How to make a welding inverter with your own hands. Tips for repairing switching power supplies DIY inverter power supplies

A little about the application and design of the UPS

An article has already been published on the site, which talks about the design of the UPS. This topic can be somewhat supplemented with a short story about repairs. The abbreviation UPS is often referred to. To avoid any discrepancies, let us agree that in this article this is a Switching Power Supply.

Almost all switching power supplies used in electronic equipment are built according to two functional circuits.

Fig.1. Functional diagrams of switching power supplies

As a rule, fairly powerful power supplies, such as computer ones, are made using a half-bridge circuit. Power supplies for powerful stage UMZCHs and welding machines are also manufactured using a push-pull circuit.

Anyone who has ever repaired amplifiers with a power of 400 watts or more knows very well how much they weigh. We are talking, naturally, about UMZCH with a traditional transformer power supply. UPSs for televisions, monitors, and DVD players are most often made according to a circuit with a single-ended output stage.

Although in reality there are other types of output stages, which are shown in Figure 2.

Fig.2. Output stages of switching power supplies

Only the power switches and the primary winding of the power transformer are shown here.

If you look closely at Figure 1, it is easy to see that the entire circuit can be divided into two parts - primary and secondary. The primary part contains a network filter, a network voltage rectifier, power switches and a power transformer. This part is galvanically connected to the AC network.

In addition to the power transformer, switching power supplies also use decoupling transformers, through which control pulses of the PWM controller are supplied to the gates (bases) of power transistors. In this way, galvanic isolation from the secondary circuit network is ensured. In more modern schemes, this decoupling is carried out using optocouplers.

The secondary circuits are galvanically isolated from the network using a power transformer: the voltage from the secondary windings is supplied to the rectifier, and then to the load. Voltage stabilization and protection circuits are also powered from the secondary circuits.

Very simple switching power supplies

They are performed on the basis of a self-oscillator when there is no master PWM controller. An example of such a UPS is the Taschibra electronic transformer circuit.

Fig.3. Electronic transformer Taschibra

Similar electronic transformers are produced by other companies. Their main purpose is . A distinctive feature of this scheme is its simplicity and small number of parts. The disadvantage is that without a load this circuit simply does not start, the output voltage is unstable and has a high level of ripple. But the lights still shine! In this case, the secondary circuit is completely disconnected from the supply network.

It is quite obvious that repairing such a power supply comes down to replacing transistors, resistors R4, R5, sometimes VDS1 and resistor R1, which acts as a fuse. There is simply nothing else to burn in this scheme. Given the low price of electronic transformers, more often than not, a new one is simply purchased, and repairs are done, as they say, “for the love of art.”

Safety First

Since there is such a very unpleasant juxtaposition of the primary and secondary circuits, which during the repair process you will definitely have to touch with your hands, even if by accident, then some safety rules should be recalled.

You can touch the switched-on source with only one hand, and in no case with both at once. Anyone who works with electrical installations knows this. But it is better not to touch at all, or only after disconnecting from the network by pulling the plug from the socket. Also, you should not solder anything while the source is on or simply twist it with a screwdriver.

In order to ensure electrical safety on power supply boards, the “dangerous” primary side of the board is outlined with a fairly wide stripe or shaded with thin strips of paint, usually white. This is a warning that touching this part of the board with your hands is dangerous.

Even a switched-off switching power supply can be touched with your hands only after some time, at least 2...3 minutes after turning off: the charge on high-voltage capacitors is retained for quite a long time, although in any normal power supply there are discharge resistors installed in parallel with the capacitors. Remember how in school they offered each other a charged capacitor! Killing, of course, will not kill, but the blow is quite sensitive.

But the worst thing is not even this: well, just think, it stung a little. If you immediately test the electrolytic capacitor with a multimeter after turning it off, then it is quite possible to go to the store for a new one.

When such a measurement is anticipated, the capacitor must be discharged, at least with tweezers. But it is better to do this using a resistor with a resistance of several tens of kOhms. Otherwise, the discharge is accompanied by a bunch of sparks and a fairly loud click, and such a short circuit is not very useful for the capacitor.

And yet, during repairs you have to touch the switched-mode power supply, at least to take some measurements. In this case, an isolation transformer, often called a safety transformer, will help protect your loved one as much as possible from electric shock. You can read how to make it in the article.

In a nutshell, this is a transformer with two windings for 220V, with a power of 100...200W (depending on the power of the UPS being repaired), the electrical diagram is shown in Figure 4.

Fig.4. Safety transformer

The winding on the left in the diagram is connected to the network; a faulty switching power supply is connected to the right winding through a light bulb. The most important thing with this connection is that you can safely touch any end of the secondary winding with ONE hand, as well as the entire element of the primary circuit of the power supply.

About the role of the light bulb and its power

Most often, repairs to a switching power supply are carried out without an isolating transformer, but as an additional safety measure, the unit is turned on through a 60...150W light bulb. By the behavior of the light bulb, you can, in general, judge the state of the power supply. Of course, such inclusion will not provide galvanic isolation from the network; it is not recommended to touch it with your hands, but it may well protect against smoke and explosions.

If, when plugged into the network, the light bulb lights up at full intensity, then you should look for a fault in the primary circuit. As a rule, this is a broken power transistor or rectifier bridge. During normal operation of the power supply, the light bulb first flashes quite brightly (), and then the filament continues to glow faintly.

There are several opinions about this light bulb. Some say that it does not help get rid of unforeseen situations, while others believe that the risk of burning a newly sealed transistor is much reduced. We will adhere to this point of view and use a light bulb for repairs.

About collapsible and non-demountable housings

Most often, switching power supplies are made in housings. Suffice it to recall computer power supplies, various adapters that plug into an outlet, chargers for laptops, mobile phones, etc.

In the case of computer power supplies, everything is quite simple. Several screws are unscrewed from the metal case, the metal cover is removed and, please, the entire board with the parts is already in your hands.

If the case is plastic, then you should look on the back side, where the power plug is located, for small screws. Then everything is simple and clear, unscrew and remove the cover. In this case, we can say that we were just lucky.

But lately everything has been moving along the path of simplifying and reducing the cost of designs, and the halves of the plastic case are simply glued together, and quite firmly. One friend told me how he took a similar block to some workshop. When asked how to disassemble it, the craftsmen said: “Aren’t you Russian?” Then they took a hammer and quickly split the body into two halves.

In fact, this is the only way to disassemble plastic glued cases. You just need to hit it carefully and not very fanatically: under the influence of blows to the body, the tracks leading to massive parts, for example, transformers or chokes, can be broken.

It also helps to insert a knife into the seam and lightly tap it with the same hammer. True, after assembly traces of this intervention remain. But even if there are minor marks on the case, you won’t have to buy a new unit.

How to find a diagram

If in previous times almost all domestically produced devices were supplied with circuit diagrams, modern foreign electronics manufacturers do not want to share their secrets. All electronic equipment is equipped only with a user manual, which shows which buttons to press. Circuit diagrams are not included with the user manual.

It is assumed that the device will work forever or that repairs will be carried out at authorized service centers where repair manuals, called service manuals, are available. Service centers do not have the right to share this documentation with everyone, but, thank goodness for the Internet, these service manuals can be found for many devices. Sometimes this can be done free of charge, that is, for nothing, and sometimes the necessary information can be obtained for a small amount.

But even if you couldn’t find the required circuit, you shouldn’t despair, especially when repairing power supplies. Almost everything becomes clear upon careful examination of the board. This powerful transistor is nothing more than an output switch, and this microcircuit is a PWM controller.

In some controllers, the powerful output transistor is “hidden” inside the chip. If these parts are large enough, then they have full markings, from which you can find the technical documentation (data sheet) of the microcircuit, transistor, diode or zener diode. It is these parts that form the basis of switching power supplies.

It is somewhat more difficult to find datasheets for small-sized SMD components. Full markings do not fit on a small case; instead, a code designation of several (three, four) letters and numbers is placed on the case. Using this code, using tables or special programs, again found on the Internet, it is possible, although not always, to find reference data for an unknown element.

Measuring instruments and tools

To repair switching power supplies, you will need the tool that every radio amateur should have. First of all, these are several screwdrivers, side cutters, tweezers, sometimes pliers and even the hammer mentioned above. This is for plumbing and installation work.

For soldering work, of course, you will need a soldering iron, preferably several, of varying power and dimensions. A regular soldering iron with a power of 25...40 W is quite suitable, but it is better if it is a modern soldering iron with a thermostat and temperature stabilization.

To solder multi-lead parts, it’s good to have on hand, if not a super expensive one, then at least a simple inexpensive soldering gun. This will allow you to solder multi-pin parts without much effort and destruction of printed circuit boards.

To measure voltages, resistances and, somewhat less frequently, currents, you will need a digital multimeter, even if not a very expensive one, or a good old pointer tester. You can read about the fact that it is too early to write off a pointer device and what additional capabilities it provides that modern digital multimeters do not have.

Can provide invaluable assistance in repairing switching power supplies. Here, too, it is quite possible to use an old, even not very broadband, cathode-ray oscilloscope. If, of course, it is possible to purchase a modern digital oscilloscope, then that is even better. But, as practice shows, when repairing switching power supplies you can do without an oscilloscope.

Actually, when repairing, there are two possible outcomes: either repair it or make it even worse. Here it is appropriate to recall Horner’s law: “Experience grows in direct proportion to the number of equipment disabled.” And although this law contains a fair amount of humor, in the practice of repairs things are exactly this way. Especially at the beginning of the journey.

troubleshooting

Switching power supplies fail much more often than other electronic equipment components. First of all, it is affected by the fact that there is a high mains voltage, which after rectification and filtering becomes even higher. Therefore, the power switches and the entire inverter cascade operate in very difficult conditions, both electrically and thermally. Most often, faults lie in the primary circuit.

Faults can be divided into two types. In the first case, failure of a switching power supply is accompanied by smoke, explosions, destruction and charring of parts, sometimes of printed circuit board tracks.

It would seem that the option is the simplest, you just need to change the burnt parts, restore the tracks, and everything will work. But when trying to determine the type of microcircuit or transistor, it turns out that the part markings have disappeared along with the housing. It is impossible to find out what was here without a diagram, which is often not at hand. Sometimes the repair ends at this stage.

The second type of malfunction is quiet, as Lyolik said, without noise and dust. The output voltages simply disappeared without a trace. If this switching power supply is a simple network adapter like a charger for a cell phone or laptop, then first of all you should check the serviceability of the output cord.

Most often, a break occurs either near the output connector or at the exit from the housing. If the unit is connected to the network using a cord with a plug, then first of all you should make sure that it is in working order.

After checking these simplest circuits, you can already go into the wilds. For these wilds, let's take the power supply circuit of the 19-inch LG_flatron_L1919s monitor. Actually the fault was quite simple: it turned on yesterday, but today it doesn’t turn on.

Despite the apparent seriousness of the device - after all, a monitor, the power supply circuit is quite simple and clear.

After opening the monitor, several swollen electrolytic capacitors (C202, C206, C207) were discovered at the output of the power supply. In this case, it is better to change all the capacitors at once, six in total. The cost of these parts is cheap, so you shouldn’t wait for them to swell too. After this replacement, the monitor started working. By the way, such a malfunction is quite common in LG monitors.

Swollen capacitors triggered the protection circuit, the operation of which will be discussed a little later. If after replacing the capacitors the power supply does not work, you will have to look for other reasons. To do this, let's look at the diagram in more detail.

Fig 5. Power supply of the LG_flatron_L1919s monitor (click on the picture to enlarge)

Surge filter and rectifier

The mains voltage is supplied to the rectifier bridge BD101 through the input connector SC101, fuse F101, and filter LF101. The rectified voltage through the thermistor TH101 is supplied to the smoothing capacitor C101. This capacitor produces a constant voltage of 310V, which is supplied to the inverter.

If this voltage is absent or much less than the specified value, then you should check the mains fuse F101, filter LF101, rectifier bridge BD101, capacitor C101, and thermistor TH101. All of these details can be easily checked using a multimeter. If you suspect capacitor C101, then it is better to replace it with a known good one.

By the way, the mains fuse doesn't just blow. In most cases, replacing it does not restore normal operation of the switching power supply. Therefore, you should look for other reasons leading to the blown fuse.

The fuse should be installed at the same current as indicated on the diagram, and in no case should the fuse be “powered up”. This may lead to even more serious problems.

Inverter

The inverter is made according to a single-cycle circuit. The PWM controller chip U101 is used as a master oscillator, to the output of which the power transistor Q101 is connected. The primary winding of transformer T101 (pins 3-5) is connected to the drain of this transistor through inductor FB101.

Additional winding 1-2 with rectifier R111, D102, C103 is used to power the PWM controller U101 in steady state operation of the power supply. The PWM controller is started when turned on by resistor R108.

Output voltages

The power supply produces two voltages: 12V/2A to power the backlight inverter and 5V/2A to power the logical part of the monitor.

From winding 10-7 of transformer T101 through the diode assembly D202 and filter C204, L202, C205, a voltage of 5V/2A is obtained.

Winding 8-6 is connected in series with winding 10-7, from which, using the diode assembly D201 and filter C203, L201, C202, C206, C207, a constant voltage of 12V/2A is obtained.

Overload protection

Resistor R109 is connected to the source of transistor Q101. This is a current sensor, which is connected through resistor R104 to pin 2 of the U101 chip.

When there is an overload at the output, the current through transistor Q101 increases, which leads to a voltage drop across resistor R109, which is supplied through resistor R104 to pin 2CS/FB of microcircuit U101 and the controller stops generating control pulses (pin 6OUT). Therefore, the voltage at the output of the power supply disappears.

It was this protection that was triggered when the electrolytic capacitors were swollen, which were mentioned above.

Protection level 0.9V. This level is set by the reference voltage source inside the microcircuit. A zener diode ZD101 with a stabilization voltage of 3.3V is connected in parallel with resistor R109, which protects the 2CS/FB input from overvoltage.

A voltage of 310V from capacitor C101 is supplied to pin 2CS/FB through a divider R117, R118, R107, which ensures that protection against increased network voltage is triggered. The permissible range of mains voltage at which the monitor operates normally is in the range of 90…240V.

Output voltage stabilization

Made on an adjustable zener diode U201 type A431. The output voltage of 12V/2A through the divider R204, R206 (both resistors with a tolerance of 1%) is supplied to the control input R of the zener diode U201. As soon as the output voltage becomes 12V, the zener diode opens and the PC201 optocoupler LED lights up.

As a result, the optocoupler transistor opens (pins 4, 3) and the controller supply voltage through resistor R102 is supplied to pin 2CS/FB. The pulses at the 6OUT pin disappear, and the voltage at the 12V/2A output begins to drop.

The voltage at the control input R of the zener diode U201 drops below the reference voltage (2.5V), the zener diode is locked and turns off the optocoupler PC201. Pulses appear at the 6OUT output, the 12V/2A voltage begins to increase and the stabilization cycle is repeated again. The stabilization circuit is built in a similar way in many switching power supplies, for example, in computer ones.

Thus, it turns out that three signals are connected to the input 2CS/FB of the controller using a wired OR: overload protection, protection against overvoltage of the network and the output of the output voltage stabilizer circuit.

This is where it’s appropriate to remember how you can check the operation of this stabilization loop. For this purpose it is enough to turn OFF!!! from the power supply network, supply 12V/2A voltage from the regulated power supply to the output.

It is better to connect to the output of the PC201 optocoupler with a pointer tester in resistance measurement mode. As long as the voltage at the output of the regulated source is below 12V, the resistance at the output of the optocoupler will be high.

Now we will increase the voltage. As soon as the voltage exceeds 12V, the arrow of the device will sharply drop in the direction of decreasing resistance. This indicates that the zener diode U201 and optocoupler PC201 are working properly. Therefore, output voltage stabilization should work fine.

In exactly the same way, you can check the operation of the stabilization loop of computer switching power supplies. The main thing is to understand what voltage the zener diode is connected to.

If all of the above checks were successful, and the power supply does not start, then you should check transistor Q101 by removing it from the board. If the transistor is working properly, the U101 chip or its wiring is most likely to blame. First of all, this is the electrolytic capacitor C105, which is best checked by replacing it with a known good one.

Most modern electronic devices practically do not use analog (transformer) power supplies; they are replaced by pulsed voltage converters. To understand why this happened, it is necessary to consider the design features, as well as the strengths and weaknesses of these devices. We will also talk about the purpose of the main components of pulsed sources and provide a simple example of an implementation that can be assembled with your own hands.

Design features and operating principle

Of the several methods of converting voltage to power electronic components, two that are most widespread can be identified:

1. Analog, the main element of which is a step-down transformer, in addition to its main function, it also provides galvanic isolation.
2. Impulse principle.

Let's look at how these two options differ.

PSU based on a power transformer

Let's consider a simplified block diagram of this device. As can be seen from the figure, a step-down transformer is installed at the input, with its help the amplitude of the supply voltage is converted, for example, from 220 V we get 15 V. The next block is a rectifier, its task is to convert the sinusoidal current into a pulsed one (the harmonic is shown above the symbolic image). For this purpose, rectifying semiconductor elements (diodes) connected via a bridge circuit are used. Their operating principle can be found on our website.

The next block performs two functions: it smoothes the voltage (a capacitor of appropriate capacity is used for this purpose) and stabilizes it. The latter is necessary so that the voltage does not “drop” when the load increases.

The given block diagram is greatly simplified; as a rule, a source of this type has an input filter and protective circuits, but this is not important for explaining the operation of the device.

All the disadvantages of the above option are directly or indirectly related to the main design element - the transformer. Firstly, its weight and dimensions limit miniaturization. In order not to be unfounded, we will use as an example a step-down transformer 220/12 V with a rated power of 250 W. The weight of such a unit is about 4 kilograms, dimensions 125x124x89 mm. You can imagine how much a laptop charger based on it would weigh.

Secondly, the price of such devices is sometimes many times higher than the total cost of the other components.

Pulse devices

As can be seen from the block diagram shown in Figure 3, the operating principle of these devices differs significantly from analog converters, primarily in the absence of an input step-down transformer.

Figure 3. Block diagram of a switching power supply

Let's consider the operating algorithm of such a source:

• Power is supplied to the network filter; its task is to minimize network noise, both incoming and outgoing, that arises as a result of operation.
• Next, the unit for converting sinusoidal voltage into pulsed constant voltage and a smoothing filter come into operation.
• At the next stage, an inverter is connected to the process; its task is related to the formation of rectangular high-frequency signals. Feedback to the inverter is carried out through the control unit.
• The next block is IT, it is necessary for automatic generator mode, supplying voltage to the circuit, protection, controller control, as well as the load. In addition, the IT task includes ensuring galvanic isolation between high and low voltage circuits.

Unlike a step-down transformer, the core of this device is made of ferrimagnetic materials, this contributes to the reliable transmission of RF signals, which can be in the range of 20-100 kHz. A characteristic feature of IT is that when connecting it, the inclusion of the beginning and end of the windings is critical. The small dimensions of this device make it possible to produce miniature devices; an example is the electronic harness (ballast) of an LED or energy-saving lamp.

• Next, the output rectifier comes into operation, since it works with high-frequency voltage; the process requires high-speed semiconductor elements, so Schottky diodes are used for this purpose.
• At the final phase, smoothing is performed on an advantageous filter, after which voltage is applied to the load.

Now, as promised, let’s look at the operating principle of the main element of this device – the inverter.

How does an inverter work?

RF modulation can be done in three ways:

• pulse-frequency;
• phase-pulse;
• pulse width.

In practice, the last option is used. This is due both to the simplicity of implementation and to the fact that PWM has a constant communication frequency, unlike the other two modulation methods. A block diagram describing the operation of the controller is shown below.

The operating algorithm of the device is as follows:

The reference frequency generator generates a series of rectangular signals, the frequency of which corresponds to the reference one. Based on this signal, a sawtooth U P is formed, which is supplied to the input of the comparator K PWM. The UUS signal coming from the control amplifier is supplied to the second input of this device. The signal generated by this amplifier corresponds to the proportional difference between U P (reference voltage) and U RS (control signal from the feedback circuit). That is, the control signal UUS is, in fact, a mismatch voltage with a level that depends on both the current on the load and the voltage on it (U OUT).

This implementation method allows you to organize a closed circuit that allows you to control the output voltage, that is, in fact, we are talking about a linear-discrete functional unit. Pulses are generated at its output, with a duration depending on the difference between the reference and control signals. Based on it, a voltage is created to control the key transistor of the inverter.

The process of stabilizing the output voltage is carried out by monitoring its level; when it changes, the voltage of the control signal U PC changes proportionally, which leads to an increase or decrease in the duration between pulses.

As a result, the power of the secondary circuits changes, which ensures stabilization of the output voltage.

To ensure safety, galvanic isolation between the power supply and feedback is necessary. As a rule, optocouplers are used for this purpose.

Strengths and weaknesses of pulsed sources

If we compare analog and pulse devices of the same power, the latter will have the following advantages:

• Small size and weight due to the absence of a low-frequency step-down transformer and control elements that require heat removal using large radiators. Thanks to the use of high-frequency signal conversion technology, it is possible to reduce the capacitance of the capacitors used in the filters, which allows the installation of smaller elements.
• Higher efficiency, since the main losses are caused only by transient processes, while in analog circuits a lot of energy is constantly lost during electromagnetic conversion. The result speaks for itself, increasing efficiency to 95-98%.
• Lower cost due to the use of less powerful semiconductor elements.
• Wider input voltage range. This type of equipment is not demanding in terms of frequency and amplitude; therefore, connection to networks of various standards is allowed.
• Availability of reliable protection against short circuits, overload and other emergency situations.

The disadvantages of pulse technology include:

The presence of RF interference is a consequence of the operation of the high-frequency converter. This factor requires the installation of a filter that suppresses interference. Unfortunately, its operation is not always effective, which imposes some restrictions on the use of devices of this type in high-precision equipment.

Special requirements for the load, it should not be reduced or increased. As soon as the current level exceeds the upper or lower threshold, the output voltage characteristics will begin to differ significantly from the standard ones. As a rule, manufacturers (even recently Chinese ones) provide for such situations and install appropriate protection in their products.

Scope of application

Almost all modern electronics are powered from blocks of this type, as an example:

Assembling a switching power supply with your own hands

Let's consider the circuit of a simple power supply, where the above-described principle of operation is applied.

Designations:

• Resistors: R1 – 100 Ohm, R2 – from 150 kOhm to 300 kOhm (selectable), R3 – 1 kOhm.
• Capacities: C1 and C2 – 0.01 µF x 630 V, C3 -22 µF x 450 V, C4 – 0.22 µF x 400 V, C5 – 6800-15000 pF (selectable), 012 µF, C6 – 10 µF x 50 V, C7 – 220 µF x 25 V, C8 – 22 µF x 25 V.
• Diodes: VD1-4 - KD258V, VD5 and VD7 - KD510A, VD6 - KS156A, VD8-11 - KD258A.
• Transistor VT1 – KT872A.
• Voltage stabilizer D1 - microcircuit KR142 with index EH5 - EH8 (depending on the required output voltage).
• Transformer T1 - a w-shaped ferrite core with dimensions 5x5 is used. The primary winding is wound with 600 turns of wire Ø 0.1 mm, the secondary (pins 3-4) contains 44 turns Ø 0.25 mm, and the last winding contains 5 turns Ø 0.1 mm.
• Fuse FU1 – 0.25A.

The setup comes down to selecting the values ​​of R2 and C5, which ensure excitation of the generator at an input voltage of 185-240 V.

Preface

I would like to warn dear readers of this article in advance that this article will have a form and content that is not entirely familiar to readers. Let me explain why.

The material presented to your attention is absolutely exclusive. All devices that will be discussed in my articles are developed, modeled, configured and brought to mind by me personally. Most often, it all starts with an attempt to put some interesting idea into practice. The path can be very thorny, and sometimes takes quite a long time, and what the final result will be, and whether there will be one at all, is not known in advance. But practice confirms that the one who walks will master the road... and the results sometimes exceed all expectations... And how fascinating the process itself is - words cannot express it. I must admit that I (like everyone else, it should be noted) do not always have enough knowledge and skills, and wise and timely advice is welcome and helps to bring the idea to its logical conclusion. Here's the specifics...

This article is addressed not so much to beginners, but rather to people who already have the necessary knowledge and experience, who are also interested in walking untrodden paths, and for whom standard approaches to solving problems are not so interesting... It is important to understand that this is not material for thoughtless repetition, but rather - the direction in which you need to move... I don’t promise readers great details about obvious, well-known and understandable things in electronics..., but I promise that the main ESSENCE will always be well covered.

About the inverter

The inverter that will be discussed was born in exactly the manner described above... Unfortunately, I cannot, without violating the rules for publishing these articles, cover in detail how it came into being, but I assure you that the circuits of the two extreme versions of the inverter are not yet available anywhere have been published...Moreover, the penultimate version of the scheme is already practically in use, and the extreme one (I hope the most perfect of them) has not yet been mocked up only on paper, but I have no doubt about its functionality, and its production and testing will take only a couple of days...

Getting acquainted with the microcircuit for the half-bridge inverter IR2153 made a good impression - a fairly small current consumed by the power supply, the presence of a dead time, built-in power control... But it has two significant drawbacks - there is no ability to adjust the duration of the output pulses and a rather small driver current... (in reality it is not stated in the datasheet, but it is unlikely to be more than 250-500 mA...). It was necessary to solve two problems - to figure out how to implement voltage regulation of the inverter, and how to increase the current of the power switch drivers...

These problems were solved by introducing field-effect transistors into the optical driver circuit, and differentiating circuits at the outputs of the IR2153 microcircuit (see Fig. 1)

Fig.1

A few words about how adjusting the pulse duration works. Pulses from the outputs of IR2153 are supplied to differentiating circuits consisting of elements C2, R2, optical driver LED, VD3-R4 - optocoupler transistor..., and elements C3, R3, optical driver LED, VD4-R5 - optocoupler transistor... The elements of the differentiating circuits are designed in this way that, with the feedback optocoupler transistor closed, the duration of the pulses at the outputs of the optical drivers is almost equal to the duration of the pulses at the outputs of IR2153. At the same time, the voltage at the inverter output is maximum.

At the moment when the voltage at the inverter output reaches the stabilization voltage, the optocoupler transistor begins to open slightly... this leads to a decrease in the time constant of the differentiating circuit, and, as a consequence, to a decrease in the duration of the pulses at the output of the optical drivers. This ensures voltage stabilization at the inverter output. Diodes VD1, VD2 eliminate the negative surge that occurs during differentiation.

I deliberately do not mention the type of optical drivers. That's why the optical driver of a field-effect transistor is a big separate topic for discussion. Their range is very large - dozens... if not hundreds of types... for every taste and color. To understand their purpose and their features, you need to study them yourself.

The presented inverter has another important feature. Let me explain. Since the main purpose of the inverter is to charge lithium batteries (although any batteries can be used, of course), measures had to be taken to limit the current at the inverter output. The fact is that if you connect a discharged battery to the power supply, the charging current can exceed all reasonable limits... To limit the charging current to the level we need, a shunt Rsh is introduced into the control electrode circuit TL431... How does it work? The minus of the battery being charged is connected not to the minus of the inverter, but to the upper terminal of the circuit Rsh... When current flows through Rsh, the potential on the control electrode TL431... increases, which leads to a decrease in the voltage at the output of the inverter, and, as a consequence, to limiting the charging current. As the battery charges, the voltage on it increases, but after it, the voltage at the inverter output also increases, tending to the stabilization voltage. In short, it’s a simple and outrageously effective contraption. By changing the Rsh rating, it is easy to limit the charge current at any level we need. That is why the Rsh rating itself is not announced... (the reference value is 0.1 Ohm and below...), it is easier to select it experimentally.

Anticipating many instructive comments about the “correctness” and “incorrectness” of the principles of charging lithium batteries, I kindly ask you to refrain from such comments and take my word for it that I am more than aware of how this is done... This is a large, separate topic... and within the framework it will not be discussed in this article.

A few words about the IMPORTANT features of setting up the signal part of the inverter...

To check the functionality and configure the signal part of the inverter, you need to apply +15 Volts to the power supply circuit of the signal part from any external power supply and check with an oscilloscope the presence of pulses on the gates of the power switches. Then, it is necessary to simulate the operation of the feedback optocoupler (by applying voltage to the optocoupler LED) and make sure that in this case an ALMOST complete narrowing of the pulses on the gates of the power switches occurs. At the same time, it is more convenient to connect the oscilloscope probes not in the standard way, otherwise - the signal wire of the probe to one of the gates of the power switch, and the common wire of the oscilloscope probe to the gate of another power switch... This will make it possible to see the pulses of different half-cycles simultaneously... (what is in the neighboring in half cycles we will see pulses of opposite polarity, it does not matter here). Now the MOST important thing is to make sure (or achieve) that when the feedback optocoupler is ON, the control pulses do NOT narrow to zero (remain of a minimum duration, but do not lose their rectangular shape...). In addition, it is important, by selecting resistor R5 (or R4), to ensure that the pulses in adjacent half-cycles are the SAME duration... (the difference is quite likely due to the difference in the characteristics of the optical drivers). See Fig.2

Fig.2

After this hassle, connecting the inverter to a 220 Volt network will most likely go without any problems. When setting up, it is very advisable to connect a small load (5 W car light bulb) to the inverter output... Due to the non-zero minimum duration of control pulses, without load, the voltage at the inverter output may be higher than the stabilization voltage. This does not interfere with the operation of the inverter, but I hope to get rid of this unpleasant moment in the next version of the inverter.

An important thing about the printed circuit board design is that it has a number of features...

For the last few years I have been using boards designed for ala-planar mounting of elements... That is, all elements are located on the side of the printed conductors. In this way, ALL elements of the circuit are soldered... even those that were not originally intended for planar mounting. This significantly reduces the labor intensity of manufacturing. In addition, the board has a completely flat bottom part and it becomes possible to place the board directly on the radiator. This design significantly simplifies the process of replacing elements during setup and repair. Some connections (the most inconvenient ones for printed wiring) are made with insulated mounting wire. This is quite justified, since it allows you to significantly reduce the size of the board.

The printed circuit board design itself (see Fig. 3) is rather the BASIS for your particular design. Its final design will need to be adjusted to suit the optical drivers you use. It should be borne in mind that different optical drivers have DIFFERENT housings, and the numbering and assignment of pins may differ from that shown in the diagram in this article. The presented board has already gone through about ten modifications regarding the signal part. Correction of the signal part, sometimes very significant, does not take much time at all.

Fig.3

I do not plan to provide an exact list of elements within the framework of this article. The reason is simple - the main goal of all this fuss is to make a useful thing with minimal labor out of the maximum available elements. That is, collect from what you have. By the way, if the output voltage of the inverter is not planned to be more than twenty volts, then any transformer from a computer power supply (assembled using a half-bridge circuit) can be used as an output transformer. The photo below is a general view of the assembled inverter, so that you have an idea of ​​what it looks like (it’s better to see once than to hear a hundred times). I beg you to be lenient with the build quality, but I simply have no choice - I only have two hands... You solder the current version, but in your head the next option is almost ripe... And otherwise - there is no way... - you can’t jump over the step.. .

Yes, that’s what I forgot to mention – questions will probably arise about the power of the inverter. I will answer this way - the maximum power of such an inverter is difficult to estimate in absentia..., it is determined mainly by the power of the power elements used, the output transformer and the maximum peak current of the output of the optical drivers. At high powers, the design itself, the damper circuits of the power switches will begin to have a big influence..., you will need to use synchronous rectifiers instead of diodes at the output... In short, this is a completely different story, much more difficult to implement... As for the described inverter, I use it to charge LiFePO4 battery with a voltage of 21.9 Volts (capacity - 15A/h) with a current of 7-8 Amperes... This is the line where the temperature of the radiator and transformer is within reasonable limits and no forced cooling is required... For my taste - cheap and cheerful..

I do not plan to talk about this inverter in more detail within the framework of this article. It is not possible to cover everything (and it takes so much time, it should be noted...), so it would be more reasonable to discuss the issues that have arisen in a separate topic on the soldering iron forum. There I will listen to all wishes and criticisms, and answer questions.

I have no doubt that many people may not like this approach. And many are sure that everything has already been invented before us... I assure you, this is not so...

But that's not the end of the story. If there is interest, then we can continue the conversation... because there is another, extreme version of the signal part. ...I hope it will be continued.

Additions dated June 25, 2014

This is how it turns out this time too - the ink on the article has not yet dried, but very interesting ideas have already appeared on how to make the signal part of the inverter more perfect...

I would like to warn you that all drawings marked with the signature “project” in a fully assembled inverter have NOT been checked! But if the performance of individual fragments of the circuit was tested on a breadboard, and their performance was confirmed, I will make a special reservation.

The operating principle of the modified signal part is still based on the differentiation of pulses from the IR2153 microcircuit. But from the point of view of the correct construction of electronic circuits, the approach here is more competent.

A couple of clarifications - the actual differentiating circuits now include C2, R2, R4 and C3, R3, R5 plus diodes VD1, VD2 and a feedback optocoupler. Diodes that eliminate negative emissions arising during differentiation are excluded..., since they are not necessary - field-effect transistors allow the supply of a gate-source voltage of +/-20 Volts. Differentiated pulses, changing their duration under the influence of the feedback optocoupler, enter the gates of transistors T1, T2, which turn on the LEDs of the optical drivers...

This scheme has been tested on a breadboard. It showed good performance and great flexibility in configuration. I highly recommend it for use.

The photo below shows a fragment of a circuit diagram with a modified signal part and a drawing of a printed circuit board with corrections for the modified signal part...

To be continued...

Update from 06.29.14

This is what the extreme version of the signal part of the inverter looks like, which I mentioned at the beginning of the article. Finally, I found the time to make its layout and look at its work in reality... I looked... and yet - yes, it is he who will be appointed as the most perfect of the proposed... The scheme can be called successful because all the elements in it perform the functions for which and are intended from birth.

This version of the controller uses a different, more familiar, method of changing the duration of controls. Pulses from the outputs of IR2153 are converted from rectangular to triangular shape by integrating circuits R2,C2 and R3,C3. The generated triangular pulses are supplied to the inverting inputs of the dual comparator LM393. The non-inverting inputs of the comparators receive voltage from the divider R4, R5. Comparators compare the current value of the triangular voltage with the voltage from the divider R4, R5, and at moments when the value of the triangular voltage exceeds the voltage from the divider R4, R5, a low potential appears at the outputs of the comparators. This leads to the optical driver LED turning on... INCREASING the voltage from the divider R4, R5 leads to a DECREASE in the pulse duration at the outputs of the comparators. This is what will make it possible to organize feedback of the inverter output with the pulse duration shaper, and thereby ensure stabilization and control of the inverter output voltage. When the feedback optocoupler is triggered, the optocoupler transistor opens slightly, the voltage from the divider R4,R5 increases, which leads to a decrease in the duration of the control pulses..., while the output voltage decreases... The value of the resistor R6* determines the degree of influence of the feedback circuit on the duration of the generated pulses ... - the smaller the value of the resistor R6*, the shorter the duration of the pulses when the feedback optocoupler is triggered... When setting up, changing the value of the resistor R6* allows you to ensure that the duration of the generated pulses at the moment the feedback optocoupler is triggered will tend to (or be equal - here it's not scary) to zero. The pictures below will help you understand the essence of how comparators work.

A few words about what is important when setting up. The setup procedure itself is quite simple, but don’t even try to do it without an oscilloscope... It’s tantamount to trying to drive blindfolded... The peculiarity (and this is rather its advantage than a disadvantage) is that it allows you to generate impulses with any ratio of durations in adjacent channels... You need to understand that the shaper can either change (introduce or eliminate completely) the duration of the dead time between the pulses of adjacent channels, but even form them in such a way that the pulses of adjacent channels will “overlap” each other ..., which, of course, is unacceptable... Your task is to monitor the pulses at the output of the drivers with an oscilloscope, changing the value of the resistor R4*, set the non-inverting inputs of the comparators to such a voltage that pulses separated by dead time 1 will be generated at the outputs of the drivers -2 μS (the wider the dead time, the lower the risk of through currents).

Then, it is necessary to turn on the feedback optocoupler, and, by changing the value of resistor R6*, select it such that the duration of the generated ones decreases to zero. During this procedure, it will not be harmful to control the MOMENT OF DISAPPEARANCE of the generated impulses. It is very desirable that the complete disappearance of the generated pulses occur SIMULTANEOUSLY... Non-simultaneous disappearance is possible if the parameters of the integrators R2,C2 and R3,C3 are very different. This can be cured by a small change in the values ​​of the elements of one of the integrators. I did it practically. For convenience, temporarily, instead of the optocoupler-R6* transistor circuit, I connected a 20 Kohm potentiometer, and set the pulse duration to the point of disappearing. The difference in the duration of the generated pulses turned out to be negligible... But I also eliminated it by installing an additional capacitor (only 30 pF) in parallel with capacitor C3.

A few words about the operating features of optical drivers... During setup, it turned out that optical drivers work better with a higher LED current. Moreover, there is another important nuance - the optical driver LED consumes more current not during the entire pulse duration, but only in fairly short periods (1-2 µS), coinciding in time with the positions of the pulse fronts. This is important, as it allows us to understand that the average current consumed by the optodriver LED is really not high at all. These considerations determine the choice of the value of resistor R7. The actually measured PEAK current of the optodriver LED, with the nominal value indicated on the diagram, is 8-10 mA.

A diode (VD5) has been added to the circuit in the circuit in the power supply circuit of the lower driver. Let me explain why. The optodrivers I use have a built-in power control system. Due to the fact that a diode is always used in the power circuit of the upper driver, the supply voltage of the upper driver is always slightly lower than the supply voltage of the lower driver. Therefore, when the supply voltage decreases, the pulses from the output of the upper driver disappear a little earlier than the lower one. To bring closer the moments when the drivers are turned off, the VD5 diode was introduced. You should always pay close attention to these moments...

Here, it’s time to note that this driver can be used (after a slight change in the logic of the comparator) together with conventional (non-optical) half-bridge drivers. For those who don’t understand what we’re talking about, look, for example, at what IR2113 is. There are a lot of similar ones..., and their use may turn out to be even more preferable than optical ones... But this is a topic for the next addition to the article... I don’t promise that I will test their work in practice, but at least at the level of circuit diagrams of several options - no problem....

That's it - there are a lot of beeches - but in reality the setup comes down to selecting two resistors. I would like to especially note that this driver is NOT critical to its power supply - in the power range of the IR2153 microcircuit (9-15 Volts), it works absolutely adequately. The disappearance of pulses from the outputs of IR2153 when its power supply decreases (at the moment the unit is turned off), leads to the closing of the power switches.

A couple more tips - you shouldn’t try to replace the IR2153 with some analogue on discrete elements - it’s not productive... In reality, it’s possible, but it’s simply not reasonable - the number of parts will increase significantly (in the original - there are only three of them..., much less). In addition, you will have to resolve issues regarding the behavior of the analogue when turned on and off (and they will definitely be). Fighting this will further complicate the scheme, and the meaning of this idea will be nullified...

For those who are interested in this topic, I am attaching, for convenience, drawings of printed circuit boards adjusted for this driver. Among them is the shaper itself in the form of a submodule... - it’s more convenient to start your first acquaintance with them. I would SPECIALLY emphasize that if you decide to try to configure the driver autonomously (without connecting power switches), remember that when setting up, you need to connect the “virtual” common of the upper driver with a real common wire (otherwise, the upper driver will have no power).

Although I did not plan further changes to the inverter, it should be noted that the presence of only one duration adjustment circuit will make it easy to introduce any current protection into it. This is a separate interesting topic, and we may return to it later...

In conclusion of this addition, let me remind you that from birth, the main purpose of the inverter is to charge lithium batteries. It is endowed with special, very important properties by its use in the Rsh circuit... For those who do not understand its purpose, I recommend once again delving into the section of the article in which it is discussed.

If we do not use Rsh (jumper), we will have a regular inverter with voltage stabilization (but without any current protection, of course...).

List of radioelements

Designation	Type	Denomination	Quantity	Note	Shop	My notepad
	Power Driver and MOSFET	IR2153	1			To notepad
	Voltage reference IC	TL431	1			To notepad
T1, T2	Field-effect transistor		2			To notepad
VD1-VD6	Diode		6			To notepad
VD7, VD8	Rectifier diode	FR607	2			To notepad
VD9	Diode bridge	RS405L	1			To notepad
	Optocoupler		1			To notepad
	Optical driver		2			To notepad
C1	Capacitor	3900 pF	1			To notepad
C2, C3, C10	Capacitor	0.01 µF	3			To notepad
C4		100 µF 25 V	1			To notepad
C5, C6	Capacitor	1 µF	2			To notepad
S7, S12	Capacitor	1000 pF	2			To notepad
S8, S9	Electrolytic capacitor	150 µF 250 V	2			To notepad
C11	Electrolytic capacitor	1000 µF	1			To notepad
R1	Resistor	5.1 kOhm	1			To notepad
R2, R3	Resistor	1.3 kOhm	2			To notepad
R4, R5	Resistor	110 Ohm	2			To notepad
R6, R7	Resistor	10 ohm	2			To notepad
R8, R9	Resistor	10 kOhm	2			To notepad
R10, R15	Resistor	3.9 kOhm	2	R10 0.5 W.		To notepad
R11	Resistor	3 kOhm	1	0.5 W		To notepad
R12	Resistor	51 Ohm	1	1 W		To notepad
R13, R14	Resistor	100 kOhm	2			To notepad
R16, R18	Resistor	1 kOhm	2			To notepad
R17	Resistor	7.76 kOhm	1			To notepad
Rsh	Resistor	0.1 Ohm or less	1			To notepad
	Transformer		1	From a computer power supply		To notepad
	Inductor		1			To notepad
F1	Fuse	2 A	1			To notepad
	Master oscillator. Option #2.
	Power Driver and MOSFET	IR2153	1			To notepad
T1, T2	MOSFET transistor	2N7002	2			To notepad
	Optocoupler		1			To notepad
	Optical driver		2			To notepad
VD1-VD3	Diode		3			To notepad
C1	Capacitor	2200 pF	1

A DIY welding inverter made from a computer power supply is becoming increasingly popular among both professionals and amateur welders. The advantages of such devices are that they are comfortable and lightweight.

The use of an inverter power source allows you to qualitatively improve the characteristics of the welding arc, reduce the size of the power transformer and thereby lighten the weight of the device, makes it possible to make adjustments smoother and reduce spatter during welding. The disadvantage of an inverter-type welding machine is its significantly higher price than its transformer counterpart.

In order not to overpay large sums of money in stores for welding, you can make one. To do this, you need a working computer power supply, several electrical measuring instruments, tools, basic knowledge and practical skills in electrical work. It would also be useful to acquire relevant literature.

If you are not confident in your abilities, then you should go to the store for a ready-made welding machine, otherwise, with the slightest mistake during the assembly process, there is a risk of getting an electric shock or burning all the electrical wiring. But if you have experience in assembling circuits, rewinding transformers and creating electrical appliances with your own hands, you can safely begin the assembly.

Operating principle of inverter welding

The welding inverter consists of a power transformer that reduces the network voltage, stabilizer chokes that reduce current ripple, and an electrical circuit block. For circuits, MOSFET or IGBT transistors can be used.

The principle of operation of the inverter is as follows: alternating current from the network is sent to the rectifier, after which the power module converts direct current into alternating current with increasing frequency. Next, the current enters the high-frequency transformer, and the output from it is the welding arc current.

Return to contents

Tools required to make an inverter

To assemble a welding inverter from a power supply with your own hands, you will need the following tools:

• soldering iron;
• screwdrivers with different tips;
• pliers;
• wire cutters;
• drill or screwdriver;
• crocodiles;
• wires of the required cross-section;
• tester;
• multimeter;
• consumables (wires, solder for soldering, electrical tape, screws and others).

To create a welding machine from a computer power supply, you need materials to create a printed circuit board, getinaks, and spare parts. To reduce the amount of work, you should go to the store for ready-made electrode holders. However, you can make them yourself by soldering crocodiles to wires of the required diameter. It is important to observe polarity when doing this work.

Return to contents

The procedure for assembling the welding machine

First of all, to create a welding machine from a computer power supply, you need to remove the power source from the computer case and disassemble it. The main elements that can be used from it are a few spare parts, a fan and standard case plates. It is important to take into account the cooling operating mode. This determines what elements need to be added to ensure the necessary ventilation.

The operation of a standard fan, which will cool the future welding machine from a computer unit, must be tested in several modes. This check will ensure the functionality of the element. To prevent the welding machine from overheating during operation, you can install an additional, more powerful cooling source.

To control the required temperature, a thermocouple should be installed. The optimal temperature for operating the welding machine should not exceed 72-75°C.

But first of all, you should install a handle of the required size on the welding machine from a computer power supply for carrying and ease of use. The handle is installed on the top panel of the block using screws.

It is important to choose screws that are optimal in length, otherwise too large ones may affect the internal circuit, which is unacceptable. At this stage of work, you should worry about good ventilation of the device. The placement of elements inside the power supply is very dense, so a large number of through holes should be arranged in it in advance. They are performed with a drill or screwdriver.

Next, you can use multiple transformers to create an inverter circuit. Typically, 3 transformers such as ETD59, E20 and Kx20x10x5 are chosen. You can find them in almost any radio electronics store. And if you already have experience creating transformers yourself, then it’s easier to do them yourself, focusing on the number of turns and the performance characteristics of the transformers. Finding such information on the Internet will not be difficult. You may need a current transformer K17x6x5.

It is best to make homemade transformers from getinax coils; the winding will be enamel wire with a cross-section of 1.5 or 2 mm. You can use 0.3x40 mm copper sheet, after wrapping it in durable paper. Thermal paper from a cash register (0.05 mm) is suitable; it is durable and does not tear so much. The crimping should be done from wooden blocks, after which the entire structure should be filled with “epoxy” or varnished.

When creating a welding machine from a computer unit, you can use a transformer from a microwave oven or old monitors, not forgetting to change the number of turns of the winding. For this work, it would be useful to use electrical engineering literature.

As a radiator, you can use PIV, previously cut into 3 parts, or other radiators from old computers. You can purchase them in specialized stores that disassemble and upgrade computers. Such options will pleasantly save time and effort in searching for suitable cooling.

To create a device from a computer power supply, you must use a single-cycle forward quasi-bridge, or “oblique bridge”. This element is one of the main ones in the operation of the welding machine, so it is better not to save on it, but to purchase a new one in the store.

Printed circuit boards can be downloaded on the Internet. This will make recreating the circuit much easier. In the process of creating the board, you will need capacitors, 12-14 pieces, 0.15 microns, 630 volts. They are necessary to block resonant current surges from the transformer. Also, to make such a device from a computer power supply, you will need capacitors C15 or C16 with the brand K78-2 or SVV-81. Transistors and output diodes should be installed on radiators without using additional gaskets.

During operation, you must constantly use a tester and a multimeter to avoid errors and to assemble the circuit faster.

After manufacturing all the necessary parts, they should be placed in the housing and then routed. The temperature on the thermocouple should be set to 70°C: this will protect the entire structure from overheating. After assembly, the welding machine from a computer unit must be pre-tested. Otherwise, if you make a mistake during assembly, you can burn all the main elements, or even get an electric shock.

On the front side, two contact holders and several current regulators should be installed. The device switch in this design will be a standard computer unit toggle switch. The body of the finished device after assembly requires additional strengthening.