THE BELL

There are those who read this news before you.
Subscribe to get the latest articles.
Email
Name
Surname
How would you like to read The Bell
No spam

Good day everyone! I’m looking at diagrams on the Internet of pulse power supplies and ... And I don’t understand! Tolley the authors do not read the "Datasheet" for the components, or do they specifically discourage the desire to assemble a UPS ??? . We look at the description of IR2153: "an improved version of IR2153 -2155, the list of improvements comes down to noise protection ... We read: the recommended load capacitance is 1000 pF, the power is 0.650 W (short-term)! So this is the data on IR2151 !!! And so we have: IR2153 can control keys with a capacitive load of 1n=1000pf!Look at the "datasheet" of the keys.IR740 is 1450pf.One and a half times higher than the recommended.Now the voltage.Recommended maximum voltage of the keys is 600v(v)!And the keys have 400v.Well, yes, this is more than 310 V! However, everyone who has come across industrial UPS circuits is well aware that the keys are set to a voltage of at least 600 V. Only in Chinese circuits sometimes burnt-out 500 V appear. I hope I explained it clearly?! As for the current of the key, and resistance the key in the open state. This has little effect on the power of the UPS. Let me explain. For a switching power supply, the current is limited by passing through the load and, as a rule, does not exceed 2-3 A in an impulse. In an impulse! We look at the "datasheet" of the keys and see: at a crystal temperature of 100 gr. current with a large margin for IR740. However, in this case it is for the key minus! The higher the key current, the longer the switching time (see the graph in the same place) and, of course, the lower the pulse steepness, which means that the efficiency is less than the maximum (75%). Accordingly, this key will work, but badly!!! As a result of the above: this combination leads to burnout of both the keys and the driver! Anyone who wants to repeat this scheme is doomed to a handful of burnt parts! I am wrong? Read the comments for similar charts. The question follows: you are so smart, so what would you advise? I advise everyone who wants to have a simple UPS assembly, take the circuit from the description and recommendation of the IR Company - the IR2153 driver with keys for a current of 4-5 a and max. voltage 600-900 V with a capacitance of the control electrode of not more than 1000 pF. Example STP5NK600C and similar MOSFET triodes. Now about the resistance in the open state for the key: indeed, the larger it is, the stronger the heating of the key. Someone will say and less efficiency. In this case, the efficiency is not 100% and the influence of the resistance is very small. So what affects efficiency? The efficiency is affected by the UPS circuit itself, for efficiency up to 94% we assemble a resonant UPS. Efficiency up to 75% - with the right keys on IR2153!. you have little such efficiency? Hm. What about a pulse transformer? How does it limit efficiency? Has anyone already counted? Losses at frequencies above 50 kHz increase many times, although up to 50 kHz the losses are not zero. We are looking at industrial circuits: winding pulse transformers is a very capricious task, two identically wound transformers have different inductances! What is this? And this is what it is! Each IT has its own optimal operating frequency. And how about you? That's all - read on and see the diagrams of the UPS of TVs, powerful amplifiers, and other factory electrical appliances. Good luck to you!

The power supply is built on a floor-bridge circuit based on the IR2153 chip. At the output of this block, you can get any voltage you need, it all depends on the parameters of the secondary winding of the transformer.

Let's take a closer look at the switching power supply circuit.

The power of the power supply with just such components is about 150 watts.

Mains alternating voltage through a fuse and a thermistor is supplied to the diode rectifier.

After the rectifier, there is an electrolytic capacitor, which, at the moment the unit is connected to the network, will be charged with a large current, the thermistor just limits this current. A capacitor is needed with a voltage of 400-450 volts. Further, a constant voltage is supplied to the power switches. At the same time, power is supplied to the IR2153 chip through a limiting resistor and a rectifier diode.

You need a powerful resistor, at least 2 watts, it is better to take a 5-watt one. The supply voltage for the microcircuit is additionally smoothed out by a small electrolytic capacitor with a capacity of 100 to 470 microfarads, preferably 35 volts. The microcircuit begins to generate a sequence of rectangular pulses, the frequency of which depends on the value of the components of the timing circuit, in my case, the frequency is in the region of 45 kHz.

A rectifier with a midpoint is installed at the output. Rectifier in the form of a diode assembly in the TO-220 package. If the output voltage is planned within 40 volts, then diode assemblies soldered from computer power supplies can be used.

The voltage boost capacitor is designed for the correct operation of the upper field switch, the capacitance depends on which transistor is used, but on average 1 uF is enough for most cases.

Before starting, you need to check the operation of the generator. For these purposes, about 15 volts of direct voltage is supplied from an external power source to the indicated pins of the microcircuit.
Next, the presence of rectangular pulses on the gate of the field keys is checked, the pulses must be completely identical, of the same frequency and filling.
The first start-up of the power supply must be done through a 220-volt safety incandescent lamp with a power of about 40 watts, be extremely careful not to touch the board during operation, after disconnecting the unit from the mains, wait a few minutes until the high-voltage capacitor is discharged through the appropriate resistor.
It is very important to point out that this circuit does not have short circuit protection, therefore any short circuits, even short ones, will lead to the failure of the power switches and the IR2153 chip, so be careful.

The main component of the power supply under consideration is the IR2153 chip (driver). This driver is available in two versions - IR2153 and IR2153D. The letter D means that the microcircuit is equipped with a diode designed to power the control stage of the upper switch. Thus, if the IR2153D driver is used in the circuit, then diode D2 does not need to be installed. The frequency of generation of this power supply is set by resistor R4 and capacitor C6 connected to the outputs of the microcircuit RT (leg 2) and CT (leg 3). The optimal generation frequency of the microcircuit is a frequency of 40 - 70 kHz, it is for this range that the core of the Tr1 transformer is selected. A feature of the microcircuit is the ability to stop generation by shorting the CT output to minus. This principle is used to organize the protection of the microcircuit from a short circuit at the output of this power source.

Schematic diagram of a switching power supply on IR2153

The principle of operation of the power supply

!
In this article, together with Roman (the author of the YouTube channel "Open Frime TV"), we will assemble a universal power supply on the IR2153 chip. This is a kind of "Frankenstein", which contains the best qualities from different schemes.

The Internet is full of power supply circuits on the IR2153 chip. Each of them has some positive features, but the author has not yet met a universal scheme. Therefore, it was decided to create such a scheme and show it to you. I think you can go straight to it. So, let's figure it out.


The first thing that catches your eye is the use of two high voltage capacitors instead of one for 400V. Thus we kill two birds with one stone. These capacitors can be obtained from old computer power supplies without spending money on them. The author specially made several holes in the board for different sizes of capacitors.








If the block is not available, then the prices for a pair of such capacitors are lower than for one high-voltage one. The capacitance of the capacitors is the same and should be at the rate of 1 uF per 1 W of output power. This means that for 300 watts of power output you will need a pair of 330uF capacitors.




Also, if we use this topology, there is no need for a second decoupling capacitor, which saves us space. And that is not all. The voltage of the decoupling capacitor should no longer be 600 V, but only 250 V. Now you can see the sizes of 250V and 600V capacitors.




The next feature of the circuit is the power supply for the IR2153. Everyone who built blocks on it faced unrealistic heating of the supply resistors.




Even if they are set from a break, a lot of heat is released. An ingenious solution was immediately applied, using a capacitor instead of a resistor, and this gives us the fact that there is no heating of the element by supply.


The author of this homemade product saw such a decision from Yuri, the author of the YouTube channel "Red Shade". Also, the board is equipped with protection, but in the original version of the circuit it was not.






But after tests on the layout, it turned out that there was too little space to install the transformer and therefore the circuit had to be increased by 1 cm, this gave extra space on which the author installed the protection. If it is not needed, then you can simply put jumpers instead of a shunt and do not install the components marked in red.




The protection current is regulated using this trimming resistor:


The shunt resistor values ​​vary depending on the maximum output power. The more power, the less resistance needed. For example, for power below 150 W, 0.3 ohm resistors are needed. If the power is 300 W, then we need 0.2 Ohm resistors, well, at 500 W and above, we put resistors with a resistance of 0.1 Ohm.


This block should not be assembled with a power higher than 600 W, and you also need to say a few words about the operation of the protection. She hiccups here. The trigger frequency is 50 Hz, this is because the power is taken from the AC, therefore the latch is reset at the mains frequency.




If you need a latched option, then in this case the power supply of the IR2153 chip must be taken constant, or rather from high-voltage capacitors. The output voltage of this circuit will be taken from a full-wave rectifier.


The main diode will be a Schottky diode in the TO-247 package, choose the current for your transformer.


If there is no desire to take a large case, then in the Layout program it is easy to change it to TO-220. There is a 1000 uF capacitor at the output, it is enough for any currents, since at high frequencies the capacitance can be set less than for a 50 hertz rectifier.




It is also necessary to note such auxiliary elements as snubbers (Snubber) in the transformer piping;


smoothing capacitors;


as well as a Y-capacitor between the grounds of the high and low sides, which dampens noise on the output winding of the power supply.


There is an excellent video about these capacitors on YouTube (the author attached the link in the description under his video (link SOURCE at the end of the article)).


You can not skip the frequency-setting part of the circuit.


This is a 1 nF capacitor, the author does not recommend changing its value, but he put a tuning resistor in the driving part, there were reasons for this. The first of them is the exact selection of the desired resistor, and the second is a small adjustment of the output voltage using the frequency. And now a small example, let's say you are making a transformer and you see that at a frequency of 50 kHz the output voltage is 26V, and you need 24V. By changing the frequency, you can find a value at which the output will be the required 24V. When installing this resistor, we use a multimeter. We clamp the contacts into crocodiles and rotate the resistor knob, we achieve the desired resistance.




Now you can see the 2nd breadboards on which the tests were carried out. They are very similar, but the protection board is slightly larger.


The author made mock-ups in order to order the manufacture of this board in China with peace of mind. In the description under the author's original video, you will find an archive with this board, schematic and seal. There will be two scarves and the first and second options, so you can download and repeat this project.

After ordering, the author was looking forward to the board, and now they have arrived. We open the package, the boards are packed well enough - you won’t find fault. We visually inspect them, everything seems to be fine, and immediately proceed to soldering the board.








And now she is ready. Everything looks like this. Now let's quickly go through the main elements not previously mentioned. First of all, these are fuses. There are 2 of them, on the high and low side. The author used such round ones, because their sizes are very modest.




Next we see the filter capacitors.


You can get them from an old computer power supply. The author wound the choke on the t-9052 ring, 10 turns with a 0.8 mm 2 wire, but you can use a choke from the same computer power supply.
Diode bridge - any, with a current of at least 10 A.


There are also 2 resistors on the board to discharge the capacitance, one on the high side, the other on the low side. Tell in:

For a long time I was worried about the topic of how you can use a power supply from a computer as a power amplifier. But remodeling the power supply is still fun, especially a pulsed one with such a dense installation. Although I am accustomed to all kinds of fireworks, I really didn’t want to scare my family, and it’s also dangerous for myself.

In general, the study of the issue led to a fairly simple solution, requiring no special details and almost no adjustment. Collected-turned-works. Yes, and I wanted to practice etching printed circuit boards using photoresist, since recently modern laser printers have become greedy for toner, and the usual laser-ironing technology did not work out. I was very pleased with the result of working with the photoresist - for the experiment, I etched the inscription on the board with a line 0.2 mm thick. And she turned out great! So, enough preludes, I will describe the scheme and the process of assembling and adjusting the power supply.

The power supply is actually very simple, almost all of the parts left after disassembling the not-so-good impulse from the computer are assembled - from those that are not “reported” to. One of these parts is a pulse transformer, which can be used without rewinding in a 12V power supply, or recalculated, which is also very simple, for any voltage, for which I used the Moskatov program.

Block diagram of a switching power supply:

The following were used as components:

ir2153 driver - a microcircuit used in pulse converters to power fluorescent lamps, its more modern counterpart is ir2153D and ir2155. In the case of using ir2153D, the VD2 diode can be excluded, since it is already built into the microcircuit. All microcircuits of the 2153 series already have a built-in 15.6V zener diode in the power circuit, so you should not bother too much with the device of a separate voltage regulator to power the driver itself;

VD1 - any rectifier with a reverse voltage of at least 400V;

VD2-VD4 - "high-speed", with a short recovery time (no more than 100ns) for example - SF28; In fact, VD3 and VD4 can be excluded, I did not set them;

as VD4, VD5 - a dual diode from a computer power supply "S16C40" is used - this is a Schottky diode, you can put any other, less powerful one. This winding is needed to power the ir2153 driver after the switching converter starts up. You can exclude both diodes and winding if you do not plan to remove power more than 150W;

Diodes VD7-VD10 - powerful Schottky diodes, for a voltage of at least 100V and a current of at least 10 A, for example - MBR10100, or others;

transistors VT1, VT2 - any powerful field, the output depends on their power, but you should not get carried away here much, as well as remove more than 300W from the unit;

L3 - wound on a ferrite rod and contains 4-5 turns of 0.7mm wire; This chain (L3, C15, R8) can be excluded altogether, it is needed to slightly facilitate the operation of transistors;

The L4 inductor is wound on a ring from the old group stabilization inductor of the same power supply from the computer, and contains 20 turns each, wound with a double wire.

Capacitors at the input can also be supplied with a smaller capacity, their capacity can be roughly selected based on the power output of the power supply, approximately 1-2 microfarads per 1 W of power. Do not get carried away with capacitors and put capacitances greater than 10,000 microfarads on the output of the power supply, as this can lead to a "salute" when turned on, since they require significant current to charge when turned on.

Now a few words about the transformer. The parameters of the pulse transformer are determined in the Moskatov program and correspond to an E-shaped core with the following data: S0 = 1.68 sq. cm; Sc = 1.44 sq. cm; Lav.l. = 86cm; Conversion frequency - 100kHz;

The resulting calculated data:

Winding 1- 27 turns 0.90mm; voltage - 155V; Wound in 2 layers with a wire consisting of 2 cores of 0.45 mm; The first layer - inner contains 14 turns, the second layer - outer contains 13 turns;

winding 2- 2 halves of 3 turns with a wire of 0.5 mm; this is a “self-powered winding” for a voltage of about 16V, it is wound with a wire so that the winding directions are in different directions, the middle point is brought out and connected to the board;

winding 3- 2 halves of 7 turns, wound with the same stranded wire, first - one half in one direction, then through the insulation layer - the second half, in the opposite direction. The ends of the windings are brought out into the "braid" and connected to a common point on the board. The winding is designed for a voltage of about 40V.

In the same way, you can calculate the transformer for any desired voltage. I have assembled 2 such power supplies - one for the amplifier on the TDA7293, the second - for 12V to power all kinds of crafts - is used as a laboratory one.

Power supply for the amplifier for voltage 2x40V:

12V switching power supply:

Power supply assembly in the case:

A photo of testing a switching power supply - that for an amplifier using a load equivalent of several MLT-2 resistors of 10 ohms, included in a different sequence. The goal was to get data on power, voltage drop and voltage difference in the arms +/- 40V. As a result, I got the following parameters:

Power - about 200W (I no longer tried to shoot);

voltage, depending on the load - 37.9-40.1V in the entire range from 0 to 200W

Temperature at maximum power 200W after a test run for half an hour:

transformer - about 70 degrees Celsius, diode radiator without active blowing - about 90 degrees Celsius. With active blowing, it quickly approaches room temperature and practically does not heat up. As a result, the radiator was replaced, and in the following photos the power supply is already with a different radiator.

When developing the power supply, materials from the vegalab and radiokot sites were used, this power supply is described in great detail on the Vega forum, there are also options for a block with short circuit protection, which is not bad. For example, with an accidental short circuit, the track on the board in the secondary circuit instantly burned out

Attention!

The first power supply should be turned on through an incandescent lamp with a power of not more than 40W. When you first turn on the network, it should flash for a short time and go out. It shouldn't glow at all! In this case, you can check the output voltages and try to lightly load the unit (no more than 20W!). If everything is in order, you can remove the light bulb and start testing.

THE BELL

There are those who read this news before you.
Subscribe to get the latest articles.
Email
Name
Surname
How would you like to read The Bell
No spam