How to repair and modify a Chinese-made 12-volt switching power supply

I want to start with the fact that I came into my hands with several burnt out and already “repaired” 220/12 V power supplies by someone. All the units were of the same type - HF55W-S-12, therefore, having entered the name in the search engine, I hoped to find a circuit . But apart from photographs of the appearance, parameters and prices for them, I found nothing. Therefore, I had to draw the circuit myself from the board. The diagram was drawn not to study the principle of operation of the power supply, but solely for repair purposes. Therefore, the network rectifier is not drawn, I also did not saw the pulse transformer and I do not know where the tap is made (start-end) on the 2nd winding of the transformer. Also, C14 -62 Ohm should not be considered a typo - there are markings on the board for an electrolytic capacitor (+ is shown in the diagram), but everywhere in its place there were resistors with a nominal value of 62 Ohms.

When repairing such devices, they need to be connected through a light bulb (100-200 W incandescent lamp, in series with the load), so that in the event of a short circuit in the load, the output transistor does not fail and the tracks on the board do not burn out. And your household will feel safer if the lights in the apartment don’t suddenly go out.
The main malfunction is the breakdown of Q1 (FJP5027 - 3 A, 800 V, 15 MHz) and, as a consequence, the breakage of resistors R9, R8 and the failure of Q2 (2SC2655 50 V\2 A 100 MHz). They are highlighted in color in the diagram. Q1 can be replaced with any transistor suitable for current and voltage. I installed BUT11, BU508. If the load power does not exceed 20 W, you can even install J1003, which can be found on the board from a burnt-out energy-saving lamp. One block was completely missing VD-01 (Schottky diode STPR1020CT -140 V\2x10 A), I installed MBR2545CT (45 V\30 A) instead, which is typical, it does not heat up at all at a load of 1.8 A (we used a 21 car lamp W\12 V). And within a minute of operation (without a radiator), the original diode heats up so much that it is impossible to touch it with your hand. I checked the current consumed by the device (with a 21 W lamp) with the original diode and with the MBR2545CT - the current (consumed from the network, I have a voltage of 230 V) dropped from 0.115 A to 0.11 A. The power decreased by 1.15 W, I believe that this is exactly how much was dissipated on the original diode.
There was nothing to replace Q2 with, so I found the C945 transistor at hand. I had to “power up” it with a circuit with a KT837 transistor (Figure 2). The current remained under control and when comparing the current with the native circuit on 2SC2655, there was an even reduction in power consumption with the same load at 1 W.

As a result, with a load of 21 W and when operating for 5 minutes, the output transistor and rectifier diode (without a radiator) heat up to 40 degrees (slightly warm). In the original version, after a minute of operation without a radiator, they could not be touched. The next step to increasing the reliability of blocks made according to this scheme is to replace the electrolytic capacitor C12 (prone to drying out of the electrolyte over time) with a conventional non-polar - non-electrolytic one. The same nominal value of 0.47 µF and a voltage of at least 50 V.
With such characteristics of the power supply, you can now safely connect LED strips without fear that the efficiency of the power supply will worsen the efficiency of LED lighting.

LEDs are replacing types of light sources such as fluorescent and incandescent lamps. Almost every home already has LED lamps; they consume much less than their two predecessors (up to 10 times less than incandescent lamps and 2 to 5 times less than CFLs or energy-saving fluorescent lamps). In situations where a long light source is needed, or it is necessary to organize illumination of a complex shape, it is used.

LED strip is ideal for a number of situations; its main advantage over individual LEDs and LED matrices is power supplies. They are easier to find for sale in almost any electrical goods store, unlike drivers for high-power LEDs, and besides, the selection of a power supply is carried out only by power consumption, because The vast majority of LED strips have a supply voltage of 12 Volts.

While for high-power LEDs and modules, when choosing a power source, you need to look for a current source with the required power and rated current, i.e. take into account 2 parameters, which complicates the selection.

This article discusses typical power supply circuits and their components, as well as tips for repairing them for novice radio amateurs and electricians.

Types and requirements for power supplies for LED strips and 12 V LED lamps

The main requirement for a power source for both LEDs and LED strips is high-quality voltage/current stabilization, regardless of mains voltage surges, as well as low output ripple.

Based on the type of design, power supplies for LED products are divided into:

Sealed. They are more difficult to repair; the body cannot always be carefully disassembled, and the inside may even be filled with sealant or compound.

Non-hermetic, for indoor use. Better amenable to repair, because... The board is removed after unscrewing several screws.

By type of cooling:

Passive air. The power supply is cooled due to natural air convection through the perforations of its case. Disadvantage is the inability to achieve high power while maintaining weight and size indicators;

Active air. The power supply is cooled using a cooler (a small fan, as installed on PC system units). This type of cooling allows you to achieve more power in the same size with a passive power supply.

Power supply circuits for LED strips

It is worth understanding that in electronics there is no such thing as a “power supply for an LED strip”; in principle, any power supply with a suitable voltage and a current greater than that consumed by the device will be suitable for any device. This means that the information described below applies to almost any power supply.

However, in everyday life it is easier to talk about a power supply according to its purpose for a specific device.

General structure of a switching power supply

Switching power supplies (UPS) have been used to power LED strips and other equipment for the last decades. They differ from transformer ones in that they operate not at the frequency of the supply voltage (50 Hz), but at high frequencies (tens and hundreds of kilohertz).

Therefore, for its operation, a high-frequency generator is needed; in cheap power supplies designed for low currents (units of amperes), a self-oscillator circuit is often found; it is used in:

electronic transformers;

electronic ballasts for fluorescent lamps;

mobile phone chargers;

cheap UPS for LED strips (10-20 W) and other devices.

A diagram of such a power supply can be seen in the figure (click on the picture to enlarge):

Its structure is as follows:

The OS includes an optocoupler U1, with its help the power part of the oscillator receives a signal from the output and maintains a stable output voltage. There may be no voltage in the output part due to a break in the VD8 diode, often this is a Schottky assembly and must be replaced. A swollen electrolytic capacitor C10 also often causes problems.

As you can see, everything works with a much smaller number of elements, the reliability is appropriate...

More expensive power supplies

The circuits that you will see below are often found in power supplies for LED strips, DVD players, radio tape recorders and other low-power devices (tens of watts).

Before moving on to considering popular circuits, familiarize yourself with the structure of a switching power supply with a PWM controller.

The upper part of the circuit is responsible for filtering, rectifying and smoothing ripples of the mains voltage 220, essentially similar to both the previous type and subsequent ones.

The most interesting thing is the PWM block, the heart of any decent power supply. A PWM controller is a device that controls the duty cycle of an output signal based on a user-defined setpoint or current or voltage feedback. PWM can control both load power using a field (bipolar, IGBT) switch, and a semiconductor controlled switch as part of a converter with a transformer or inductor.

By changing the width of the pulses at a given frequency, you also change the effective value of the voltage, while maintaining the amplitude, you can integrate it using C- and LC-circuits to eliminate ripple. This method is called Pulse Width Modeling, that is, modeling a signal using the pulse width (duty factor/duty factor) at a constant frequency.

In English it sounds like a PWM-controller, or Pulse-Width Modulation controller.

The figure shows bipolar PWM. Rectangular signals are control signals on transistors from the controller; the dotted line shows the shape of the voltage in the load of these switches - the effective voltage.

Higher-quality power supplies of low average power are often built on integrated PWM controllers with a built-in power switch. Advantages over self-oscillator circuit:

The operating frequency of the converter does not depend on either the load or the supply voltage;

Better stabilization of output parameters;

Possibility of simpler and more reliable adjustment of the operating frequency at the stage of design and modernization of the unit.

Below are several typical power supply circuits (click on the picture to enlarge):

Here RM6203 is both a controller and a key in one housing.

The same thing, but on a different chip.

Feedback is carried out using a resistor, sometimes an optocoupler connected to an input called Sense (sensor) or Feedback (feedback). Repair of such power supplies is generally similar. If all the elements are working properly, and the supply voltage is supplied to the microcircuit (Vdd or Vcc leg), then the problem is most likely in it, more accurately looking at the output signals (drain, gate leg).

Almost always, you can replace such a controller with any analogue with a similar structure; to do this, you need to check the datasheet against the one installed on the board and the one you have and solder it, observing the pinout, as shown in the following photographs.

Or here is a schematic representation of the replacement of such microcircuits.

Powerful and expensive power supplies

Power supplies for LED strips, as well as some power supplies for laptops, are made on the UC3842 PWM controller.

The scheme is more complex and reliable. The main power component is transistor Q2 and transformer. During repairs, you need to check the filtering electrolytic capacitors, the power switch, Schottky diodes in the output circuits and output LC filters, the supply voltage of the microcircuit, otherwise the diagnostic methods are similar.

However, more detailed and accurate diagnostics are only possible using an oscilloscope; otherwise, checking for short circuits on the board, soldering of elements and breaks will cost more. Replacing suspicious nodes with known working ones can help.

More advanced models of power supplies for LED strips are made on the almost legendary TL494 chip (any letters with the numbers “494”) or its analogue KA7500. By the way, most AT and ATX computer power supplies are built on these same controllers.

Here is a typical power supply diagram for this PWM controller (click on the diagram):

Such power supplies are highly reliable and stable.

Brief verification algorithm:

1. We power the microcircuit according to the pinout from an external power source of 12-15 volts (12 leg is plus, and 7 leg is minus).

2. A voltage of 5 Volts should appear on the 14 legs, which will remain stable when the power supply changes, if it “floats” - the microcircuit needs to be replaced.

3. There should be a sawtooth voltage at pin 5; you can “see” it only with the help of an oscilloscope. If it is not there or the shape is distorted, we check compliance with the nominal values ​​of the timing RC circuit, which is connected to pins 5 and 6; if not, in the diagram these are R39 and C35, they must be replaced; if nothing has changed after that, the microcircuit has failed.

4. There should be rectangular pulses at outputs 8 and 11, but they may not exist due to the specific feedback implementation circuit (pins 1-2 and 15-16). If you turn off and connect 220 V, they will appear there for a while and the unit will go into protection again - this is a sign of a working microcircuit.

5. You can check the PWM by short-circuiting the 4th and 7th legs, the pulse width will increase, and short-circuiting the 4th to 14th legs, the pulses will disappear. If you get different results, the problem is in MS.

This is the most brief test of this PWM controller; there is a whole book about repairing power supplies based on them, “Switching Power Supplies for IBM PC.”

Although it is dedicated to computer power supplies, there is a lot of useful information for any radio amateur.

Conclusion

The circuitry of power supplies for LED strips is similar to any power supplies with similar characteristics; they can be repaired, modernized, and adjusted to the required voltages quite well, of course, within reasonable limits.

In the past about power supplies, I mentioned that I ordered two units for review, today I will talk about the second test subject.
In its own way it is interesting and well made, but not without its shortcomings.
All more detailed information, as always, is under the cut.

I already had a power supply with the same power and the same voltage, but in this case these power supplies are radically different, which prompted me to take it for the test.
The review will be in the same format as always, but the comments and conclusions will be completely different.

I’ll start today in an unusual way, with packaging :)))
The power supply, like last time, had its own cardboard “house”. But this time there was a marking on the package - Led power supply, although it has nothing to do with powering the LEDs since it works as a source of voltage, not current, but in this case it does not matter much.
There is also a power marking on the side, and I immediately noticed that at first it was highlighted - 150 Watts, then crossed out and marked - 180 Watts, but we will return to this later.

The first distinguishing feature of this power supply is its form factor. The power supply is made on the basis of a U-shaped aluminum chassis that acts as a radiator; usually power supplies are made in the form of an L-shaped chassis with a perforated casing.
This design should improve the cooling of power elements and reduce the size of the block, but the heating test will come later.

The dimensions of the power supply are very modest, length 200mm, width 59mm, height 36mm.

At the ends of the block there are connectors for connecting 220 Volt power + grounding and 12 Volt output.
The output terminals are made double, with two contacts for each polarity.
This is caused by a rather large output current, up to 15 Amps; in this option it is more convenient to connect the load.

Each terminal block has a protective cover. In a previous review of the 180 Watt power supply, I was asked if the lid opened all the way, as a person had problems with it.
The lid, although it has rather tight latches, opens at 90 degrees.

The manufacturer claims the following characteristics:
Input voltage - 110/220 Volts ± 15% (which is strange since the power supply does not have a voltage switch)
Output voltage - 12 Volts
Output current - 15 Amperes.

Since there was nothing else interesting outside, I climbed inside.
The unit is extremely easy to disassemble; there are four screws on the sides, unscrewing which you can easily remove the top cover.
The first thing that caught my eye was that the power supply was assembled using single-cycle circuitry.
In my personal opinion, a power supply with a power of 180 Watts assembled according to this scheme is already on the border of good and evil.
The fact is that at low powers such a circuit works perfectly, but at high powers push-pull, bridge or half-bridge ones already “rule the roost” (this circuitry is used in most computer power supplies).
This BP is located approximately on the border of the division of “spheres of influence”.

The power supply survived the first power-up quite normally, which in itself is pleasant :)
Initially it was set to 12.21 Volts (only later did I understand why).
The adjustment range is not very large, minimum 11.75, maximum 12.63.
After checking the adjustment range, I set the power supply to the stated 12 Volts.

Several photos of the main components of the power supply.
1. Surge filter, this time there is a thermistor that protects against current surge when the power supply is turned on, there is a place for a protective varistor, but they “forgot” to solder it.
2. The input capacitor has a capacity of 150 μF, looks more like a proprietary one, and is designed for a maximum temperature of 105 degrees. If it weren’t for the reduced capacity, then I would say that it’s excellent, but otherwise it’s just good.
3. The high-voltage transistor is pressed using an L-shaped plate. there is a paste, and it looks like silicone.
4. Two diode assemblies are installed at the output, also pressed with a metal plate through the paste, but to the other wall of the housing.

Let's look further. The board is screwed onto one mounting screw, inserted into the slots of the housing itself, and can only be inserted and removed together with the dielectric insert.
You can see that the board is almost empty; only large elements are installed on top of it.
This is how branded power supplies are usually made (at least that’s what I remember).

PCB.

A couple of more detailed photos of the printed circuit board.
The secondary side, precise resistors are used, this is good, the wiring of the feedback circuit is also interesting, it is clear that they still thought about the routing. By the way, the power supply is made by the same manufacturer as the previous one for 24 Volts.

Primary side.

An unknown to me was used as a PWM controller.
But I noticed that the manufacturer placed a ceramic capacitor parallel to the electrolyte in the power circuit of this microcircuit. This happens quite rarely, but in vain.
The current measuring shunt is made in the form of six resistors connected in parallel.

The circuit diagram is slightly different from the previous power supply.
In the diagram, some positions have a designation of the form - 22 (11) and a serial number of the element consisting of several numbers. This means that several parallel elements are installed; the total value is given in brackets.

Individual photographs of the main components of the power supply.
1. Input power filter elements, noise suppression capacitor and inductor.
2. A thermistor to limit the starting current and a diode bridge, this time a 4 Ampere 600 Volt diode bridge.
3. Additional noise suppression capacitors, correct Y type.
4. High voltage transistor. The transistor is in an insulated housing, designed for current up to 12 Amperes and voltage up to 650 Volts. In my opinion, it could have been installed more powerfully, but the test showed that everything was fine with it.

1. The interwinding capacitor is also of the correct Y type, which is rare these days.
Next to it there is an empty space for installing the same capacitor, connecting the minus of the output circuit to the power supply housing, but it was also “forgotten”. I won’t say that it is very important, but it would not be superfluous.
2. Output diode assemblies, no questions, the parameters correspond to the output current and voltage of the power supply.
A few words about the transformer. Made correctly, it is clear that the primary winding is made of two wires and is divided into two parts (this is desirable to improve the connection between the windings). The output winding is made of four wires, although at such currents a Litz winding looks better.

The output capacitors are made up of five pieces. Before the choke, three pieces of 1000 μF for 25 Volts are installed, after which there are two pieces of 1000 μF for 16 Volts. I think it was worth installing all capacitors at 25 Volts at least. And ideally, before the choke, 35 Volts, after - 25 Volts, but this is rarely found even in branded power supplies.
The output choke has been detuned; the location allows you to install a choke with a higher inductance and designed for a higher current. I would recommend replacing it with a more suitable one.
A small measurement of the capacitance of the capacitors showed that the indicated and actual capacitance corresponded.

Well, actually, we’re done with the review of the design and elemental base, now we can safely move on to testing.
For this purpose, the same “stand” was assembled as in the previous review. It included:
Experimental power supply.
Electronic load
Oscilloscope
Multimeter
Non-contact thermometer

The testing methodology is almost standard.
Switch on, load, warm up for 20 minutes, increase load current, warm up for 20 minutes, etc. until we hit the maximum current, or until the power supply makes its last squeak.
The oscilloscope probe divider was in the 1:1 position, the oscilloscope division value was set to 0.1 Volt.
1. First check at idle speed, output voltage is 11.98 Volts.
2. Increasing the load current to 3 Amps, the voltage dropped sharply to 11.65 Volts.

After I saw that the output voltage dropped sharply under a relatively light load, I immediately remembered that it was originally set to 12.21 Volts.
Apparently the load resistors located at the output of the block do not quite cope with their function and the output voltage rises at idle.
I had to adjust the output voltage to 11.99 Volts at a current of 3 Amps.
I didn't touch the regulator anymore.

1. Load current 6 Amperes, voltage 12 Volts, ripples occur with a voltage of about 0.4 Volts
2. Load current 9 Amperes, voltage 11.92 Volts, the range of ripples has remained almost unchanged, but they have become more frequent.

1. Load current 12 Amperes, voltage 11.84 Volts, ripple voltage about 0.5 Volts
2. The load current is about 14 amperes (the load no longer provides), the voltage has dropped to 11.8 Volts, but the ripple has already increased quite significantly and amounted to 0.65 Volts.

As I wrote above, data on the temperature of the components was taken every 20 minutes.
The first value is idle after approximately 20-30 seconds of running under a current of 10 Amps (this is how it happened), the next ones were taken before the next increase in current.
The last value is an additional 20-minute warm-up to assess the dynamics of temperature growth. The total test time was 2 hours.
Temperatures measured:
High voltage transistor, transformer, output diodes, output capacitors.
The value with the higher temperature was taken as the temperature of the output diode (one assembly had a temperature several degrees higher).


At almost the maximum load current, the power supply noticeably overheats, so during operation you should not count on a current of more than 12 Amps.

At the end of the experiment, I took a picture of the heating of the entire power supply as a whole; unfortunately, I don’t have a thermal imager yet, so that’s the only way.

Resume.
Pros
Nice and well thought out design.
Having a power filter with the correct types of capacitors.
Most of the components are selected correctly and according to the power of the power supply.

Cons
High ripple level can be improved by replacing the output choke.
Large heating at maximum current, unfortunately, cannot be corrected by simple modifications.

My opinion. At the very beginning, I wrote that I would return to talk about the power of the power supply, which was originally indicated on the packaging. I believe that the initially indicated 150 Watt is the power at which this power supply can operate quite safely.
I was pleased with the good design, the presence of a full-fledged power filter, and the correct capacitors (affects safety). But the high temperature upset me, and while it is common for semiconductors, it is dangerous for the transformer and output capacitors.
The capacitor capacity, in my opinion, is somewhat underestimated and is also more suitable for a power of 150 watts rather than 180.
The total result is a completely normal, well-made power supply with a power of 144 Watts, or, in other words, 12 Volts 12 Amps.

I hope that the review was useful and will allow you to make the right choice.

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

I'm planning to buy +42 Add to favorites I liked the review +60 +114

Have you ever wanted to turn on the TV, stereo or other equipment when you are in the car or relaxing in nature? An inverter should solve this problem. It converts 12 V DC to 120 V AC. Depending on the power of the Q1 and Q2 transistors used, as well as how “big” transformer T1 is, the inverter can have an output power from 1 W to 1000 W.

Schematic diagram

List of elements

Element	Qty	Description
		Tantalum capacitors 68 µF, 25 V
		Resistors 10 Ohm, 5 W
		Resistors 180 Ohm, 1 W
		Silicon diodes HEP 154
		npn transistors 2N3055 (see "Notes")
		24 V transformer with a tap from the middle of the secondary winding (see "Notes")
		Wires, housing, socket (for output voltage)

Notes

1. Transistors Q1 and Q2, as well as transformer T1, determine the output power of the inverter. With Q1, Q2 = 2N3055 and T1=15A, the inverter has an output power of 300 Watts. To increase power, the transistors and transformer must be replaced with more powerful ones.
2. The easiest and cheapest way to get a large transformer is to rewind the transformer from a microwave oven. These transformers have an output power of up to 1000 watts and are of good quality. Go to a repair shop or look at a junkyard and pick out the largest microwave. The larger the oven, the larger the transformer. Remove the transformer. Do this carefully, do not touch the terminal of the high voltage capacitor, which may still be charged. You can check the transformer, but they are usually fine. Careful not to damage the primary winding, remove the secondary (2000V) winding. Leave the primary one in place. Now wind 24 turns of enameled wire over the primary winding with a tap from the middle of the winding. The diameter of the wire will depend on the current you require. Insulate the winding with electrical tape. The transformer is ready. Choose more powerful transistors Q1 and Q2. The listed 2N3055 parts are rated at only 15A.
3. Remember that when powering a powerful load, the circuit consumes a huge amount of current. Don't let your battery die.
4. Since the output voltage of the converter is 120V, it must be placed in a housing.
5. Only tantalum capacitors must be used as C1 and C2. Conventional electrolytic capacitors overheat and explode due to constant overcharging. The capacitor capacity can only be 68 µF - no change.
6. There may be some difficulties in running this scheme. If there is an error in the installation of the circuit, the design of the transformer, or if the components are incorrectly replaced, the converter may not work.
7. If you want to get a voltage of 220/240 V at the output of the converter, you need to use a transformer with a primary winding of 220/240 V (according to the circuit, it is secondary). The rest of the circuit remains unchanged. The current that the inverter will draw from a 12 V source at an output voltage of 240 V will be twice as much as at a voltage of 120 V.

Switching power supplies (SMPS) are usually quite complex devices, which is why novice radio amateurs tend to avoid them. However, thanks to the proliferation of specialized integrated PWM controllers, it is possible to construct designs that are quite simple to understand and repeat, with high power and efficiency. The proposed power supply has a peak power of about 100 W and is built according to the flyback topology (flyback converter), and the control element is the CR6842S microcircuit (pin-compatible analogues: SG6842J, LD7552 and OB2269).

Attention! In some cases, you may need an oscilloscope to debug the circuit!

Specifications

Block dimensions: 107x57x30 mm (dimensions of the finished block from Aliexpress, deviations are possible).
Output voltage: versions for 24 V (3-4 A) and 12 V (6-8 A).
Power: 100 W.
Ripple level: no more than 200 mV.

On Ali it is easy to find many options for ready-made blocks according to this scheme, for example, by queries like "Artillery power supply 24V 3A", "Power supply XK-2412-24", "Eyewink 24V switching power supply" and the like. On amateur radio portals this model has already been dubbed “folk” due to its simplicity and reliability. Circuitry options 12V and 24V differ slightly and have an identical topology.

Example of a finished power supply from Ali:

Pay attention! In this power supply model, the Chinese have a very high percentage of defects, so when purchasing a finished product, before turning it on, it is advisable to carefully check the integrity and polarity of all elements. In my case, for example, the VD2 diode had the wrong polarity, which is why after three starts the unit burned out and I had to change the controller and key transistor.

The methodology for designing SMPS in general, and this particular topology in particular, will not be considered here in detail, due to too much information - see separate articles.

Switching power supply with a power of 100W on the CR6842S controller.

Purpose of input circuit elements

We will consider the block diagram from left to right:

F 1	Regular fuse.
5D-9	The thermistor limits the current surge when the power supply is turned on. At room temperature it has a small resistance, which limits current surges; when current flows, it heats up, which causes a decrease in resistance, and therefore does not subsequently affect the operation of the device.
C 1	Input capacitor to suppress asymmetrical noise. It is permissible to increase the capacitance slightly; it is desirable that it be an interference-suppressing capacitor like X2 or had a large (10-20 times) margin of operating voltage. For reliable interference suppression, it must have low ESR and ESL.
L 1	Common mode filter to suppress symmetrical interference. It consists of two inductors with the same number of turns, wound on a common core and connected in phase.
KBP307	Rectifier diode bridge.
R5, R9	Circuit required to run CR6842. Through it, the primary charge of capacitor C 4 is carried out to 16.5V. The circuit must provide a trigger current of at least 30 µA (maximum, according to the datasheet) over the entire input voltage range. Also, during operation, this chain controls the input voltage and compensates for the voltage at which the key closes - an increase in the current flowing into the third pin causes a decrease in the threshold voltage for closing the key.
R 10	Timing resistor for PWM. Increasing the value of this resistor will reduce the switching frequency. The nominal value should be in the range of 16-36 kOhm.
C 2	Smoothing capacitor.
R 3, C 7, VD 2	A snubber circuit that protects the key transistor from reverse emissions from the primary winding of the transformer. It is advisable to use R 3 with a power of at least 1W.
C 3	A capacitor that shunts the interwinding capacitance. Ideally, it should be Y-type, or it should have a large margin (15-20 times) of operating voltage. Serves to reduce interference. The rating depends on the parameters of the transformer; it is undesirable to make it too large.
R 6, VD 1, C 4	This circuit, powered from the auxiliary winding of the transformer, forms the controller’s power circuit. This circuit also affects the operating cycle of the key. It works as follows: for correct operation, the voltage at the seventh pin of the controller must be in the range of 12.5 - 16.5 V. The voltage of 16.5 V at this pin is the threshold at which the key transistor opens and energy begins to be stored in the transformer core (at this time the microcircuit is powered from C 4). When it drops below 12.5V, the microcircuit turns off, so capacitor C 4 must provide power to the controller until energy is supplied from the auxiliary winding, so its rating should be sufficient to keep the voltage above 12.5V while the key is open. The lower limit of the C 4 rating should be calculated based on the controller consumption of about 5 mA. The time of the private key depends on the charging time of this capacitor to 16.5V and is determined by the current that the auxiliary winding can supply, while the current is limited by resistor R 6 . Among other things, through this circuit the controller provides overvoltage protection in the event of failure of the feedback circuits - if the voltage exceeds 25V, the controller will turn off and will not start working until the power from the seventh pin is removed.
R 13	Limits the gate charge current of the key transistor and also ensures its smooth opening.
VD 3	Transistor gate protection.
R 8	Pulling the shutter to the ground performs several functions. For example, if the controller is turned off and the internal pull-up is damaged, this resistor will ensure rapid discharge of the transistor gate. Also, with correct board layout, it will provide a shorter gate discharge current path to ground, which should have a positive effect on noise immunity.
BT 1	Key transistor. Installed on the radiator through an insulating gasket.
R 7, C 6	The circuit serves to smooth out voltage fluctuations across the current-measuring resistor.
R 1	Current measuring resistor. When the voltage on it exceeds 0.8V, the controller closes the key transistor, thus regulating the open key time. In addition, as mentioned above, the voltage at which the transistor will be closed also depends on the input voltage.
C 8	Feedback optocoupler filter capacitor. It is permissible to increase the denomination a little.
PC817	Opto-isolation of the feedback circuit. If the optocoupler transistor closes, this will cause an increase in voltage at the second terminal of the controller. If the voltage on the second pin exceeds 5.2V for longer than 56 ms, this will cause the key transistor to close. This provides protection against overload and short circuit.

In this circuit, the 5th pin of the controller is not used. However, according to the datasheet for the controller, you can attach an NTC thermistor to it, which will ensure that the controller turns off in case of overheating. The stabilized output current of this pin is 70 μA. The temperature protection response voltage is 1.05V (the protection will turn on when the resistance reaches 15 kOhm). The recommended thermistor rating is 26 kOhm (at 27°C).

Pulse transformer parameters

Since a pulse transformer is one of the most difficult elements of a pulse block to design, calculating a transformer for each specific block topology requires a separate article, so there will not be a detailed description of the methodology here, however, to repeat the described design, the main parameters of the transformer used should be indicated.

It should be remembered that one of the most important rules when designing is the correspondence between the overall power of the transformer and the output power of the power supply, so first of all, in any case, choose cores that are suitable for your task.

Most often, this design is supplied with transformers made on cores of type EE25 or EE16, or similar. It was not possible to collect enough information on the number of turns in this SMPS model, since different modifications, despite similar circuits, use different cores.

An increase in the difference in the number of turns leads to a reduction in switching losses of the key transistor, but increases the requirements for its load capacity in terms of maximum drain-source voltage (VDS).

For example, we will focus on standard cores of type EE25 and the maximum induction value Bmax = 300 mT. In this case, the ratio of turns of the first-second-third winding will be equal to 90:15:12.

It should be remembered that the indicated turns ratio is not optimal and the ratios may need to be adjusted based on test results.

The primary winding should be wound with a conductor no thinner than 0.3 mm in diameter. It is advisable to make the secondary winding with a double wire with a diameter of 1 mm. A small current flows through the auxiliary third winding, so a wire with a diameter of 0.2 mm will be quite sufficient.

Description of output circuit elements

Next, we will briefly consider the output circuit of the power supply. In general, it is completely standard and differs minimally from hundreds of others. Only the feedback chain on the TL431 may be interesting, but we will not consider it in detail here, because there is a separate article about feedback chains.

VD 4	Dual rectifier diode. Ideally, select one with a voltage/current margin and a minimum drop. Installed on the radiator through an insulating gasket.
R 2 , C 12	Snubber circuit to facilitate diode operation. It is advisable to use R2 with a power of at least 1W.
C 13, L 2, C 14	Output filter.
C 20	Ceramic capacitor, RF output shunt capacitor C 14.
R 17	Load resistor providing no-load load. It also discharges the output capacitors in the event of startup and subsequent shutdown without load.
R 16	Current limiting resistor for LED.
C 9, R 20, R 18, R 19, TLE431, PC817	Feedback circuit on a precision power supply. Resistors set the operating mode of the TLE431, and PC817 provides galvanic isolation.

What can be improved

The above circuit is usually supplied ready-made, but if you assemble the circuit yourself, nothing prevents you from improving the design a little. Both input and output circuits can be modified.

If in your outlets the ground wire has a connection to a quality ground (and is not simply not connected to anything, as is often the case), you can add two additional Y-capacitors, each connected to its own network wire and ground, between L 1 and the input capacitor C 1. This will ensure balancing of the potentials of the network wires relative to the housing and better suppression of the common-mode component of the interference. Together with the input capacitor, two additional capacitors form the so-called. "protective triangle".

After L 1 it is also worth adding another X-type capacitor, with the same capacity as C 1.

To protect against high-amplitude surge voltages, it is advisable to connect a varistor (for example, 14D471K) in parallel with the input. Also, if you have ground, for protection in the event of an accident on the power supply line, in which instead of phase and zero, phase falls on both wires, it is advisable to create a protective triangle of the same varistors.

When the voltage rises above the operating voltage, the varistor reduces its resistance and current flows through it. However, due to the relatively low speed of varistors, they are not able to bypass voltage surges with a rapidly rising edge, therefore, for additional filtering of fast voltage surges, it is advisable to also connect a bidirectional TVS suppressor (for example, 1.5KE400CA) in parallel with the input.

Again, if there is a ground wire, it is advisable to add two more Y-capacitors of small capacity to the output of the block, connected according to the “protective triangle” circuit in parallel with C 14.

To quickly discharge capacitors when the device is turned off, it is advisable to add a megaohm resistor in parallel to the input circuits.

It is advisable to shunt each electrolytic capacitor via RF with small-capacity ceramics located as close as possible to the capacitor terminals.

It would be a good idea to also install a limiting TVS diode at the output - to protect the load from possible overvoltages in case of problems with the unit. For the 24V version, for example 1.5KE24A is suitable.

Conclusion

The scheme is simple enough to repeat and stable. If you add all the components described in the “What Can Be Improved” section, you will get a very reliable and low-noise power supply.