For a long time, InWin cases were equipped with power supplies manufactured by FSP Group (initially they were also labeled as SPI, Sparkle Power Inc.), but a few years ago InWin decided to open its own production of power supplies. On the this moment these models are both installed in InWin cases and sold separately from them - and, of course, the good reputation of the InWin brand has led to buyers' interest in new power supplies.
Below I bring to your attention the testing of five models of InWin power supplies of three different series. In each of the series there are two models that differ from each other only in the presence or absence of a passive PFC, all other parameters are identical, and there is no point in describing them twice - therefore, some of the blocks are grouped in pairs.
InWin IW-ISP300A2-0 and IW-ISP300A3-1
These two power supplies actually differ from each other only in the presence of a passive PFC in the A3-1 model, so below I will consider them together - according to the results of measurements, only the power factor differed in some way.
The stabilizer of the first block is made on the IW1688 chip, the second - on the SG6105D, however, exactly the same printed circuit boards and strapping components make you think that the IW1688 is nothing more than a relabeled SG6105D.
The heatsinks are quite thin, only about 2mm thick, with little finning along the entire height. One corner is cut out of the radiator with key transistors - in its place in the A3-1 model there is a passive PFC choke, which is attached to the top cover of the unit. A standard two-section mains filter is installed at the input of the unit, capacitors at the input of the high-voltage rectifier are 470 μF each.
A somewhat incomprehensible situation arises with the power of the block. On the one hand, the InWin website for the ISP300A2-0 model clearly indicates a power of 300W. On the other hand, as you can see in the picture above, it says in black and white: "+3.3V & +5V & +12V = 235W (Max)". On the remaining voltages - and these are two negative voltages and a standby power source - you can dial another 21W, but no more; in total, the maximum power of the unit turns out to be 250W, but not 300W.
The same conclusion follows from the maximum permissible load currents - they exactly correspond to the recommendations of the standard for 250-watt power supplies. Thus, the conclusion is unequivocal - this unit is designed for a power of 250W. A completely similar situation is observed with the ISP300A3-1 block.
The blocks have a standard set of connectors for their class:
20-pin ATX connector on a 41 cm cable;
4-pin ATX12V connector on a 43 cm cable;
a cable with two hard drive power connectors, 24 cm long from the block to the first connector and another 15 cm to the second;
a cable with two power connectors for hard drives and one for a drive, 24 cm long to the first connector and then 15 cm between the connectors;
a cable with one hard drive power connector and one floppy drive connector, 24 cm long to the first connector plus 15 cm to the second.
In addition to the lack of power connectors for S-ATA hard drives (which, in general, is completely normal for an inexpensive ATX12V 1.2 block), it is worth noting relatively short wires - in large cases, 24 cm hard drive power cables may not be enough.
The cross-load characteristics of the blocks are not ideal, but quite good - the blocks will quite confidently "hold" the average computer. A little surprising is the low voltage stability of + 3.3V - usually it fluctuates within 2-3%, here the entire 5% range has passed, but in any case there should not be any problems with it.
Voltage ripples under full load (250W) are pronounced, but do not exceed the permissible limits - their range on the + 5V bus is 30 mV at the maximum allowable 50 mV, and on the + 12V bus - 80 mV at the maximum allowable 120 mV. There are no low-frequency ripples (at double the mains frequency, that is, 100 Hz) in the unit.
The unit has one 80 mm fan Top Motor DF1208SH. There is an adjustment of the speed of its rotation, but it works quite inefficiently - the change in speed occurs almost abruptly with an increase in load over 150W. Thus, at a low load (less than 150W), the unit will be very quiet, but as it increases, the noise it produces will increase dramatically - the fan accelerates to almost three thousand rpm.
The efficiency of both power supplies is at an average level - about 75%, but the power factor, of course, is noticeably different - for a block with passive PFC, it almost reaches 0.8.
These two power supplies produce a somewhat ambiguous impression. On the one hand, they are quite neatly assembled and demonstrate good parameters, but, on the other hand, the small length of the wires and the manufacturer's desire to overestimate the allowable power of the blocks by one step confuses. However, in any case, they are perfect for computers of junior and medium configurations.
InWin IW-ISP350J2-0
This unit, which is one step higher in the InWin model line, differs from its predecessors both in its electrical parameters, and in terms of design - firstly, it complies with the ATX12V 1.3 standard (the main difference from version 1.2 is the maximum allowable current on the +12V bus increased to 18A), and secondly, it is made with a 12-cm fan, which should provide more quiet operation of the unit. The fan grill protrudes a lot from the case of the unit, which may prevent its installation in some cases (for example, in HEC / Compucase / Ascot cases, the grill will rest against the stiffener, preventing the unit from falling into place).
The unit is made according to a typical scheme, on the IW1688 stabilizer and without any additional stabilization of the output voltages. Network filter assembled entirely, at the input of the block - two capacitors of 560 microfarads, the shape of the radiators has changed - they have become thicker, and the fins are represented by four short fins, two on each side of the radiator. Despite the location of the fan on the top cover, the unit has ventilation holes on the front wall - through them part of the warm air will be blown back into the computer case.
We tested a model without power factor correction, but there is also a version with passive PFC - IW-ISP350J3-1. As with the ISP300 series blocks discussed above, there are no other differences between J2-0 and J3-1.
In this case, the manufacturer also slightly deceives buyers - it would seem that from the name of the unit and the information on the manufacturer's website it follows that its power is 350W, but the label clearly indicates that this is not the case. In fact, the maximum long-term load power of the unit is 300W, which immediately follows from the fact that the maximum allowable load power of the +5V, +12V and +3.3V buses should not exceed 285W.
The load currents of the unit slightly exceed the requirements of the standard - according to admissible currents+ 5V and + 12V buses, it complies with the old ATX12V 1.2 standard, while in the newer, version 1.3, these currents have been reduced.
The unit is equipped with the following connectors:
20-pin ATX connector on a 40 cm cable;
4-pin ATX12V connector, 42 cm flex cable;
two cables with one S-ATA power connector, two power connectors for P-ATA hard drives, 42 cm long to the first connector (S-ATA), 8 cm to the second and plus 20 cm to the third;
one cable with two P-ATA power connectors for the hard drive and one for the drive, 25 cm long to the first connector, 15 cm to the second and another 20 cm to the third.
As you can see, not only two S-ATA connectors have appeared in the block, but the length of the wires has also noticeably increased.
The block holds the load well on the +12V bus, but things are worse with a large load of +5V, to the point that it could not reach the limit value of 200W at all - the voltages went beyond the permissible limits already when the load on this bus was less than 150W . As with its predecessors, the voltage of + 3.3V also depends relatively strongly on the load.
The range of output voltage ripples has grown a little - however, this is not surprising, since the ratings of the filter parts at its output are the same as for the models of the ISP300 series, but the load is already somewhat higher. However, the ripple does not go beyond the permissible limits.
The fan speed control works just as discretely - the speed changes from minimum to maximum with a jump at a load power of about 170W, and at maximum speed it is difficult to call the unit quiet, its 12-cm fan spins up to 2000 rpm, and the noise of the air flow becomes more than tangible.
In terms of its efficiency, the unit practically does not differ from the ISP300A2-0 considered above.
In fact, the block is a slightly more powerful (however, it should be noted once again that its real power is not 350, but 300W) version of the IW-ISP300 series discussed above with a 12 cm fan. Its parameters are at a good level, but the block can be called quiet only when working in low-power systems - if the load exceeds 170W, the fan jumps to maximum speed.
InWin IW-P430J2-0 and IW-P430J3-1
By marking the blocks, we can already conclude that these are models with 12 cm fans (letter "J"), one of which is equipped with a passive PFC (index "3-1"). By appearance and the declared characteristics, the blocks are very similar to the IW-ISP350J2-0 discussed above, with the exception of only a larger load capacity. As with the ISP350, the downside of the case is the heavily protruding fan grille. In principle, of course, you can always change it yourself, however, since the case does not have recesses for the grille mounting bolts, the new grille will have to be placed inside, between the fan and the case, otherwise it will protrude noticeably outward.
Layout printed circuit board block differs, although not fundamentally, from the ISP350J2-0, however, the used element base, and the circuit design is the same. The mains filter is fully assembled, the capacitance of the capacitors at the input of the unit is 820 microfarads each, the radiators are thick, with four short perpendicular ribs each.
Compared to its predecessors, the length of part of the trains has greatly increased. The unit is equipped with:
connectors ATX (20-pin) and ATX12V (4-pin) on cables 45 cm long;
a cable with two power connectors for hard drives and one for a drive, 45 cm long to the first connector, 15 cm to the second and another 10 cm to the third;
a cable with two hard drive power connectors with a distance from the case to the first connector increased to 75 cm;
a cable with three power connectors for hard drives, 60 cm long to the first connector, 15 cm to the second and another 10 cm to the third (unlike the first cable, the third connector here is also for powering the hard drive, not the drive);
a cable with two S-ATA power connectors for hard drives, 70 cm long to the first connector and plus 15 cm to the second.
In total, the block has seven power connectors for P-ATA hard drives and two power connectors for S-ATA hard drives, and the length of the wires is enough even for a very large case. Of the minuses, it can only be noted that for the entire almost meter long train there are only a couple of nylon ties.
As in the case of the previous blocks, the real power is not 430, but only 350W.
The ATX12V 1.3 standard does not specify blocks with a power of more than 300W, so the comparison in the table above is given with a 300W block. As you can see, in comparison with the ISP350J2-0, only the load capacity of the + 5V bus has grown, and even then only by a dozen watts. Thus, the advantages of these blocks will only be shown with a balanced load, when a large total power is evenly distributed between all the output rails of the power supply.
But the stability of the output voltages of the blocks turned out to be noticeably better - they perfectly endured the high load on the + 5V bus. The voltage of + 3.3V also changes quite noticeably here, but on average it is closer to the nominal one - if on previous blocks the nominal value was reached at a small load, here - at an average one.
Apparently, in compensation for good stability, the ripples increased even more, and their amplitude on the + 5V bus already slightly exceeds the allowable limit of 50 mV. When the load power is reduced to 300W, the ripple level decreases so much that it falls within the allowed limits.
With fan speed control, the IW-P430 series has the same problem as the previously reviewed units - the speed changes abruptly from minimum to maximum, except that the power at which the jump occurs has increased by a hundred watts. At the same time, the maximum speed has also increased - it reaches 2300 rpm, which is quite a lot for a 12-cm fan, you can't call the unit quiet at such speeds. This speed control, by the way, also explains the polar points of view among buyers on the noise of InWin power supplies - if the load power is low, then the unit is really quite quiet, but when working close to the maximum, it can easily become the noisiest element of the computer.
The efficiency indicators of the blocks differ little from those for the models considered above - the efficiency is about 75%, slightly changing depending on the load power, and the power factor is about 0.68 ... 0.7 for a block without PFC and 0.75 .. .0.78 for block with PFC. Regarding the latter, one can only repeat the idea that I have repeatedly expressed - passive power factor correction only allows the manufacturer to fit into the European requirements for the composition of harmonics in the current consumed by the device (switching power supplies without PFC do not fit these requirements at all, and therefore they are not sold in Europe can), but no more than that - the actual power factor changes rather weakly.
So, in fact, the IW-P430J2-0 and IW-P430J3-1 blocks differ from their younger counterparts only quantitatively, but not qualitatively - the maximum allowable load power and the number of connectors and the length of the wires on which they are located have slightly increased.
Conclusion
As I wrote above, for a long time, power supplies manufactured by FSP Group were installed in cases sold under the InWin brand - and therefore, when InWin began producing its own power supplies, it was a natural reaction of many users to compare them with FSP products.Alas, this comparison is clearly not in favor of InWin - FSP Group products compare favorably in terms of width model range(suffice it to mention that there are still no ATX12V 2.0 models among InWin blocks, while the THN series from FSP Group showed excellent results in our tests) and characteristics. Of the minuses, it is worth noting enough high level ripple, which increases with load power, stepped fan speed control, short wires on all models except the older one... There are no high power units among InWin products - the older model is designed for 350W.
However, the output power marking deserves a separate discussion - judging by it, InWin decided to follow the path of half-named Chinese manufacturers who like to name the power supply in the spirit of "ATX-500W" and attribute "Max. output power: 300W" in small letters. For all five blocks I tested, the figure in the model name, as well as the power explicitly indicated on the manufacturer's website, turned out to be one step higher than the actual power of the blocks. Moreover, on the labels of some of the blocks, additional markings are indicated, for example, "ATX12V300WP4", which, it would seem, should be decoded as "ATX12V 300W power supply that meets the power requirements of systems on Intel Pentium 4" - however, there is also another inscription, "+3.3V & +5V & +12V = 235W (Max)", from which it clearly follows that the unit is designed for a power of 250W (the missing 15W are recruited due to negative output voltages and an on-duty source), but not 300 W. In fairness, I must say that I tried to start the IW-P430J2-0 unit at a power of 430 W - it did not fail for half an hour of operation, but the radiators heated up so that I did not dare to continue the experiment.
However, if we compare the blocks produced by InWin not with the products of the FSP Group, but with less famous manufacturers, then they already look quite worthy, thanks to accurate manufacturing and very good parameters. Thus, if you are faced with a choice between an InWin and FSP block, then most likely you should give preference to FSP products, but if less respectable companies appear as the second option, of which there are many on the market, then, undoubtedly, InWin power supplies deserve close attention. attention. They will be especially good for computers of small and medium power.
Appendix
Load characteristics of the tested blocks: download.Program for viewing them: download.
Power SupplyIW- ISP300 J2-0
It is such a source that is installed in this case, so to speak, a standard 300-watt power supply, although the manufacturer honestly writes + 3.3V & + 5V & + 12V = 235W (max) on the sticker.
Those. 300 W is the maximum short-term power. The iron itself, from which the source is made, is quite thin, an order of magnitude worse than that from which the body as a whole is made. On the rear wall there is an input connector, a power switch and a 110V / 220 V network switch. I do not recommend you forget about the latter. To improve ventilation, a lot of holes are located along the entire surface of the rear wall. However, some may be confused about the location of the main power supply cooler. It is fixed on the bottom wall and is much larger in size than a conventional fan. From above, everything is decorated with a fashionable now chrome-plated grille. Big sizes allow you to reduce the rotation speed, and, consequently, the whole system will work quieter. The fan is marked as FD1212-S3142E DC 12V 0.32A - as you can see, the current consumption is quite solid. Inside, everything is standard for a medium-class 300W source.
The overall quality of the installation can be rated at four on a five-point system. At the input there are two impressive capacities of 470 microfarads x 200 V.
All power elements that experience strong heating are installed on fairly massive radiators. In any case, during testing, the heating was not very noticeable. The transformers used are also impressive in size, which is natural for such a declared power. The output is also set pretty a large number of filter containers. The master oscillator is assembled on an IW 1688 chip, it is marked as IN WIN and a brand name is applied to the case.
In general, all the details of the input filter (namely, the Chinese like to save on them) are installed in their places, even a capacitance of 0.33 uF is soldered to the input connector. But the fact remains, and there are still a significant number of unsoldered elements on the board. Having studied the topology of the board, and based on the fact that given source there are modifications (for example, IW-ISP300A2-0), it seems to me that this is not a fact of saving. It's just that the manufacturer makes different power supplies using the same type of boards, and somewhere some details are simply not put on the circuitry. This is just an assumption, but it looks like the truth, which, as you know, is difficult to get to the bottom of. Naturally, a simple statement of facts cannot satisfy us, so we will test the source.
Power supply testing
|
The dependence of the output voltage on the magnitude of the load |
Ripple (at 40% power rating) |
||||||
During this test, we will study the main parameters of the "feeder" and their dependencies. To do this, we will connect a powerful load with variable resistance to the tires that are most often loaded (+ 5V and + 12V), and we will control the current and voltage at the output using measuring instruments. Honestly, system monitoring and other things I believe much less than calibrated instruments. The test results can be seen in the table below.
According to her data, it is easy to say that the + 5V bus showed a good result. At light loads, the output voltage was fully consistent with the nominal value. At maximum load, the voltage naturally decreased. However, the deviation did not exceed 11%, which is a good result. But the voltage drop across the + 12V bus was much more significant, and it deviated from the nominal value by more than one volt. However, in percentage terms, this amounted to 8.75%. Of course, such a result can by no means be considered an achievement, but in general it looks pretty good. What surprised me was the weak heating during operation, even at almost maximum power, I didn’t have to think about any overheating. There are no problems with filters, both with input and output. The value of the variable component at the output does not exceed ~36mV on the +12V bus, and ~24mV on the +5V bus at a load power of 40% of the rated power. This value cannot be called critical. In general, I can rate this source as a "strong four". With its use, you can safely assemble a low-power computer, all indicators indicate that if all the necessary conditions are met, then there will be no problems. Of course, for lovers of sophisticated systems and overclocking, it is not entirely suitable. However, this case is an example of a well-balanced solution for building a home or office PC, and the power supply installed in it fully corresponds to this class.
Output
The tested case showed excellent results. It perfectly combines a good design (although this is subjective), excellent workmanship, high functionality. It is very convenient to assemble a computer based on it due to the presence of all kinds of pleasant devices. Everything is done to be able to perform any operation in the shortest possible time. And if we take into account the presence of competent additional cooling and a high-quality power supply, then this thing generally seems very tempting and competitive.
Testing equipment provided by the company
Again an upgrade, again a problem with the power supply. Like last time, not enough power. It would seem nothing, you can buy a new one. But such a block costs decent money. As always, they all go to more "important" parts - the processor, video card, memory ... Oh, how you don't want to spend money. But, there is nothing to do, you have to buy a new power supply. And there remains an old, useless, completely serviceable block. Sometimes even a few from previous upgrades. But only the power of the 12 V lines is not enough! Everything else is sufficient.
Why not combine several blocks into one more powerful one? They did that in the early 2000s. It is easy to ensure the synchronous switching on of two units - it is enough to connect their "ground" wires and the PS_ON (green) contacts of the 20-pin connectors. Drives and hard drives were hung on one block, and everything else on the other. Then it helped. But now the main power consumption is shared between the video card and the processor. These are 12 volt lines.
advertising
Now, if you use two old blocks and load only 12-volt lines from them, the voltage will be skewed and the stability of these same voltages will be violated. This is due to the fact that in old blocks not each voltage is stabilized separately, but the average value between 5 and 12 V. The voltage skew occurs due to the uneven distribution of the load on the +12 V and +5 V buses. Moreover, with the prevailing consumption of 12 V it just goes down, and 5 V goes up. Even if this phenomenon does not occur, old block on the 12 V line gives at best a third of the power. In modern conditions, this is not enough. And the efficiency of such a system will be small.This can be avoided by modifying the second power supply so that it stabilizes only the 12 V line and gives all its power to it. In 2004, I wrote on this topic. It described a way to remove only the voltage skew. This is no longer enough. Now everything looks different.
A few years ago, additional power supplies for video cards appeared on sale: FSP VGA Power,. Correct solution. The power of the old block is almost always more than enough to power the motherboard and processor, but for the video card ... Not anymore.
A typical computer rarely requires a power supply more powerful than 450W, but everything changes when it comes to performance gaming systems. A modern top-end video card consumes fairly. And there are video cards with two GPUs. And they can also be combined into SLI or CrossFire ... It's already good to have two independent +12 V power lines with a current of 30 A, which allows you to organize SLI or CrossFire without loading the main power supply of the system.
The use of several blocks is possible because manufacturers began to equip motherboards processor power connectors that are not electrically connected to the 20-pin ATX connector. Additional power connectors also exist on video cards. They can also be powered from a separate source. Unfortunately, such devices have not received wide distribution. Why? I think it's about the price. It's easier to add a little more and buy a full block.
Background to this article: there were many laudatory responses on the Internet about the conversion of the POWER MAN IW-P350 computer PSU into a 13.8V 20A transceiver power supply, after which UA4NFK purchased this power supply (Power Man model NO: IW-P430J2-0 is written on the case ( Fig. 1), but on the IW-P350W board (Fig. 2), which suggests the withdrawal of "extra" money from Russian buyers). But with recommendations for rework, it turned out to be a bummer, at best, they offered to remake it for money. I had to figure it out and help.
Fig.1
Rice. 2
Schematic found on the internet IW-P300A2-0 R1.2 DATA SHEET VER. 27.02.2004 dated pv2222 (at) mail.ru 90 percent coincided with the real power supply, the documentation for the SQ6105 processor (a complete analogue is installed on this board - IW1688) was also found, so we could start. After analyzing the circuit and documentation for the processor, in order to obtain a current of 22-24A at a voltage of 13.8V, it was decided to use a 5-volt rectifier (as it has the most powerful transformer winding) with the replacement of a full-wave rectifier circuit with a bridge one. The two missing diodes in the bridge were taken from the free ones, from +3 and +12V rectifiers. Additionally, a 2200 uF capacitor at 16V and eight resistors RR1 - RR8 were required.
Initial circuit diagram
This is what it looks like after the makeover.
Modified transceiver power supply circuit diagram (click on the link to enlarge)
Fig.3
Fig.4
Fig.5
Fig.6
Schematic modification
Before I take on the alteration, I want to warn you that in the process of alteration, you can easily fall under life-threatening voltage, as well as burn the power supply. You must be qualified.
1. We disassemble the PSU case, turn off the fan, solder the wire from the board going to the socket on the 220V case, remove the 110/220V switch and solder the wires coming from it (so as not to accidentally switch and burn the PSU). We remove the board from the case.
2. We solder the plug with the cord to the pads on the 220V board. The fee must be completely exempt from metal case and lie on a dielectric surface. We find the R66 resistor on the board, coming from pin 1 of the SG6105 MS (a complete analogue is installed on this board - IW1688) and solder a 330 Ohm resistor to the case (RR1 on Fig 6). By this we imitate the constantly pressed power button of the computer. We will turn off and turn on the PSU with a power switch on the PSU case. We connect the load in the form of a 12V 0.5-2A light bulb in the output of the PSU + 12V (black - ground, yellow wires + 12V), turn on the PSU in the network, check the performance of the PSU - the bulb should burn brightly. We check the voltage on the light bulb with a tester - approximately + 12V.
3. Disconnect the PSU from the 220V network. We disable the analysis by the SQ6105 processor plus 5 volts - we cut the track coming from pin 3 of the SQ6105, and we connect pin 3 to pin 20 with a jumper or a 100-220 Ohm resistor (RR5 on Fig 6). All resistors can be taken with a minimum power of 0.125 W or less. We turn on the power supply unit in the network (to check the correctness of the actions performed), the light should be on.
4. Disconnect the PSU from the 220V network. We turn off the analysis by the SQ6105 processor plus 3 volts - we cut the track near pin 2 and solder two resistors, 3.3 kOhm from pin 2 to the case (RR7 on Fig 6), 1.5 kΩ from pin 2 to pin 20 (RR6 on Fig 6). We turn on the power supply to the network, if it does not turn on, it is necessary to select the resistors more accurately in order to get +3.3V at pin 2.
5. Disconnect the PSU from the 220V network. We turn off the analysis by the SQ6105 processor minus 5 and 12 volts - we solder R44 (near pin 6), and pin 6 itself is connected to the case through a 33 kOhm resistor (more precisely 32.1 kOhm) (RR8 on Figure 5). We turn on the PSU in the network, if it does not turn on, it is necessary to select a resistor more accurately.
6. Disconnect the PSU from the 220V network. We solder extra parts - L3, L3A, L4, L5, C15, C12, R20, R18, R19, C11, C12, Q11, D27, D18, D28, Q7, R33, R34, RC, C28, R29, R32, RA, DA, D8, Q6, L9, C20, C21, D16, D17, L7, C16, C17, U1, D19, R41, R64, C42. Instead of C20, C21 we put 1500 (2200) uF at 16V (one is soldered, the other must be bought).
7. We fasten the soldered diode assemblies to the radiator through insulating heat-conducting pads (Fig.3, Fig.4). We connect all the anodes (the extreme terminals of the assemblies) together with a thick red wire, cut off at one end from the secondary winding T1 - the second end of this wire remains soldered in the old place, near the ground (black) wires coming from the PSU. We connect the cathodes of the assemblies (middle terminals): one - to T1 terminals 8.9 in the hole from L3, the second - to T1 terminals 10.11 in the hole L3A ( Fig.3, Fig.4). We replace R40 with 47 kOhm (RR2 with Fig 6), put VR1 in the middle position. To power the fan circuit (it is not on the circuit), we bridge the tracks + 5V and + 12V ( Fig 7). We solder all the extra wires coming from the board, leaving only all the red ones (this is now + 13.8V) (in the photo these wires are changed to yellow), twist or twist them into one wire, and the same number of black wires (this is now -13.8V ), they can also be twisted or woven. You can replace them with one thicker wire, with a cross section of at least 6 square.
Fig.7
8. We connect the load (12V 0.5-2A light bulb) to the PSU output - 13.8V. We turn on the PSU in the network. We measure the voltage on the light bulb with a tester and carefully adjust VR1 to the required value. To obtain an adjustment range of 12.0 - 13.97V, I had to parallelize RR2 with a 1.0 MΩ RR3 resistor (RR3 on Fig 6).. To
9. Disconnect the PSU from the 220V network. To obtain a current cutoff of 25-27A, we reduce R8 by paralleling it with a 6.2 kΩ resistor (RR4 in Fig. 6). We rearrange the fan in the case on the contrary ( Fig.9), he used to drive air into the PSU, now he will blow it out. If it will work noisily, you can lower the speed by including a diode in the red wire of the fan power supply or several series. We bite the blinds on one side of the case with wire cutters through one, to improve cooling ( Fig.8). We fasten the board into the case, solder the wires to the plug from the 220V board, attach the fan, assemble the case.
Fig.8
Fig.9
10. We check for a light bulb, if everything is fine, turn it off and change the load to 0.45 Ohm. I took about 21 meters of a double field - each wire is about 0.9 ohms. The skein of the field worker was lowered into a bucket of water. I controlled the current through an ammeter at 30 amperes.
11. At a current of 22A, in an hour of operation, a bucket of water will noticeably warm up. If after an hour everything works, there is hope for a long-term and trouble-free operation of the PSU! It remains to protect it from surges in the 220V network and install thyristor surge protection at the PSU output, although the latter is very unlikely.
In conclusion, a few positive points: the voltage of 13.8V on the board drops by 0.03 V under a load of 22A, T1, T6 heat up very weakly, a radiator with a diode bridge is stronger. After the alteration, protections remain: for current 25-27A, for voltage - in case of a drop of less than 12V, for excess of more than 15V, for overheating of a radiator with a diode bridge.
