Pulse power supply circuits for beginner radio amateurs. How a simple and powerful switching power supply works. Schematic diagram of connection between DSN-VC288 and lm2596

The steadily increasing demand for these products testifies to how urgently repairers and radio amateurs need laboratory power supplies for their workshops. Since in practice it is often necessary to have several power supplies, and the prices for finished products are too high, not to mention the scarcity laboratory power supplies, then it is profitable to independently manufacture such power supplies.

This article discusses a very simple and easily repeatable power supply design that does not contain scarce or expensive components, which allows it to be manufactured by any radio amateur. The presence in this power supply of the function of stabilizing the current in the load seriously expands the possibilities of using it in practice. And not only during repair and laboratory work, but also charging a wide variety of batteries. If we take into account that the power supply in question provides the ability to smoothly adjust the stabilized current in the load (from the minimum value to the maximum), then the scope of use of this power supply becomes very extensive.

The complicated circuitry of most modern power supplies (PSUs) prevents its practical implementation, since its implementation requires time and material resources, and this in our time is perhaps the main factor hindering self-production complex structures. This design does not require large expenses.

The voltage stabilizer (SV) circuit can operate both in voltage stabilization mode and in current stabilization mode.

The output stabilized voltage is set within 0...18 V. The output stabilized current (in current stabilization mode) is set within 0...14 A).

The main disadvantage of many circuits with current stabilization is that after disconnecting the MV from the power supply, a constant voltage appears at the output, close in value to the input voltage of the MV! And while the battery of oxide capacitors of the MV bridge rectifier is being discharged (through the MV load), it is unknown what can happen in the equipment connected to the output of such a power supply. The most unpleasant thing is that there is not even a mention of this negative phenomenon, nor of possible options for eliminating this shortcoming. In the circuit diagram of this SN, a circuit design option for solving this problem is proposed.

When developing this design, the following principles were taken into account:

1. You should not strive to complicate the schemes of your structures, if you do not forget about possible repair work, which, sooner or later, will still have to be carried out.

2. It is better to spend money on purchasing more modern components if they simplify the design of the power supply rather than having to go through the trouble of manufacturing complex power supply designs using a large number of components.

3. It makes sense to manufacture several copies of the power supply unit, even if they are not required in the foreseeable future. At a minimum, you need to have several power supplies made for different output currents and voltages. The widespread use of one powerful power supply leads to its accelerated failure.

Power supply diagram

Taking into account the above, a circuit was developed on the basis of which power supplies of various powers can be manufactured. The diagram of the SN under consideration is shown in Fig. 1. The basis of this CH is operational amplifier(O-Amp) type LM358N.

These op-amps have become very common due to their ability to operate in a special mode with a unipolar supply voltage. Last but not least, the spread of these op amps was facilitated by their widespread use in a variety of designs of small-sized digital multimeters.

The voltage stabilizer itself is made on half of this op-amp DA1. 1. The second op-amp DA1.2 has protection for the output current MV.

Let's consider the purpose of the main elements of the circuit and the features of the ratings of some of its parts. As can be seen from Fig. 1, the op-amp is powered directly from one common rectifier of the power supply. Thanks to the use of this type of op-amp, it was possible to avoid complicating the MV circuit as a whole. That is, due to the absence of the necessary negative (relative to the common power bus) voltage source to power the op-amp, it was possible to further simplify the MV circuit. Thanks to the use of a programmable (precision) zener diode (chip) type TL431, it was possible to simplify the circuit of the reference voltage source (VS). It turned out to be possible to refuse any stable current generators (GCT) powering this zener diode.

The reference voltage is removed from the reference voltage, made on an IC of type TL431 (VD1) and from the variable resistor R4, which is a regulator of the CH output voltage, and is supplied to the non-inverting input of the op-amp DA1. 1. Part of the output voltage is supplied to the inverting input (pin 2 of the op-amp) DA1.1, which is removed from the resistor voltage divider R8R6.

From the same ION, the voltage is also removed to the MV electronic protection unit, which is made on the second half of the LM358N (DA1.2) and through the resistor voltage divider R11R14 is supplied to the variable resistor R12, which is a regulator for setting the required value of the maximum MV output current.

Thus, the inputs of this op-amp are connected to a powerful resistor R17, which is a current sensor for the MV protective unit. The magnitude of the CH limiting current (output stable current CH) depends on the voltage on the motor of variable resistor R12 and on the resistance of resistor R17.

The greater the value of this voltage and the lower the resistance of the current sensor resistor R17, the greater the value of the output current CH will be.

The circuit based on op-amp DA1.2 is a voltage comparator that compares the reference voltage on the motor of the variable resistor R12 with the voltage drop on the current sensor - R17. More precisely, the comparator compares these voltages in magnitude, and depending on which of the voltages is larger in magnitude, the voltage value at the output of this op-amp also changes. When the output current is below the operating threshold of the comparator (depending on the position of the resistor R12 slider), then the voltage at the inverting input of the op-amp is less than that at the resistor R17, and therefore at the non-inverting input of the op-amp. At the output of the op-amp, there is a low voltage (no more than 0.1...0.2 V), insufficient to open transistors VT3 and VT4. In this case, the HL1 LED, which is an indicator of the operation of the protection unit, does not light up and the protection does not have any effect on limiting the MV output current.

As soon as the voltage at the current sensor R17 exceeds the voltage at the inverting input of the op-amp (approximately the value of the bias voltage of the op-amp), the comparator will change its state and a large voltage will appear at its output, approaching in value the supply voltage of the op-amp (minus approximately 1.5V) . The protection transistor VT3 will turn on and, with its open collector-emitter junction, it will close the connection point of resistors R9 R10 to the common wire of the CH circuit. Powerful base composite transistor VT1-VT2 turns out to be de-energized and connected to the common CH wire. Since the emitter VT2, one way or another (using an external load CH or through a stable current generator on transistor VT5) is already connected to the common wire of the CH circuit, the composite transistor is forcibly closed. Depending on the situation (MV load), on the magnitude of the output current and voltage, the output produces a voltage stabilization mode or a current stabilization (limitation) mode.

As can be seen from the diagram, only some types of op-amps will be able to operate normally in a similar mode with a unipolar supply voltage, since a conventional op-amp will require setting the “midpoint” of the supply voltage at its inputs, which will certainly lead to the appearance of about half of its supply voltage at the output of the op-amp. This, in turn, will disrupt the functioning of the protection as a whole. Obviously, the easiest way to adapt a conventional op-amp in this circuit is to use a bipolar supply voltage for it.

With single-pole power supply, op-amps such as, for example, LM324N can also work. There are four op amps in one case of this op amp. According to the source, the internal circuitry of the op-amps in question is similar. On LM324N you can also try to assemble this CH according to the diagram in Fig. 1. The main requirement for the op-amp in the DA1.2 comparator circuit is that its output has a minimum voltage when the protection is not turned on. In principle, similar requirements are put forward in relation to the op amp of the SN DA1.1 itself. Only by fulfilling this requirement can you ensure reliable locking of the protective transistor VT3. Very important comments are in order here.

Real “pitfalls” await us in the process of purchasing foreign components, including those with LM358N type ICs, where defects can be very diverse. Many defects of these op-amps appear only after they are installed in a working structure. If experiments are carried out with such copies of the LM358N, then people often attribute failures during prototyping (practical design) to other facts, for example, to “crude” (imperfect) circuitry in the designs used. But in fact, the used LM358N had a “hidden” defect and simply failed. It is very important to test the LM358N before installing it on the PCB.

The most common defect in op-amps such as the LM358N is a complete (obvious) malfunction of one of the two op-amps, when, for example, there is no voltage at the output of one op-amp. It does not appear for any combination of voltages at the op-amp inputs. This is the most typical situation. There were also instances of LM358N in which output voltage exceeded the “zero” value and ranged from zero to several volts. Less common were LM358N instances with “uncontrolled” (by inputs) output voltages from 1 V and up to almost the full value of the LM358N supply voltage.

Hidden and unexpected defects of the LM358N are those in which the output stage of the LM358N fails, most often a “break” of the output stage, and before the output current of the LM358N reaches 5 mA. It was clearly noted that op-amps cease to fail if the output current of the LM358N is limited to the level of 3A. It became obvious that there was a point in further minimizing the output current of the LM358N. Without doubting that it makes sense to always use an op-amp with its output current not exceeding 3mA.

Using transistors VT3 and VT4 in the CH circuit (Fig. 1), we achieved a solution to the described problem LM358N.

Recommendations that save the low-quality output stage of the LM358N from probable failure are also used in relation to the DA1.1 op-amp, where its output stage operates on a fairly high-resistance load, represented by resistor R9, the right terminal of which is connected to the common wire if the protection is triggered. This case is the most “heavy” for the DA1.1 output stage, but this mode of operation of the op-amp occurs only when the MV is operating in the GTS mode. In normal MV operation mode, the load on op-amp DA1.1 is further reduced (the load resistance increases). Now the op-amp operates on the total resistance of resistors R9, R10 and the input resistance of the composite Darlington transistor VT1, VT2. The last component is formed by the base current VT1, VT2, which is insignificant at the MV load current, for which the MV circuit was originally designed (up to ZA).

The base current of transistor VT1 does not exceed hundreds of microamps in the most unfavorable set of circumstances, when the MV load current is maximum and the DC gain of the Darlington transistors is minimal. It was the large gain of these transistors, with an appropriate margin, that made it possible to radically increase the resistance of resistor R9 without fear of a significant violation of the characteristics of the CH.

The proposed construction of the MV circuit has one more positive quality, which consists in reliable operation of the protection unit. The situation is such that op-amp DA1.1, involved in the voltage regulation circuit, does not participate in the current regulation (limitation) loop (circuit).

This excludes the DAI.1 op-amp from the protection path, which has a beneficial effect on the performance of the protection as a whole. In the case when DAI.1 is controlled by comparator DA1.2 with current limitation, the situation will be different, not in favor of the above.

Capacitor C1, which closes the inverting input of the op-amp with its output, is an indispensable attribute in this CH circuit. Without it, the stable operation of the comparator, as well as the entire SN, will be disrupted. As a result, the comparator circuit is self-excited.

This phenomenon also has an impact on the circuit of the MV itself, even when the operating threshold of the comparator is far from the current value at the output of the MV.

Something similar is true in relation to the correction circuits of the op-amp DA1.1, namely in relation to the binding elements of the op-amp R7, C2.

In no way should we forget that the op-amp correction circuits connected between the input and output of the op-amp can pose a serious load on the output stage of the op-amp. The load is reactive, i.e. As the frequency increases, the load on the op-amp output increases. In our case, in relation to the LM358N, these correction circuits are a real threat to the output stage. Why is a rather high-resistance resistor R7 installed in series with the correction capacitor C2 in the circuit (Fig. 1). Here, such a small capacitance is not enough as in the comparator circuit on DA1.2.

If for some reason this circuit is not installed, the normal functioning of the MV circuit will be disrupted. The above is true with a small caveat. At direct current, the CH can work quite decently without the correction circuit R7C2. Stability can also be maintained when the MV operates at low frequencies (tens to hundreds of hertz), but with increasing frequency in the load, with pulsed current consumption by the load, the situation can change dramatically. However, even at low frequencies, “traces” of self-excitation will already appear at the output of DA1.1, i.e. This is a bit of a catch, since at the output of the MV itself everything may look quite decent, and it will be difficult to diagnose anything at the output of the MV with an oscilloscope.

If the amplitude of these pulsations is insignificant, then, as a rule, they are not paid attention to. Often, to observe RF generation in the MV, an oscilloscope with a wider bandwidth is needed (at least 10 MHz, and sometimes a 50 MHz device is required).

With the pulsed nature of the load, the situation changes dramatically, and from the hidden “sub-excitation” mode, op-amp DA1.1 can (depending on the frequency and parameters of the pulse signal) switch to the most common self-excitation mode, when the “additive” from op-amp DA1 begins to be superimposed on the test pulse . 1. This phenomenon is usually already clearly visible on the oscilloscope screen. That is why it never hurts to test any design under an impulse load. Only if we detect and obtain a “ringing” (“excitation”) mode of a stationary nature, will we be able to evaluate its parameters and eliminate it.

Often the problem occurs in a certain limited frequency range, in a particular operating mode, or with a load of a certain nature.

Some clarifications are also needed here. It's about not only about the technical consequences that result from a pulsed load at the MV output, but also, first of all, we mean a violation of operating modes (in the form of self-excitation, etc.) in the MV circuit itself.

This clarification is necessary in order to avoid confusion with those disturbances that result from a pulsed load only at the MV output, without disturbing the operating modes directly in the MV circuit.

The design of the SN can be arbitrary, it all depends on the parts used and the capabilities of the radio amateur. It should be remembered that the MV supply voltage of 30 V is close to the maximum permissible for the LM358N, the maximum permissible for which is 32 V. If you need to obtain a higher value of the MV output voltage, then some changes must be made to the MV circuit, which will be discussed below further stated.

The MV circuit in Fig. 1 allows you to use almost any available small-sized network transformer for the appropriate voltage, without resorting to winding additional windings. The choice of transformer depends entirely on the MV parameters.

Power supply parts

Resistors: R1 - 2.7 kOhm; R2, R5-R7, R15, R16 - 10 kOhm; R3 - 5.1 kOhm; R4, R12 - 33 kOhm; R8, R9 - 15 kOhm; RIO, R20, R21 - 4.7 kOhm; R11 - 10 kOhm; R13 - 1 kOhm (selected); R14 - 620 Ohm; R17 - 0.12 Ohm; R18, R19 - 30 kOhm; R22 - 30 Ohm.

Resistor R1 type MLT - 0.5 W; R4, R12 - SPZ-23v-A - 0.25 W; R11 - SPZ-38v; R17 - powerful (5 W) wire made foreign.

Capacitor CI, C2 - K10-176; C4 - 470 uF x 25V - K50-29V.

SN designs also contain several more capacitors, which are not shown in Fig. 1. One capacitor is soldered parallel to the supply pins of the LM358N (pins 4 and 8), its capacitance is in the range of 0.068...0.1 μF (ceramic). And the second capacitor is soldered parallel to the output terminals CH, its capacitance was selected in the range of 4.7...10 μF (1-2 pcs. K73-17x63V).

Oxide capacitor SZ (100 µF x 63 V) imported. It is soldered in parallel with the standard oxide capacitors of the bridge rectifier. Another such oxide capacitor is soldered parallel to the anode-cathode terminals of the TL431.

The ION - VD1 - TL431 microcircuit can be replaced with another integrated voltage stabilizer (taking into account its maximum permissible input voltage), without forgetting about the deterioration of the TKN in the ION. It is acceptable to use a precision zener diode, for example, D818E, but you must remember that the stability of such an ION will be entirely determined by the stability of the current through it. It is necessary to use a highly stable GTS (instead of resistor R1) if the D818E is powered from the main MV rectifier.

In the case where increased requirements are placed on the ION on the TL431 regarding the stability of the ION voltage, resistor R1 also needs to be replaced by the GTS. In this case, GST was performed according to the simplest scheme on one field effect transistor, type KPZZD, in the source circuit of which a 510 Ohm resistor is installed (selected to achieve a GST current approximately equal to 2 mA). The field-effect transistor must satisfy two important requirements: a voltage (drain-source and gate-drain) of at least 25 V and an initial drain current of at least 2 mA. This GST can be replaced with a bipolar version, assembled similarly to the GTS circuit in Fig. 1 on transistor VT5, with the only difference that the resistance of resistor R22 was increased 10-15 times until the required GTS current was obtained, and instead of a medium or high power transistor in the new GST used low-power KT315B (G), as well as BC547S or KT3102 with any letter index.

The GTS circuit when powering the TL431 is especially helpful when the MV was manufactured for a current of 6A or more, since with a high MV current, increased voltage drops appear on the rectifier from which the circuit of our ION is powered. It is intended to minimize the current instability through the ION caused by these voltage sags. scheme additional GTS.

Hence the importance of all circuit “little things” without exception.

Transistors VT3 and VT4 type KT315G (for the installation of which the printed circuit board CH is designed) or any other silicon with ike.max of at least 35 V and h2ia of at least 100. Foreign transistor BC547S was used as VT1. These transistors, despite their low cost, have a large and stable, almost constant gain (usually about 500) with collector currents ranging up to 50 mA. It can be replaced with any similar one, for example from the KT3102 series (h2ia no less than 200 and ike.max no less than 35 V). Transistor VT2 type KT827 with any letter index. Instead, you can use its analogue, assembled on two transistors: KT8101 and KT817 (or KT815) according to the internal Darlington circuit of the KT827 itself.

The situation is that inside the KT827 there are not only resistors that shunt the base-emitter junctions of both transistors, but also two diodes, an important function of which is performed by a diode that protects the collector-emitter junction of the more powerful transistor (KT8101) from voltage of the opposite polarity.

If the KT827 is replaced by a KT829 transistor or a foreign transistor BDX53C (analogue of KT829), the maximum CH current must be halved (to 1.5A). Transistor GTS VT5 type KT815, KT817, KT819 with letter indices B or G. It can be replaced with other similar ones, for example KT802, KT803, KT805, KT808, etc.

LED HL1 - foreign, cheap price, red light, HL2 - also cheap foreign, green light.

Power supply circuit board

One of the variants of the stabilizer printed circuit board is shown in Fig. 2 and Fig. 3.

The goal was not to create a miniature board, so there is a lot of free space on it. This makes it easier to draw the board using conventional methods, for example, using nitro paint.

Transistor VT2 was installed on an effective heat sink with a cooling surface within the range of 1500...2000 cm2, if the design did not use forced cooling (fan blowing). In the latter case, the heat removal area was 5-6 times smaller. The current source transistor VT5 was installed on a small plate radiator with an area of ​​25 cm2. The GTS circuit elements are located outside the board.

All power supply designs are equipped with systems for eliminating current surges in the power supply circuit (the primary winding of the CT), which were assembled according to the diagram.

You can test MV on alternating current (with a dynamic MV load) using a very simple circuit using a powerful field-effect transistor of the IRFZ48N type, which is controlled (switched) by the output signal of the measuring generator (GZ-112). The diagram and description of this design are given in the article.

Setting up the power supply

The scheme is being set up in two stages. They start with the circuit on DAI.1, and then begin to set up the protection system. Although you can do the opposite, since a MV circuit without protection becomes vulnerable to current overloads and short circuits in the load.

For the element ratings indicated on the diagram, the output voltage MV is 18 V. If necessary, it is adjusted by selecting the ratings of resistors R3 (R2) or R6 (R7). It is a little easier to change the voltage value of the ION than to change the MV circuit. If you need to have increased stability of the MV voltage, then these resistors must be precision.

Configuring the protection unit begins with selecting and setting the maximum protection current. To facilitate this procedure, the printed circuit board is provided with the installation of a 10 kOhm tuning resistor, type SPZ-38v, instead of a constant resistor R11.

If you use a different value for resistor R17 (for example, 0.1 Ohm instead of 0.12 Ohm), you may also have to select resistor R14.

For a maximum protection current equal to 3A, the value of the ION voltage for protection (on resistor R14) should be 450 mV.

As a simple guideline, when recalculating the protection unit, we were guided by the following reasoning. Since the ION voltage across resistor R14 determines the maximum protection current, this voltage must always be greater than the voltage drop across the current sensor R17 at maximum current. Naturally, this ION voltage should be with a reserve.

It must be remembered that fairly stable resistors should be used as R17. Otherwise, if the resistance of resistor R17 begins to change as it warms up, then the value of the protection current set by resistor R12 will also change. Therefore, in order to reduce the instability of resistance R17, it is necessary to reduce its temperature, for which they use a resistor with a reserve of power dissipation or use several resistors of the same type, for example, four identical resistors, which are connected in parallel and in series, so that the total resistance of four resistors is equal to the resistance of one resistor. The total maximum power quadruples, and the resistance stability when exposed to temperature also increases significantly, since the power dissipation across each resistor is reduced by a factor of four. For the same reason, stable resistors should be used as resistors Rll, R14 and R17.

As you can see, the circuit can be adapted to any value of the MV load current. If it is necessary to implement a more accurate setting of the protection current at low current values, then it will be necessary to introduce a sub-range in which the ION voltage will change in a limited range. For a protection current of 0...300 mA, the ION voltage was 0 - 50...70 mV, which significantly increases the convenience of MV operation with low-power loads.

Of great interest is the possibility of increasing the current in the load. The maximum CH current can be doubled by parallel connection of another KT827 type transistor. For this purpose, the collectors of both transistors (VT2 and additional) are connected in parallel, but the emitters and bases of both transistors must be separated from each other.

The fact is that it is impossible to equally distribute the collector currents of both KT827s using emitter resistors alone. Therefore, both in the base circuits and in the emitter circuits it is necessary to include equalizing resistors personally for each instance of KT827. For two copies of KT827, the maximum output current of the MV protection was set to 6...7A, which is already sufficient in most practical cases.

It should be remembered that when the radiator is blown, the temperature will be significantly lower than that of a massive radiator without such cooling, therefore, the real (at a specific temperature of the KT827) maximum permissible power dissipation of the KT827 during blowing will be greater.

In addition, the use of blowing fans allows you to get a serious gain in terms of weight and size indicators due to very “modest” and inexpensive radiators to purchase. Considering the excessively high prices for massive radiators, we also gain in material costs, since coolers today can be purchased at low prices.

Good day, forum users and site guests. Radio circuits! Wanting to put together a decent, but not too expensive and cool power supply, so that it has everything and it doesn’t cost anything. In the end, I chose the best, in my opinion, circuit with current and voltage regulation, which consists of only five transistors, not counting a couple of dozen resistors and capacitors. Nevertheless, it works reliably and is highly repeatable. This scheme has already been reviewed on the site, but with the help of colleagues we managed to improve it somewhat.

I assembled this circuit in its original form and encountered one unpleasant problem. When adjusting the current, I can’t set it to 0.1 A - at least 1.5 A at R6 0.22 Ohm. When I increased the resistance R6 to 1.2 Ohm - the current at short circuit the result was at least 0.5 A. But now R6 began to heat up quickly and strongly. Then I used a small modification and got a much wider current regulation. Approximately 16 mA to maximum. You can also make it from 120 mA if you transfer the end of the resistor R8 to the T4 base. The bottom line is that before the resistor voltage drops, a drop in the B-E junction is added and this additional voltage allows you to open T5 earlier, and as a result, limit the current earlier.

Based on this proposal, I conducted successful tests and eventually received a simple laboratory power supply. I'm posting a photo of mine laboratory block power supply with three outputs, where:

• 1-output 0-22v
• 2-output 0-22v
• 3-output +/- 16V

Also, in addition to the output voltage regulation board, the device was supplemented with a power filter board with a fuse block. What happened in the end - see below.

Our smaller friends (the Chinese) have flooded the electronics market, but they are not always conscientious, but many expensive models of computer power supplies are decent in their class. But still, most power supplies, as I call them, are castrated, that is, when the printed circuit board was designed for some elements, and others are soldered into it, and not all of them, especially for input filters, they are almost never found in cheap models.

ATX block diagram

The main disadvantage of all cheap power supplies

In general, everything is within normal limits.
Short voltage surges are noticeable. As the load increases, emissions increase. The consequence is glitches in memory and other digital elements of the PC. Note that the load of 30% is the majority of PCs not burdened with more than one HDD. Those who have a simple video card and a CPU that consumes no more than 15W.

Second drawback

The theory says that UPSs are very critical to load current instability. In our case, this drawback manifests itself in all its glory. This is what the +12V voltage oscillogram looks like under dynamic load.

On Fig.2 section No. 1 – static load. Section No. 2 – HDD in read/write mode. Characteristic dips in the +12V supply voltage. The magnitude and duration of the dip depends on the filter parameters of the power supply and the power of the HDD. Consequence: due to instability of the +12V power bus, the hard drive begins to slam its heads on the “pancakes”. Bad things appear. Glitches in devices powered from the +12V bus (ISA cards, COM ports)

How to deal with it

Let's consider filter power supply.

Fig.3 Filter (what it is)

In most AT units, the filter for the +5V power bus consists of two 1000µFx10V electrolytic capacitors. For the +12V power bus, one capacitor is 1000μFx16V. For switching power supplies, the capacitance of filter capacitors is taken at the rate of 500..1000 μF per 1A load current. In our case, for the +5V bus we get a maximum load current of 4A. For the +12V power bus, the maximum load current will be 2A.
In most cases, an emergency does not occur. But when using even one HDD of the IBM DPTA 7200RPM type (or with similar power consumption), the above glitches were observed.

Fig.4 Filter. (what it should be)

For this scheme ( Fig.4) the following parameters are valid: +5V bus – maximum dynamic load current 20A.
+12V bus – maximum dynamic load current 8A.

Electrolytic capacitors eliminate current instability. Ceramic (2.2 µF 3..6 pcs.) eliminate pulse voltage surges. A series with low resistance for pulsed currents is recommended (I think that’s what it’s called). Each company labels them differently. From what you can get in St. Petersburg - for example, Hitano, EXR series, operating temperature up to 105 Celsius. For +5V - two things 2200uF or 3300uF 6.3 or 10V (you need to look at the dimensions, power supply manufacturers squeeze the space very much). I can’t recommend anything regarding ceramics. From what I have seen, only TKE and accuracy differ (for example +80 -50%). I think this is not important in filters of this kind. Here, the larger the capacity, the better. It's probably better to take SMD (unpackaged) and solder from the back of the board directly to the conductors. Regarding the coils in output filters: if you don’t have winding experience, it’s better not to experiment. If you can buy it, you can try it. Or unsolder it from a dead power supply. With output coils you need to be very careful. Check the block only by loading it on resistors.

After upgrading the filter, look at the oscillogram

After upgrading the filter, the oscillogram of the +5V bus

This is what the voltage “surface” of a branded power supply looks like under load. There are voltage surges, but they are insignificant (much less than the permissible norm) and practically do not increase with increasing load. The total capacity (my version) of electrolytic capacitors is 6800 μF. 1.5 µF ceramic capacitors (whatever was at hand). For interest, we tested an ATX power supply from PowerMan from an InWin A500 case -The oscillogram is similar, but there are no voltage spikes.

On Fig.6 section 2 corresponds to dynamic load.
The filter capacity is one capacitor 4700 μFx25V (HDD in read/write mode). The maximum interference is no more than 100mV. The PowerMan ATX power supply showed approximately the same result.

Safety/reliability of the high-voltage part of the power supply

Mains voltage oscillogram

Operation of several PCs without a filter

Someone will say: “well, we don’t care whether our PC connects to the network or not. Well, we saved on a surge protector, so what.” Perhaps the following oscillogram will convince you.

On R Andp.9 section No. 1 – work of a powerful hammer drill. Section No. 2 – switching on a powerful inductive consumer (for example, a refrigerator or vacuum cleaner). I'll turn it onAn inductive load is always accompanied by a powerful voltage surge. The surge voltage is calculated using the following formula:

Where: - contact resistance at the moment of opening. - circuit circuit resistance 220V. - mains voltage (220V).

It is not difficult to guess that the numerator is always greater than the denominator.On the oscillogram ( Fig.9) section 2 - there is a “dip” in the mains voltage lasting 20..500 ms (typical for connecting consumers with a reactive nature of resistance to the network). UPS saves you from short voltage dips (minimum uninterruptible power supply turn-on time is 4 ms). It's good if it exists. It may be necessary to increase the capacity of the high voltage DC filter (by Fig.10– electrolytes 680x250V).Usually installed 220x200V. Atpower consumption 100W reservecapacity (220x200V) is enough for 70..100ms. If you increase the capacity to 680..1000μFx200V, then do not forget to replace the diode assembly RS205 (2A 500V) with RS507 (5A 700V)!!! Be sure to have a 4.7 ... 10 Ohm 10A thermistor. They usually save money on thermistors. Set the usual resistance 1 Ohm, 1 Watt

Surge filter + rectifier

Of all the elements in the filter circuit of a conventional power supply, there is only a PS405L thermistor and a fuse (the most necessary). Sometimes a symmetrical transformer is installed (5mH in the diagram). Of course - an RS205 rectifier and a high-voltage DC filter (2 electrolytes 220x200V).

Increased efficiency

Replacing powerful key transistors

We will be replacing the imported bipolar KSE13007 (or NT405F, 2SC3306) with our Soviet field device KP948A.

circuit diagram for switching on a field-effect transistor.

This option is suitable for ATX power supplies, because The block starts fromefficient low-power power source. This scheme is not suitable for AT blocks. Therefore, I left the transistor wiring as is, adding a 15V zener diode (as shown in the diagram Fig.11). It is not necessary to install zener diodes, because the forward voltage at the gate does not exceed 1V (direct diode), and its reverse breakdown voltage does not exceed 10V, Capacitors 1μFx50v ( Fig.12) it is worth installing ceramic ones (if the goal is to increase reliability), the drying out of these electrolytes (especially near a hot radiator) is the main reason for the failure of the power supply, since the power transistors are not turned off sharply enough.

I don’t know why - but it works for me. The power drop on transistors is reduced by 3..5 Watts. Although I still left the zener diodes. As a result, it stops heating.

Rectifier diodes

We install powerful rectifier diodes on normal radiators. A CPU heatsink will do - cut it in half. One half is a +5V rectifier. The second is for the +12V rectifier. It is also recommended to replace the power diode assemblies with our Soviet KD2998A diodes. Radiators - enlarge. All! Now you can remove the fan from the power supply. In this case, normal heat exchange inside the housing is disrupted. But if this is a power supply for a router, then there is nothing special to heat up inside the case. If this is a file server - then at your own peril and risk. Although Manowar Manowar claims that he has a converted ATX power supply loaded with 2HDD 7200RPM + ULF and the whole thing works without a fan.

The basis was the CODEGEN-300X power supply (like 300W, well, you understand the Chinese 300). The brain of the power supply is the PWM controller KA7500 (TL494...). These are the only ones I had to redo. The PIC16F876A will control the PWM switch, it is also used to control and set the output voltage and current, information is displayed on the LCD WH1602(...), adjustment is carried out using buttons.
One person helped make the program good man(IURY, site "Cat", which is radio), for which many thanks to him!!! The archive contains a circuit diagram, a board, and a program for the controller.

We take a working power supply (if not working, then we need to restore it to working condition).
We roughly determine where everything will be located. We choose a place for the LCD, buttons, terminals (sockets), power indicator...
We have decided. Making markings for the LSD “window”. We cut it out (I cut it with a small 115mm grinder), maybe someone with a Dremel, someone by drilling holes, and then adjusting it with a file. In general, it is more convenient and accessible for everyone. It should look something like this.

We are thinking about how we will mount the display. Can be done in several ways:
a) connect to the connector control board;
b) do it through a false panel;
c) or...
Or... directly solder 4 (3) M2.5 screws to the case. Why M2.5 and n M3.0? The LSD has holes 2.5mm in diameter for mounting.
I soldered 3 screws, because when soldering the fourth, the jumper is unsoldered (you can see it in the photo). Then you solder the jumper - the screw disappears. Just very close distance. I didn’t bother - I left 3 pieces.

Soldering done phosphoric acid. After soldering, everything must be washed thoroughly with soap and water.
Let's try out the display.

Let's study the circuit, namely everything regarding the TL494 (KA7500). Everything that concerns legs 1, 2, 3, 4, 13, 14, 15, 16. We remove all the wiring near these terminals (on the main power supply board), and install the parts according to the diagram.

We remove everything unnecessary on the main power supply board. All details regarding +5, -5, -12, PG, PS - ON. We leave only everything related to +12 V and standby power supply +5V SB. It is advisable to find a diagram for your power supply so as not to delete anything unnecessary. In the power supply circuit +12 volts - we remove the original electrolytes and replace them with something similar in capacity, but with an operating voltage of 35-50 volts.
It should look something like this.

To enlarge, click on the diagram

Looking at the characteristics of the existing power supply (sticker on the case) - for 12V, the output current should be 13A. Wow that looks good!!! Let's look at the board, what forms 12V, 13A??? Ha, two FR302 diodes (according to the datasheet 3A!). Well, let the maximum current be 6A. No, this doesn’t suit us, we need to replace it with something more powerful, and with a reserve, so we set 40CPQ100 - 40A, Uarb = 100V.

There were some kind of insulating gaskets, rubberized fabric (something similar) on the radiator. I tore it off and washed it. I supplied our domestic mica.
I installed longer screws. I squeezed more mica under one from behind. I decided to supplement the unit with an indicator for overheating of the heat sink on the MP42. A germanium transistor is used here as a temperature sensor

The heat sink overheat indicator circuit is assembled using four transistors. KT815, KT817 was used as a stabilizer transistor, and a two-color LED was used as an indicator.

I didn't draw the printed circuit board. I think that there should not be any particular difficulty in assembling this unit. How the unit is assembled can be seen in the photo below.

We make a control board. ATTENTION! Before connecting your LCD, study the datasheet for it!! Especially conclusions 1 and 2!

We connect everything according to the diagram. We install the board in the power supply. You also need to isolate the main board from the case. I did all this using plastic washers.

Setting up the circuit.

1. All adjustments to the power supply must be carried out only through an incandescent lamp 60 - 150 W, connected to the break of the network cable.
2. Isolate the power supply housing from GND, and connect the circuit that was formed through the housing with wires.
3.Iizm (U15) - the output current is set (the correctness of the indicator readings) using the standard A meter.
Uizm (U14) - the output voltage is set (the correctness of the indicator readings), according to the standard V meter.
Uset_max (U16) - sets the MAX output voltage

The maximum output current of this power supply is 5 amperes (or rather 4.96A), limited by the firmware.
It is not advisable to set the maximum output voltage for this power supply to more than 20-22 volts, since in this case the probability of breakdown of power transistors increases due to the lack of PWM control limit by the TL494 microcircuit.
To increase the output voltage to more than 22 volts, it is necessary to rewind the secondary winding of the transformer.

The trial run was successful. On the left is a two-color indicator of heat sink overheating (cold radiator - green LED, warm - orange, hot - red). On the right is the power supply indicator.

Installed a switch. The base is fiberglass, covered with self-adhesive "Oracle".

The final. What happened at home.

Source: http://vprl.ru

Wide adjustment range Output reference voltage……5V +-05%

Peculiarities :

• Full range of PWM control functions
• Output sink or sink current of each output…..200mA
• Can be operated in push-pull or single-stroke mode
• Built-in double pulse suppression circuit
• Wide adjustment range
• Output reference voltage…………………………………….5V +-05%
• Easy to organize synchronization

General description :

1114EU3/4 – TL494

Specifically designed for UPS construction, the TL493/4/5 ICs provide the designer with advanced capabilities when designing UPS control circuits. The TL493/4/5 includes an error amplifier, a built-in variable oscillator, a dead-time comparator, a control trigger, a 5V precision ionizer, and an output stage control circuit. The error amplifier produces a common mode voltage in the range of –0.3...(Vcc-2) V. The dead time comparator has a constant offset that limits the minimum dead time duration to about 5%.

It is possible to synchronize the built-in generator by connecting pin R to the reference voltage output and applying an input ramp voltage to pin C, which is used for synchronous operation of several UPS circuits.

Independent output drivers on transistors provide the ability to operate the output stage using a common emitter circuit or an emitter follower circuit. The output stage of the TL493/4/5 microcircuits operates in single-cycle or push-pull mode with the ability to select the mode using a special input. The built-in circuit monitors each output and prohibits the issuance of a double pulse in push-pull mode.

Devices with the suffix L guarantee normal operation in the temperature range -5...85С, with the suffix C guarantee normal operation in the temperature range 0...70С.

Structural scheme:

Case pinout:

Parameter Limits:

Supply voltage…………………………………………………………….41V

Amplifier input voltage………………………………………...(Vcc+0.3)V

Collector output voltage…………………………………………...41V

Collector output current………………………………………………….…250mA

Total power dissipation in continuous mode……………………….1W

Operating ambient temperature range:

With suffix L………………………………………………………………………………-25..85С

With suffix C………………………………………………………………..0..70С

Storage temperature range………………………………………..-65…+150С

Functional Description:

The TL494 chip is a PWM controller for a switching power supply, operating at a fixed frequency, and includes all the blocks necessary for this. The built-in ramp voltage generator requires only two external components R and C to set the frequency. The generator frequency is determined by the formula:

Modulation of the output pulse width is achieved by comparing the positive sawtooth voltage obtained at capacitor C with two control signals (see timing diagram). The NOR gate drives output transistors Q1 and Q2 only when the on-chip flip-flop clock line is in the logic LOW state. This occurs only during the time when the amplitude of the ramp voltage is higher than the amplitude of the control signals. Consequently, an increase in the amplitude of the control signals causes a corresponding linear decrease in the width of the output pulses. Control signals refer to the voltages produced by the dead time adjustment circuit (pin 4), error amplifiers (pins 1, 2, 15, 16) and the circuit feedback(conclusion 3).

The dead time adjustment comparator input has an offset of 120mV, which limits the minimum dead time at the output the first 4% of the ramp voltage cycle duration. This results in a maximum duty cycle of 96% when pin 13 is grounded and 48% when pin 13 is referenced.

It will increase the duration of the dead time at the output by applying a constant voltage in the range of 0..3.3V to the dead time adjustment input (pin 4). The PWM comparator adjusts the width of the output pulses from the maximum value determined by the dead time adjustment input to zero when the feedback voltage changes from 0.5 to 3.5V. Both error amplifiers have a common-mode input range of –0.3 to (Vcc-2.0)V and can be used to read voltage or current values ​​from the output of a power supply. The outputs of the error amplifiers have an active High level voltage and combined by the OR function at the non-inverting input of the PWM comparator. In this configuration, the amplifier that requires minimal time to turn on the output dominates the control loop. During the discharge of capacitor C, a positive pulse is generated at the output of the dead time adjustment comparator, which clocks the trigger and blocks the output transistors Q1 and Q2. If a reference voltage is applied to the operating mode selection input (pin 13), the trigger directly controls two output transistors in antiphase (push-pull mode), and the output frequency is equal to half the generator frequency. The output driver can also operate in single-ended mode, where both transistors turn on and off simultaneously, and when a maximum duty cycle of less than 50% is required. This is desirable when the transformer has a ringing winding with a clamping diode used to suppress transients. If high currents are required in single-ended mode, the output transistors can be operated in parallel. To do this, you need to short the input of the OTS operating mode selection to ground, which blocks the output signal from the trigger. The output frequency in this case will be equal to the generator frequency.

The TL494 has a built-in 5.0V reference that can provide up to 10mA of current to bias external circuit components. The reference voltage has an error of 5% in the operating temperature range from 0 to 70C.

DIRECTORY. Dodeka Publishing House. 1997

Yesterday I was testing a charger on a microcontroller, made on the basis of ATX, everything worked until it started beeping and suddenly, without any sign, died a heroic death. During the first inspection I couldn’t find a fault, so I went to Google and asked and this is what it gave me.

Fig.1 Typical ATX power supply circuit

Checking the high voltage part of the ATX power supply

First, we check: a fuse, a protective thermistor, coils, a diode bridge, high-voltage electrolytes, power transistors T2, T4, the primary winding of the transformer, control elements in the base circuit of power transistors.
Power transistors usually burn out first. It is better to replace with similar ones: 2SC4242, 2SC3039, KT8127(A1-B1), KT8108(A1-B1), etc. Elements in the base circuit of power transistors (check resistors for open circuits). As a rule, if a diode bridge burns out (the diodes short-circuit), then high-voltage electrolytes fly out from the alternating current entering the circuit. Usually the bridge is RS205 (2A 500V) or worse. Recommended - RS507 (5A 700V) or equivalent. Well, the fuse is always the last one to burn.
And so: all non-working elements are replaced. You can begin to safely test the power part of the unit. To do this, you will need a transformer with a 36V secondary winding. We connect as shown in Fig. 2. The output of the diode bridge should have a voltage of 50..52V. Accordingly, at each high voltage electrolyte there will be half of 50..52V. Between the emitter and collector of each power transistor there should also be half of 50..52V.

Checking the standby power supply

The standby power supply powers the TL494CN and +5VSB. As a rule, T11, D22, D23, C30 fail. You should also check the primary and secondary windings of the transformer.

Checking the control circuit

To do this you will need a stabilized 12V power supply. We connect the UPS under test to the circuit as shown in the diagram in Fig. 1 and look at the presence of oscillograms at the corresponding terminals. Take oscilloscope readings relative to the common wire.

Checking power transistors

In principle, there is no need to check operating modes. If the first two points are passed, then the power supply can be considered 99% serviceable. However, if the power transistors have been replaced with other analogues or if you decide to replace bipolar transistors for field ones (for example, KP948A, the pinout is the same), then you need to check how the transistor handles transient processes. To do this, you need to connect the unit under test as shown in Fig. 2. Disconnect the oscilloscope from the common wire! The oscillograms on the collector of the power transistor are measured relative to its emitter (as shown in Fig. 5, the voltage will vary from 0 to 51V). In this case, the process of transition from low to high level should be instantaneous (or almost instantaneous), which largely depends on the frequency characteristics of the transistor and damper diodes (in Fig. 5 FR155. analogue 2D253, 2D254). If the transition process occurs smoothly (there is a slight slope), then most likely within a few minutes the radiator of the power transistors will become very hot. (during normal operation, the radiator should be cold).

Checking the output parameters of the power supply

After all the above work, it is necessary to check the output voltages of the unit. Voltage instability under dynamic load, intrinsic ripple, etc. You can, at your own peril and risk, plug the unit under test into the working system board or assemble the diagram Fig. 6.

This circuit is assembled from PEV-10 resistors. Mount the resistors on an aluminum radiator (a 20x25x20 channel is very suitable for these purposes). Do not turn on the power supply without a fan! It is also advisable to blow on the resistors. Observe ripples with an oscilloscope directly at the load (peak to peak should be no more than 100 mV, in the worst case 300 mV). In general, it is not recommended to load the power supply with more than 1/2 of the declared power (for example: if it is indicated that the power supply is 200 Watts, then load no more than 100 Watts).

Scheme pulse stabilizer not much more complicated than the usual one used in transformer power supplies, but more difficult to configure.

Therefore, insufficiently experienced radio amateurs who do not know the rules of working with high voltage(in particular, never work alone and never adjust a switched-on device with two hands - only one!), I do not recommend repeating this scheme.

In Fig. Figure 1 shows an electrical circuit of a pulse voltage stabilizer for charging cell phones.

Rice. 1 Electrical diagram pulse voltage stabilizer

The circuit is a blocking oscillator implemented on transistor VT1 and transformer T1. Diode bridge VD1 rectifies the alternating mains voltage, resistor R1 limits the current pulse when turned on, and also serves as a fuse. Capacitor C1 is optional, but thanks to it the blocking generator operates more stably, and the heating of transistor VT1 is slightly less (than without C1).

When the power is turned on, transistor VT1 opens slightly through resistor R2, and a small current begins to flow through winding I of transformer T1. Thanks to inductive coupling, current also begins to flow through the remaining windings. At the upper (according to the diagram) terminal of winding II there is a small positive voltage, through the discharged capacitor C2 it opens the transistor even more strongly, the current in the transformer windings increases, and as a result the transistor opens completely, to a state of saturation.

After some time, the current in the windings stops increasing and begins to decrease (transistor VT1 is completely open all this time). The voltage on winding II decreases, and through capacitor C2 the voltage at the base of transistor VT1 decreases. It begins to close, the voltage amplitude in the windings decreases even more and changes polarity to negative.

Then the transistor turns off completely. The voltage on its collector increases and becomes several times greater than the supply voltage (inductive surge), however, thanks to the chain R5, C5, VD4, it is limited to a safe level of 400...450 V. Thanks to the elements R5, C5, generation is not completely neutralized, and through for some time the polarity of the voltage in the windings changes again (according to the principle of operation of a typical oscillating circuit). The transistor begins to open again. This continues indefinitely in a cyclical mode.

The remaining elements of the high-voltage part of the circuit assemble a voltage regulator and a unit for protecting transistor VT1 from overcurrent. Resistor R4 in the circuit under consideration acts as a current sensor. As soon as the voltage drop across it exceeds 1...1.5 V, transistor VT2 will open and close to the common wire of the base of transistor VT1 (forcibly close it). Capacitor SZ speeds up the reaction of VT2. Diode VD3 is necessary for the normal operation of the voltage stabilizer.

The voltage stabilizer is assembled on one chip - an adjustable zener diode DA1.

To galvanically isolate the output voltage from the network voltage, an optocoupler VOL is used. The operating voltage for the transistor part of the optocoupler is taken from winding II of transformer T1 and smoothed by capacitor C4. As soon as the voltage at the output of the device becomes greater than the nominal one, current will begin to flow through the zener diode DA1, the optocoupler LED will light up, the collector-emitter resistance of phototransistor VOL2 will decrease, transistor VT2 will open slightly and reduce the amplitude of the voltage at the base of VT1.

It will open weaker, and the voltage on the transformer windings will decrease. If the output voltage, on the contrary, becomes less than the nominal voltage, then the phototransistor will be completely closed and transistor VT1 will “swing” in full force. To protect the zener diode and LED from current overloads, it is advisable to include a resistor with a resistance of 100...330 Ohms in series with them.

Setting up
First stage: it is recommended to connect the device to the network for the first time through a 25 W, 220 V lamp, and without capacitor C1. The resistor R6 slider is set to the bottom (according to the diagram) position. The device is turned on and off immediately, after which the voltages on capacitors C4 and Sb are measured as quickly as possible. If there is a small voltage across them (according to the polarity!), then the generator has started, if not, the generator does not work, you need to look for errors on the board and installation. In addition, it is advisable to check transistor VT1 and resistors R1, R4.

If everything is correct and there are no errors, but the generator does not start, swap the terminals of winding II (or I, but not both at once!) and check the functionality again.

Second stage: turn on the device and control with your finger (not the metal pad for the heat sink) the heating of the VTI transistor, it should not heat up, the 25 W light bulb should not light up (the voltage drop across it should not exceed a couple of volts).

Connect some small low-voltage lamp to the output of the device, for example, rated for a voltage of 13.5 V. If it does not light, swap the terminals of winding III.

And at the very end, if everything works fine, check the functionality of the voltage regulator by rotating the slider of the trimming resistor R6. After this, you can solder in capacitor C1 and turn on the device without a current-limiting lamp.

The minimum output voltage is about 3 V (the minimum voltage drop at the DA1 pins exceeds 1.25 V, at the LED pins - 1.5 V).
If you need a lower voltage, replace the zener diode DA1 with a resistor with a resistance of 100...680 Ohms. The next setup step requires setting the device output voltage to 3.9...4.0 V (for a lithium battery). This device charges the battery with an exponentially decreasing current (from approximately 0.5 A at the beginning of the charge to zero at the end (for lithium battery with a capacity of about 1 A/h this is acceptable)). In a couple of hours of charging mode, the battery gains up to 80% of its capacity.

About details
A special design element is a transformer.
The transformer in this circuit can only be used with a split ferrite core. The operating frequency of the converter is quite high, so only ferrite is needed for transformer iron. And the converter itself is single-cycle, with constant magnetization, so the core must be split, with a dielectric gap (one or two layers of thin transformer paper are laid between its halves).

It is best to take a transformer from an unnecessary or faulty similar device. In extreme cases, you can wind it yourself: core cross-section 3...5 mm2, winding I-450 turns with a wire with a diameter of 0.1 mm, winding II-20 turns with the same wire, winding III-15 turns with a wire with a diameter of 0.6 ...0.8 mm (for output voltage 4...5 V). When winding, strict adherence to the winding direction is required, otherwise the device will work poorly or not work at all (you will have to make an effort when setting it up - see above). The beginning of each winding (in the diagram) is at the top.

Transistor VT1 - any power of 1 W or more, collector current of at least 0.1 A, voltage of at least 400 V. Current gain b2b must be greater than 30. Transistors MJE13003, KSE13003 and all other type 13003 of any company are ideal. As a last resort, domestic transistors KT940, KT969 are used. Unfortunately, these transistors are designed for a maximum voltage of 300 V, and at the slightest increase in the mains voltage above 220 V they will break through. In addition, they are afraid of overheating, i.e. they need to be installed on a heat sink. For KSE13003 and MGS13003 transistors, a heat sink is not needed (in most cases, the pinout is the same as for domestic KT817 transistors).

Transistor VT2 can be any low-power silicon, the voltage on it should not exceed 3 V; the same applies to diodes VD2, VD3. Capacitor C5 and diode VD4 must be designed for a voltage of 400...600 V, diode VD5 must be designed for the maximum load current. The diode bridge VD1 must be designed for a current of 1 A, although the current consumed by the circuit does not exceed hundreds of milliamps - because when turned on, a rather powerful surge of current occurs, and you cannot increase the resistance of the resistor Ш to limit the amplitude of this surge - it will heat up very much.

Instead of the VD1 bridge, you can install 4 diodes of type 1N4004...4007 or KD221 with any letter index. Stabilizer DA1 and resistor R6 can be replaced with a zener diode, the voltage at the output of the circuit will be 1.5 V greater than the stabilization voltage of the zener diode.

The "common" wire is shown in the diagram for graphical purposes only and should not be grounded and/or connected to the device chassis. The high voltage part of the device must be well insulated.

Decor
The elements of the device are mounted on a board made of foil fiberglass laminate in a plastic (dielectric) case, in which two holes are drilled for indicator LEDs. A good option (used by the author) is to design the device board in a housing from a used A3336 battery (without a step-down transformer).

Source: http://shemotechnik.ru

Title: Switching power supplies. Theoretical foundations of design and guidance for practical application

Number of pages: 272

Publisher: M.: Publishing House "Dodeka-XXI", nep. from English, series “Power Electronics”

The year of publishing: 2008

Description

Switching power supplies (SMPS) are quickly replacing outdated linear power supplies due to their high performance, improved voltage regulation and small size. The book discusses in detail the fundamental theoretical principles and design methods of switching power supplies and provides information that will not only help engineers optimize the selection of commercial power supplies for their projects, but will also allow them to develop their own original SMPS circuits. The book is aimed at readers who want to delve deeper into the essence of the operation of switching power supplies and their design, without getting into the mathematical jungle.

Particular attention is paid to the selection of appropriate components, such as chokes and transformers, to ensure safe and reliable operation of the SMPS circuits. The examples of original projects proposed by the author illustrate certain compromises that must necessarily be made when developing switching power supplies. Both mains power supplies and DC/DC converters are considered.
The book covers all the basic switching power supply circuits, including flyback, forward, bridge, buck, boost, and combination circuits. As examples, practical circuits of a 220-volt network switching power supply and a 110-volt uninterruptible power supply are given.

All electronic repair technicians know the importance of having a laboratory power supply, which can be used to obtain various voltage and current values ​​​​for use in charging devices, powering, testing circuits, etc. There are many varieties of such devices on sale, but Experienced radio amateurs are quite capable of making a laboratory power supply with their own hands. For this, you can use used parts and housings, supplementing them with new elements.

Simple device

The simplest power supply consists of just a few elements. Beginner radio amateurs will find it easy to design and assemble these lightweight circuits. The main principle is to create a rectifier circuit to produce direct current. In this case, the output voltage level will not change; it depends on the transformation ratio.

Basic components for a simple power supply circuit:

1. A step-down transformer;
2. Rectifier diodes. You can connect them using a bridge circuit and get full-wave rectification, or use a half-wave device with one diode;
3. Capacitor for smoothing ripples. Electrolytic type with a capacity of 470-1000 μF is selected;
4. Conductors for mounting the circuit. Their cross section is determined by the magnitude of the load current.

To design a 12-volt power supply, you need a transformer that would lower the voltage from 220 to 16 V, since after the rectifier the voltage decreases slightly. Such transformers can be found in used computer power supplies or purchased new ones. You can come across recommendations about rewinding transformers yourself, but at first it is better to do without it.

Silicon diodes are suitable. For devices of small power, ready-made bridges are available for sale. It is important to connect them correctly.

This is the main part of the circuit, not yet quite ready for use. It is necessary to install an additional zener diode after the diode bridge to obtain a better output signal.

The resulting device is a regular power supply without additional functions and is capable of supporting small load currents, up to 1 A. However, an increase in current can damage circuit components.

To get a powerful power supply, it is enough to install one or more amplifier stages on transistor elements TIP2955.

Important! To provide temperature regime diagrams on powerful transistors it is necessary to provide cooling: radiator or ventilation.

Adjustable power supply

Voltage-regulated power supplies help solve more complex tasks. Commercially available devices differ in control parameters, power ratings, etc. and are selected taking into account the planned use.

A simple adjustable power supply is assembled according to the approximate diagram shown in the figure.

The first part of the circuit with a transformer, diode bridge and smoothing capacitor is similar to the circuit of a conventional power supply without regulation. You can also use a device from an old power supply as a transformer, the main thing is that it matches the selected voltage parameters. This indicator for the secondary winding limits the control limit.

How the scheme works:

1. The rectified voltage goes to the zener diode, which determines the maximum value of U (can be taken at 15 V). The limited current parameters of these parts require the installation of a transistor amplifier stage in the circuit;
2. Resistor R2 is variable. By changing its resistance, you can get different output voltage values;
3. If you also regulate the current, then the second resistor is installed after the transistor stage. It is not in this diagram.

If a different regulation range is required, it is necessary to install a transformer with the appropriate characteristics, which will also require the inclusion of another zener diode, etc. The transistor requires radiator cooling.

Any measuring instruments for the simplest regulated power supply are suitable: analog and digital.

Having built an adjustable power supply with your own hands, you can use it for devices designed for different operating and charging voltages.

Bipolar power supply

The design of a bipolar power supply is more complex. Experienced electronics engineers can design it. Unlike unipolar ones, such power supplies at the output provide voltage with a plus and minus sign, which is necessary when powering amplifiers.

Although the circuit shown in the figure is simple, its implementation will require certain skills and knowledge:

1. You will need a transformer with a secondary winding divided into two halves;
2. One of the main elements are integrated transistor stabilizers: KR142EN12A - for direct voltage; KR142EN18A – for the opposite;
3. A diode bridge is used to rectify the voltage; it can be assembled using separate elements or using a ready-made assembly;
4. Variable resistors are involved in voltage regulation;
5. For transistor elements, it is imperative to install cooling radiators.

A bipolar laboratory power supply will also require the installation of monitoring devices. The housing is assembled depending on the dimensions of the device.

Power supply protection

The simplest method of protecting a power supply is to install fuses with fuse links. There are fuses with self-recovery that do not require replacement after blowing (their life is limited). But they do not provide a full guarantee. Often the transistor is damaged before the fuse blows. Radio amateurs have developed various circuits using thyristors and triacs. Options can be found online.

To make a device casing, each craftsman uses the methods available to him. With enough luck, you can find a ready-made container for the device, but you will still have to change the design of the front wall in order to place control devices and adjusting knobs there.

Some ideas for making:

1. Measure the dimensions of all components and cut the walls from aluminum sheets. Apply markings on the front surface and make the necessary holes;
2. Fasten the structure with a corner;
3. The lower base of the power supply unit with powerful transformers must be reinforced;
4. For external treatment, prime the surface, paint and seal with varnish;
5. The circuit components are reliably insulated from the external walls to prevent voltage on the housing during a breakdown. To do this, it is possible to glue the walls from the inside with an insulating material: thick cardboard, plastic, etc.

Many devices, especially large ones, require the installation of a cooling fan. It can be made to operate in constant mode, or a circuit can be made to automatically turn on and off when the specified parameters are reached.

The circuit is implemented by installing a temperature sensor and a microcircuit that provides control. For cooling to be effective, free access of air is necessary. This means that the back panel, near which the cooler and radiators are mounted, must have holes.

Important! During assembly and repair electrical devices we must remember the danger of defeat electric shock. Capacitors that are under voltage must be discharged.

It is possible to assemble a high-quality and reliable laboratory power supply with your own hands if you use serviceable components, clearly calculate their parameters, use proven circuits and the necessary devices.

Video

!
Today we will assemble a powerful laboratory power supply. It is currently one of the most powerful on YouTube.

It all started with the construction of a hydrogen generator. To power the plates, the author needed a powerful power supply. Buying a ready-made unit like DPS5020 is not our case, and our budget did not allow it. After some time, the scheme was found. Later it turned out that this power supply is so versatile that it can be used absolutely everywhere: in electroplating, electrolysis, and simply for powering various circuits. Let's go over the parameters right away. Input voltage is from 190 to 240 volts, output voltage is adjustable from 0 to 35 V. Output rated current is 25A, peak current is over 30A. Also, the unit has automatic active cooling in the form of a cooler and current limitation, which is also short circuit protection.

Now, as for the device itself. In the photo you can see the power elements.

It's time to move on to printed circuit boards. First, let's look at the power supply board.

Next, let's look at the converter board. Here, too, you can slightly adjust the placement of elements. The author had to move the second output capacitor upward, since it did not fit. You can also add another jumper, this is at your discretion.
Now we move on to etching the board.

When everything is ready, we proceed to attach the parts to the surface of the radiator, but since the flanges of the active elements have contact with one of the terminals, it is necessary to isolate them from the body with substrates and washers.

We will mount it with M3 screws, and for better thermal transfer we will use non-drying thermal paste.
When we have placed all the heating parts on the radiator, we solder previously uninstalled elements onto the converter board, and also solder the wires for resistors and LEDs.

Now you can test the board. To do this, we apply voltage from a laboratory power supply in the region of 25-30V. Let's do a quick test.

You can also adjust the temperature at which the cooler operates. We perform calibration using a tuning resistor.
The thermistor itself must be secured to the radiator. All that remains is to wind the transformer for the power supply on this giant core:

Now the hardest part is placing everything inside the case. First of all, we connect the two radiators into one and secure it.
We will connect the power lines with a 2-millimeter core and a wire with a cross-section of 2.5 square.

There were also some problems with the fact that the radiator occupies the entire back cover, and it is impossible to route the wire there. Therefore, we display it on the side.