PSUs

This time last year Custom PC took the market by storm as the first publication in the UK to scientifically labs-test PSUs. We set ourselves a tough act to follow, so this year we've performed even more in-depth tests on three times as many PSUs as we tested last year.

But why should you be so concerned about which PSU powers your PC? Let me put it this way: there's practically no point in buying a new CPU or graphics card if your current PSU isn't up to the job. Modern components require substantially more power than their predecessors, and without a sufficient, stable power supply, your graphics card's GPUs may be automatically underclocked or your PC could become unstable and even crash. Even worse, too much or too little voltage may fry a component.

So, with power demands on the increase, we've rounded up 29 different PSUs, divided into two categories: up to 500W, and 501W and above. But whatever your PC's power requirements and budget, you should look out for just one thing: a PSU that delivers a stable voltage on each of its rails without overheating, blowing up or needing to be cooled by a fan that sounds like a tropical storm.

To find out which PSUs are the quietest and most stable, which can really do what their manufacturers claim, and which six blew up during the 100-per-cent load stress test, read on.

WHAT A PSU DOES

Ultimately, a PSU is simply an AC to DC converter - converting 240V AC to the voltages required by the components inside a PC. However, while a laptop PSU only has to output a single voltage (typically around 16V), a desktop PSU has to output four different voltages (known as rails): 3.3V, 5V, 12V and 5V standby.

WATT TO LOOK FOR

For most mid-range PCs with a couple of hard disk drives and optical drives, and a single 3D card, a PSU of between 350 and 500W will suffice. However, if you're building a very high-end SLI system, using a dual-core CPU and top-end 3D cards, you should think about getting a more powerful unit.

Aside from power, you also need to consider which connectors your PC will need. Many new motherboards have 24-pin ATX connectors instead of the older 20-pin connectors, while most dual-CPU and dual-core CPU motherboards require an extra 8-pin EPS12V connector too.

Unfortunately, the trouble with selecting an appropriate PSU is the limited information provided by the manufacturers, which doesn't really tell you how good the PSU really is.

Sure, each PSU is rated as capable of outputting a certain wattage - for example, 500W - but this really doesn't tell you much, as a lot of manufacturers list the wattage in an extremely confusing manner.

Many simply add up the wattage over the primary rails (3.3V, 5V and 12V) and list this as the total. However, no PSU is capable of the maximum wattage from each rail simultaneously, so this method gives an overly generous impression of the PSU's capabilities. Make sure that you read the label carefully to find out how many watts can be produced in parallel by each rail. More importantly, check the current rating on the 12V rail, as this is used to power modern CPUs and graphics cards. For example, Nvidia recommends a minimum of 26A on the 12V rail(s) for a GeForce 7800 GTX.

Unfortunately, looking at the label won't tell you anything about voltage stability, which is probably the most important factor in a PSU. Producing a stable voltage on each rail is a pretty tough task, and some PSUs are much better at it than others. This is especially true of 500W+ PSUs, as most of these will have two or more 12V rails. Voltage stability is important because some components may not start up if the voltage is too low, or could burn out if the voltage is too high. And you certainly don't want to risk damaging your new £300 CPU or graphics card, all because you bought a nefariously labelled PSU.

Fortunately, the Intel ATX spec, which you'll find at www.formfactors.org, lays down in black and white the physical and electrical characteristics necessary to adhere to the spec. If you read last year's PSU Labs test, in which we exposed several dodgy PSUs, then you may be surprised to hear that there are still lots of iffy power supplies available that don't meet the ATX spec.

The ATX spec cites that there can be a maximum 5 per cent variance above or below the voltage on the 3.3V, 5V, 12V and 5V standby rails, and up to 10 per cent on the -12V rail. To save you working this out for yourself, the table below shows the minimum and maximum voltages for each rail.

HOW WE TESTED

Since Custom PC was the first publication in the UK to carry out a scientific Labs test of PSUs, we though it imperative to maintain our market-leading status and go one step further this year. Instead of testing 12 PSUs, we sent 31 little square boxes off to Maxpoint in Hamburg, the only company in Europe that possesses the equipment required to properly load-test modern PSUs. As before, the tests were carried out solely by the Custom PC editorial team (namely James) on industry-standard Fast Auto FA-828ATE and Statron 3229 equipment.

To test each PSU, we programmed the load testers to drain the amount of power that each manufacturer claims its PSU can deliver. The voltage of each rail was tested at 50, 75 and 100 per cent loads to see if it was within the ATX spec. We then left each PSU running at 100 per cent load for 30 minutes to find out if it could produce stable voltages over an extended period. This meant a set of between 20 and 40 gruelling tests per PSU, with each test getting steadily harder as we drained more power from the PSU.

Let's take the generic 600W PSU as an example. Its manufacturer claims that it can output 29A on the 3.3V rail, 45A on the 5V rail and 22A on the 12V, and has a total combined power of 600W. As it only has a single 12V rail, this meant we had to carry out 20 individual measurements to test its voltage stability.

As you can see, this PSU failed to provide sufficient voltage on the 5V rail when we tried to draw 75 per cent from it (Test 2). The failed test measurement is highlighted in red to make it easier to spot. This means that components drawing power from the 5V rail may fail to work properly. Even more seriously, the PSU switched itself off (permanently) when we tried to draw the full 100 per cent claimed load (Test 3).

An area of increasing interest is the overall efficiency of the PSU itself, or, in other words, how much power it requires from the mains to produce the requested load. Efficiency is very important because, like an internal combustion engine, any wasted energy is dissipated as heat, which the cooling system has to get rid of before it reduces the effectiveness of the electronic components inside the PSU, or catches fire. This means that a less efficient PSU will require a more powerful and noisy cooling system. An efficient PSU, on the other hand, will run cooler and draw less power, so it will help to keep your electricity bill down. The ATX spec cites that at 50 per cent load, efficiency must be 72 per cent or higher, and at 100 per cent load, 70 per cent or higher. The efficiency figures are also quoted on the test sheet.

It's also important to test the efficiency of the Power Factor Correction (PFC) circuitry of a PSU. The PFC distributes the power between the different rails, so an inefficient PFC means that a lower fraction of the PSU's available power actually gets distributed across each rail. And because components drain different amounts of power on different rails over time, it's very important for a PSU to have good PFC. The PFC of the generic 600W PSU is also listed at the end of each test on the Test Sheet.

THE SCORES

Although there was a grand total of 1,030 test measurements to analyse, the process was made slightly easier by splitting the PSUs into two categories - up to 500W, and 501W and above. It would be pointless to compare a quad 12V rail 850W PSU with a generic 400W, so the two categories of PSUs were scored separately.

The voltage stability tests take pride of place in the scoring box and account for 55 per cent of each PSU's overall score. The scores shown are the number of test measurements that are within the limits of the ATX spec and have been converted into a percentage to allow PSUs with a different number of rails to be compared.

Obviously, it's only really worth seriously considering a PSU that has perfect (100 per cent) stability. Unfortunately, we don't have the space to print all the voltage stability measurements, but the graphs on p84 and p94-96 show the voltages during the 100 per cent load test.

The Features score includes the overall efficiency and PFC efficiency as measured during the 50, 75 and 100 per cent load test scenarios, the number, type and quality of the cables, plus how noisy the cooling fan(s) are at 100 per cent load. The Value score is a combination of the Stability and Features scores and the total output (in watts) divided by the price. These three scores were then added together to provide the Overall score of each PSU.