Poly-Modding: Does it actually work?

Antec SmartPower SP-400

The Original Capacitors

_

Like all Antec SmartPowers, this one originally came with Fuhjyyu capacitors, which is one of the worst brands available. As a result, they never last longer than three years before the capacitors fail. Unfortunately, I don’t have any pictures of this power supply with it’s original capacitors. It was re-capped with equivalent Japanese electrolytics long ago. With the exception of the 12V rail, all of the rails have two capacitors which filter them. The 12V rail uses a single 3300µF 16V capacitor, the 5V uses a 1000µF (replaced with a 1200µF) and a 2200µF capacitor, both rated at 10V, the 3.3V rail uses a 3300µF and a 4700µF capacitor (both rated for 6.3V) and the 5Vsb rail uses two 1000µF 6.3V capacitors.

The replacement capacitors

_

Power supplies are generally optimized to work best with a certain capacitance and ESR, and significantly changing the properties of the capacitors can, in theory, adversely affect the performance. For this reason, I’ve never believed that power supplies should be poly-modded. Additionally, 3300µF polymer capacitors simply don’t exist, and neither do 200V ones, so that means I’ve had to significantly lower the capacitance on most of the rails and will be leaving the two big capacitors alone. Personally, I don’t think this is going to work well, but I’ve been wrong before.

The Testing

Each poly-modded power supply will be run through two tests. The first will be a performance test on my load tester at 100% load. To pass, the voltages and ripple have to be in spec. The second test involves running the power supply for 24 hours in a power-hungry and hot running Pentium D based computer which will be running Prime95 in its ‘In-Place FFT’ torture test mode, which is designed for maximum heat generation and power consumption. The PC will be housed in a cheap and rather badly ventilated case, so it will get hot. To pass this test, the computer must still be running 24 hours later with no lock-ups, blue screens of death or other errors. Admittedly, I do prefer testing power supplies on the load tester, but it’s simply not practical for me to sift through 24 hours worth of data just to make sure that the power supply was behaving itself. If the power supply fails the first test, I won’t run it through the second, since it’s already failed the testing and could damage my testing PC.

After the results of the first test, I had high hopes for this unit, since the performance was very similar to how it was originally. However, my hopes were short-lived, because my test PC wouldn’t even boot with the poly-modded SP-400 installed. This unit is clearly not designed to work properly with these capacitors. Let’s see if our next contender can do better.

Pages: 1 2 3 4 5 6 7 8 9 10 11