Power supply unit based on a ready-made adjustable DC-DC converter. The device of computer power supplies and the method of testing them Dc converter in the power supply

09.02.202127.05.2017

DC / DC converters are widely used to power various electronic equipment. They are used in computing devices, communication devices, various control and automation schemes, etc.

Transformer power supplies

In traditional transformer power supplies, the voltage of the supply network with the help of a transformer is converted, most often reduced, to the desired value. The undervoltage is smoothed out by a capacitor filter. If necessary, a semiconductor stabilizer is installed after the rectifier.

Transformer power supplies are usually equipped with linear stabilizers. Such stabilizers have at least two advantages: they are low cost and an insignificant number of parts in the strapping. But these advantages are eaten away by low efficiency, since a significant part of the input voltage is used to heat the regulating transistor, which is completely unacceptable for powering portable electronic devices.

DC / DC converters

If the equipment is powered by galvanic cells or batteries, then voltage conversion to the required level is possible only with the help of DC / DC converters.

The idea is quite simple: direct voltage is converted into alternating voltage, as a rule, with a frequency of several tens or even hundreds of kilohertz, increases (decreases), and then is rectified and fed into the load. Such converters are often called pulse converters.

An example is a boost converter from 1.5V to 5V, just the output voltage of a computer USB. A similar low-power converter is sold on Aliexpress.

Figure: 1. Converter 1.5V / 5V

Pulse converters are good because they have high efficiency, within 60..90%. Another advantage of switching converters is a wide range of input voltages: the input voltage can be lower than the output voltage or much higher. Generally DC / DC converters can be divided into several groups.

Classification of converters

Downward, in English terminology step-down or buck

The output voltage of these converters, as a rule, is lower than the input voltage: without special losses for heating the regulating transistor, a voltage of only a few volts can be obtained at an input voltage of 12 ... 50V. The output current of such converters depends on the demand of the load, which in turn determines the converter circuitry.

Another English name for the down converter is chopper. One of the translation options for this word is interrupter. In the technical literature, a buck converter is sometimes referred to as a "chopper". Let's just remember this term for now.

Increasing, in English terminology step-up or boost

The output voltage of these converters is higher than the input voltage. For example, with an input voltage of 5V, a voltage of up to 30V can be obtained at the output, and, moreover, its smooth regulation and stabilization is possible. Quite often, boost converters are called boosters.

Universal converters - SEPIC

The output voltage of these converters is held at a given level at the input voltage both above the input voltage and below. Recommended in cases where the input voltage can vary significantly. For example, in a car, the battery voltage can vary within 9 ... 14V, but it is required to obtain a stable voltage of 12V.

Inverting converters - inverting converter

The main function of these converters is to obtain an output voltage of reverse polarity relative to the power supply. Very convenient in cases where bipolar power is required, for example.

All of the mentioned converters can be stabilized or non-stabilized, the output voltage can be galvanically connected to the input voltage or have a galvanic voltage isolation. It all depends on the specific device in which the converter will be used.

To move on to a further story about DC / DC converters, you should at least in general terms understand the theory.

Chopper down converter - buck type converter

Its functional diagram is shown in the figure below. The arrows on the wires show the directions of the currents.

Fig. 2. Functional diagram of the chopper stabilizer

The input voltage Uin is fed to the input filter capacitor Cin. The VT transistor is used as a key element; it carries out high-frequency current switching. It can be either. In addition to these details, the circuit contains a discharge diode VD and an output filter - LCout, from which the voltage is supplied to the load Rн.

It is easy to see that the load is connected in series with the VT and L elements. Therefore, the circuit is sequential. How does the voltage drop?

Pulse Width Modulation - PWM

The control circuit produces rectangular pulses with a constant frequency or constant period, which are essentially the same thing. These pulses are shown in Figure 3.

Fig. 3. Control pulses

Here t is the pulse time, the transistor is open, tp is the pause time, - the transistor is closed. The ratio ti / T is called the duty cycle duty cycle, it is denoted by the letter D and is expressed in %% or simply in numbers. For example, with D equal to 50%, it turns out that D \u003d 0.5.

Thus, D can vary from 0 to 1. At a value of D \u003d 1, the switch transistor is in a state of full conductance, and at D \u003d 0, in a cutoff state, it is, simply put, closed. It is easy to guess that at D \u003d 50%, the output voltage will be equal to half the input voltage.

It is quite obvious that the regulation of the output voltage occurs by changing the width of the control pulse t and, in fact, by changing the coefficient D. This regulation principle is called (PWM). In almost all switching power supplies, it is with the help of PWM that the output voltage is stabilized.

In the diagrams shown in Figures 2 and 6, the PWM is "hidden" in rectangles labeled "Control circuit", which performs some additional functions. For example, it can be a soft start of the output voltage, remote switching on or short-circuit protection of the converter.

In general, converters are so widely used that manufacturers of electronic components have established the production of PWM controllers for all occasions. The assortment is so great that just to list them you will need a whole book. Therefore, it never occurs to anyone to assemble converters on discrete elements, or as they often say in "loose powder".

Moreover, ready-made converters of low power can be bought on Aliexpress or Ebay for an insignificant price. At the same time, for installation in an amateur design, it is enough to solder the input and output wires to the board, and set the required output voltage.

But back to our figure 3. In this case, the coefficient D determines how long it will be open (phase 1) or closed (phase 2). For these two phases, the diagram can be represented by two figures. The figures DO NOT SHOW those elements that are not used in this phase.

Fig. 4. Phase 1

When the transistor is open, the current from the power source (galvanic cell, battery, rectifier) \u200b\u200bpasses through the inductive choke L, the load Rн, and the charging capacitor Cout. In this case, a current flows through the load, the capacitor Cout and the choke L store energy. The current iL INCREASES GRADUALLY, the influence of the inductance of the choke affects. This phase is called pumping.

After the voltage across the load reaches a predetermined value (determined by the setting of the control device), the transistor VT closes and the device goes to the second phase - the discharge phase. The closed transistor is not shown in the figure at all, as if it did not exist. But this only means that the transistor is off.

Fig. 5. Phase 2

When the transistor VT is closed, energy replenishment in the choke does not occur, since the power supply is disconnected. Inductance L tends to prevent a change in the magnitude and direction of the current (self-induction) flowing through the choke winding.

Therefore, the current cannot stop instantly and is closed through the "diode-load" circuit. Because of this, the VD diode is called the discharge diode. Typically this is a fast Schottky diode. After the control period has elapsed, phase 2 of the circuit switches to phase 1, the process is repeated again. The maximum voltage at the output of the considered circuit can be equal to the input voltage, and nothing more. To get the output voltage higher than the input, boost converters are used.

For now, it is only necessary to recall the actual value of the inductance, which determines the two modes of operation of the chopper. In case of insufficient inductance, the converter will operate in the discontinuous current mode, which is completely unacceptable for power supplies.

If the inductance is large enough, then the work takes place in the mode of continuous currents, which allows using the output filters to obtain a constant voltage with an acceptable level of ripple. Boost converters also work in the continuous current mode, which will be discussed below.

To slightly increase the efficiency, the discharge diode VD is replaced by a MOSFET transistor, which is opened by the control circuit at the right time. Such converters are called synchronous. Their use is justified if the power of the converter is large enough.

Boost step-up or boost converters

Step-up converters are used mainly for low-voltage power supply, for example, from two or three batteries, and some components require 12 ... 15V voltage with low current consumption. Quite often, a boost converter is briefly and clearly called the word "booster".

Fig. 6. Functional diagram of the boost converter

The input voltage Uin is fed to the input filter Cin and is fed to the series-connected L and the switching transistor VT. A diode VD is connected to the junction point of the coil and the drain of the transistor. The load Rн and the shunt capacitor Cout are connected to the other terminal of the diode.

The VT transistor is controlled by a control circuit that generates a stable frequency control signal with an adjustable duty cycle D, in the same way as described above when describing the chopper circuit (Fig. 3). The VD diode blocks the load from the switch transistor at the right times.

When the switch transistor is open, the right-hand side of the L coil is connected to the negative pole of the power supply Uin. The growing current (the influence of inductance affects) from the power source flows through the coil and the open transistor, energy is accumulated in the coil.

At this time, the VD diode blocks the load and the output capacitor from the key circuit, thereby preventing the output capacitor from discharging through the open transistor. The load at this moment is fed by the energy stored in the capacitor Cout. Naturally, the voltage across the output capacitor drops.

As soon as the voltage at the output becomes slightly lower than the specified one (determined by the settings of the control circuit), the key transistor VT is closed, and the energy stored in the choke through the diode VD recharges the capacitor Cout, which supplies the load. In this case, the EMF of the self-induction of the coil L is added to the input voltage and is transferred to the load, therefore, the output voltage is greater than the input voltage.

When the output voltage reaches the set stabilization level, the control circuit opens the transistor VT, and the process is repeated from the energy storage phase.

Universal converters - SEPIC (single-ended primary-inductor converter or converter with asymmetrically loaded primary inductance).

Such converters are mainly used when the load has little power, and the input voltage changes relative to the output voltage up or down.

Fig. 7. Functional diagram of the SEPIC converter

Very similar to the boost-converter circuit shown in Figure 6, but with additional elements: capacitor C1 and coil L2. It is these elements that ensure the operation of the converter in the voltage reduction mode.

SEPIC converters are used in cases where the input voltage varies widely. An example is the 4V-35V to 1.23V-32V Boost Buck Voltage Step Up / Down Converter Regulator. It is under this name that a converter is sold in Chinese stores, the circuit of which is shown in Figure 8 (to enlarge, click on the figure).

Fig. 8. Schematic diagram of the SEPIC converter

Figure 9 shows the appearance of the board with the designation of the main elements.

Fig. 9. External view of the SEPIC converter

The figure shows the main parts in accordance with figure 7. Note that there are two coils L1 L2. On this basis, you can determine that this is a SEPIC converter.

The input voltage of the board can be in the range of 4… 35V. In this case, the output voltage can be adjusted in the range of 1.23 ... 32V. The operating frequency of the converter is 500KHz. With a small size of 50 x 25 x 12mm, the board provides power up to 25 W. Maximum output current up to 3A.

But a remark should be made here. If the output voltage is set at 10V, then the output current cannot be higher than 2.5A (25W). With an output voltage of 5V and a maximum current of 3A, the power will be only 15W. The main thing here is not to overdo it: either do not exceed the maximum permissible power, or do not go beyond the permissible current.

The participants in our today's article are the older models of power supplies from three different manufacturers, and the older ones are not just in terms of their power, but in terms of the technologies invested in them. All three companies claim that the power supplies discussed below represent the pinnacle of progress, they use the latest circuitry solutions, some for the first time. Of course, we could not pass by this: what today is the pinnacle of progress tomorrow will become a completely massive decision.

Since the innovations promised to us are not cosmetic, like the colors of the fans that are fashionable this season, but at the level of circuitry and the principles of operation of switching power supplies, we will consider the most interesting points as detailed as possible. Unfortunately, for many authors this means either the verbatim reprinting of advertising brochures (as a result, for example, Seasonic blocks in the list of key characteristics include the "honeycomb structure of the ventilation grill", although, according to the most modest estimates, it is at least 95% of the commercially available power supplies of middle and upper classes), or unsoldering the unit into its component parts with a listing of the types of transistors and diodes (this is a laborious task, but in general it is not of practical use - the brands of transistors are interesting only to people who repair these units). In order not to be like the majority, we will try to consider precisely the features of the circuitry of the new blocks that distinguish them from yesterday's solutions - at the level of schematic diagrams (of course, somewhat simplified relative to the real block), from which it would be possible to understand not so much what kind transistors are in this block, how much what for they stand exactly like this and why the manufacturer calls this an advantage.

To understand the relevant parts of the article (in the description of each block they are highlighted in the subsection "Circuitry"), some familiarity with electronics is required - at least, an idea of \u200b\u200bthe essence of circuit diagrams and the work of individual parts in them. For specialists and simply radio amateurs who want to get acquainted in detail with the theory and practice of implementing the described solutions, we will also provide links to the relevant articles.

If you are just picking a good power supply, and therefore you are not interested in the features of the circuitry, then the schematic diagrams can simply be scrolled through - all other sections, including the testing itself, are performed in accordance with our usual methodology.

Testing technique

A description of the testing methodology of the equipment we use, as well as a brief explanation of what these or those passport parameters or the parameters of the power supplies we measure in practice mean in practice, can be found at the following link: “ Power supply testing method". If you feel that you are not well versed in the numbers and terms that the article abounds in, please read the relevant sections of this description, we hope it will clarify many questions.

For a complete list of the models that have been in our laboratory, please follow the link “ Tested Power Supply Catalog».

Antec Signature SG-850

Despite the fact that Signature PSUs are not the most powerful in the Antec product line, they are officially considered senior models. The manufacturer promises us stability, power, silence and high efficiency - Signature units are certified for compliance with the "80 PLUS Bronze" standard (efficiency is not less than 82% at load from 20% to 100%).

The block manufacturer is Delta Electronics.

Packaging and delivery set

The box is made quite original: it is made of thick black cardboard with a bright yellow stripe in the middle. On the top surface there is a gold inscription "Antec", on the side, looking closely, you can see the inscription "Signature 850 watt power supply" extruded in black on black. There is no other information on the box.

Inside, in addition to the unit itself, we find an installation manual, a set of removable cables, a power cord and four screws.

Appearance

The block is made in a body painted with black matte paint. Body length - 180 mm. It is interesting that the golden inscription "Antec" is not painted, but is made on a separate plate glued into the recess of the block cover.

On the back panel we find four connectors for removable cables - two for video cards and two for peripherals. Next to each of the connectors is indicated the +12 V line to which it is connected.

Circuit design

Signature is arranged according to a slightly unconventional scheme - its electronics are separated into two full-size boards located facing each other on opposite sides of the block.

On the first board, we find an input filter (top left in the picture), a standby power supply (bottom left) and an active PFC along with a rectifier and high voltage smoothing capacitors (right side of the board). Due to the fact that a whole large board was given for these schemes, the arrangement of the parts turned out to be noticeably more spacious than in most blocks of comparable power.

On the left side of the board, you can see an electromagnetic relay (rectangular piece in a brown case) - this is one of the ways to increase efficiency, which has recently been practiced in most expensive units. The task of the relay is simple: it completely disconnects the high voltage from the active PFC input if the unit is turned off. This increases the reliability of the unit (the parts are not energized at all), and slightly reduces its consumption in the "sleep" mode, when only the standby source is working.

The second board contains a power transformer and its switch (transistors on a small heat sink), rectifiers (diode assemblies on a long heat sink running through the entire board), output LC filters, a control circuit for output voltages and currents, as well as two DC / DC converters (" DC-DC converters "in the English-language literature), engaged in obtaining voltages of +3.3 V and +5 V from +12 V.

Since such converters in the foreseeable future will be regularly encountered in various reviews of power supplies, and in advertisements from almost all manufacturers, let's dwell on why they are needed.

Let's start with history and basic theory. The simplest pulse converter looks something like this:

The high-voltage part (to the left of the transformer T1) is shown conditionally, the input voltage value of 400 V is indicated for blocks with active PFC, in blocks without it it is lower, about 310 V. The high-voltage part is built according to the "forward converter" scheme, currently being very popular among power supply designers.

The PWM controller ("PWM control") controls the transistor Q1, switching it with a frequency of the order of several tens of kilohertz; the transformer T1 is connected to the transistor, which lowers the voltage and isolates the low-voltage circuits of the unit from the high-voltage ones. Current pulses through the diode D1, left according to the scheme, charge the capacitors C1-C3 of the output filter and the choke L1 (if the energy is accumulated in the capacitors in the form of an electric field, then in the choke it is in the form of a magnetic field), while the current passes through the load connected to the unit. Between the pulses, the choke is discharged through the right diode of the D1 assembly, while the current again passes through the load. The L2 choke has a small inductance and is only needed to suppress high-frequency interference.

Due to the presence of capacitors, the voltage across the load fluctuates within small limits, rising during the arrival of pulses and decreasing between them. However, if the pulses become shorter, then the average voltage begins to decrease, and vice versa - in this way, we get the opportunity to control the output voltage of the unit, changing the duration of the transistor Q1 on at each pulse. By introducing feedback from the output of the unit on the PWM controller, we can not only control the output voltage, but make the controller keep it constant.

NB: you can briefly get acquainted with the different types of switching power supplies in English in the article " Switching Power Supply Topology Review"(PDF, 1.09 Mb), as well as in the diagram" Power Supply Topologies Poster"(PDF, 143 KB).

However, we have several voltages in the computer power supply unit - and which one do you want to keep? Let's say we started the toy - the video card worked at full capacity, the load on the +12 V bus increased, the voltage at this block output dropped, the PWM controller tried to pull it to the previous level ... and thereby simultaneously increased the voltage at the +5 V output ...

Initially, a group stabilization scheme was used in computer power supplies to obtain several more or less stable output voltages from one transformer:

Group stabilization

To more or less balance the different outputs, an L1 inductor is introduced into the design of the unit, the so-called group stabilization inductor - several windings are wound on one core, one piece for each output voltage. With an increase in the current through one winding, a negative voltage is induced in the rest, which partly compensates for the increase in the output voltages of the corresponding buses described above.

As a result, we get a power supply with several outputs, which, despite the presence of only one regulating element (the PWM controller and the transistor Q1 controlled by it), maintains all output voltages at a more or less constant level. Nevertheless, with a strong imbalance of the load in such a circuit, the voltages begin to noticeably depart from the nominal value.

Magnetic amplifier

To obtain more stable output voltages, several years ago, additional stabilizers were used in the middle and upper-level power supplies according to the so-called magnetic amplifier circuit, it is also a circuit with a saturable core. More precisely, such stabilizers have been used on the +3.3 V bus for a very long time, and recently they have spread to the +5 V bus, as a result of which all three main output voltages received independent stabilization.

In a circuit with a magnetic amplifier, the group stabilization choke is divided into two completely separate chokes, L2 and L3, which no longer have anything to do with voltage stabilization. But in front of one of them a special design choke L1 appeared, the behavior of which can be controlled using a controller ("MagAmp control"), which is a conventional low-power linear voltage regulator. The choke has a specific effect - it shortens the pulses coming from the transformer T1, and the value of this shortening can change in real time:

Before and after choke L1

And the shorter the pulses, the lower the voltage at the unit output. Accordingly, the second winding of the transformer T1 should be wound with a margin of the number of turns, and we will “remove” the excess voltage using the inductor of the magnetic amplifier L1.

As a result, we get two separate regulators: the main PWM controller focuses only on the +12 V output and keeps the voltage on it stable, regardless of the other outputs, and the additional magnetic amplifier regulates the +5 V. What's nice, the circuit is not only simple , but also effective - the energy loss in the magnetic amplifier is close to zero.

NB: more details about the operation of magnetic amplifiers can be read in English in the article “ Magnetic Amplifier Control for Simple, Low-Cost, Secondary Regulation"(PDF, 1.5 MB).

Although the magnetic amplifier itself is the L1 choke, the easiest way to recognize the power supplies with it is by the large and perfectly visible L2 and L3. L1 is much smaller and is usually located next to the power transformer.

Despite the ability of magnetic amplifiers to maintain stable output voltages of the unit within ± 3% of the nominal at any load, they have a number of disadvantages. Firstly, additional chokes (L2 and L3) are rather cumbersome, and you cannot get rid of them - they play a very important role in forward converters: they accumulate the energy transmitted through the transformer, which is then given to the load. Secondly, each output voltage of the unit requires its own winding on the transformer T1, which complicates its design and manufacture - especially taking into account what powers are now required to fit into the given dimensions.

DC-DC converters, with which we started this conversation, are a replacement for magnetic amplifiers:

DC-DC converter

The converter in this case is formed by transistors Q2, Q3 and a choke L2. In fact, this is a completely independent forward pulse converter, which has its own PWM controller and is able to lower the +12 V voltage to any desired level, be it +5 V or +3.3 V. Unlike the main converter of the unit, it does not have a transformer - it is already isolated from the high-voltage part.

This scheme has several advantages at once. Firstly, DC-DC converters are powered from a constant voltage of +12 V and do not require a separate transformer winding - accordingly, the design of the T1 transformer is greatly simplified, only one secondary winding remains on it. Secondly, they can operate at significantly higher frequencies than the main converter of the unit, and therefore the size of the choke L2 and the capacity of the filter capacitors at the output are reduced, as a result of which space is saved inside the power supply. Thirdly, they have their own independent controller, and therefore, as in the case of magnetic amplifiers, the unit's output voltages are regulated independently of each other, which ensures their excellent stability.

Why are DC / DC converters used only now, and only in the most expensive units? The reason is simple - they are expensive: a PWM controller microcircuit, several transistors ... However, semiconductor components are gradually becoming cheaper, and the above advantages in the form of a simplified power transformer T1 and a smaller occupied volume help to save a little - and now DC-DC converters have become economically profitable at least in blocks of the highest price category. A couple of years later, and they will descend into the blocks of the middle class, as previously happened with magnetic amplifiers.

What are the advantages of using DC-DC converters for the user? In general, practically none. It is possible to find out whether they are used in this particular unit without looking inside, but it is quite difficult - at least it will require a good oscilloscope. They are interesting and convenient for development engineers, and they began to be used because their price dropped to a reasonable level.

Are DC / DC converters the newest invention? Of course no. Any electronic engineer whose work at least somehow concerns switching power supplies will draw you a couple of basic circuits on the nearest napkin, without even thinking - not to mention the fact that we have met blocks with such converters before, starting with blocks SilverStone and ending 1500 watt Xigmatek and Thermaltake.

In the case of Antec Signature, we find two boards with DC / DC converters between two heatsinks. One board provides a voltage of +5 V, the other - +3.3 V, both of them are powered from the main source, designed for an output voltage of +12 V. The photo shows the chokes of the converters - you can appreciate their modest size.

At the output of the unit, capacitors of the KZE and KZH series manufactured by United Chemi-Con are used.

The assembly quality of the unit can be described as excellent: exemplary soldering, reliable fastening of all large-sized parts, neat wiring. There is nothing to find fault with.

Loops and connectors

a motherboard power cable with a 20 + 4-pin connector, 54 cm long;
a processor power cable with an 8-pin connector, 55 cm long;
processor power cable with 4-pin connector, 56 cm long;

two video card power cables with one 6 + 2-pin connector on each, 55 cm long;
a ribbon cable with three power connectors for PATA hard drives and one for a floppy drive, 54 + 14 + 14 + 14 cm long;

a ribbon cable with three power connectors for SATA hard drives, 53 + 15 + 15 cm long.
two 8-pin connectors for additional loops;
two 6-pin connectors for additional loops.

two power cables for video cards with one 6-pin connector on each, 55 cm long;
two loops with three power connectors for PATA hard drives on each, 54 + 14 + 14 cm long.
two loops with three power connectors for SATA hard drives on each, 57 + 15 + 15 cm long.

The set of connectors is sufficient, but nothing more: you can connect a pair of video cards and six hard drives to the unit without using adapters (there are nine SATA connectors in total, but one cable will go to power the optical drive, and in most cases it will not reach the basket with hard drives. ). It should be noted that with power loops for hard drives Antec sets a good example for some other manufacturers: there are four loops for two connectors on the block, two with PATA and two with SATA power connectors, so the user can choose what is more important to him.

Passport parameters

Signature SG-850 is designed for long-term load power up to 829 W, and 780 W of them can be delivered via the +12 V bus, divided into four virtual lines. The parameters are absolutely adequate, they do not cause the slightest internal protest.

Paired with a UPS

Paired with the APC SmartUPS SC 620, the unit worked with a load of up to 380 W when powered from an outlet, but for the transition to batteries to be successful, the load had to be reduced to 350 W. At the same time, bubbling sounds were periodically heard from the UPS, so it is impossible to call them absolutely stable working together with the unit.

Stability of voltage

The result of the cross-load test is quite typical for units with independent voltage stabilization - +12 V holds perfectly at any load balance, +3.3 V deviates by less than 3%, and only the deviation of +5 V slightly exceeds 3%, and then only at the maximum load on the block. Recall that the standard deviation is up to 5%, so the Signature showed an excellent result in this test.

Everything is fine on the +12 V bus, but at +5 V and +3.3 V there are noticeable ripples, and some peaks exceed the permissible limit of 50 mV. However, nothing really critical can be seen on the oscillogram.

Pay attention to how different the ripple looks on low-voltage buses and +12 V - this is a consequence of the fact that the latter is provided by the main converter of the unit, and the former have their own pulse stabilizers operating at high frequency.

The unit uses a Nidec Beta SL fan, model D08A-12PS3-06AH1 - unfortunately not available on the Nidec website. Despite the high power of the unit, the fan has a modest size of 80x80x25 mm. It is four-wire, with PWM speed control, which should provide a wide operating speed range.

Indeed, the fan speed varies depending on the load on the unit more than three times. At loads up to 400 W, it rotates at about 700 rpm, while the unit is completely silent. Then the speed begins to grow according to a law close to linear, but noticeable noise can be called only at a load above 650 W. In general, the unit can be safely attributed to the quietest among those on sale, especially when operating at light load.

Thus, Antec Signature once again refutes the thesis that a large fan is required for quiet operation. The main thing is not the size, but the ability to correctly design the cooling.

However, it should be noted that in cases with a top-mounted power supply and insufficiently effective internal volume blowing under heavy load, the unit can be additionally heated by hot air, which will further increase the fan speed - and then it will start emitting a noticeable whistle. Therefore, from the point of view of silence, it is better to choose cases with a lower location of the power supply - but this advice applies, however, not only to Antec Signature.

Efficiency and power factor

The efficiency of the unit is very good, in a wide power range - from 300 to 600 W - it confidently keeps above 88%, and even at full load it decreases only to 86%. The power factor also did not disappoint, about half of the graph it fluctuates around 0.99.

Duty source + 5Vsb

The standby source in Signature is designed for a current of up to 3 A, and at full load its voltage is only 0.1 V lower than nominal, which is completely within the permissible range.

Conclusion

Well, Antec Signature made a very good impression: excellent build quality, excellent electrical parameters, a full-fledged set of cables and connectors, as well as such quiet operation at low loads that one can only wonder how this is possible on a block with only an 80 mm fan. ... Unsurprisingly, Antec considers the Signature Series to be the best among its products, despite not being the most powerful.

Perhaps the only drawback of this unit can be called its price: at the time of preparation of the article in Moscow retail for the 850-W version we reviewed, we would have to pay about 10 thousand rubles, and for 650-W - about 8 thousand. But if that price doesn't seem too high to you, then Antec Signature will not disappoint you.

Enermax Revolution 85+ ERV850EWT

As a representative of Enermax itself described this block, “we tried to put as many new products inside the block as its body could fit” - and in this light the name “Revolution” becomes clear. The unit is certified for compliance with the 80PLUS Silver standard (efficiency is higher than 85% in the range of loads from 20% to 100%), moreover, when operating in a 220 V network, its efficiency can exceed 90%. The unit is capable of operating at any load, including zero, to avoid problems with systems that have arbitrarily effective mechanisms to reduce consumption during periods of inactivity. In addition, Enermax talks a lot about the specific technologies used in the block - but we will write about the most interesting of them below, in the discussion of the block circuitry.

So far, we can say that, judging by the description, Revolution 85+ is more than a worthy opponent for the Antec Signature just reviewed.

Packaging and delivery set

The unit is supplied in a regular cardboard box of a rather large size, which contains the name of the series, the unit's power and main characteristics.

Inside the box, in addition to the unit itself, there are an installation manual (in 11 languages), a set of removable cables and a bag for storing them, as well as a power cord and four screws.

Appearance

We will learn more about how the Enermax engineers worked, but the designers definitely did their best. The block is painted with a rough (not just matte, but rough) gray paint, and a separate red plate is placed under the fan. Such a combination of colors, as well as the very fact that the block is not just painted, but assembled from parts of different colors, makes a good impression - and this is not about assessing the quality of the metal, assembly, and so on (it should be noted, all this is at a height), namely about the first sight.

On the back panel there are ten connectors for connecting removable cables - a record number, it should be noted. Six of them are for peripherals, four are for additional power cables for video cards and processor. Interestingly, the latter are made with a large margin in the number of contacts, which allows not only powering from one loop through a pair of terminal connectors, but also talking about the possibility of supporting future video cards with new connector standards and high load currents. Of course, in the event that such video cards actually appear, you will have to purchase them separately, after finding them on sale.

In terms of form factor, the connectors are similar to the typical Molex Mini-Fit Jr., but they are equipped with latches on the sides to fix the cables (for the standard Mini-Fit Jr., as you can see from the example of, say, the motherboard power connector, the latch is located in the center of the long side). This is reliable, but not very convenient, since with all the cables connected, it is difficult to squeeze the latches located between the rows of connectors.

The block body length is 190 mm.

Circuit design

The block is built according to a two-transformer circuit, but a little atypical - this is the so-called synchronized transformer circuit. The idea of \u200b\u200busing two transformers is not new in itself - with a high power unit, a single transformer is difficult to fit into the required dimensions, so it is logical to divide it into two half power. The problem that arises immediately is how to distribute the load so as not to get into a situation where one transformer is overloaded and the other is idle. This problem was observed in Enermax Galaxy DXX units, in which, for stable operation, the load had to be connected so that each of the two transformers operated at a power of at least several tens of watts.

And this is where the transformer timing circuit comes in:

In the diagram, it is shown in a simplified form, both in high-voltage (in fact, each of the transformers is controlled by two transistors, which allows Enermax to speak of a "quadruple converter"), and in low-voltage (instead of conventional diodes, Revolution 85+ uses transistors, about which we will talk more about below) parts, but for understanding the essence this is irrelevant.

So, we have one PWM controller that controls two forward converters at once - Q1-T1 and Q2-T2. He does it in such a way that transistors Q1 and Q2 turn on strictly alternately.

Each of the transformers has its own rectifier, as well as its own choke, in which energy is accumulated, but after the chokes, the two circuits are combined into one, in which the usual smoothing capacitors are located. Since the transistors Q1 and Q2 open and close in antiphase, the pulses to the chokes L1 and L2 come in antiphase.

As a result, the operation of the circuit on the oscillogram will look like this:

The pulses, alternately coming from each of the transducers, are added without intersecting in time, as a result of which at point "3" (that is, in fact, at the output of the power supply unit) the oscillogram looks exactly as if we had one a converter, but operating at double the frequency. Not only did we solve the problem of balancing the load between the transformers - the circuit is built in such a way that each of them always provides exactly half of the current power, and from the point of view of the load, the unit does not differ at all from the one-transformer - but also "free" received a doubling of the output frequency filter. And the higher the frequency, the less capacitors and chokes can be made, which finally smooth out the voltage ripples.

At the same time, it is difficult to raise the frequency of a single converter just like that - you need expensive high-frequency transistors, expensive materials for the transformer core ... Here Enermax engineers killed several birds with one stone: they entered the transformers in size into the unit, and ensured the ideal load distribution between them, and the output filter has made the job twice as easy.

NB: read more about synchronized transformers in English in the article “ Interleaving power stages - not just for buck converters any more».

Is this technology new? In power supplies, yes. But it was not developed by Enermax and not now, as can be seen from the article from the previous paragraph, dated 2004.

At the output of the unit, two identical chokes are clearly visible - one per transformer ...

The UCC28220 microcircuit located on a small daughter board is used as a synchronous converter controller.

But the features of the circuitry of Revolution 85+ do not just end there, but rather just begin. Taking a look at the radiator, on which the diode assemblies of the output rectifier are usually located (they were designated D1 and D2 above in the diagram), we will find that there are no diode assemblies at all! Instead of them there are IRFB3307 field-effect transistors:

The fact is that the Revolution 85+ uses the so-called synchronous rectifiers, in which diodes are replaced by transistors. What for?

Let's take a look at the characteristics of a typical Schottky diode, which are usually used in power supplies - STMicro S60L40C (PDF, 55 kB). We are interested in the graph of the dependence of the power dissipated on the diode on the current, that is, Fig. 1 on the second page: at a constant current of 20 A, more than 8 watts will be wasted on the diode - dissipated in the form of heat. This happens due to the fact that when a current flows through the diode, a small voltage drops on it, on the order of several tenths of a volt. Tenths of a volt multiplied by tens of amperes - you get units of watts.

What does a diode do in a rectifier? It opens in one direction of current and closes in the other. Let's replace it with a transistor, which we will control so that it mimics the operation of a diode - let it be the aforementioned IRFB3307 (PDF, 357 kB). In the open state, the resistance of its channel is only 5 mΩ, therefore, at a current of 20 A, power will be released P \u003d I²R \u003d 20² × 0.005 \u003d 2 W. More than four times less than on a conventional diode! Of course, this is an ideal case, but it gives an idea of \u200b\u200bthe scale of the savings.

Well, getting the transistors to switch at the right moments is already a matter of technology. In the simplest case, their gates are connected directly to the transformer windings:

If it is required to obtain a higher control efficiency of transistors and, accordingly, lower energy losses, then one or another synchronous rectifier controller is introduced into the circuit:

NB: more details about the use of synchronous rectifiers can be read in English in the article “ The Implication of Synchronous Rectifiers to the Design of Isolated, Single-Ended Forward Converters"(PDF, 433 KB).

Is the synchronous rectifier circuit for power supplies new? Yes, no doubt - there have been no such models in our laboratory so far. Is it an Enermax invention? "For example, we can confidently predict that sooner or later synchronous rectifiers will be used in the secondary circuits of computer power supplies - there is nothing particularly new in this technology, it has been known for a long time, it is simply too expensive for now, and the benefits it provides do not cover the costs." - I wrote back in 2006. Well, now the time has come.

Taking another look at the internals of the Enermax Revolution 85+, we notice an unusually large number of different small components on the board with output connectors ...

These are our old acquaintances, already discussed in detail in the description of the Antec Signature circuitry - DC-DC converters, with the help of which +5 V and +3.3 V are obtained from +12 V. Enermax engineers fully used their advantages, removing these converters from the main board of the unit and placing them entirely next to the connectors - indeed, what the back wall is in vain is empty.

Anpec APW7073 microcircuits are used as converters controllers. Power transistors are also located nearby, and they are heated so weakly that they do not even need a radiator - its role is successfully performed by the copper foil of the board, to which the transistors are soldered.

On the reverse side of the board, there are chokes (one for each converter) and smoothing capacitors. There are also connectors for removable cables nearby, to which, among other things, the voltage converters generated and are supplied.

There are other tricks in Enermax Revolution 85+ - for example, here is a jumper that connects two parts of the same track, and is needed to reduce its total resistance, and therefore energy losses. But they are no longer so principled, and therefore much less interesting.

Loops and connectors

The unit is equipped with the following cables and connectors:

a motherboard power cable with a 24-pin connector, 53 cm long;
processor power cable with 8-pin connector, 59 cm long;
processor power cable with 4 + 4-pin connector, 59 cm long;
two video card power cables with one 6 + 2-pin connector on each, 59 cm long;
fan tachometer loop, 55 cm long;
four connectors for video cards power cables;
six connectors for drive power cables.

The kit also comes with:

two power cables for video cards with two 6 + 2-pin connectors on each, 50 cm long;
ribbon cable with three power connectors for PATA hard drives and one for a floppy drive, 45 + 10 + 10 + 10 cm long;
ribbon cable with three power connectors for PATA hard drives, 45 + 10 + 10 cm long;

three loops with four power connectors for SATA hard drives on each, 46 + 10 + 10 + 10 cm long.

Well, although a fair amount of the connectors on the power supply unit remained unused, the resulting set is still impressive: six 6 + 2-pin (sic!) Power connectors for video cards, six power connectors for PATA hard drives, 12 power connectors for SATA hard drives ... , the owner of Revolution 85+ will have no problems with connecting anything in the foreseeable future.

Passport parameters

The total power of the unit is 850 W, while the manufacturer promises that with such a load, the unit can work at an air temperature of 50 ° C for an unlimited time. On the +12 V bus, the unit can deliver up to 840 W; this bus is divided into six lines - virtual, since, despite the presence of two transformers, as already mentioned above, from the point of view of the connected load, Revolution 85+, due to the peculiarities of its circuitry, does not differ in any way from ordinary single-transformer blocks and does not impose any specific restrictions.

Paired with a UPS

Paired with the APC BackUPS SC 620, it worked with a load of up to 385 W on mains power and up to 350 W on batteries. The transfer to batteries was going well, the UPS was absolutely stable, emitting only a slight hum.

Stability of voltage

Independent stabilization provides the expectedly excellent result: voltages of +12 V and +3.3 V are within the 3% tolerance, the voltage deviation of +5 V is only slightly higher than 3%, and even then - at extreme loads.

As you can see, the diagram on both axes is built from zero - the block is really capable of operating stably with no load.

Output voltage ripple

The picture is similar to what we have already seen in the Antec Signature: quite noticeable high-frequency oscillations with separate narrow bursts. However, there is nothing critical.

Fan speed control

The unit uses a Globe Fan RL4Z, 135x135x25 mm. This is a normal fan with a three-wire connection, in contrast to the 4-wire models of Enermax MODU82 + and PRO82 + units. The unit has a tachometer output that can be connected to the motherboard and control the fan speed from the BIOS or using the appropriate utilities.

The fan speed is kept at about 700 rpm with a load of up to 550 W, after which it begins to grow linearly. Nevertheless, even at maximum load it only reached 1120 rpm, which makes the Revolution 85+ a very quiet PSU.

In addition, the manufacturer assures that the design of the block case with the metal edges bent inward along the perimeter of the fan opening further reduces the noise level by 1-2 dB. Unfortunately, we do not yet have the ability to measure the noise level directly with an accuracy that suits us.

After turning off the unit, the fan continues to rotate at low speed for 45 seconds.

Efficiency and power factor

Record result - Enermax Revolution 85+ became the first power supply in our laboratory to exceed the 90% efficiency barrier! And the main role in this, most likely, was played by the synchronous rectifier described in detail above - the unit in which it is used in the main converter also enters our laboratory for the first time.

Duty source + 5Vsb

Unlike most competitors, in Revolution 85+ the standby source allows itself to be loaded with a current of up to 5 A. It copes with its task without any problems: at full load, the output voltage is 4.87 V with a minimum allowable 4.75 V.

Conclusion

It is difficult to say whether it is a revolution or an evolution - but Enermax engineers managed to create something really new and original from a technical point of view: this unit is able to surprise not only with its appearance, but also with its electronic filling. Now the capabilities of Revolution 85+ may even seem redundant - just look at the number of unused connectors for connecting removable cables - however, Enermax emphasizes that they tried to create a platform designed not only for today, but also for the foreseeable future, accommodating all the best that allows modern electronics. And it looks like they succeeded.

In addition to the moral satisfaction from owning one of the most technically advanced power supplies, Enermax Revolution 85+ provides its owner with good workmanship, excellent electrical parameters, a rich set of loops and quiet operation in the entire load range, up to a maximum of 850 watts.

The main disadvantage is, as you probably already guessed, the price - at the moment in Moscow, Revolution 85+ with a power of 850 W can be purchased for 12 thousand rubles; however, on the day of the preparation of the material, only one store offered it, so in the future the price may fall. Enermax itself recommends a cost of 309 US dollars or 229 euros, excluding taxes.

Seasonic M12D SS-850EM

Although Seasonic does not make as ambitious statements as Enermax, its M12D power supply unit can compete with Revolution 85+ in a number of parameters - it is also certified according to the 80PLUS Silver standard, and its efficiency can reach 90%. By the way, Seasonic has recently presented a line of units of lower power, certified for the even tougher "80PLUS Gold".

Among other features of the M12D that are interesting for us in the context of today's article, we can note the use of DC-DC converters, which we wrote about in detail above, talking about Antec Signature.

Packaging and delivery set

The unit comes in a small box in a bright orange and black paint, on the back of which are the main characteristics. In the kit you will find removable cables, a bag for storing them, an installation manual, a power cord, bolts and a sticker on the case of the system unit.

Appearance

The unit is made in the usual case for Seasonic products, painted in matte black. Only the silver-blue label on the fan grille stands out as a bright spot.

There are six Molex Mini-Fit Jr. connectors on the back of the unit. for connecting removable cables, four for peripherals and two for video cards. It is not very convenient that all connectors are made of the same color; however, since they have a different number of contacts, it is impossible to accidentally confuse the loop.

Circuit design

At first glance, the block looks quite ordinary, differing from the previous Seasonic models only by a different form of radiators - two of them acquired very wide “petals”, the third, on which the active PFC power elements are located, became smaller and almost lost against the general background.

The unit is made according to a single-transformer circuit, the Champion Micro CM6802 chip acts as an active PFC controller and the main stabilizer (PDF, 338 kB).

However, the most interesting thing for us is hidden behind the wires - a narrow aluminum plate, standing on the side of a large radiator, turns out to be the own radiator of the DC-current converter, providing voltages of +3.3 V and +5 V.

Seasonic kindly provided us with this converter, so we didn't have to disassemble the unit to the bottom to admire it:

On the front side there are two chokes - one for each output voltage - and filter capacitors. Pay attention to the dimensions: it's no wonder that manufacturers preferred to use such compact boards instead of additional chokes for magnetic amplifiers.

The reverse side of the module is closed with a heatsink - without it, the converter, tightly clamped in a narrow gap between the loops and the large block heatsink, would overheat. However, the radiator can be easily removed:

Before us are two familiar Anpec APW7073 PWM controllers, as well as seven transistors - this is the control and power parts of the DC-DC converter. It is completely autonomous: you can take such a scarf separately from the power supply, connect +12 V to it - and it will work, giving out +5 V and +3.3 V.

At the output of the unit - more precisely, in the 12-volt bus - United Chemi-Con KZE capacitors are used.

Loops and connectors

The unit is equipped with the following cables and connectors:

a motherboard power cable with a 20 + 4-pin connector, 52 cm long;
processor power cable with 8-pin connector, 56 cm long;
processor power cable with 4-pin connector, 52 cm long;
two power cables for video cards with one 6 + 2-pin connector on each, 59 cm long;
a ribbon cable with three power connectors for SATA hard drives, 32 + 14 + 14 cm long;
two connectors for video cards power cables;
four connectors for drive power cables.

The kit also comes with:

two power cables for video cards with two 6 + 2-pin connectors on each, 55 cm long;
ribbon cable with three power connectors for PATA hard drives, 55 + 15 + 15 cm long;
ribbon cable with three power connectors for PATA hard drives, 45 + 15 + 15 cm long;
cable with two power connectors for PATA hard drives, length 36 + 15 cm;
cable with three power connectors for SATA hard drives, length 55 + 15 + 15 cm;
ribbon cable with three power connectors for SATA hard drives, 45 + 15 + 15 cm long;
ribbon cable with two power connectors for SATA hard drives, 35 + 15 cm long;
adapter from one power connector of a PATA hard drive to two power connectors of drives, 15 cm long.

While the length of the list might seem like the M12D is the richest power supply we've covered in today's article, this is not entirely true. He is the most comfortable one. Instead of a pack of identical cables, the manufacturer has attached a set, in which each cable has its own length, so you can choose the most suitable depending on the case used.

However, there is no reason to complain about the number: six power connectors for video cards, 12 for SATA hard drives, eight for PATA hard drives ... it's hard to imagine a system that won't be enough.

Passport parameters

The unit is designed for a long-term load power of up to 850 W, of which 840 W can be delivered via the +12 V bus, divided into two virtual lines.

Paired with a UPS

Paired with the APC SmartUPS SC 620 UPS, the unit operates with a load of up to 360 W when powered both from the mains and from batteries. Transfer to battery is normal and UPS is stable.

Stability of voltage

Alas, despite the independent stabilization of the output voltages, the M12D turned out to be the only block in today's article that showed a result different from ideal: the +5 V voltage changed quite noticeably depending on the load, and as a result went beyond the permissible limits.

However, in a real computer everything will be in order - the problem is observed only with a heavy load on this bus, and this does not happen in modern systems, where almost all consumption is at +12 V.

Output voltage ripple

But the result on pulsations of the block is excellent - they are practically not visible on any of the three monitored buses, except that at +12 V individual bursts slip.

Fan speed control

The unit has a Sanyo Denki San Ace 120 fan, standard size 120x120x25 mm. Apart from the slightly unusual shape of the blades, we noticed another interesting thing:

No, we are not talking about an open bearing - it will still not work to lubricate it, and this is not necessary, since ball bearings have grease inside and does not leak out. We are talking about recesses running along the ring, two of which are partially filled with brown varnish. Does the fan manufacturer additionally balance each fan by adding a drop of varnish on the side of the impeller, which will be lighter due to imperfection of the technological process? Incredible, but it hurts too much.

At a load of up to 500 W, the fan speed was kept at 800 rpm, after which it began to grow. At low and medium loads, the unit works very quietly, but at about 650 W the sound of the fan becomes clearly visible, and at a power close to maximum, it is simply loud. Nevertheless, since it is problematic to assemble a computer whose other components would be quiet under such a load, the M12D can rightfully be considered a low-noise power supply.

Efficiency and power factor

Eh, just a little short of the record! The unit achieved an impressive efficiency of 90% (at a load power of about 400 W), but still one percent fell short of the level shown by Enermax Revolution 85+. Nevertheless, the result is excellent, and so far only one PSU has been in our laboratory that can block it.

Duty source + 5Vsb

The standby source of the unit is designed for a current of up to 3 A, and copes with its task without problems: the voltage is kept within normal limits.

Conclusion

Although after Revolution 85+ the Seasonic block looks not so impressive, in fact it practically does not lag behind competitors in the face of Enermax and Antec - even if one of them is technologically perfect, but what are we, users, to that perfection, if the output parameters are still about are the same?

For the latter, the M12D can be blamed only for not very stable voltages. There are no other complaints about this unit: low ripple level, highest efficiency, an excellent set of loops and quiet operation at small and medium loads make it an excellent choice for a home computer.

At the time of writing, the Seasonic M12D was not on sale in Moscow. The MSRP for the unit is $ 299.

Conclusion

In general, it was difficult to expect bad results from the blocks, which three eminent manufacturers consider the best models in their product lines. Indeed, Antec Signature, Enermax Revolution 85+ and Seasonic M12D did not show any serious technical flaws: powerful, well-made, with good electrical parameters and quiet operation, perfect for high-end computers, including those equipped with two -three video cards. In general, there is nothing more to say here - no matter which block of the three you choose, it will not disappoint you. As a minus, they can be written down - and, again, all three at once - except that a considerable price.

If we talk about the latest technologies, then here the Enermax Revolution 85+ stands out sharply - this is the first power supply unit among those who have visited our tests, which was able to demonstrate an efficiency of more than 90%. A two-transformer circuit with perfect balancing and the ability to work with any load from zero watts, a synchronous rectifier on the + 12V bus (for the first time in our practice!), Independent DC / DC converters - Enermax engineers really invested heavily in the development of this unit. If you are interested in power electronics and want to look at the road ahead for power supplies in the near future, the Revolution 85+ is a good example.

Two other models, Antec Signature and Seasonic M12D, are more common in circuitry: their developers instead of revolutionary innovations preferred to hone already known and used technologies (we even saw DC-DC converters "live" more than two years ago). They failed to catch up with Enermax in terms of the parameters demonstrated, but the lag is also small - the efficiency of these units is less by 1-3%, the fans are a little noisier under heavy load, and there is no difference at all for other points.

In general, having paid so much attention to the power supply circuitry, we wanted to convey two thoughts to you. Firstly, computer power supplies do not stand still, they are developing and improving, and this development lies not only in changing the shape of the ventilation grill holes and the color of the fan illumination. New controllers appear, operating frequencies increase, some circuitry solutions are replaced by others ... There is practically nothing in common between the two power supplies released ten years apart, although, at first glance, the parts are about the same color and shape. Secondly, despite this, it is worth being somewhat skeptical about the statements of manufacturers about the newest, just invented and patented technologies from all sides. These or those new nodes in mass-produced power supplies appear when it turns out to be economically beneficial. Take, for example, magnetic amplifiers: they have been used for a long time as regulators of the +3.3 V bus, you will find such a stabilizer in any decent 250-watt ATX-unit of the end of the last century, but only in recent years, when power supplies have grown dramatically the load capacity of the +12 V bus, the use of two magnetic amplifiers - what we call "independent stabilization of output voltages" - made sense. The same thing happens with other technologies: they exist, but until some time the return on their use simply does not cover the costs.

What can we expect in the future? Well, for example, digital programmable PWM controllers, the algorithm of which allows you to adapt "on the fly" to different types of load. They already exist, but they are still far from widespread use in power supplies, due to both imperfect technology and high cost. And this, of course, is not the only example.

General concept of DC DC converters

Linear regulators used in transformer power supplies maintain a constant output voltage due to a circuit element, such as a transistor, on which excess voltage is deposited. The control system constantly monitors the output voltage and corrects for its drop across this element.

Linear stabilizers have several advantages:

lack of interference;
low price and ease of use.

But such a device is not without its drawbacks:

excess voltage is converted to heat;
there is no way to increase the voltage.

Pulse-type dc to dc converters are circuits capable of converting one voltage level to another using coils and capacitors, temporarily storing energy in them and discharging them in such a way as to obtain the final desired signal levels.

The principle of operation of a pulse converter

The basis for the operation of many converters is the phenomenon of self-induction. Let's say there is an inductor through which a direct current flows. If the current flow is suddenly interrupted, in the magnetic field induced around the coil, an EMF of self-induction appears and, accordingly, a voltage with reverse polarity at its terminals.

Important! By controlling the current and switching time of the circuit, the self-induction voltage can be adjusted.

A switching converter is an electronic circuit containing a coil that is cyclically connected to a power source and turned off.

If the induced voltage is added to the input voltage, a boost converter is obtained;
When the coil is turned on so that the voltage induced in it is subtracted from the MT voltage, there will be a voltage lowering circuit.

Since the coil requires cyclic charging, the circuit needs a capacitor that will filter the signal and maintain a constant output voltage.

Important! The filtering is not perfect - the output voltage is always pulsed. Excessive levels of this interference can cause circuit malfunction, such as halting the microcontroller.

Pulse converter parameters

Main technical characteristics of devices indicated by the manufacturer:

Output voltage. Can be fixed (non-adjustable) or set within a specific range. In case of possible deviations, the manufacturer should indicate their limits, for example, 5V +/- 0.2V;
Maximum output current;
Input voltage;
Efficiency. It is understood as the ratio of output power to input power. The difference between the two is the heat loss. The indicator is expressed as a percentage. The closer to 100%, the better.

Important! Efficiency also depends on the working conditions. Therefore, you should carefully study the notes to the manufacturers' catalogs in search of graphs. It may turn out that a very expensive converter has worse parameters than a much cheaper one, which is optimized for operation with a different supply voltage.

The input voltage, depending on the type of inverter, can be:

below the output if the circuit is boosting;
higher than the output if the converter is buck;
higher or lower, but within the range (sepic).

Boost converters are indispensable when you need to raise the voltage. Let's say the device is equipped with a 3.6V lithium-ion battery and an LCD display designed to supply 5V.

Important! In general, increasing voltage is less effective than lowering it. Therefore, it is better to have a high voltage source that will be reduced to proper voltage than vice versa.

In the case of the third configuration, the input voltage can fluctuate, the decision to increase or decrease it is made by the circuit itself in order to obtain a stable signal at the output. These converters are ideal for applications where the supply voltage is close to the desired voltage. Although the range of regulation can be large. For example, the input is 4-35 V, the output is 1.23-32 V.

Since the power loss is small, the dc dc voltage converter is well suited for low voltage battery applications. It is useful, for example, when the control electronics are powered by 5 V and the actuators are powered by a 12 V battery.

If we assume that the control electronics takes a current of 200 mA, then the power consumption will be 5 V x 200 mA \u003d 1 W. When using the 7805 regulator to reduce the voltage, the power consumed from the battery will be 12 V x 200 mA \u003d 2.4 W. Power that the receiver will not receive, 1.4 W, is converted to heat. The heating of the stabilizer will be significant.

In the case of a switching converter with 90% efficiency, the power consumed from the battery is 1.11 W. Losses - only 0.11 W. The module temperature will rise almost imperceptibly.

In addition to the three types of dc dc converters, there are also inverting ones that change the polarity of the output signal. Such a circuit is needed to power operational amplifiers.

Pulse width modulation

Pulse width modulation (PWM) is a type of signal used to change the amount of power sent to a load. It is widely used in digital circuits that need to emulate an analog signal.

The generated pulses are rectangular, the relative width of which can vary in comparison with the period. The result of this ratio is called the duty cycle, and its units are presented as a percentage:

D \u003d t / T x 100%, where:

D - duty cycle;
t is the time when the signal is positive;
T - period.

The duty cycle is varied so that the average value of the signal is the approximate voltage to be obtained. By changing the value of D, you can control a key transistor, which is used in almost all switching converter circuits.

The fundamental circuit consists of an inductor, a capacitor, a diode, and a key transistor. The transistor is used to switch a signal with a high frequency and is controlled by PWM. The duty cycle D sets the opening and closing times of the transistor.

When the transistor is on, current flows through the coil, load resistor, and capacitor. Energy is accumulated in the choke and capacitor, and the current does not increase abruptly, but gradually. At this time, the diode is locked;
When a given voltage level is reached, which determines the control parameters of the transistor, the transistor is locked, but due to the EMF of self-induction in the choke, the current begins to flow along the circuit formed with the participation of an open diode, since the polarity on the coil has changed. In this case, the current decreases slowly at a rate Uout / L.

By adjusting the transistor control, you can get the required voltage level, but not higher than the input.

Boost converter

Its circuit contains the same elements as the step-down device, but their connection is different. The opening of the transistor is still controlled by the PWM settings.

When the transistor is open, current flows through the inductor and the transistor. The coil current increases at the rate Vin / L and it stores energy. The diode is closed at this stage to prevent the output capacitor from discharging through the transistor, which, in turn, feeds the load resistance;
When the voltage drops below a certain level, the transistor is closed by a control signal. The diode opens and the output capacitor is recharged. The input voltage is added to the voltage generated across the coil and the output is higher;
When the specified voltage limits are reached, the thyristor opens again and the cycle repeats.

In SEPIC converters, the circuit is built on a combined principle. Another choke and capacitor are installed in it. Components L1 and C2 work to increase the voltage, L2 and C1 to lower the voltage.

Voltage converter with galvanic isolation

Isolated dc dc converters are required in a wide range of applications, including power measurement, industrial programmable logic controllers (PLCs), isolation bipolar transistor (IGBT) power supplies, and more. They are used to provide galvanic isolation, improve safety and noise immunity.

Depending on the accuracy of the output voltage regulation,dc dc galvanically isolated converters fall into three categories:

adjustable;
unregulated;
semi-adjustable.

For such devices, the input circuit is isolated from the output. The simplest forward converter circuit has two isolated circuits: in one there is a key transistor and a transformer, in the other - an inductor, a capacitor, and a load resistance. The transistor receives a pulse control signal with a duty cycle D.

When the transistor is on, VD is passing current and D1 is off. The current flows in a loop through the coil, capacitor and load. Energy is accumulating in the coil;
When the transistor is turned off, the voltage on the transformer windings changes sign, so VD closes, and D1 begins to pass current that flows along the circuit between the coil, D1, the capacitor and the load resistance. The output voltage will be:

Uout \u003d (w2 / w1) x D, where w2, w1 is the number of turns of two transformer windings.

This is how a single-ended forward converter circuit works. There are flyback circuits and push-pull circuits, with energy supplied to the output during both conversion cycles. To reduce losses, MOS transistors are used instead of diodes.

Video

Today I will write not only about the product that I tested, but also about how it sometimes happens when you plan one thing, but for some reason it turns out completely different.
In general, who is interested, please, under cat.

Recently, a colleague ksiman laid out the "halves" of this converter, the same scarves, only without an indication device, so these reviews partly complement each other.
In the comments, I mentioned that I am also planning to review this board. The review said that everything ended not very well (or rather, quite badly). For me, everything was not very smooth either, although it ended better, but more on that later, but for now I will turn to the review of my version of this DC-DC converter.

In general, I saw such a small DC-DC converter and wanted to feel what it is. I ordered it for a review, received it after a while, but somehow there was no time to deal with it and I generally put it aside for now.

After a while, I finally got my hands on it, took a number of photographs, felt, examined.
He came in a small sealed bag.

It is small in itself, smaller than a matchbox.
In this case, the manufacturer claims the following characteristics:
Input voltage: 5V-30V
Output voltage: 0.8V-29V
Output current: 5A maximum (Requires a heat sink for currents over 3A)
Conversion efficiency: 95% (maximum)
Conversion frequency: 300KHz
Output ripple: 50mV (maximum)
Working temperature: -40 ℃ to + 85 ℃
Size: 51 x 26.3 x 114

On the sides are connectors for connecting to the power supply and to the load.
The assembly is neat, I won't say anything bad here.

At the top there are two trimming resistors, one regulates the current, the second, respectively, the voltage.
The current is adjustable in the range of 0.06-5.5 Amperes.
Voltage in the range 0.82-30 Volts
Also near the trimming resistors there is a red LED indicating the transition to the current stabilization mode.

The reverse side of the board can be said to be "bare", there is only a shunt in the form of a 50mOhm resistor.
By the way, I will immediately note that in devices of this type, where heat from the microcircuit is transferred to the board, for better heat transfer, it is generally customary to make many transitions with metallization between the sides of the board. Unfortunately, this has not been done here. Therefore, installing a radiator from the back is ineffective.

As I wrote above, the converter consists of two boards. DC-DC converter is no different from the converter from the above by me. The difference between these two modifications is that an indication board was attached to mine.
Moreover, it is connected through the mounting racks.
The left two are the input of the converter board, the right ones, respectively, to the output.
This connection allows you to monitor the output voltage and measure the flowing current.
The design is very convenient and simple.

The converter is assembled using the XL4005E1 PWM controller. This PWM controller is rated for 5 Amps of output current and input voltage up to 32 Volts.
Judging by the datasheet, it is a very good microcircuit, but as practice has shown, it is very "delicate".
Also worth noting is the SK86 diode, judging by its maximum current of 8 amperes. To be honest, I do not understand how it can dissipate the power that is released on it at such a current.
But in any case, the manufacturer supplied a rather powerful diode, and they often put something worse.

This photo shows the part responsible for adjusting the current limiting and indication of the end of the charge (two small LEDs are visible on the right).
The power supply diagram can be seen in Ksiman's colleague, for which many thanks to him :)

There are two indicators at the top.
The upper one, in blue, displays the output voltage, up to 10 Volts displays in the 1.23 format, above 10 Volts - 23.4. The last digit displays the character - V
The lower indicator, red, displays the output current in the 1.23 format, the last digit displays the symbol - A.
On the left there is an RX-TX connector. This was one of the reasons why I ordered this board, I wanted to try to connect it to the computer, but alas, nothing came of it :(
The purpose of the right connector is completely incomprehensible to me.

The board is assembled, let's say, on a C grade, it seems to be normal, but some inaccuracy is clearly visible.

The board contains:
Microcontroller
Shift register for indicator control
Presumably operational amplifier sgm8592y
Voltage stabilizer 7130H

And now a little nuance. This is the second board, the first died a heroic death during testing and preparation of the review. I can't say exactly what she died of, but it looked like this - The input voltage is about 28-29 Volts, a 10 Ohm resistor is attached to the output, I gradually increase the voltage across the resistor using the trimming resistor of the board, then a small click and the input voltage at the output , breakdown of the power transistor.
Perhaps a marriage, perhaps some kind of ripple or something else, but I would not advise raising the input voltage too much, although the datasheet indicates 32 Volts and the maximum 35 Volts.
Better to limit at 25-27 volts.
After that, I ordered a second board, since a lot had already been done to prepare for the review.

When first turned on, the board is set to an output voltage of about 5 volts. The current is about 1 Ampere.
In the photo, the board is connected to a 24 Volt power supply from my recent one.
If you unscrew the voltage adjustment trimmer to the maximum, then the output voltage at idle is equal to the input voltage.

There seems to be nothing especially to paint on the board, so I'll move on to testing.
The testing will be attended by:
The monitored board.
at 24 volts.
Contactless

Electronic
Pen and paper :)

The testing methodology was as follows:
The heating and pulsations of the output voltage were measured at the following set voltages of 5-10-15-20 Volts; at each voltage, load currents of 1-2-3 Amperes were set.
First, the characteristics were measured at 5 Volts, under a current of 1-2-3 Amperes, with an interval of 10 minutes, after which the board cooled down to room temperature and the cycle was repeated, but with the next voltage. A total of 12 measurements came out.
Problems were added by dynamic indication, I had to take a bunch of pictures in order to then choose one that shows the maximum number of indicator digits. In general, the display has a rather low switching frequency, flickering is a little but noticeable.
The first check is at idle, there are practically no ripples.
The oscilloscope probe divider is in the 1: 1 position.

More detailed test results

3.5 Volts 3 Amperes
4.10 Volt 1 Ampere

5.10 Volts 2 Amperes
6.10 Volts 3 Amperes

7.15 Volt 1 Ampere
8.15 Volts 2 Amperes

9.15 Volts 3 Amperes
10.20 Volts 1 Ampere

11.20 Volts 2 Amperes
12.20 Volts 3 Amperes

The entire verification cycle took about 3.5 hours.
The obtained temperature conditions:
The temperature of the PWM controller, diode, choke and output capacitor was monitored.
When I tested it, I decided to check for 3 Amperes, as it was written on the store page, I decided that I would sleep, so I would sleep, there would be a couple of these. But the experiment showed that the converter came out and the mikruha did not go into protection, the maximum temperature reached for the PWM controller was 110.2 degrees.

A little about the application of the board

In the photo above, you can see the factory 24 Volt power supply. But since there was an epic with the reordering of the board, then, as you know, I started working on this device a long time ago, and I did not have a factory power supply, so I had to do it myself.
And the factory PSU, according to my estimates, did not really fit into the case I chose, although it is much easier to use the factory one.
I have already described the power supply unit of my design in one of the, this is the same board, but some elements are installed more / more powerful. If interested, I can post the diagram here with all the changes.
Thoughts aloud, maybe it is worth starting the production of constructors ..... :)

I prepared such a "constructor" for assembly :)

Since initially I still counted on about 25-28 Volts and 3 Amperes, then the power supply unit was made with a margin, watts by 90-100. And since one of the key elements, the size of which directly depends on the power, is a transformer, then I chose it with a margin.
True, the board was not designed for this size, but with some tweaks I still stuck it in :)

I got such a neat transformer.

Another problem was that I needed to achieve a minimum thickness in the area of \u200b\u200bthe low-voltage part so that the elements of the power supply did not interfere with the converter board.
Because of this, some of the elements had to be put down.
The board turned out to be a little ugly, but all the elements correspond to the calculated power, this was more important to me.
The radiator of the output diode was an aluminum plate, standing along the long side, for safety, I insulated it in the area of \u200b\u200bthe feedback optocoupler.
He is not in this photo yet.
The radiator of the PWM controller is cut off from a special profile (I bought it with a meter, the board is straced for two types of radiators)

The power supply turned out to be much larger than the converter board.

But even here, not everything was simple.
I had some of the elements in stock, like any thrifty radio amateur, and some of the elements had to be bought.
The PWM controller chip is also on the shopping list.
The program for calculating a pulsed power supply unit recommended that I use TOP249. But somehow it so happened that the store where I usually buy was closed and I went to another, but there was no 249, but it was 250, it is a little more powerful. I thought it’s okay, I’ll buy it.
When I first turned on the power supply, it showed no signs of life, at all.
The only thing that happened was the voltage of 5 volts at the control leg of the PWM controller, it should be there, but the PWM controller did not start.
Since I collected quite a few different power supplies, I knew perfectly well that the rest of the circuit was in perfect order, and in case of disturbances in the rest of it, it behaves differently, making attempts to start. But it was quiet here.
Having rummaged in stocks, I found a weaker PWM controller, TOP247, put it on and the power supply unit started up with half a kick.
It turns out that I bought a fake. If there is someone from Kharkov, then I can tell you where you should NOT buy.
Moreover, the fake mikruha has laser marking, and the normal one has paint marking.

In general, having recovered another problem, I proceeded to further assembly.
I collected everything you need in a pile, terminals, variable resistors and handles for them, wires, a power switch.

The voltage regulation resistor is connected with two wires, the current - three.
Since the above experiment showed that the board does not normally give even 3 Amperes, I decided to limit it to 2 Amperes, but I wanted 3 :(
To do this, I put a constant 5.1 kΩ resistor in parallel with the extreme contacts of the variable resistor. The maximum adjustment turned out to be about 2.3 Amperes.
I also limited the voltage adjustment range, and in the same way, but set the nominal value to 51K, it turned out about 26 Volts.
At the same time, the above operations slightly stretched the adjustment scale and became more convenient to use,

Then I marked and reamed / cut out all the necessary holes for the indicator, variable resistors, terminals, power cable and switch.

At the last moment, I almost forgot to connect the wires to the board. The fact is that I thought to glue the board, so I can't connect the wires later.

The board, resistors and terminal blocks are installed. The great honor of the insides is literally back to back, but everything fit :)

The wires to the power supply are soldered just before installation.
If it was a factory power supply, it would be more convenient, there are already terminals there.

We tighten the input wires with ties so that they do not climb to the radiator, arrange the rest and can be closed.

That's it, the power supply is almost ready, there is a lack of dark glass for the indicator.
In fact, the readings are better read than it turned out in the photo. With the flash, you can see the switched off segments, and without the flash, the indicator starts to dazzle, so I didn't get a better photo, sorry.
I did not sign the management, in principle I did everything as logically as possible, the blue indicator is the voltage, respectively, it is regulated by a variable with a blue handle, similarly to the current.
I brought the indication of the current limiting mode to the panel, I did not display two LEDs with the indication of the charge mode, I do not see the point in them.

The current limitation turned out to be at the level of 2.23 Amperes, I think that in this mode the board will work without problems.
At first I wanted to attach a radiator to the board, but then I realized the whole pointlessness of this idea, since the choke is also heated, which must be increased, and the diode with the microcircuit, and heat is poorly transferred to the reverse side of the board.

By the way, about the choke, theoretically this cooling board should have produced 30 Volts 5 Amperes, that's 150 watts. Formally, this is half of my laboratory 300 Watt power supply, but if you go into it and roughly compare the dimensions of the power elements, then the difference, as they say, is obvious. This board, even theoretically, will not be able to deliver 5 Amperes, except with a different choke and at a low output voltage.

And so the summary:
pros.
Neat workmanship, not great, but quite good.
The converter has been tested for currents up to 3 Amperes, albeit with higher temperatures.
The accuracy of current and voltage measurements is quite good, it did not cause any special complaints.
Low ripple level, maximum recorded about 60mV at a frequency of 300KHz.
Compact design.

Minuses.
Large heating at currents over 2-2.5 Amperes.
You should be careful about the excess of the input voltage or put a protective suppressor at the input.
The choke is wound with a thin wire

In my opinion, at currents up to 2 Amperes it is possible to operate quite normally. It was a little upsetting that I could not figure out the RF / TX signals. It is quite possible to modify the converter with "little blood", rewind the choke with a thicker wire with a decrease in the number of turns by 1.5 times, or replace it with a more powerful one (this is better). Replace the diode with a more powerful one, or even better, even take it out, at least to the reverse side of the board, the thermal mode of operation will improve.
The declared efficiency of 95% is hardly achievable, but I think that the real one is somewhere nearby, but with a big reservation, for a certain operating mode. At a current of 3 Amperes, about 4 watts of heat (roughly) was released on the board, which gives us a very low efficiency at 5 volts output. With an increase in the output voltage, the efficiency gradually increases, although StepDown should not have such a steep relationship.
In general, what can I say, I spent money on spare parts, a lot of time assembling the power supply board, assembling all this together, but as a result I got a power supply with characteristics:
The output voltage is 0.85-24 Volts.
The output current is 0.06-2.25 Amperes.
Sparsely, but has the right to life, just a power supply unit could not have made such a power.

Hope the information I provided was helpful.

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

Power supply unit based on a ready-made adjustable DC-DC converter. The device of computer power supplies and the method of testing them Dc converter in the power supply

Testing technique

Antec Signature SG-850

Enermax Revolution 85+ ERV850EWT

Seasonic M12D SS-850EM

Conclusion

Other materials on this topic

DC-DC module control.

Calibration of voltmeter readings.

General concept of DC DC converters

The principle of operation of a pulse converter

Pulse converter parameters

Pulse width modulation

Boost converter

Voltage converter with galvanic isolation

Video

Testing technique

Antec Signature SG-850

Enermax Revolution 85+ ERV850EWT

Seasonic M12D SS-850EM

Conclusion

Other materials on this topic

DC-DC module control.

Calibration of voltmeter readings.

General concept of DC DC converters

The principle of operation of a pulse converter

Pulse converter parameters

Pulse width modulation

Boost converter

Voltage converter with galvanic isolation

Video

Powerful DIY auto amplifier

DIY LED lamps

Development of methods for troubleshooting and restoration of a failed subunit (module) with identification of a failure, adjustment and replacement of a failed element and performance monitoring