The 2012 el-cheapo Power Supply Roundup

Sun Pro ATX-550W

Yes, I know what you’re all thinking: “You already tested one of these last year. Why are you doing it again?” Well, this one feels a bit lighter than the last one, and so I opened the cover and sure enough, this one is a bit more gutless than the last one. Sun Pro is obviously not very consistent with their designs.

_

Externally, this power supply is exactly the same as the one I blew up last year. The “low noise design” bullet point at the bottom of the label could hardly be further from the truth. Although the internals are somewhat different from last year’s unit, this one uses the same loud fan and also lacks a speed controller. The exclamation mark on the top is there to indicate that this one was the lower quality variant, not a faulty unit.

Load Testing

Test 1 (120.53W Load – Cold)

Rail Load Voltage Ripple
12V 4.94A 12.35V 54.2mV
5V 4.99A 4.99V 49.8mV
3.3V 10.24A 3.38V 4.4mV
−12V 0A −12.11V 32.2mV
5Vsb 0A 5.07V 4.0mV
AC Power 162.23W
Efficiency 74.30%
Power Factor 0.6

Test 2 (203.84W Load – Cold)

Rail Load Voltage Ripple
12V 9.89A 12.36V 72.8mV
5V 9.8A 4.9V 60.8mV
3.3V 10.09A 3.33V 6.2mV
−12V 0.1A −12.38V 70.4mV
5Vsb 1A 5.01V 7.2mV
AC Power 275.53W
Efficiency 73.98%
Power Factor 0.58

 

Test 3 (256.79W Load – Cold)

Rail Load Voltage Ripple
12V 14.56A 12.13V 93.4mV
5V 9.76A 4.88V 74.0mV
3.3V 9.94A 3.28V 7.0mV
−12V 0.11A −12.77V 86.2mV
5Vsb 0.99A 4.96V 9.0mV
AC Power 353.5W
Efficiency 72.64%
Power Factor 0.58

Test 4 (303.53W Load – Cold)

Rail Load Voltage Ripple
12V 17A 12.14V 116.2mV
5V 14.13A 4.71V 81.8mV
3.3V 9.64A 3.18V 8.0mV
−12V 0.11A −13.41V 111.0mV
5Vsb 0.98A 4.88V 27.4mV
AC Power 440.84W
Efficiency 68.85%
Power Factor 0.58

Now that is nothing short of appalling. All I could pull was 120W with no load at all on the 5Vsb and −12V rails (which is an unrealistic scenario in a PC) with the ripple in spec. In a more realistic load scenario, the 5V rail’s ripple was too high and the ripple on the 12V rail, although still in spec, was getting very close to the limit, at 116mV. This unit was also very inefficient. The efficiency never even reached 75%, despite having 230V to play with. The voltage regulation was also sub-par. The 5V rail was below the minimum 4.75V allowed in ATX specifications during Test 4. The 3.3V rail’s regulation wasn’t very good either. The power supply exploded when I asked for 350W load.

Rail Test 3 (256.79W) Test 4 (305.53W)
12V
5V
3.3V
−12V
5Vsb

A Look Inside

_

The primary side is identical to last year’s unit. The input filtering consists of two common-mode chokes, two X capacitors, three non-safety-rated ceramic capacitors (including the one after the rectifier) and three MOVs (one directly across the AC input, two across the primary capacitors). It has the same 470µF JEE primary capacitors and the same 12A D304X switching transistors. The four diodes are rated for only 2A, which is theoretically insufficient for a 550W power supply, although that’s specified as the “average” current and each diode only conducts on one half of the AC cycle so they might survive 3 or 4A. Such small diodes may also be stressed by the inrush current when plugging the unit in.

The secondary side, though, is quite different from last year’s unit, and this is what really lets this power supply down. There are fewer capacitors, and no PI filtering coils, which explains why the ripple was so much higher. Bolted to the heat sink, the 12V rectifier is a MOSPEC U20C20C, rated for 20A at a case temperature of 100°C. Not great, but good enough for the 17A claimed by the label. The 5V rectifier is an S30D45CS Schottky, rated at 30A. Um, weren’t we supposed to have 38A on this rail? The 3.3V uses linear regulation from the 5V rail, which explains why the 3.3V ripple was so low, as linear regulators are generally very good at cleaning up ripple. The transistor used is a UT40N03G, rated at 40A. The secondary capacitors are all Sapcon brand.

_

The fan is also the same as last year’s unit, and so it was also very noisy.

Specifications and Conclusions

Real Wattage 120W
OEM Sun Pro
PFC None
Price Unknown
ATX Connector type 20+4 pin
Worst-case voltage regulation (12v, 5v, 3.3v) 3%, 5.8%, 3.6%
Worst-case ripple (12v, 5v, 3.3v) 116.2mV, 74.0mV, 8.0mV
Worst-case efficiency 68.85%
Input filtering Adequate – but not safety rated
CPU Connector ATX12v (4 pin)
PCIe Connectors None
Molex (Peripheral) Connectors 4
FDD Power connectors 2
SATA Power connectors None

 

Pros: At least it has input filtering

Cons: Exploded at 350W load, horribly inefficient, no SATA connectors, Poor ripple suppression, Poor voltage regulation, Noisy, Poor build quality

Score: 1/10

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