A transformer is a necessary element of any welding source. It reduces the network voltage to the arc voltage level, and also provides galvanic isolation of the network and the welding circuit. It is known that the dimensions of a transformer are determined by its operating frequency, as well as the quality of the magnetic core material.

Note.

As the frequency decreases, the dimensions of the transformer increase, and as the frequency increases, they decrease.

Transformers of classical sources operate at a relatively low network frequency. Therefore, the weight and dimensions of these sources were mainly determined by the mass and volume of the welding transformer.

Recently, various high-quality magnetic materials have been developed that make it possible to somewhat improve the weight and size parameters of transformers and welding sources. However, a significant improvement in these parameters can only be achieved by increasing the operating frequency of the transformers. Since the frequency of the mains voltage is standard and cannot be changed, it is possible to increase the operating frequency of the transformer using a special electronic converter.

Block diagram of inverter welding source

A simplified block diagram of an inverter welding source (IWS) is shown in rice. 1. Let's look at the diagram. The mains voltage is rectified and smoothed, and then supplied to the electronic converter. It converts direct voltage into high frequency alternating voltage. High-frequency alternating voltage is transformed using a small-sized high-frequency transformer, then rectified and fed into the welding circuit.

Transformer types

The operation of the electronic converter is closely related to the magnetization reversal cycles of the transformer. Since the ferromagnetic material of the transformer core has nonlinearity and is saturated, the induction in the transformer core can only grow to a certain maximum value Vm.

After reaching this value, the core must be demagnetized to zero or remagnetized in the opposite direction to the value – Bm. Energy can be transmitted through a transformer:

• in the magnetization cycle;
• in the magnetization reversal cycle;
• in both cycles.

Definition.

Converters that provide energy transfer in one cycle of transformer magnetization reversal are called single-cycle.

Accordingly, converters that provide energy transfer in both magnetization reversal cycles of the transformer are called two-stroke.

Single-ended forward converter

Advantages of single-ended converters. Single-cycle converters are most widely used in cheap and low-power inverter welding sources designed to operate from a single-phase network. Under conditions of sharply variable load, such as the welding arc, single-cycle converters compare favorably with various push-pull converters:

• they do not require balancing;
• they are not susceptible to such a disease as through currents.

Therefore, to control this converter, a simpler control circuit is required compared to what would be required for a push-pull converter.

Classification of single-cycle converters. According to the method of transferring energy to the load, single-cycle converters are divided into two groups: forward and flyback ( rice. 2). In forward converters, energy is transferred to the load at the moment of the closed state, and in flyback converters - at the moment of the open state of the key transistor VT. In this case, in a flyback converter, energy is stored in the inductance of the transformer T during the closed state of the switch and the switch current has the shape of a triangle with a rising edge and a steep cutoff.

Note.

When choosing the type of ISI converter between forward and flyback, preference is given to a forward single-ended converter.

Indeed, despite its great complexity, a forward converter, unlike a flyback converter, has high power density. This is explained by the fact that in a flyback converter a triangular current flows through the key transistor, and in a forward converter, a rectangular current flows. Consequently, at the same maximum switch current, the average current value of a forward converter is twice as high.

Main advantages flyback converter is:

• lack of a choke in the rectifier;
• possibility of group stabilization of several voltages.

These advantages provide an advantage to flyback converters in various low-power applications, such as power supplies for various household television and radio equipment; as well as service power supplies for the control circuits of the welding sources themselves.

Transformer of a single-transistor forward converter (SFC), shown on rice. 2, b, has a special demagnetizing winding III. This winding serves to demagnetize the core of the transformer T, which is magnetized during the closed state of the transistor VT.

At this time, the voltage on winding III is applied to diode VD3 in blocking polarity. Due to this, the demagnetizing winding does not have any influence on the magnetization process.

After turning off the transistor VT:

• the voltage on winding III changes its polarity;
• diode VD3 is unlocked;
• the energy accumulated in the transformer T returns to the primary power source Up.

Note.

However, in practice, due to insufficient coupling between the transformer windings, part of the magnetizing energy is not returned to the primary source. This energy is usually dissipated in the VT transistor and damping circuits (on rice. 2 not shown), degrading the overall efficiency and reliability of the converter.

Oblique bridge. This disadvantage is not present in two-transistor forward converter (DFC), which is often called "oblique bridge" (rice. 3, a). In this converter (due to the introduction of an additional transistor and diode), the primary winding of the transformer is used as a demagnetizing winding. Since this winding is completely connected to itself, the problems of incomplete return of magnetization energy are completely eliminated.

Let us consider in more detail the processes occurring at the moment of magnetization reversal of the transformer core.

A common feature of all single-ended converters is that their transformers operate in conditions with one-way magnetization.

Magnetic induction B (in a transformer with one-way magnetization) can only vary within the range from maximum Bm to residual Br, describing a partial hysteresis loop.

When transistors VT1, VT2 of the converter are open, the energy of the power source Up is transferred to the load through transformer T. In this case, the transformer core is magnetized in the forward direction (section a-b on rice. 3, b).

When transistors VT1, VT2 are locked, the current in the load is maintained by the energy stored in the inductor L. In this case, the current is closed through the diode VD0. At this moment, under the influence of the EMF of winding I, the diodes VD1, VD2 open, and the demagnetization current of the transformer core flows through them in the opposite direction (section b-a on rice. 3, b).

The change in induction ∆B in the core occurs practically from Bm to Br and is significantly less than the value ∆B = 2·Bm possible for a push-pull converter. Some increase in ∆B can be obtained by introducing a non-magnetic gap into the core. If the core has a non-magnetic gap δ, then the residual induction becomes less than Br. If there is a non-magnetic gap in the core, the new value of the residual induction can be found at the intersection point of a straight line drawn from the origin at an angle Ѳ to the magnetization reversal curve (point B1 on rice. 3, b):

tgѲ= µ 0 · l c/δ,

where µ 0 – magnetic permeability;

l c – length of the average magnetic field line of the magnetic core, m;

δ – length of non-magnetic gap, m.

Definition.

Magnetic permeability – this is the ratio of induction B to tension H for a vacuum (also valid for a non-magnetic air gap) and is a physical constant, numerically equal to µ 0 = 4π·10 -7 H/m.

The value tgѲ can be considered as non-magnetic gap conductivity, reduced to the length of the core. Thus, introducing a non-magnetic gap is equivalent to introducing a negative magnetic field strength:

Н1 = -В1/ tgѲ.

Push-pull bridge converter

Advantages of push-pull converters. Push-pull converters contain more elements and require more complex control algorithms. However, these converters provide lower input current ripple and greater output power and efficiency from the same discrete key component power.

Scheme of a push-pull bridge converter. On rice. 4, a shows a diagram of a push-pull bridge converter. If we compare this converter with single-ended ones, then it is closest to a two-transistor forward converter ( rice. 3) . A push-pull converter is easily converted into it if you remove a pair of transistors and a pair of diodes located diagonally (VT1, VT4, VD2, VD3 or VT2, VT3, VD1, VD4).

Thus, a push-pull bridge converter is a combination of two single-cycle converters operating alternately. In this case, energy is transferred to the load during the entire period of operation of the converter, and the induction in the transformer core can vary from -Bm to +Bm.

As in the DPP, diodes VD1-VD4 serve to return the energy accumulated in the leakage inductance Ls of the transformer T to the primary power source Up. MOSFET internal diodes can be used as these diodes.

Operating principle. Let us consider in more detail the processes occurring at the moment of magnetization reversal of the transformer core.

Note.

A common feature of push-pull converters is that their transformers operate in conditions with symmetrical magnetization reversal.

Magnetic induction B, in the core of a transformer with symmetrical magnetization reversal, can vary from negative -Bm to positive +Bm maximum induction.

In each half-cycle of the DMP operation, two keys located diagonally are open. During pause, all transistors of the converter are usually closed, although there are control modes when some transistors of the converter remain open during pause.

Let's focus on the control mode, according to which all DMP transistors are closed during a pause.

When transistors VT1, VT4 of the converter are open, the energy of the power source Up is transferred to the load through transformer T. In this case, the transformer core is magnetized in the conventional reverse direction (section b-a in Fig. 4, b).

During a pause, when transistors VT1, VT4 are closed, the current in the load is maintained by the energy stored in the inductor L. In this case, the current is closed through the diode VD7. At this moment, one of the secondary windings (IIa or IIb) of transformer T is short-circuited through an open diode VD7 and one of the rectifier diodes (VD5 or VD6). As a result of this, the induction in the transformer core remains virtually unchanged.

After the pause is completed, transistors VT2, VT3 of the converter open, and the energy of the power source Up is transferred to the load through transformer T.

In this case, the transformer core is magnetized in the conventional forward direction (section a-b on rice. 4). During a pause, when transistors VT2, VT3 are closed, the current in the load is maintained by the energy stored in the inductor L. In this case, the current is closed through the diode VD7. At this moment, the induction in the transformer core remains virtually unchanged and is fixed at the achieved positive level.

Note.

Due to the fixation of inductions in pauses, the core of the transformer T is capable of reversing magnetization only when the diagonally located transistors are open.

In order to avoid one-sided saturation under these conditions, it is necessary to ensure equal open time of the transistors, as well as symmetry of the power circuit of the converter.

The power part of our homemade inverter-type semi-automatic welding machine is based on an asymmetric bridge circuit, or, as it is also called, an “oblique bridge”. This is a single-ended forward converter. The advantages of such a scheme are simplicity, reliability, minimal number of parts, and high noise immunity. Until now, many manufacturers produce their products using the “oblique bridge” design. You also cannot do without disadvantages - these are large pulse currents from the power supply, lower efficiency than in other circuits, and large currents through the power transistors.

Block diagram of a forward converter “oblique bridge”

The block diagram of such a device is shown in the figure:

Power transistors VT1 and VT2 operate in the same phase, i.e. they open and close at the same time, therefore, compared to a full bridge, the current through them is twice as large. The TT transformer provides current feedback.
You can learn more about all types of inverter converters for welding machines from the book.

Description of the inverter circuit

Semiautomatic inverter welding machine, operating in MMA (arc welding) and MAG (special wire welding in a gas environment) modes.

Control board

The following inverter components are installed on the control board: a master oscillator with a galvanic isolation transformer, current and voltage feedback units, a relay control unit, a thermal protection unit, and an “anti-stick” unit.

Master oscillator

The current control unit (for MMA mode) and the master oscillator (OG) are assembled on LM358N and UC2845 microcircuits. UC2845 was chosen as the MG, rather than the more common UC3845 due to the more stable parameters of the former.

The generation frequency depends on the elements C10 and K19, and is calculated by the formula: f = (1800/(R*C))/2, where R and C are in kilo-ohms and nanofarads, the frequency is in kilohertz. In this circuit, the frequency is 49KHz.

Another important parameter is the fill factor, calculated using the formula Kzap = t/T. It cannot be more than 50%, and in practice it is 44-48%. It depends on the ratio of denominations C10 and R19. If the capacitor is taken as small as possible, and the resistor as large as possible, then Kzap will be close to 50%.

The generated SG pulses are fed to the VT5 switch, which operates on the T1 galvanic isolation transformer (TGR), wound on an EE25 core, used in electronic units for starting fluorescent lamps (electronic ballasts). All windings are removed and new ones are wound according to the diagram. Instead of the IRF520 transistor, you can use any of this series - IRF530, 540, 630, etc.

Current feedback

As mentioned earlier, for arc welding What is important is a stable output current, for semi-automatic - a constant voltage. Current feedback is organized on the TT current transformer; it is a ferrite ring of size K 20 x 12 x 5, placed on the lower (according to the diagram) terminal of the primary winding of the power transformer. Depending on the primary winding current T2, the pulse width of the master oscillator decreases or increases, maintaining the output current unchanged.

Voltage feedback

Welding semi-automatic inverter type requires voltage feedback; for this, in MAG mode, switch S1.1 supplies the voltage from the output of the device to the output voltage adjustment unit, assembled on elements R55, D18, U2. Powerful resistor K50 sets the initial current. And with contacts S1.2, the key on the transistor VT1 short-circuits the regulator R2 to the maximum current, and the key VT3 disables the “anti-stick” mode (switching off the GB when the electrode sticks).

Thermal protection block

A homemade semi-automatic welding machine includes an overheating protection circuit: this is provided by a unit on transistors VT6, VT7. Temperature sensors at 75 degrees C (there are two of them, normally closed, connected in series) are installed on the radiator of the output diodes and on one of the radiators of the power transistors. When the temperature is exceeded, transistor VT6 shorts pin 1 of the UC2845 to ground and disrupts the generation of pulses.

Relay control unit

This block is assembled on a DD1 CD4069UB microcircuit (analogous to 561LN2) and a VT14 BC640 transistor. These elements provide the following operating mode: when you press the button, the gas valve relay is immediately turned on, after about a second, the VT17 transistor allows the generator to start and at the same time the pull-out mechanism relay turns on.

The relays that control the “pulling” and the gas valve, as well as the fans, are powered by the stabilizer on the MC7812 mounted on the control board.

Power unit based on HGTG30N60A4 transistors

From the TGR output, pulses pre-generated by drivers on transistors VT9 VT10 are supplied to power switches VT11, ME12. “Snubbers” are connected parallel to the collector-emitter terminals of these transistors - chains of elements C24, D47, R57 and C26, D44, R59, which serve to keep powerful transistors in the range of permissible values. In the immediate vicinity of the keys there is a capacitor C28, assembled from 4 capacitors of 1 micron x 630v. Zener diodes Z7, Z8 are necessary to limit the voltage at the switch gates to 16 volts. Each transistor is installed on a radiator from a computer processor with a fan.

Power transformer and rectifier diodes

The main element of the semiautomatic welding circuit is the powerful output transformer T2. It is assembled on two E70 cores, N87 material from EPCOS.

Calculation of a welding transformer

The turns of the primary winding are calculated according to the formula: N = (Upit * timp)/(Badd * Ssec),
where Upit = 320V – maximum supply voltage;
timp = ((1000/f)/2)*K – pulse duration, K = (Kzap*2)/100 = (0.45*2)/100 = 0.9 timp = ((1000/49)/2 )*0.9 = 9.2;
Vdop = 0.25 – permissible induction for the core material;
Ssection = 1400 – core section.
N = (320 * 9.2)/(0.25 * 1400) = 8.4, rounded to 9 turns.
The ratio of secondary to primary turns should be approximately 1/3, i.e. we wind 3 turns of the secondary winding.

The power transformer can be wound on a different standard size; the turns are calculated using the above formula. For example, for a core 2 x E80 at f = 49Khz, turns in the primary: 16, secondary: 5.

Selecting the cross-section of the wires of the primary and secondary windings, winding the transformer

We select the wire cross-section at the rate of 1mm.kv = 10A output current. This device should produce approximately 190A under load, so we take the secondary cross-section of 19mm.kv (a bundle of 61 wires with a diameter of 0.63mm). The cross section of the primary is selected 3 times smaller, i.e. 6mm.sq. (harness of 20 wires with a diameter of 0.63 mm). The cross-section of the wire depending on its diameter is calculated as: S = D²/1.27 where D is the diameter of the wire.

Winding is carried out on a frame made of 1mm PCB, without side cheeks. The frame is mounted on a wooden frame according to the dimensions of the core. The primary winding is wound (all turns in one layer). Then 5 layers of thick transformer paper, with the secondary winding on top. The coils are compressed with plastic ties. Then the frame with windings is removed from the mandrel and impregnated with varnish in a vacuum chamber. The chamber was made from a liter jar with a tight lid and a hose attached to the suction tube of the compressor from the refrigerator (you can simply dip the trans in varnish for a day, I think it will also become saturated).

DIY WELDING MACHINE

OVERVIEW OF WELDING INVERTER DIAGRAMS AND DESCRIPTION OF OPERATING PRINCIPLE

Let's start with a fairly popular welding inverter circuit, often called the Bramaley circuit. I don’t know why this name was attached to this scheme, but Barmaley’s welding machine is often mentioned on the Internet.
There were several options for the Barmaley inverter circuit, but their topology is almost the same - a forward single-ended converter (quite often called an “oblique bridge”, for some reason), controlled by a UC3845 controller.
Since this controller is the main one in this circuit, let’s start with the principle of its operation.
The UC3845 chip is produced by several manufacturers and is part of the UC1842, UC1843, UC1844, UC1845, UC2842, UC2843, UC2844, UC2845, UC3842, UC3843, UC3844, and UC3845 series of chips.
The microcircuits differ from each other in the supply voltage at which they start and self-lock, in the operating temperature range, as well as in small circuit changes that allow the duration of the control pulse in the XX42 and XX43 microcircuits to be increased to 100%, while in the XX44 and XX45 series microcircuits the duration of the control pulse cannot exceed 50%. The pinout of the microcircuits is the same.
An additional 34...36 V zener diode is integrated into the microcircuit (depending on the manufacturer), which allows you not to worry about exceeding the supply voltage when using the microcircuit in a power supply with a VERY wide range of supply voltages.
Microcircuits are available in several types of packages, which significantly expands the scope of use

The microcircuits were initially designed as controllers for controlling the power switch of a single-cycle medium-power power supply, and this controller was equipped with everything necessary to increase its own survivability and the survivability of the power supply it controls. The microcircuit can operate up to frequencies of 500 kHz, the output current of the final driver stage is capable of developing a current of up to 1 A, which in total allows you to design fairly compact power supplies. The block diagram of the microcircuit is shown below:

On the block diagram, an additional trigger is highlighted in red, which does not allow the duration of the output pulse to exceed 50%. This trigger is only installed on the UCx844 and UCx845 series.
In microcircuits made in packages with eight pins, some pins are combined inside the chip, for example VC and Vcc, PWRGND and GROUND.

A typical switching power supply circuit for the UC3844 is shown below:

This power supply has indirect secondary voltage stabilization, since it controls its own power supply generated by the NC winding. This voltage is rectified by diode D3 and serves to power the microcircuit itself after it starts, and after passing through the divider on R3 it goes to the input of the error amplifier, which controls the duration of the control pulses of the power transistor.
As the load increases, the amplitude of all output voltages of the transformer decreases, which also leads to a decrease in the voltage at pin 2 of the microcircuit. The logic of the microcircuit increases the duration of the control pulse, more energy accumulates in the transformer and, as a result, the amplitude of the output voltages returns to the original value. If the load decreases, the voltage at pin 2 increases, the duration of the control pulses decreases, and again the amplitude of the output voltage returns to the set value.
The chip has an integrated input for organizing overload protection. As soon as the voltage drop across the current-limiting resistor R10 reaches 1 V, the microcircuit turns off the control pulse at the gate of the power transistor, thereby limiting the current flowing through it and eliminating overload of the power supply. Knowing the value of this control voltage, you can regulate the protection operation current by changing the value of the current-limiting resistor. In this case, the maximum current through the transistor is limited to 1.8 amperes.
The dependence of the magnitude of the flowing current on the value of the resistor can be calculated using Ohm’s law, but it’s too lazy to pick up a calculator every time, so after calculating once, we’ll simply enter the results of the calculations into the table. Let me remind you that you need a voltage drop of one volt, therefore the table will only indicate the protection operation current, resistor values ​​and their power.

I, A	1	1,2	1,3	1,6	1,9	3	4,5	6	10	20	30	40	50
R, Ohm	1	0,82	0,75	0,62	0,51	0,33	0,22	0,16	0,1	0,05	0,033	0,025	0,02
								2 x 0.33		2 x 0.1	3 x 0.1	4 x 0.1	5 x 0.1
P,W	0,5	1	1	1	1	2	2	5	5	10	15	20	25

This information may be needed if the welding machine being designed is without a current transformer, and control will be carried out in the same way as in the basic circuit - using a current-limiting resistor in the source circuit of the power transistor or in the emitter circuit, when using an IGBT transistor.
A switching power supply circuit with direct control of the output voltage is offered in the datasheet for the chip from Texas Instruments:

This circuit controls the output voltage using an optocoupler; the brightness of the optocoupler LED is determined by an adjustable zener diode TL431, which increases the coefficient. stabilization.
Additional transistor elements have been introduced into the circuit. The first one imitates a soft start system, the second one increases thermal stability by using the base current of the introduced transistor.
It will not be difficult to determine the tripping current of the protection of this circuit - Rcs is equal to 0.75 Ohm, therefore the current will be limited to 1.3 A.
Both the previous and this power supply circuits are recommended in the datasheets for the UC3845 from Texas Instruments; in the datasheets of other manufacturers, only the first circuit is recommended.
The dependence of frequency on the values ​​of the frequency-setting resistor and capacitor is shown in the figure below:

The question may involuntarily arise - WHY ARE SUCH DETAILS NEEDED AND WHY ARE WE TALKING ABOUT POWER UNITS WITH A POWER OF 20...50 WATT??? THE PAGE WAS ANNOUNCEED AS A DESCRIPTION OF A WELDING MACHINE, AND HERE ARE SOME POWER SUPPLY UNITS...
In the vast majority of simple welding machines, the UC3845 microcircuit is used as a control element, and without knowledge of the principle of its operation, fatal errors can occur that contribute to the failure of not only a cheap microcircuit, but also quite expensive power transistors. In addition, I am going to design a welding machine, and not stupidly clone someone else’s circuit, look for ferrites, which I may even have to buy, in order to replicate someone else’s device. No, I’m not happy with this, so we take the existing circuit and refine it to suit what we need, to suit the elements and ferrites that are available.
That is why there will be quite a lot of theory and several experimental measurements, and that is why in the table of protection resistor ratings, resistors connected in parallel (blue cell fields) are used and the calculation is made for currents of more than 10 amperes.
So, the welding inverter, which most sites call the Barmaley welder, has the following circuit diagram:

INCREASE

In the upper-left part of the diagram there is a power supply for the controller itself and, in fact, ANY power supply with an output voltage of 14...15 volts and providing a current of 1...2 A can be used (2 A is so that fans can be installed more powerful - the device uses computer fans and according to the scheme there are as many as 4 of them.
By the way, I even managed to find a collection of answers on this welding machine from some forum. I think this will be useful for those who are planning to purely clone the circuit. LINK TO DESCRIPTION.
The arc current is adjusted by changing the reference voltage at the input of the error amplifier; overload protection is organized using current transformer TT1.
The controller itself operates on an IRF540 transistor. In principle, any transistor with not very high gate energy Qg (IRF630, IRF640, etc.) can be used there. The transistor is loaded onto control transformer T2, which directly supplies control pulses to the gates of power IGBT transistors.
To prevent the control transformer from being magnetized, it is equipped with a demagnetizing winding IV. The secondary windings of the control transformer are loaded onto the gates of power transistors IRG4PC50U through a rectifier using 1N5819 diodes. Moreover, the control circuit contains IRFD123 transistors that force the closing of the power section, which, when the polarity of the voltage on the windings of transformer T2 changes, opens and absorbs all the energy from the gates of the power transistors. Such closing accelerators facilitate the current mode of the driver and significantly reduce the closing time of power transistors, which in turn reduces their heating - the time spent in linear mode is significantly reduced.
Also, to facilitate the operation of power transistors and suppress impulse noise that occurs when operating an inductive load, chains of 40 Ohm resistors, 4700 pF capacitors and HFA15TB60 diodes are used.
For final demagnetization of the core and suppression of self-induction emissions, another pair of HFA15TB60 is used, installed to the right according to the diagram.
A half-wave rectifier based on a 150EBU02 diode is installed on the secondary winding of the transformer. The diode is shunted by an interference suppression circuit using a 10 Ohm resistor and a 4700 pF capacitor. The second diode serves to demagnetize the inductor DR1, which accumulates magnetic energy during the forward stroke of the converter, and during the pause between pulses releases this energy to the load due to self-induction. To improve this process, an additional diode is installed.
As a result, the output of the inverter does not produce a pulsating voltage, but a constant one with a small ripple.
The next sub-modification of this welding machine is the inverter circuit shown below:

I didn’t really delve into what was complicated about the output voltage; I personally liked the use of bipolar transistors as closing the power section more. In other words, both field and bipolar devices can be used in this node. In principle, this was implied by default, the main thing is to close the power transistors as quickly as possible, and how to do this is a secondary question. In principle, using a more powerful control transformer, you can dispense with the closing transistors - it is enough to apply a small negative voltage to the gates of the power transistors.
However, I was always confused by the presence of a control transformer in the welding machine - well, I don’t like winding parts and, if possible, I try to do without them. The search for welder circuits continued and the following welding inverter circuit was dug up:

INCREASE

This circuit differs from the previous ones in the absence of a control transformer, since the opening and closing of power transistors occurs by specialized IR4426 driver microcircuits, which in turn are controlled by 6N136 optocouplers.
There are a couple more goodies implemented in this scheme:
- an output voltage limiter made on the PC817 optocoupler has been introduced;
- the principle of stabilizing the output current is implemented - the current transformer is not used as an emergency one, but as a current sensor and takes part in adjusting the output current.
This version of the welding machine guarantees a more stable arc even at low currents, since as the arc increases, the current begins to decrease, and this machine will increase the output voltage, trying to maintain the set value of the output current. The only drawback is that you need a biscuit switch for as many positions as possible.
Another diagram of a welding machine for self-production also caught my eye. The output current is stated to be 250 amperes, but this is not the main thing. The main thing is to use the rather popular IR2110 chip as a driver:

INCREASE

This version of the welder also uses output voltage limitation, but there is no current stabilization. There is one more embarrassment, and quite a serious one. How is capacitor C30 charged? In principle, during the pause, the core should be pre-demagnetized, i.e. The polarity of the voltage on the windings of the power transformer must be changed and so that the transistors do not fly off, diodes D7 and D8 are installed. It seems that for a short time a voltage of 0.4...0.6 volts less than the common wire should appear at the upper terminal of the power transformer; this is a fairly short-term phenomenon and there are some doubts that the C30 will have time to charge. After all, if it does not charge, the upper arm of the power section will not open - there will be no place for the boost voltage of the IR2110 driver to come from.
In general, it makes sense to think about this topic more thoroughly...
There is another version of the welding machine, made according to the same topology, but it used domestic parts and in large quantities. The circuit diagram is shown below:

INCREASE

The first thing that catches your eye is the power part - 4 pieces of IRFP460 each. Moreover, the author in the original article claims that the first version was assembled on an IRF740, 6 pieces per arm. This is truly a “need for cunning invention.” Here you should immediately make a memorization - both IGBT transistors and MOSFET transistors can be used in the welding inverter. In order not to get confused with definitions and pinouts, we embroider a drawing of these same transistors:

In addition, it makes sense to note that this circuit uses both limiting the output voltage and a current stabilization mode, which is regulated by a 47 Ohm variable resistor - the low resistance of this resistor is the only drawback of this implementation, but if you wish, you can find one, and increasing this resistor to 100 Ohms is not critical , you will just need to increase the limiting resistors.
Another version of the welding machine caught my eye while studying foreign sites. This device also has current regulation, but it is not done in a very ordinary way. The current control pin is initially supplied with a bias voltage and the higher it is, the less voltage is required from the current transformer, therefore, the less current will flow through the power section. If the bias voltage is minimal, then to achieve the limiter operation current, a higher voltage from the CT will be required, which is only possible when a large current flows through the primary winding of the transformer.
The schematic diagram of this inverter is shown below:

INCREASE

In this welding machine circuit, electrolytic capacitors are installed at the output. The idea is certainly interesting, but this device will require electrolytes with a small ESR, and at 100 volts such capacitors are quite problematic to find. Therefore, I will refuse to install electrolytes, and will install a couple of MKP X2 5 µF capacitors, used in induction cookers.

WE ASSEMBLE YOUR WELDING MACHINE

WE BUY PARTS

First of all, I’ll say right away that assembling a welding machine yourself is not an attempt to make the machine cheaper than a store-bought one, since in the end it may turn out that the assembled machine will be more expensive than the factory one. However, this idea also has its advantages - this device can be purchased on an interest-free loan, since it is not at all necessary to buy the entire set of parts at once, but make purchases as free money appears in the budget.
Again, studying power electronics and assembling such an inverter yourself provides invaluable experience that will allow you to assemble similar devices, sharpening them directly to your needs. For example, assemble a starting-charger with an output current of 60-120 A, assemble a power source for a plasma cutter - although a specific device, it is a VERY useful thing for those working with metal.
If it seems to someone that I have fallen into Ali’s advertising, then I will say right away - yes, I am advertising Ali, because I am satisfied with both the price and quality. With the same success, I can advertise the sliced ​​loaves of the Ayutinsky bakery, but I buy the black bread from Krasno-Sulinsky. I prefer condensed milk and recommend it to you, “Cow from Korenovka”, but cottage cheese is much better than the Tatsinsky dairy plant. So I’m ready to advertise everything that I tried myself and liked.

To assemble the welding machine, you will need additional equipment that is necessary for assembling and setting up the welding machine. This equipment also costs some money, and if you are really going to deal with power electronics, then you will need it later, but if assembling this device is an attempt to spend less money, then feel free to abandon this idea and go to the store for a ready-made welding inverter.
I buy the vast majority of components from Ali. You have to wait from three weeks to two and a half months. However, the cost of the components is much cheaper than in a radio parts store, to which I still have to travel 90 km.
Therefore, I’ll immediately make a short instruction on how best to buy components on Ali. I will give links to the parts used as they are mentioned, and I will give them to the search results, because there is a possibility that in a couple of months some seller will not have this product. I will also give prices for the mentioned components for comparison. Prices will be in rubles at the time of writing this article, i.e. mid-March 2017.
By clicking on the link to the search results, first of all, it should be noted that the sorting is done by the number of purchases of a particular product. In other words, you already have the opportunity to see exactly how much of this product a particular seller sold and what reviews they received for these products. The pursuit of a low price is not always correct - Chinese entrepreneurs try to sell ALL products, so sometimes there are relabeled elements, as well as elements after dismantling. Therefore, look at the number of reviews about the product.

If the same components are available at a more attractive price, but the number of sales from this seller is not large, then it makes sense to pay attention to the total number of positive reviews about the seller.

It makes sense to pay attention to the photographs - the presence of a photograph of the product itself indicates the responsibility of the seller. And in the photo you can clearly see what kind of markings there are, this often helps - the laser and paint markings are visible in the photo. I buy power transistors with laser markings, but I bought IR2153 with paint markings - the microcircuits are working.
If power transistors are chosen, then quite often I do not disdain transistors from dismantling - they usually have a fairly decent price difference, and for a device that you assemble yourself, you can use parts with shorter legs. It’s not difficult to distinguish the details even from a photo:

Also, several times I ran into one-time promotions - sellers without a rating generally put up some components for sale at VERY ridiculous prices. Of course, the purchase is made at your own peril and risk. However, I made a couple of purchases from similar sellers and both were successful. The last time I purchased MKP X2 5 µF capacitors for 140 rubles, 10 pieces.

The order arrived quite quickly - a little over a month, 9 pieces of 5 µF, and one of exactly the same size at 0.33 µF 1200 V. I did not open a dispute - I have all the capacitances for induction toys at 0.27 µF and how I would even need 0.33 uF. And the price is too ridiculous. I checked all the containers - they were working, I wanted to order more, but there was already a sign - THE PRODUCT IS NO LONGER AVAILABLE.
Before this I took dismantling IRFPS37N50, IRGP20B120UD, STW45NM50 several times. All transistors are in good working order, the only thing that was somewhat disappointing was that on the STW45NM50 the legs were remolded - on three transistors (out of 20) the leads literally fell off when I tried to bend them to fit my board. But the price was too ridiculous to be offended by anything - 20 pieces for 780 rubles. These transistors are now used as replacement transistors - the case is cut down to the terminal, the wires are soldered and filled with epoxy glue. One is still alive, two years have passed.

The issue with power transistors is still open, but connectors for the electrode holder will be needed for any welding machine. The search was long and quite active. The thing is that the difference in price is very confusing. But first, about the marking of connectors for the welding machine. Ali uses European markings (well, that’s how they write it), so we’ll dance from their markings. True, a chic dance will not work - these connectors are scattered across various categories, ranging from USB connectors, BLOW TORCHES and ending with OTHER.

And in terms of the name of the connectors, not everything is as smooth as we would like... I was VERY surprised when I typed DKJ35-50 into the search bar on Google Chrome and the WIN XP OS and got NO RESULTS, but the same query on the same Google Chrome , but WIN 7 gave at least some results. Well, first, a small sign:

DKZ	DKL	DKJ
			MAX CURRENT, A	DIAMETER ANSWER/ PLUG, MM	SECTION WIRES, MM2
DKZ10-25	DKL10-25	DKJ10-25	200	9	10-25
DKZ35-50	DKL35-50	DKJ35-50	315	13	35-50
DKZ50-70	DKL50-70	DKJ50-70	400	13	50-70
DKZ70-95	DKL70-95	DKJ70-95	500	13	70-95

Despite the fact that the holes and plugs of the 300-500 ampere connectors are the same, they are actually capable of conducting different currents. The fact is that when turning the connector, the plug part rests against the end of the mating part, and since the diameters of the ends of more powerful connectors are larger, a larger contact area is obtained, therefore the connector is able to pass more current.

SEARCHING FOR CONNECTORS FOR WELDING MACHINES
SEARCH DKJ10-25	SEARCH DKJ35-50	SEARCH DKJ50-70
SOLD BOTH RETAIL AND IN SETS

I bought DKJ10-25 connectors a year ago and this seller no longer carries them. Just a couple of days ago I ordered a pair of DKJ35-50. I bought it. True, I had to first explain to the seller - the description says that the wire is 35-50 mm2, and in the photo it is 10-25 mm2. The seller assured that these are connectors for 35-50 mm2 wire. We'll see what he sends - there's time to wait.
As soon as the first version of the welding machine passes the tests, I will begin to assemble the second version with a much larger set of functions. I won’t be modest - I’ve been using a welding machine for more than six months now AuroraPRO INTER TIG 200 AC/DC PULSE(there is exactly the same one named “CEDAR”). I really like the device, and its capabilities simply caused a storm of delight.

But in the process of mastering the welding machine, several shortcomings emerged that I would like to eliminate. I won’t go into detail about what exactly I didn’t like, since the device is really not bad, but I want more. That’s why I actually started developing my own welding machine. The Barmaley-type device will be a training device, and the next one will have to surpass the existing Aurora.

WE DETERMINE THE PRINCIPLE DIAGRAM OF THE WELDING MACHINE

So, having looked at all the circuit options that deserve attention, let’s start assembling our own welding machine. First you need to decide on a power transformer. I will not buy w-shaped ferrites - ferrites from line transformers are available and there are quite a lot of the same ones. But the shape of this core is quite peculiar, and the magnetic permeability is not indicated on them...
You will have to make several test measurements, namely, make a frame for one core, wind about fifty turns on it and, putting this frame on the cores, select those with the same inductance as possible. In this way, cores will be selected that will be used to assemble a common core consisting of several magnetic cores.
Next, you will need to find out how many turns need to be wound on the primary winding so that the core does not go into saturation and uses the maximum overall power.
To do this, you can use the article by Biryukov S.A. (DOWNLOAD), or you can, based on the article, build your own stand to test the saturation of the core. The second method is preferable for me - for this stand I use the same microcircuit as for the welding machine - UC3845. First of all, this will allow me to “touch” the microcircuit in person, check the adjustment ranges, and by installing a socket for microcircuits in the stand, I will be able to check these microcircuits immediately before installing them in the welding machine.
We will assemble the following diagram:

Here is an almost classic UC3845 connection circuit. VT1 contains a voltage stabilizer for the microcircuit itself, since the range of supply voltages of the stand itself is quite large. Any VT1 in a TO-220 package with a current of 1 A and a K-E voltage above 50 V.
Speaking of supply voltages, you need a power supply with a voltage of at least 20 volts. The maximum voltage is no more than 42 volts - this is still a safe voltage for working with bare hands, although it is better not to go above 36. The power supply must provide a current of at least 1 ampere, i.e. have a power of 25 W and above.
It is worth considering here that this stand operates on the booster principle, so the total voltage of the zener diodes VD3 and VD4 should be at least 3-5 volts higher than the supply voltage. It is highly not recommended to exceed the difference by more than 20 volts.
As a power supply for the stand, you can use a car charger with a classic transformer, not forgetting to put a pair of 1000 μF 50V capacitors at the charging output. We set the charging current regulator to maximum - the circuit will not take more than necessary.
If you don’t have a suitable power supply and there is nothing to assemble it from, then you can PURCHASE A READY POWER SUPPLY, you can choose one in a plastic case or a metal one. Price from 290 rubles.
Transistor VT2 serves to regulate the voltage supplied to the inductance, VT3 generates pulses on the inductance under study, and VT4 acts as a device that demagnetizes the inductance, so to speak, an electronic load.
Resistor R8 is the conversion frequency, and R12 is the voltage supplied to the inductor. Yes, yes, exactly the choke, since while we do not have a secondary winding, this piece of the transformer is nothing more than a very ordinary choke.
Resistors R14 and R15 are measuring - with R15 the microcircuit controls the current, and with both the voltage drop shape is monitored. Two resistors are used to increase the drop voltage and reduce garbage collection by the oscilloscope - terminal X2.
The choke being tested is connected to terminals X3, and the power supply voltage of the stand is connected to terminals X4.
The diagram shows what I have assembled. However, this circuit has a rather unpleasant drawback - the voltage after transistor VT2 strongly depends on the load, so in my measurements I used the position of the R12 engine, at which the transistor is completely open. If you bring this circuit to mind, then it is advisable to use a parametric voltage regulator instead of a field controller, for example, like this:

I won’t do anything else with this stand - I have an LATR and I can easily change the power supply voltage of the stand by connecting a test, ordinary transformer through the LATR. The only thing I had to add was a fan. VT4 operates in linear mode and heats up quite quickly. In order not to overheat the common radiator, I installed a fan and limiting resistors.

The logic here is quite simple - I enter the parameters of the core, do the calculation for the converter on IR2153, and set the output voltage equal to the output voltage of my power supply. As a result, for two rings K45x28x8, for secondary voltage it is necessary to wind 12 turns. Motaems...

We start with the minimum frequency - you don’t have to worry about overloading the transistor - the current limiter will work. We stand on terminals X1 with an oscilloscope, gradually increase the frequency and observe the following picture:

Next, we create a proportion in Excel to calculate the number of turns in the primary winding. The result will differ significantly from the calculations in the program, but we understand that the program takes into account both pause time and voltage drop on power transistors and rectifier diodes. In addition, an increase in the number of turns does not lead to a proportional increase in inductance - there is a quadratic dependence. Therefore, an increase in the number of turns leads to a significant increase in inductive reactance. Programs also take this into account. We will do not much differently - to correct for these parameters in our table, we introduce a decrease of 10% in the primary voltage.
Next we construct a second proportion by which it will be possible to calculate the required number of turns for secondary voltages.
Before the proportions with the number of turns, there are two more plates with which you can calculate the number of turns and the inductance of the output choke of the welding machine, which is also quite important for this device.

In this file the proportions are at SHEET 2, on SHEET 1 calculations of switching power supplies for a video about calculations in Excel. I decided to give free access after all. The video in question is here:

A text version on how to compile this table and the initial formulas.

We finished the calculations, but there was a wormhole left - the design of the stand, as simple as three kopecks, showed quite acceptable results. Can I assemble a full-fledged stand powered directly from the 220 network? But galvanic connection to the network is not very good. And removing the energy accumulated by inductance using a linear transistor is also not very good - you will need a VERY powerful transistor with a HUGE heatsink.
Okay, you don't need to think much...

We seem to have figured out how to find out the saturation of the core, let’s choose the core itself.
It has already been mentioned that I personally am too lazy to look for and buy W-shaped ferrite, so I take out my box of ferrites from line transformers and select ferrites of the same size. Then I make a mandrel specifically for one core and wind 30-40 turns on it - the more turns, the more accurate the inductance measurement results will be. I need to choose the same cores.
Having folded the resulting ones into an W-shaped structure, I make a mandrel and wind a test winding. Having recalculated the number of turns of the primary, it turns out that the overall power will not be enough - Barmalei contains 18-20 turns of the primary. I take larger cores - left over from some old blanks - and a couple of hours of stupidity begin - checking the cores according to the method outlined in the first part of the article, the number of turns is even greater than that of a quad core, but I used six sets and the size is much larger...
I'm getting into the calculation programs of "Old Man" - aka Denisenko. Just in case, I drive in a double core Ш20х28. Calculation shows that for a frequency of 30 kHz the number of turns of the primary is 13. I admit the idea that the “extra” turns are wound to prevent 100% saturation, and the gap also needs to be compensated.

Before introducing my new cores, I recalculate the area of ​​the round edges of the core and derive the values ​​​​for the supposedly rectangular edges. I do the calculation for a bridge circuit, since in a single-cycle converter ALL available primary voltages are applied. Everything seems to fit - you can get about 6000 W from these cores.

Along the way, it turns out that there is some kind of mistake in the programs - completely identical data for the cores in the two programs give different results - ExcellentIT 3500 and ExcellentIT_9 broadcast different power of the resulting transformer. The difference is several hundred watts. True, the number of turns of the primary winding is the same. But if the number of turns of the primary is the same, then the overall power should be the same. Another hour already increased stupidity.
In order not to force visitors to search for Starichka’s programs, he collected them in one collection and packaged them in one archive, which can be DOWNLOADED. Inside the archive are almost all the programs created by the Old Man that we could find. I also saw a similar collection on some forum, but I don’t remember which one.
To solve the problem that has arisen, I am re-reading Biryukov’s article again...
I turn the oscilloscope onto the resistor in the source circuit and begin to observe changes in the shape of the voltage drop across different inductances.
At small inductances, there actually is an inflection in the shape of the voltage drop across the source resistor, but on a quad core from TDKS it is linear at least at a frequency of 17 kHz, at least at 100 kHz.
In principle, you can use data from calculator programs, but hopes were placed on the stand and they really fell apart.
I slowly fold back the turns on the gear core and run it on the stand, observing the changes in the oscillograms. Really some bullshit! The current is limited by the stand even before the voltage curve begins to bend...
It’s not possible to get by with little expense - even if you increase the current limit to 1A, the voltage drop across the source resistor is still linear, but a pattern appears - having reached a certain frequency, the current limit turns off and the pulse duration begins to change. Still, the inductance is too high for this stand...
All that remains is to check my suspicions and wind a test winding of 220 volts and...
I take my monster out from the shelf - I haven’t used it for a long time.

Description of this stand with a drawing of a printed circuit board.
I understand perfectly well that assembling such a stand for the sake of assembling a welding machine is quite a labor-intensive task, so the given measurement results are only an intermediate result in order to have at least some idea of ​​what cores can be used and how. Further, during the assembly process, when the printed circuit board for the working welder is ready, I will once again double-check the results made in these measurements and try to develop a method for error-free winding of a power transformer using the finished board as a test stand. After all, a small stand is quite functional, but only for small inductances. You can, of course, try to play around with the number of turns, reducing them to 2 or 3, but even reversing the magnetization of such a massive core requires a lot of energy and you can’t get away with a 1 A power supply. The technique using the stand was rechecked using a traditional core Ш16х20, folded in half. Just in case, the dimensions of W-shaped domestic cores and recommended replacements with imported ones have been added.
So, although the situation with the cores has become clearer, just in case the results will be rechecked on a single-cycle inverter.

In the meantime, let's start making a harness for the transformer of the welding machine. You can make a tourniquet, you can glue a tape. I have always liked tapes more - they are, of course, superior to bundles in terms of labor intensity, but the winding density is much higher. Therefore, it is possible to reduce the tension in the wire itself, i.e. In the calculation, do not include 5 A/mm2, as is usually done for such toys, but, for example, 4 A/mm2. This will significantly facilitate the thermal regime and most likely make it possible to obtain a PV equal to 100%.
PV is one of the most important parameters of welding machines, PV is P duration IN inclusions, i.e. time of continuous welding at currents close to maximum. If the duty cycle is 100% at maximum current, then this automatically transfers the welding machine to the professional category. By the way, even for many professional ones, the PV is 100% only with an output current equal to 2/3 of the maximum. They save on cooling systems, but I think I’m going to make a welding machine for myself, therefore I can afford much larger areas of heat sinks for semiconductors, and make the transformer have an easier thermal regime...

A welding inverter is a fairly popular device that is necessary both in the household and in an industrial enterprise. This is not surprising, because the power sources that were used before (converters, transformers, rectifiers) had many disadvantages. Among them are weight and dimensions, high energy consumption, but a small range of control of the welding mode and low conversion frequency. By making a welding inverter using thyristors with your own hands, you will receive a powerful power supply for the necessary work. This will also help you save a lot of money, although it will still require certain labor and material costs.

Welding inverter: features and functions of the device

The inverter's job is to convert alternating mains current into its direct high-frequency counterpart.

This happens in several stages. Current flows to the rectifier unit from the network. There, after transformation, the voltage changes from alternating to constant. And the inverter performs a reverse conversion, that is, the incoming DC voltage again becomes alternating, but with a higher frequency. After this, the voltage is reduced by a transformer, and this parameter is modified into a high-frequency DC voltage through the output rectifier.

The design of the welding inverter and its features

Due to the fact that there are no heavy parts in the design of the device, it is very compact and lightweight. It includes the following components:

The design of a simple cross-coupled inverter.

• inverter;
• network and output rectifiers;
• throttle;
• high frequency transformer.

Even novice welders can work with such machines. They are used both in everyday life and in the construction industry or in car services. Due to the fact that there is adjustment of operating modes, you can cook both thin and thick metals. And the increased conditions of arc combustion and weld formation give you the opportunity to weld any alloys, ferrous and non-ferrous metals using welding inverters, using all possible welding technologies.

Benefits of using an inverter

In the field of welding equipment, such devices are in particular demand due to their many advantages and advantages. By making an inverter with your own hands, you will receive:

• the ability to weld complex non-ferrous metals and structural steels;
• protection against overheating, mains voltage fluctuations, and current overloads;
• high stability of the welding current even though the voltage may fluctuate in the network;
• high-quality seam;
• There will be virtually no spatter during welding;
• arc burning will be stabilized in a given key, even if external adverse influence is observed;
• many other useful functions.

DIY inverter circuits

Taking as a basis how the circuit is built and how the inverter conversion process itself is controlled, there are several types of devices that are the most common in use. The full bridge and half bridge options refer to two push-pull circuits, and the “oblique” bridge refers to a single-stroke circuit. A full bridge circuit, called push-pull, operates with bipolar pulses. They are fed to key transistors (which are paired), and they lock and open the electrical circuit.

Slant bridge inverter circuit.

The half-bridge circuit will differ from the previous version in that its current consumption is increased. Transistors operating on the same push-pull model act as keys. Each of them is supplied with half of the input mains voltage. The power of the inverter, in comparison with the current of a full bridge, is half the value. This scheme has its advantages in low-power devices. In addition, you can use a group of transistors, rather than one very powerful one.

The last option is an “oblique” bridge. These are inverters that operate on a single-cycle principle. Here you will be dealing with unipolar impulses. Simultaneous opening of the transistor switches will eliminate the possibility of a short circuit. But among the disadvantages of this scheme is the magnetization of the transformer magnetic circuit.

Look at one of the standard inverter circuits. This is a design designed by Yu. Negulyaev. To assemble such a device at home, you will need your desire, readiness to work and the necessary element base, which you can either find on the radio market or remove from old household appliances.

Instructions for assembling the device

Standard inverter circuit designed by Yu. Negulyaev

Take a 6mm duralumin plate. Connect all heat-emitting conductors and wires to it. Please note that here the wire does not need to be encircled with thermal insulating material. Using an old circuit (for example, a computer), you do not have to look for transistors and thyristors separately.

Next, prepare a special high-power fan (you can even use a car radiator). It will blow air over everything, including the resonant choke. Be sure to press the latter onto your base using a spacer seal.

To make the throttle device itself, take six copper cores. You can find them on the market or make them yourself from parts of an unnecessary old TV. Press the diodes to the base of the circuit, and then attach voltage regulators and insulation seals to them.

When installing the transformer, insulate the conductor bundles using electrical tape or fluoroplastic strip. Place the conductors in different directions so that they do not come into contact and cause malfunctions. You will need to install a force field on the field-effect transistor to extend the performance of your inverter. To do this, take a copper wire with a 2 mm cross-section. Having tinned it, wrap it in several layers with regular thread. This way you will protect your conductor from various damages both during soldering and welding. To secure the installation, use insulating heels. This way you will also transfer the load from the transistors to them.

Installations for automatic welding of longitudinal seams of shells - in stock!
High performance, convenience, ease of operation and reliability in operation.

Welding screens and protective curtains - in stock!
Radiation protection when welding and cutting. Big choice.
Delivery throughout Russia!

Most often, when constructing welding inverters, three main types of high-frequency converters are used: half-bridge, asymmetric bridge (or “oblique bridge”) and full bridge. Under the guise of a half bridge and a full bridge, there are resonant converters. Depending on the control system for the output parameters, converters are available with PWM (pulse width), with PFM (frequency control), with phase control, and combinations of these three. All these types of converters have their advantages and disadvantages. Let's start with a half bridge with PWM. The block diagram of such a converter is shown in Fig. 3.

This is the simplest converter from the two-stroke family, but no less reliable. The disadvantage of this circuit is that the voltage “swing” on the primary winding of the power transformer is equal to half the supply voltage. But on the other hand, this fact is a plus; you can use a smaller core without fear of entering saturation mode.

For low-power inverters (2-ZkW), such a converter is very promising. But PWM control requires special care when installing power circuits; drivers must be installed to control power transistors. Transistors of such a half-bridge operate in hard switching mode, so increased demands are placed on control signals.

There must be a “dead time” between two antiphase pulses, the absence of a pause, or its insufficient duration, always leads to the occurrence of a through current through the power transistors.

The consequences are easily predictable - failure of transistors. A very promising type of half-bridge converter is the resonant half-bridge. The block diagram of such a half-bridge is shown in Fig. 4.

The current flowing through the power circuits has a sinusoidal shape, and this removes the load from the filter capacitors.

With this design, power switches do not require drivers! An ordinary pulse transformer is enough to switch the power transistors. The quality of control pulses is not as significant as in a PWM circuit, although there should be a pause (“dead time”).

Another plus is that this circuit allows you to do without current protection and the shape of the current-voltage characteristic (volt-ampere characteristic) has an immediately falling form and does not require parametric shaping.

The output current is limited only by the magnetizing inductance of the transformer and can reach significant values ​​during a short circuit; this must be taken into account when choosing output diodes, but this property has a positive effect on the ignition and burning of the arc!

Typically, the output parameters are regulated by changing the frequency, but the use of phase control gives much more advantages and is the most promising for a welding inverter, since it allows you to bypass such an unpleasant phenomenon as the coincidence of resonance with the short-circuit mode, and the range of adjustment of the output parameters is much wider. Phase adjustment allows you to change the output current practically from 0 to Imax.

The next scheme is an asymmetric bridge, or “oblique bridge”. The block diagram of such a converter is shown in Fig. 5.

The asymmetric bridge is a single-cycle, forward-flow converter.

A converter of this configuration is very popular both among manufacturers of welding inverters and among radio amateurs. The first welding inverters were built exactly like an “oblique bridge”. Simplicity and reliability, ample opportunities for adjusting the output current, noise immunity - all this still attracts developers of welding inverters.

And although the disadvantages of such a converter are quite significant, these are large currents through transistors, high requirements for the shape of control pulses, which implies the use of powerful drivers to control power switches, high requirements for the installation of power circuits, large pulse currents place high demands on input filter capacitors, electrolytic Capacitors really don't like large pulse currents. To keep transistors in the ODZ (permissible value range), RCD chains (snubbers) are required.

But, despite all these shortcomings and low efficiency, the “oblique bridge” is still used in welding inverters to this day. Transistors T1 and T2 operate in phase, opening together and closing together. The energy is not stored in the transformer, but in the output inductor of the inductor. The duty cycle does not exceed 50%, which is why to obtain the same power with a bridge converter, double current through the transistors is required. The operation of such a converter will be examined in more detail using the example of a real welding inverter.

The next type of converter is a full bridge with PWM. Classic push-pull converter! The block diagram of the full bridge is shown in Fig. 6.

The bridge circuit makes it possible to obtain power 2 times more than a half-bridge, and 2 times more than an “oblique bridge”, with the same values ​​of currents and switching losses. This is explained by the fact that the “swing” of the voltage of the primary winding of the power transformer is equal to the supply voltage.

Accordingly, to obtain the same power, for example, with a half-bridge (in which the drive voltage is 0.5U supply), the current through the transistors will be 2 times less! Full bridge transistors operate diagonally when T1 - T3 are open, T2 - T4 are closed, and vice versa. The current transformer monitors the amplitude value of the current flowing through the switched on diagonal. You can regulate the output current of such a converter in two ways:

1) change the duration of the control pulse, leaving the cutoff voltage unchanged;

2) change the level of cutoff voltage coming from the current transformer, leaving the duration of the control pulses unchanged.

Both of these methods allow you to change the output current within a fairly wide range. The disadvantages and requirements of a full bridge with PWM are exactly the same as those of a half bridge with PWM. (See above). And finally, let's consider the most promising RF converter circuit for a welding inverter - a resonant bridge. The block diagram is shown in Fig. 7.

As it may seem at first glance, the resonant bridge circuit is not very different from a PWM bridge, and this is true. In practice, only an LC resonant circuit is additionally introduced, connected in series with the power transformer. However, the introduction of this chain completely changes the processes of power transfer. Losses are reduced, efficiency increases, the level of electromagnetic interference is reduced by orders of magnitude, and the load on the input electrolytes is reduced. As you can see, you can completely remove current protection; power transistor drivers may only be needed if MOSFET transistors with a gate capacitance greater than 5000pF are used. For IGBT transistors, one pulse transformer is sufficient.

The output current of the resonant converter can be controlled in two ways: frequency and phase. Both of them were mentioned earlier, in the description of the resonant half-bridge. And the last type of RF converter is a full bridge with a leakage choke. Its circuit is practically no different from the circuit of a resonant bridge (half-bridge), just like the LC circuit is connected in series with a transformer, only it is not resonant. C = 22 µFx63V works as a balancing capacitor, and L of the inductor acts as a reactance, the value of which linearly depends on the frequency. The control of such a converter is frequency. As the frequency increases, the resistance L increases. The current through the power transformer decreases. Simple and reliable. Most industrial inverters are built on this principle of adjusting and limiting the output current.