrites can withstand only 0.15-0.35 T, their hysteresis loop is “rectangular”, and operate at higher frequencies, so the probability of “jumping into saturation” is orders of magnitude higher.

If the magnetic circuit is saturated, the induction in it no longer grows and the EMF of the secondary windings disappears, even if the primary has already melted (remember school physics?). Now turn off the primary current. The magnetic field in soft magnetic materials (hard magnetic materials are permanent magnets) cannot exist stationary, like an electric charge or water in a tank. It will begin to dissipate, the induction will fall, and an EMF of the opposite relative to the original polarity will be induced in all windings. This effect is widely used in IIN.

Unlike saturation, the through current in semiconductor devices (simply - a draft) is definitely a harmful phenomenon. It arises due to the formation/absorption of space charges in the p and n regions; for bipolar transistors - mainly in the base. Field-effect transistors and Schottky diodes are practically free from draft.

For example, when applying / removing voltage to the diode, until the charges are collected / resolved, it conducts current in both directions. That is why the voltage loss on the diodes in the rectifiers is greater than 0.7V: at the moment of switching, part of the charge of the filter capacitor has time to drain through the winding. In a parallel doubling rectifier, the draft flows through both diodes at once.

A draft of transistors causes a voltage surge on the collector, which can damage the device or, if a load is connected, damage it with a through extra current. But even without that, a transistor draft increases dynamic energy losses, like a diode one, and reduces Device efficiency. Powerful FETs they are almost not subject to it, tk. do not accumulate charge in the base in its absence, and therefore switch very quickly and smoothly. “Almost”, because their source-gate circuits are protected from reverse voltage by Schottky diodes, which are a little, but see through.

Types of TIN

UPSs are descended from a blocking generator, pos. 1 in Fig. 6. When Uin is turned on, VT1 is ajar by current through Rb, current flows through the winding Wk. It cannot instantly grow to the limit (again, we recall school physics), an EMF is induced in the base Wb and the load winding Wn. With Wb, it forces the unlocking of VT1 through Sat. According to Wn, the current does not flow yet, does not let VD1.

When the magnetic circuit is saturated, the currents in Wb and Wn stop. Then, due to the dissipation (resorption) of energy, the induction drops, an EMF of opposite polarity is induced in the windings, and the reverse voltage Wb instantly locks (blocks) VT1, saving it from overheating and thermal breakdown. Therefore, such a scheme is called a blocking generator, or simply blocking. Rk and Sk cut off high-frequency interference, which blocking gives more than enough. Now you can remove some useful power from Wn, but only through the 1P rectifier. This phase continues until the Sb is completely recharged or until the stored magnetic energy runs out.

This power, however, is small, up to 10W. If you try to take more, VT1 will burn out from the strongest draft before blocking. Since Tr is saturated, the blocking efficiency is no good: more than half of the energy stored in the magnetic circuit flies away to heat other worlds. True, due to the same saturation, blocking to some extent stabilizes the duration and amplitude of its impulses, and its scheme is very simple. Therefore, blocking-based TIN is often used in cheap phone chargers.

Note: the value of Sat largely, but not completely, as they say in amateur reference books, determines the pulse repetition period. The value of its capacitance should be linked to the properties and dimensions of the magnetic circuit and the speed of the transistor.

Blocking at one time gave rise to a line scan of televisions with cathode ray tubes (CRT), and she is a TIN with a damper diode, pos. 2. Here, the CU, based on signals from Wb and the DSP feedback circuit, forcibly opens / closes VT1 before Tr is saturated. With VT1 locked reverse current Wk closes through the same damper diode VD1. This is the working phase: already more than in blocking, part of the energy is removed into the load. Large because at full saturation all excess energy flies away, but here this is not enough. In this way, it is possible to remove power up to several tens of watts. However, since the CU cannot operate until Tp approaches saturation, the transistor still draws heavily, the dynamic losses are high, and the efficiency of the circuit leaves much to be desired.

IIN with a damper is still alive in TVs and CRT displays, since IIN and line scan output are combined in them: a powerful transistor and Tr are common. This greatly reduces production costs. But, frankly, IIN with a damper is fundamentally stunted: the transistor and transformer are forced to work all the time on the verge of an accident. The engineers who managed to bring this circuit to acceptable reliability deserve the deepest respect, but no one can stick a soldering iron there, except for craftsmen who have passed vocational training and those with relevant experience are strongly discouraged.

Push-pull INN with a separate feedback transformer is most widely used, because. has the best quality and reliability. However, in terms of high-frequency interference, it sins terribly compared to the “analog” power supplies (with transformers on iron and CNN). Currently, this scheme exists in many modifications; powerful bipolar transistors in it are almost completely replaced by field, controlled special. IC, but the principle of operation remains unchanged. It is illustrated by the original scheme, pos. 3.

The limiting device (UO) limits the charge current of the input filter capacitances Cfin1(2). Their large value is an indispensable condition for the operation of the device, because. in one working cycle, a small fraction of the stored energy is taken from them. Roughly speaking, they play the role of a water tank or an air receiver. When charging "short" charge extra current can exceed 100A for up to 100 ms. Rc1 and Rc2 with a resistance of the order of MΩ are needed to balance the filter voltage, because the slightest imbalance of his shoulders is unacceptable.

When Sfvh1 (2) is charged, the ultrasonic launcher generates a triggering pulse that opens one of the arms (which one does not matter) of the inverter VT1 VT2. A current flows through the winding Wk of a large power transformer Tr2 and the magnetic energy from its core through the winding Wn almost completely goes to rectification and to the load.

A small part of the energy Tr2, determined by the value of Rolimit, is taken from the winding Wos1 and fed to the winding Wos2 of a small basic feedback transformer Tr1. It quickly saturates, the open shoulder closes, and due to dissipation in Tr2, the previously closed shoulder opens, as described for blocking, and the cycle repeats.

In essence, a two-stroke IIN is 2 blockings, “pushing” each other. Since the powerful Tr2 is not saturated, the draft VT1 VT2 is small, completely "sinks" in the Tr2 magnetic circuit and eventually goes into the load. Therefore, a two-stroke IMS can be built for a power of up to several kW.

Worse, if he is in XX mode. Then, during the half-cycle, Tr2 will have time to saturate and the strongest draft will burn both VT1 and VT2 at once. However, now there are power ferrites for induction up to 0.6 T on sale, but they are expensive and degrade from accidental magnetization reversal. Ferrites are being developed for more than 1 T, but in order for the IIN to reach "iron" reliability, at least 2.5 T is needed.

Diagnosis technique

When troubleshooting in an “analog” PSU, if it is “stupidly silent”, they first check the fuses, then the protection, RE and ION, if it has transistors. They ring normally - we go further element by element, as described below.

In the IIN, if it “starts up” and immediately “stalls”, they first check the UO. The current in it is limited by a powerful low resistance resistor, then shunted by an optothyristor. If the “rezik” is apparently burnt out, the optocoupler is also changed. Other elements of the UO fail extremely rarely.

If the IIN is “silent, like a fish on ice”, the diagnostics are also started with the UO (maybe the “rezik” has completely burned out). Then - UZ. In cheap models, they use transistors in the avalanche breakdown mode, which is far from very reliable.

The next step in any PSU is electrolytes. The destruction of the case and the leakage of the electrolyte are not as common as they say in Runet, but the loss of capacity happens much more often than the failure of active elements. Check electrolytic capacitors with a multimeter with the ability to measure capacitance. Below the face value by 20% or more - we lower the “dead man” into the sludge and put a new, good one.

Then there are active elements. You probably know how to ring diodes and transistors. But there are 2 tricks here. The first is that if a Schottky diode or a zener diode is called by a tester with a 12V battery, then the device may show a breakdown, although the diode is quite good. It is better to call these components with a dial gauge with a 1.5-3 V battery.

The second is powerful field workers. Above (did you notice?) It is said that their I-Z are protected by diodes. Therefore, powerful field-effect transistors seem to ring like serviceable bipolar ones, even unusable, if the channel is not completely “burned out” (degraded).

Here, the only way available at home is to replace them with known-good ones, and both at once. If a burnt one remains in the circuit, it will immediately pull a new serviceable one with it. Electronic engineers joke that powerful field workers cannot live without each other. Another prof. joke - "replacing a gay couple." This is due to the fact that the transistors of the IIN shoulders must be strictly of the same type.

Finally, film and ceramic capacitors. They are characterized by internal breaks (located by the same tester with checking the “air conditioners”) and leakage or breakdown under voltage. To “catch” them, you need to assemble a simple shemka according to Fig. 7. Step-by-step check of electrical capacitors for breakdown and leakage is carried out as follows:

• We put on the tester, without connecting it anywhere, the smallest limit for measuring direct voltage (most often - 0.2V or 200mV), detect and record the instrument's own error;
• We turn on the measurement limit of 20V;
• We connect a suspicious capacitor to points 3-4, the tester to 5-6, and to 1-2 we apply a constant voltage of 24-48 V;
• We switch the voltage limits of the multimeter down to the smallest;
• If on any tester it showed at least something other than 0000.00 (at the smallest - something other than its own error), the capacitor being tested is not good.

This is where the methodological part of the diagnostics ends and the creative part begins, where all the instructions are your own knowledge, experience and consideration.

Pair of impulses

UPS article is special, due to their complexity and circuit diversity. Here, we will first look at a couple of samples on pulse-width modulation (PWM), which allows you to get best quality UPS. There are many schemes for PWM in RuNet, but PWM is not as terrible as it is painted ...

For lighting design

You can simply light the LED strip from any PSU described above, except for the one in Fig. 1 by setting the required voltage. Well suited SNN with pos. 1 Fig. 3, these are easy to make 3, for channels R, G and B. But the durability and stability of the glow of LEDs do not depend on the voltage applied to them, but on the current flowing through them. Therefore, a good power supply for an LED strip should include a load current stabilizer; technically - a stable current source (IST).

One of the schemes for stabilizing the current of a light tape, available for repetition by amateurs, is shown in Fig. 8. It was assembled on an integral timer 555 (domestic analogue - K1006VI1). Provides a stable tape current from a power supply unit with a voltage of 9-15 V. The value of a stable current is determined by the formula I = 1 / (2R6); in this case - 0.7A. Power transistor VT3 - necessarily field, from a draft due to the charge of the base of the bipolar PWM simply will not form. The inductor L1 is wound on a ferrite ring 2000NM K20x4x6 with a 5xPE 0.2 mm bundle. Number of turns - 50. Diodes VD1, VD2 - any silicon RF (KD104, KD106); VT1 and VT2 - KT3107 or analogues. With KT361 etc. input voltage and dimming ranges will decrease.

The circuit works like this: first, the time-setting capacitance C1 is charged through the R1VD1 circuit and discharged through VD2R3VT2, open, i.e. in saturation mode, through R1R5. The timer generates a sequence of pulses with a maximum frequency; more precisely - with a minimum duty cycle. The VT3 inertialess key generates powerful pulses, and its VD3C4C3L1 strapping smoothes them to DC.

Note: the duty cycle of a series of pulses is the ratio of their repetition period to the pulse duration. If, for example, the pulse duration is 10 µs, and the gap between them is 100 µs, then the duty cycle will be 11.

The current in the load increases, and the voltage drop across R6 slightly opens VT1, i.e. switches it from the cut-off (locking) mode to the active (amplifying) mode. This creates a base current leakage circuit VT2 R2VT1 + Upit and VT2 also goes into active mode. The discharge current C1 decreases, the discharge time increases, the duty cycle of the series increases and the average current value drops to the norm specified by R6. This is the essence of PWM. At the current minimum, i.e. at maximum duty cycle, C1 is discharged through the VD2-R4 circuit - the internal timer key.

In the original design, the ability to quickly adjust the current and, accordingly, the brightness of the glow, is not provided; There are no 0.68 ohm potentiometers. The easiest way to adjust the brightness is to turn on the gap between R3 and the emitter VT2 potentiometer R * 3.3-10 kOhm after adjustment, highlighted in brown. By moving its slider down the circuit, we will increase the discharge time of C4, the duty cycle and reduce the current. Another way is to shunt the base transition VT2 by turning on the potentiometer by about 1 MΩ at points a and b (highlighted in red), less preferable, because. the adjustment will be deeper, but coarse and sharp.

Unfortunately, an oscilloscope is needed to establish this useful not only for ICT light tapes:

1. The minimum + Upit is applied to the circuit.
2. By selecting R1 (pulse) and R3 (pause), a duty cycle of 2 is achieved, i.e. the duration of the pulse must be equal to the duration of the pause. It is impossible to give a duty cycle less than 2!
3. Serve maximum + Upit.
4. By selecting R4, the nominal value of the stable current is achieved.

For charging

On Fig. 9 - a diagram of the simplest ISN with PWM, suitable for charging a phone, smartphone, tablet (a laptop, unfortunately, will not pull) from a home-made solar battery, wind generator, motorcycle or car battery, magneto flashlight-"bug" and other low-power unstable random power supplies. See the input voltage range on the diagram, it's not an error. This ISN is indeed capable of outputting a voltage greater than the input. As in the previous one, there is an effect of changing the polarity of the output relative to the input, this is generally a proprietary feature of PWM circuits. Let's hope that, after reading carefully the previous one, you will understand the work of this tiny little one yourself.

Along the way about charging and charging

Charging batteries is a very complex and delicate physical and chemical process, the violation of which at times and tens of times reduces their resource, i.e. number of charge-discharge cycles. The charger must, by very small changes in the battery voltage, calculate how much energy is received and regulate the charge current accordingly according to a certain law. That's why Charger by no means and by no means a PSU, and only batteries in devices with a built-in charge controller can be charged from conventional PSUs: phones, smartphones, tablets, and certain models of digital cameras. And charging, which is a charger, is the subject of a separate discussion.

Question-remont.ru said:

There will be sparks from the rectifier, but it's probably nothing to worry about. The point is the so-called. differential output impedance of the power supply. For alkaline batteries, it is of the order of mOhm (milliohm), for acid batteries it is even less. A trance with a bridge without smoothing has tenths and hundredths of an ohm, i.e. approx. 100 - 10 times more. And the starting current of a DC collector motor can be 6-7 or even 20 times more than the working one. Yours, most likely, is closer to the latter - fast accelerating motors are more compact and economical, and the huge overload capacity of the batteries allows you to give the engine current, how much it will eat for acceleration. A trans with a rectifier will not give as much instantaneous current, and the engine accelerates more slowly than it is designed for, and with a large armature slip. From this, from a large slip, a spark arises, and then it is kept in operation due to self-induction in the windings.

What can be advised here? First: take a closer look - how does it sparkle? You need to look at work, under load, i.e. during sawing.

If sparks dance in separate places under the brushes, it's okay. I have a powerful Konakovo drill that sparks so much from birth, and at least henna. For 24 years, I changed brushes once, washed with alcohol and polished the collector - just something. If you have connected an 18V tool to the 24V output, then some sparking is normal. Unwind the winding or extinguish the excess voltage with something like a welding rheostat (resistor approx. 0.2 Ohm for a dissipation power of 200 W) so that the motor has the rated voltage in operation and, most likely, the spark will go away. If, however, they connected to 12 V, hoping that after rectification it would be 18, then in vain - the rectified voltage under load drops a lot. And the collector electric motor, by the way, does not care whether it is powered by direct current or alternating current.

Specifically: take 3-5 m of steel wire with a diameter of 2.5-3 mm. Roll into a spiral with a diameter of 100-200 mm so that the turns do not touch each other. Lay on a non-flammable dielectric pad. Strip the ends of the wire to a shine and roll up the “ears”. It is best to immediately lubricate with graphite grease so that they do not oxidize. This rheostat is included in the break of one of the wires leading to the tool. It goes without saying that the contacts must be screw, tightened tightly, with washers. Connect the entire circuit to the 24V output without rectification. The spark is gone, but the power on the shaft has also dropped - the rheostat needs to be reduced, one of the contacts must be switched 1-2 turns closer to the other. It still sparks, but less - the rheostat is too small, you need to add turns. It is better to immediately make the rheostat obviously large so as not to screw additional sections. Worse, if the fire is along the entire line of contact of the brushes with the collector, or spark tails trail behind them. Then the rectifier needs a smoothing filter somewhere, according to your data, from 100,000 microfarads. Cheap pleasure. The “filter” in this case will be an energy storage device for engine acceleration. But it may not help - if the overall power of the transformer is not enough. Efficiency of DC collector motors approx. 0.55-0.65, i.e. trance is needed from 800-900 watts. That is, if the filter is installed, but still sparks with fire under the entire brush (under both, of course), then the transformer does not hold out. Yes, if you put a filter, then the bridge diodes must also be at triple operating current, otherwise they can fly out from the charge current surge when connected to the network. And then the tool can then be launched 5-10 seconds after being connected to the network, so that the “banks” have time to “pump up”.

And worst of all, if the tails of sparks from the brushes reach or almost reach the opposite brush. This is called a round fire. It very quickly burns out the collector to complete disrepair. There can be several reasons for the round fire. In your case, the most likely is that the motor was turned on at 12 V with rectification. Then, at a current of 30 A, the electric power in the circuit is 360 watts. Anchor slip is more than 30 degrees per revolution, and this is necessarily a continuous all-round fire. It is also possible that the motor armature is wound with a simple (not double) wave. Such electric motors are better able to overcome instantaneous overloads, but their starting current is mother, don’t worry. I can’t say more precisely in absentia, and I don’t need anything - it’s hardly possible to fix anything with my own hands. Then, probably, it will be cheaper and easier to find and purchase new batteries. But first, nevertheless, try to turn on the engine at a slightly increased voltage through a rheostat (see above). Almost always, in this way, it is possible to bring down a continuous all-round fire at the cost of a small (up to 10-15%) decrease in the power on the shaft.

Hello. I bring to your attention a review of the integrated linear adjustable voltage (or current) stabilizer LM317 at a price of 18 cents apiece. In a local store, such a stabilizer costs an order of magnitude more, which is why I was interested in this lot. I decided to check what is sold at such a price and it turned out that the stabilizer is quite high quality, but more on that below.
In the review, testing in the mode of a voltage and current stabilizer, as well as checking protection against overheating.
Interested please...

A little theory:

Review hero:

Testing:
Similar stabilizers are produced by many manufacturers, here.
The location of the legs is as follows:
1 - adjustment;
2 - exit;
3 - entrance.
We collect the simplest voltage stabilizer according to the scheme from the manual:


Here is what we managed to get with 3 positions of the variable resistor:
The results, frankly speaking, are not very good. It does not turn out to be called a stabilizer.
Next, I loaded the stabilizer with a 25 Ohm resistor and the picture completely changed:

Next, I decided to check the dependence of the output voltage on the load current, for which I set the input voltage to 15V, set the output voltage to about 5V with a trimmer resistor, and loaded the output with a variable 100 Ohm wire resistor. Here's what happened:
It was not possible to obtain a current of more than 0.8A, because the input voltage began to drop (the PSU is weak). As a result of this testing, the stabilizer with a radiator heated up to 65 degrees:

To test the operation of the current stabilizer, the following circuit was assembled:


Instead of a variable resistor, I used a constant one, here are the test results:
Current stabilization is also good.
Well, how can a review be without burning the hero? To do this, I assembled the voltage stabilizer again, applied 15V to the input, set the output to 5V, i.e. 10V fell on the stabilizer, and loaded it by 0.8A, i.e. 8W of power was allocated on the stabilizer. Removed the radiator.
The result is shown in the following video:

Outcome:

On this, let me take my leave, good luck!

The product was provided for writing a review by the store. The review is published in accordance with clause 18 of the Site Rules.