by Oleg Artamonov
11/29/2006 | 07:03 AM
A little over a year ago I reviewed power supplies from In-Win and was left somewhat disappointed (for details refer to our article called In Win Power Supply Units Roundup). The company’s model range included only out-dated ATX12V 1.2 models and not a single ATX12V 2.0 one. And the then available models didn’t perform too well in my tests. I have also heard repairmen complaining at insufficient reliability of In-Win power supplies, mostly due to a multi-piece standby source, i.e. assembled out of separate discrete components, whereas a majority of manufacturers have long come to use special-purpose integrated controllers like Power Integrations’ TOPSwitch or Fairchild’s Green FPS.
Still, I’ve got some interest in In-Win power supplies. The company occupies a popular niche of inexpensive, yet good-quality system cases (and power supplies because In-Win manufactures them on its own), so I’d want to see if the quality is up to the customers’ expectations. Moreover, In-Win has begun to ship its power supplies apart from system cases as boxed versions. And the most exciting thing is that the company has introduced new PSU series in which the model names begin with the letters IP instead of IW.
For this review I tested both boxed versions of PSUs and those installed into system cases. I didn’t spot any significant difference, however, so you can apply the results of the boxed IW-P430J2-0 to the IW-P430J2-0 that comes along with a system case and vice versa.
Labels on boxes with system cases used to indicate the specific model of the installed power supply, but the labels on some In-Win system cases only show the wattage of the PSU.
I will start out with two old models from the IW series and will then go over to the new IP line that splits into two series, AJ and Q. The Q series is remarkable since it is the only ATX12V 2.x-compliant one that In-Win is currently offering.
We have got two samples of this PSU model in our labs. One was selling in a box separate of a system case and another was extracted from an In-Win IW-J551TA case. There are only superficial differences between the two: the boxed PSU doesn’t have an input voltage switch and can only work in a 220VAC power grid whereas the PSU preinstalled in the system case is equipped with such a switch.
This model has been long in In-Win’s product range and has become hopelessly obsolete by now. It complies with the ATX12V 1.3 standard that was ratified in April 2003 as transitional from ATX12V 1.2 to 2.0. This transition is long over, and most of the major manufacturers do not offer PSUs compliant with the ATX specification earlier than 2.0 even for inexpensive microATX system cases now. But here we are dealing with a power supply that sells either separately or in a full-size ATX system case!
The internal design of the PSU is familiar to me by my earlier tests of In-Win products. It is a typical enough product with a circuit design that is steadily passing away: a half-bridge push-pull converter with a controller on an SG6105 chip (it may be remarked as In-Win IW1688 in some PSUs). The standby source is assembled out of discrete components. Active PFC is not provided for, but passive PFC can be installed.
The PSU has thick aluminum heatsinks with horizontal ribs. That’s not the best thing possible for the fan that is blowing at them from above rather than sideways. The PSU has got the following cables and connectors:
As I wrote above, the PSU is declared to comply with the ATX12V 1.3 standard. This means that it doesn’t have to provide a load current of over 18A on the +12V power rail. In other words, the maximum output power of this power supply is going to be not much higher than 216W in a modern computer where it is the +12V rail that bears most of the total load.
There is a certain oddity about this label. There’s a bold caption, “430W Output”, but the table lists quite different numbers, with a total of only 350W. So, what wattage does the IW-P430J2-0 really have? 350W or 430W?
Alas, the power supply only lasted a couple of minutes under a load of 420W. After that time, the transistors of its inverter died with a loud plop. In-Win was overoptimistic declaring an output power of 430W for this product.
At a load of 350W there is a voltage ripple of 55 millivolts on the +12V rail and 23 millivolts on the +5V rail. There’s only high-frequency pulsation here.
The cross-load diagram looks quite good. If power consumption of today’s computers were like the ATX12V 1.3 standard implies, it would fit into the “green zone”. But modern computers put a heavy load on the +12V rail, so the +5V voltage is going to be too high, although within permissible limits.
An ARX FD1212-S3142E fan is installed in this power supply.
The speed of the fan is varied from 1050-1100rpm under low load to 1800rpm at a 350W load. The PSU has moderate noisiness. Its fan is audible even at low loads, but I guess quite a lot of users wouldn’t find its noise loud or irritating.
The PSU’s efficiency is 76% at the maximum and declines to 70% (which is the allowable minimum demanded by the ATX12V 1.3 specification) at high loads. Well, this unit can’t work more than a couple of minutes at such a load anyway.
So, the IW-P430J2-0 has produced predictable results in my tests. It could be called out-dated a year ago, but now it is downright obsolete. Major PSU makers do not produce products of an ATX12V version lower than 2.0 anymore. Priced at $40, this PSU finds itself in the same category with, say, FSP’s ATX-400PAF/PNF models that have similar characteristics but are fully ATX12V 2.0 compliant and can yield their full declared power of 400W without problems.
Unfortunately, In-Win isn’t just selling a boxed version of the IW-P430J2-0, but is also installing it into many system cases. This PSU is far from a good choice if you want to build a more or less powerful computer in such a system case.
I would want to begin to describe this PSU, which belongs to the old IW series too, with its accessories. We’ve got a boxed version of this unit as we couldn’t find a system case with it. The other PSUs have an ordinary set of accessories that includes a short manual, a power cord, four screws to fasten the PSU in the system case, and a 24-20 adapter for older mainboards.
The IW-P560A2-0 offers more:
With all these accessories, the text on the box, “New Generation of Power Supplies,” can provoke nothing but a good laugh.
The PSU itself is an ordinary gray box cooled with two fans.
It resembles the above-tested IW-P430J2-0 internally, although is more densely packed. The regulator is based on an SG6105D chip, too.
The heatsinks are larger in comparison with the previous PSU model and this creates the impression of higher component density. There have also appeared an upright-standing card with various small components (particularly, with a fan speed controller) and a card with a line filter that is soldered right on the input 220VAC connector. The main components of the PSU are identical to those of the previous model, though.
The PSU is equipped with the following cables and connectors:
There’s utter confusion as concerns the wattage of this power supply. The name of the model seems to imply 560W, but the label mentions a peak load of 600W (without specifying what the term “peak” means here and what the maximum non-peak load is). The table contains a total of 600W with something which goes without the word “peak” (and I really wonder what is the purpose of declaring output power with one decimal place precision). So I have to carry out an experiment again.
The experiment ended with the following outcome: the power supply burned down in less than a minute at an output power of 540W. It was again the inverter’s transistors that proved to be the weakest link and the consequences were very picturesque:
I had unsoldered the high-voltage capacitors to make this photograph. The transistors are an impressive sight indeed with their burnt legs and pieces of plastic torn out of their cases. The surrounding components are covered in soot and copper from the legs of the transistors. Some of the copper paths on the reverse side of the PCB had just evaporated, leaving a black smudge on the PSU case.
I didn’t have a second sample of the PSU, so I had to shorten my test program. I hadn’t measured the output voltage ripple and the efficiency of the PSU before its death, but considering the similar circuit design, I think these parameters don’t differ much from those of the IW-P430J2-0.
The cross-load characteristic of this PSU isn’t perfect, but overall acceptable.
The PSU is equipped with two fans from Delta Electronics, an AFB0812H (80x80x25mm) and an AFB0812HB (80x80x15mm). They have similar speeds that are changing within a range of 1500-3000rpm, making the IW-P560A2-0 an average-noisiness model.
Telling you the truth, the purpose of this model’s existence in 2006 A.D. is a mystery to me. This power supply might make an acceptable midrange product if the manufacturer declared a wattage of about 450W for it and equipped it with normal modern connectors instead of the heap of adapters. This product would have a chance then even despite its compliance with the ATX12V 1.3 standard because it provides 25A on the +12V rail as opposed to the previous model’s 18A. The marking of this PSU doesn’t make it clear what exactly wattage it has, and installing it into a modern computer makes it necessary to use a lot of adapters, most of which are absurd and some even look alarming (yes, I mean the graphics card power adapter). The IW-P560A2-0 comes to retail at about $90 and there exist much more interesting alternatives from a number of manufacturers for that money!
The AJ series has but recently appeared on In-Win’s catalogue, but its specs are very similar to those of the A and J models I have described above. It is still the out-dated ATX12V 1.3 standard. So what is the difference between the IP-P300AJ2-0 and the old IW-P300J2-0, for example?
Different output power ratings are the single discrepancy between the IP-P300AJ2-0 and the IP-P350AJ2-0, so I will describe and test them both together. The former PSU comes in the In-Win IW-S506 system case and the latter comes in a box.
These PSUs don’t seem to differ externally from the IW-P430J2-0.
But they are dramatically different on the inside! These PSUs don’t just use another PCB, they feature quite different components. An UC3843AN chip is used as a PWM controller now. The PSU features a single-ended single-transistor converter. That’s not a very popular solution due to the necessity of using 900V transistors (a double supply voltage, i.e. over 600V, is applied to them in this circuit design). The standby source is at last based on an integrated chip with a minimum of accompanying components instead of a handful of discrete components. This Infineon ICE2A0565Z chip is a specialized PWM controller for low-power sources, up to 23W when working in a 220V power grid and up to 13V when working in the universal range of 85-265V, which is quite enough for a PSU’s standby source. A special-purpose chip from Weltrend is used as an output voltage supervisor. These changes are all welcome. In-Win has finally transitioned to using more modern components than in the PSUs described in the previous sections.
Heatsinks punched out of an aluminum bar are used now instead of extrusive heatsinks. They look cheaper but I wouldn’t claim they are inferior to the older ones. Such heatsinks are better suited to receive airflow from above.
The PSUs still don’t have active PFC (a passive PFC device can be installed optionally) and have group voltage regulation.
The junior (300W) model is equipped with these cables and connectors:
The senior model has 5cm longer CPU and mainboard cables and a non-separable 24-pin connector instead of the splittable 20+4 one. An adapter is enclosed with it to connect to older mainboards. The rest of its connectors are the same.
Unfortunately, the PSUs still belong to the out-dated ATX12V 1.3 standard, and the load current of the +12V power rail is not higher than 18A (for comparison: the typical 300W ATX12V 2.0 power supply must provide a current of 22A on that rail; the difference is bigger with PSUs of higher wattage). Otherwise, I have no questions about the specs. All the numbers match each other, and there is no mention of the word “peak”.
The 350W model differs only in the overall wattage because the load currents are the same. I want to remind you about the specifics of power consumption of modern computer systems again. Most of the power they consume comes from the +12V power rail whereas the combined power draw from the +5V and +3.3V rails is not higher than a few dozen watts. So, even if the IP-P300AJ2-0’s 12V rail is loaded fully (216W), there is still 64W left for the +5V and +3.3V rails which is more than enough for a majority of computers this power supply may be installed in. Replacing it with an IP-P350AJ2-0 doesn’t make any difference because you will reach the 216W limit on the +12V rail sooner than you reach an overall output power of at least 300W. This is an example of how the capabilities of a power supply may be limited not by its wattage but by its compliance with an out-dated version of the ATX standard. As a result, there is in fact no practical difference between these two PSUs despite their different wattage.
The senior model could work with an APC SmartUPS SC 620 at a load up to 310W irrespective of the power source (electric mains or batteries); the switching to the batteries was performed without problems. The junior model’s output power was limited by its own capabilities rather than by the UPS.
At a load of 330W there is a voltage ripple of about 9 millivolts on the +5V rail, 17 millivolts on the +12V rail, and 12 millivolts on the +3.3V rail. There is no low-frequency pulsation here. The results of the junior PSU model aren’t much different.
The two models have produced almost identical cross-load diagrams, save for the difference due to their max output power. The diagrams look good. It is only the +12V voltage that reaches the maximum allowable deflection, but it does so when there is a very high load on the +5V rail, an improbable situation for modern computers.
An ARX FD1212-S3142E fan is installed in each PSU, and its speed is steadily increasing as the temperature grows up (the sensor is fastened on one of the juts of the heatsink with diode packs). The speed is about 1100rpm at maximum, so these are average-noisiness models, too. The fan is perfectly audible at that speed. Quite a lot of users wouldn’t have anything against it, but you may want to consider other PSU models if you want to have a silent computer.
The efficiency of the PSUs is about 79%. That’s not impressive by today’s standards, yet better by a few percent than with the previous models. The power factor is about 0.65, which is a typical value for PFC-less products.
So, notwithstanding their compliance with the ATX12V 1.3 standard, the new PSUs from In-Win don’t look too bad. The company has finally transitioned to modern components, increased the efficiency, and wrote a true-to-life output power rating on the label. Still, the AJ series models can’t be anything more than entry-level power supplies due to such reasons as the old version of the ATX standard they comply with and the rather noisy fan. The IP-P300AJ2-0 (or a system case with it) can be bought for a typical office computer. The 350W model doesn’t provide any real advantages over the 300W one.
Finally we’ve come across two In-Win power supplies that are ATX12V 2.2 compliant!
Both models come to market in boxed versions. Although they are described as OEM versions on the In-Win website, I haven’t yet seen system cases selling with such PSUs. Two models with the Q index, a 300W and a 400W one, are declared to support BTX system cases, which are yet very rare and exotic. Among ATX system cases from In-Win there can be found only one high-wattage power supply. It is the above-described IW-P430J2-0 that cannot provoke much enthusiasm.
The PSUs have traditional gray-colored cases although the In-Win website shows a photograph of a dark-gray case with blue fan highlighting. Note the words “Full Range” on the PSU. This implies active PFC.
There is indeed a PFC device here. You can see its choke on the left, next to the two barrels of high-voltage capacitors. The PFC controller is based on an Infineon ICE1PCS02 chip located under that choke.
The circuit design is otherwise similar to the AJ series models described earlier in this review. The standby source is based on an ICE2A0565Z chip; the main regulator on an UC3845B chip (it is placed on a small separate card). The supervisor is based on a Weltrend WT7525 (on a separate card, too). The heatsinks have become intricately shaped and larger.
The PSUs both have an identical set of cables and connectors:
Each of the PSUs offers two +12V output circuits with a max combined current of 23.3A (the junior model) and 26.7A (the senior model). This is higher than 18A offered by older ATX12V 1.3 power supplies, but the ATX12V 2.2 standard demands even more. According to it, typical 400W and 450W power supplies must have a +12V power rail with a load capacity of 27A and 30A, respectively. In other words, these In-Win PSUs still lag one step behind the requirements of the ATX standard.
At a load of 450W the senior PSU model has a voltage ripple of 13 millivolts on the +5V rail, 32 millivolts on the +12V rail and 12 millivolts on the +3.3V rail. The junior model doesn’t differ from it much in this respect. There is no low-frequency pulsation on the output.
When working with an APC SmartUPS SC 620, the power supplies could each work under a load of 330W and 315W from the mains and batteries, respectively. The switching to the UPS’ batteries was performed without problems.
The cross-load characteristics of the power supplies coincide closely, so I only publish the senior model’s diagram. The junior model’s diagram can be made out of it by discarding the loads higher than its max output power.
The voltages are quite stable for power supplies with group voltage regulation. The PSUs cope easily with almost any load. It’s only at a catastrophic load misbalance (which can never be achieved in a real computer) that the +12V voltage exceeds the allowable limit. In the area of real-life loads it is all very pretty, except that the +5V voltage is just a bit too high.
ARX FD1212-S3142E fans are installed in the PSUs. The speed is varied from 1100 to 2000rpm. That is, the power supplies are average in terms of nosiness, mostly due to the rather high fan speed at low loads. They will suit an office computer well, but you may want to choose something quieter for your home PC.
The efficiency of these PSUs is the same 79% as with the AJ series models. It declines a little at full output power. By the way, this is another example that active PFC has no great effect on the efficiency factor: there is additional loss in the PFC circuit, but the loss in the main regulator’s primary circuits is somewhat reduced because it receives a higher voltage (400V) from the active PFC device.
The power factor is fluctuating around 0.99 through the entire load range.
So, In-Win has finally produced ATX12V 2.0 power supplies. It would be just immodest to delay any longer with their release, after all. There is nothing exceptional about these power supplies, but In-Win has never targeted the top market segment with its products. We now have to wait for the Q series power supplies to begin to be installed into the company’s system cases.
When I was beginning this test session I had five boxed In-Win PSUs at my disposal, but then I wanted to see how the PSUs differed from those they installed into system cases. So I took a few system cases that differed in the declared PSU wattage. In one of them I found an IP-P300AJ2-0 which didn’t differ much from the boxed IP-P350AJ2-0. Another system case contained an IW-P430J2-0 which didn’t differ at all from the boxed version. The third system case turned to have a power supply from FSP Group. Although it doesn’t fit too well into this review, I still want to test it. After all, it is installed in an In-Win system case, the IW-V523TA micro-ATX model.
The PSU has a plain gray case and an 80mm fan. The punched-out fan grid is fastened on only two screws, making the PSU look what it is, a regular entry-level model.
The steel of the case is thin, but the manufacturer took care that it didn’t rattle. The parts of the case are fastened together not only with four screws at the top, but also with three screws at the bottom. This makes the whole arrangement very rigid.
Inside I found a familiar platform that FSPxxx-60THN/THA and ATX-xxx60PN/PA series units are based on. These were the first power supplies from FSP to comply with the ATX12V 2.0 standard, but this platform has moved into the entry-level sector by now, leaving place to newer products. By the way, FSP has already developed two generations of ATX12V 2.0 power supplies whereas In-Win has released such units only recently.
The PSU has small heatsinks to one of which a fan speed controller card is fastened. The input capacitors have small capacitance, two by 470µF. Besides these, I can find no traces of the manufacturer’s trying to cheapen the cost of this product. The PSU is equipped with passive PFC and offers the following cables and connectors:
The cables are short, but it’s normal because the system case it is installed in is microATX rather than full-size ATX. Still, they might have made two SATA power plugs, I think.
The PSU seems to be intended to sell with a system case and doesn’t come to shops on its own. That’s logical from a marketing point of view. FSP is offering several good models in the entry-level sector (priced at $20-30), so there is no sense in undermining their sales with yet another model.
The specs of the ATX350F-C meet the ATX12V 2.0 requirements. It has two +12V lines with an allowable combined load of 23A (this is 2A lower than the recommendation for the typical 350W unit, but 1A higher than is recommended for the typical 300W unit) and a very moderate load capacity of the +5V and +3.3V rails (modern computers just don’t need much from them).
The output voltage ripple amounted to 55 millivolts on the +12V rail and 25 millivolts on the +5V rail at full load of 350W. There was both high- and low-frequency present. The latter is not a good fact, but the pulsation amplitude was two times below the acceptable maximum, so I can’t have any real complaints.
The cross-load diagram is generally the same as with In-Win’s PSUs: the +3.3V voltage is very stable while the +12V and +5V change greatly depending on load. They only violate the acceptable limits at a very strong load misbalance, though. Everything is going to be all right in a real computer, although the +5V is going to be 0.1-0.15V higher than it should be (to remind you: a deflection of 5% or 0.25V is the allowable maximum).
The PSU is equipped with an 80mm fan from Yate Loon. Products of that firm have gained a reputation of inexpensive, yet rather quiet fans. The speed is changed in a linear way from 1200 to 2700rpm, making this power supply, like the In-Win PSUs, average in terms of noisiness. The fan is barely audible at low speeds, but makes itself heard at 1800-2000rpm and higher. Its noise isn’t irritating, though.
The PSU is 76% efficient at maximum. It has a power factor of about 0.74 thanks to its passive PFC.
So, this power supply has performed expectably in my tests. I say “expectably” because I have already met and tested this platform from FSP under different names. PSUs based on it are stable and predictable, and the ATX350F-C doesn’t make an exception. Despite its manifest entry-level design it has good parameters, complies with the ATX12V 2.0 standard (something In-Win’s power supplies are yet to grow up to), and features problem-free operation. It’s not an exceptional product, but it will suit an office-class computer perfectly.
I should confess I am not satisfied with the results of In-Win’s power supplies in this test session. On one hand, I’ve got an impression the company is aware of its lagging behind the competitors and is trying to catch up by introducing a new basic platform on modern components and releasing power supply models that comply with the up-to-date ATX12V 2.2 standard. On the other hand, its product range is still predominated by old PSU models. Even among the boxed products there are two hopelessly out-dated models, one of which left me with a feeling of utter bewilderment. As for In-Win’s system cases, they come with only one ATX12V 2.0 power supply, which is manufactured not by In-Win but by FSP Group!
In-Win should make haste in putting new PSU models into production. It is bad practice to be selling not-very-cheap system cases with a power supply that is declared as 430W but whose real wattage is only 350W and that complies with an ATX specification that has been out-dated for two years already. They should take an example from FSP Group whose even junior PSU models are up to today’s standards.
My shopping advice to you is to check out beforehand what power supply is installed in the In-Win system case you want to buy. Consider those models that have the letters IP, not IW, in the name. At the current moment you have to pay $5-8 for each additional step in the PSU wattage (300-350-430W) when you buy a system case, but this test session has shown that it’s stupid to buy the system case with the IW-P430J2-0 with its false 430W. Choosing between the 300W and 350W models (IP-P300AJ2-0 and IP-P350AJ2-0, respectively), you can go for the junior one without losing in anything. If you think your computer needs a power supply with higher wattage, you have to buy one separately because In-Win doesn’t install anything really more powerful than the IP-P300AJ2-0 into its system cases as yet. I hope this situation is going to change and we’ll see modern Q series PSUs in the company’s popular system cases, too.
Among the boxed power supplies from In-Win you should pay attention to the models with the letters IP in their name, too. I can’t say they are interesting products, but they may suit an undemanding user due to their low price. They have acceptable parameters and average level of noise and are surely superior in the quality of manufacture to many even cheaper brands.