by Oleg Artamonov
07/13/2005 | 10:51 AM
In-Win Development Inc. has the reputation of a manufacturer of rather inexpensive (their consumer products range from $50 to $70 in price) yet quality system cases for home and office computers as well as for entry-level servers.
<%BANNER[article]%>In-Win’s cases had used to be equipped with power supplies from FSP Group (originally marked as “SPI, Sparkle Power Inc.”), but a few years ago In-Win launched its own power supply manufacture. These units are currently installed into In-Win cases and also sell separately. The good reputation of the In-Win brand attracts the users’ attention to their power supplies.
Below follow my tests of five PSU models from In-Win’s three different series. Each series includes two models differing in their lacking or having passive PFC, the rest of the parameters being identical. So I deal with them in twos.
The only point of difference between these two models is that the A3-1 has a passive PFC device, so I will describe them together. Moreover, the power factor was actually the only parameter the two models differed much in by the results of my tests.
The regulator of the first unit is based on an IW1688 chip, the regulator of the second on an SG6105D chip, but the identical PCBs and accompanying components make me suspect that the IW1688 is nothing else but a remarked SG6105D.
The heatsinks are only 2 millimeters thick, with some ribbing along the entire height. One corner is cut out of the heatsink with the switching transistors: a passive PFC throttle, fastened to the top panel of the PSU, is situated there in the A3-1 model. A standard two-section line filter is installed on the PSU’s input. The capacitors on the input of the high-voltage rectifier have a 470µF rating each.
There’s a rather confusing situation with the wattage of the unit. On one hand, the In-Win website clearly declares 300 watts for the ISP300A2-0 model, but on the other hand, the PSU’s own label reads “+3.3V & +5V & +12V = 235W (Max)” (see the snapshot above). The remaining voltages (the two negative voltages plus the standby voltage source) may total to 21 watts at best. So, the resulting wattage of this power supply is 250 rather than 300 watts.
The same conclusion comes from the maximum allowable currents: they exactly match the industry standard recommendations for 250W power supplies. So, there can’t be two opinions – this unit is a 250W power supply. The same strange thing about the declared wattage occurs with the ISP300A3-1 model, by the way.
The PSUs have a typical (for their class) selection of connectors:
There are no power connectors for SATA drives, but it is normal for an inexpensive ATX12V 1.2 product. A more serious drawback is that the 24-cm-long disk drive cables may be too short in large system cases.
The cross-load characteristics of the power supplies aren’t ideal, but good. These PSUs will handle a midrange computer quite confidently. The only surprising thing is the low stability of the +3.3V voltage. It usually deviates by no more than 2-3% from the norm, but here it reached 5%. Still, this shouldn’t be a problem anyway.
There’s distinct voltage ripple at full load (250W) but it never goes outside the prescribed limits: 30 millivolts on the +5V rail and 80 millivolts on the +12V rail, the allowable maximums being 50 and 120 millivolts, respectively. There is no low-frequency ripple here (at the double frequency of the power grid, i.e. at 100Hz).
The PSU is equipped with one 80mm Top Motor DF1208SH fan. The speed of this fan is controlled rather ineffectively: the speed changes with a jump as the load gets over 150W. So, the power supply is going to be very quiet at small loads (less than 150 watts), but at high loads the noise will grow up noticeably – the fan rotates at almost 3000rpm at its full speed.
The efficiency of the two units is average, about 75%, while the power factor differs, of course. It is almost 0.8 with the unit that is equipped with a passive PFC device.
So, these two power supplies left me perplexed. They are neatly assembled and have good parameters, but the short length of the cables and the manufacturer’s intention to overstate the allowable wattage of the units are disturbing. Yet anyway, they will both suit nicely for low-end and midrange computer configurations.
This model is located a step higher in the In-Win product range. It differs from its predecessors in its electrical parameters as well as in the design. First, it complies with the ATX12V 1.3 standard (which mainly differs from version 1.2 in having a higher allowable current on the +12V rail – 18 amperes, to be exact). Second, it is equipped with a 12cm fan for quieter operation. The fan’s grid protrudes a little from the PSU’s case – this may be an obstacle to installing it into some system cases (for example, this grid will press against a stiffness rib in HEC/Compucase/Ascot cases, preventing you from putting the PSU in its place).
The unit follows the typical PSU design without additional regulation of the output voltages. It is based on an IW1688 regulator. The line filter is complete. There are two 560µF capacitors on the PSU’s input. The shape of the heatsinks has changed: they have become thicker, with four short ribs (two ribs on each side of a heatsink). Although the fan is attached to the top panel, there are vent holes in the front panel of the PSU – it means that some portion of the warm air will be blown back into the computer’s system case.
We got a model without power factor correction for our tests, but a passive PFC version, IW-ISP350J3-1, is currently selling, too. Like with the above-described ISP300 series units, there’s no other difference between the J2-0 and the J3-1.
The manufacturer again misleads the user: the name of the PSU and the information on the manufacturer’s website seem to imply an output power of 350W, but the label says different. In reality, the sustained load power of this PSU is 300 watts because the maximum allowable load power on the +5V, +12V and +3.3V rails cannot exceed 285W.
The load currents of the PSU surpass the standard a little: the allowable currents on the +5V and +12V rails comply with the older ATX12V 1.2 standard, while the newer, version 1.3 standard reduced these currents.
The PSU offers the following connectors:
As you can see, the unit has got two SATA power connectors, and the cables themselves have become longer.
The unit maintains the stability of the +12V rail well enough, but performs worse at high loads on the +5V rail. The PSU couldn’t even make it to the maximum load of 200 watts because the voltages go beyond the required ranges even at 150W on this rail. Like with the previous models, the +3.3V voltage varies rather much depending on the load.
The ripple of the output voltages has become a little stronger if compared with the previous models. This is natural since the ratings of the output filters of this PSU are the same as the ones in the ISP300 series models, while the load is higher here. Still, the ripple is always within the norm.
This power supply also changes the speed of its fan with a sudden jump at about 170W load power. It’s hard to call this PSU quiet when the fan is at its maximum speed of 2000rpm.
The efficiency of this power supply is the same as that of the above-described ISP300A2-0.
So, this PSU is in fact a higher-power version of the IW-ISP300 models with a 12cm fan (again, the real wattage of this power supply is 300W rather than 350W). The parameters are good, but the unit isn’t quiet even in low-end computers – if the load is above 170W, the fan spins up immediately to its maximum rotational speed.
The model names suggest that these PSUs are equipped with 12cm fans (the letter J), and one of them has a passive PFC device (the index “3-1”). Externally and in the characteristics these units resemble the IW-ISP350J2-0 (see above), save for the higher allowable load power. Like with the ISP350, the protruding grid above the fan is a drawback. You can of course replace it yourself, but since there are no indents for the screws, the new grid must be placed inside, between the fan and the case. Otherwise, it will protrude rather far to the outside of the case, too.
The component layout on the PCB differs, although slightly, from the ISP350J2-0, but the employed elements and the circuit design are the same. The line filter is complete; the capacitors on the PSU’s input are 820µF each; the heatsinks are thick, with four short perpendicular ribs.
The cables have got longer since the previous model. The PSU offers:
So, there are seven connectors for PATA and two connectors for SATA drives here, and the cables are long enough even for a large system case. The only drawback is that there’s only a couple of nylon braces for the entire cable.
Like with the previous units, the output power is overstated: it is actually only 350 watts rather than 430 watts.
The ATX12V 1.3 standard doesn’t describe power supplies with an output power of over 300W, so the table compares the In-Win PSUs with the typical 300W unit. As you see, the load capacity of the +5V rail is the only parameter that has improved in comparison with the ISP350J2-0, and only by a dozen of watts. So, the advantages of these units will only show up at balanced loads, i.e. when the total load is evenly distributed among the output rails.
The stability of the output voltages is much better than with the above-described PSUs: these two models easily handle high loads on the +5V rail. The +3.3V voltage fluctuates rather greatly here, too, but is closer to the norm (this voltage equaled 3.3V at small loads with the PSUs described above, but at medium loads here).
As a kind of compensation for the good stability the ripple has got stronger and now exceeds the allowable maximum on the +5V rail a little. At 300W load power the ripple diminishes to meet the requirements, though.
There’s the same problem with the fan speed control here: the speed changes from lowest to highest in a jump. The only difference from the earlier-described units is that the moment of the jump is a hundred watts higher. The maximum speed of the fan is higher, too – 2300rpm. It is quite a lot for a 12cm fan, so the power supply is not quiet at such speeds. By the way, this kind of speed control explains why there exist two contradicting opinions about the noise parameters of In-Win units among the users: these power supplies are really quiet at low loads, but may become the noisiest component of the computer under high loads.
The efficiency of the PSUs is similar to the models described above – about 75%. The unit with passive PFC has a power factor of 0.75-0.78 against the PFC-less unit’s 0.68-0.7. The last thing once again confirms my point that passive power factor correction only helps the manufacturer meet the European requirements to the harmonics in the current consumed by the electronic device (switching power supplies without PFC don’t meet these requirements at all and thus cannot sell in Europe). Passive correction affects but slightly the power factor proper.
So, the IW-P430J2-0 and IW-P430J3-1 only differ from the junior models in quantity rather than quality: the maximum allowable load power, the number of the connectors and the length of the cables have all increased a little, and that’s all.
As I said above, In-Win system cases had used to come with power supplies from FSP Group. So when In-Win began to produce their own power supplies, a natural reaction from many users was to compare them with FSP Group’s alternatives.
Unfortunately, In-Win loses this comparison: FSP Group offers a much wider range of models (for example, In-Win doesn’t offer ATX12V 2.0 models as yet, while FSP Group’s THN series performed very well in my tests, for details see our article called FSP Group Power Supply Units Roundup ), and these models also have better characteristics. The disadvantages of In-Win’s units are the rather high voltage ripple that grows up at higher loads, the step-like control over the fan speed and the short cables in all models except the senior one. By the way, In-Win doesn’t offer high-wattage units, either. The senior model is only rated for 350 watts.
Well, the marking of the output power is quite another story. It seems In-Win decided to follow the example of obscure Chinese manufacturers who like to call their power supply something like “ATX-500W” and add “Max output power: 300W” in small print. The output power of all the five units tested in this review is a step below the number that stands in their model names and is declared on the manufacturer’s website.
Moreover, some labels have an additional marking. For example, we have “ATX12V300WP4” which seems to mean “an ATX12V power supply with an output power of 300 watts and compliant with the power requirements to Intel Pentium 4 platforms”. But there’s another marking nearby: “+3.3V & +5V & +12V = 235W (Max)”, which means that the PSU is rated for 250 watts (the remaining 15 watts is the sum of the negative voltages and the standby source), but not for 300 watts! I should acknowledge I launched an IW-P430J2-0 at a 430W load and the power supply lasted through half an hour of operation, but its heatsinks grew so hot that I didn’t risk to continue.
Yet, on the other hand, if you compare In-Win’s units with products from less renowned manufacturers than FSP Group, they will look much better, with their very neat assembly and good parameters. I mean that if you are choosing between power supplies from In-Win and FSP Group, you should prefer FSP, but if the alternative is from some obscure firm, then you may want to prefer In-Win. They are especially suitable for low-power and mid-power computer systems.
You can download the cross-load characteristics of the tested power supplies here and view them with this program .