Contents
Cooler Power GX850
It really sounds as though this company is trying to copy the Cooler Master GX series with that model name. Given that these units can sell for around the $60 mark, they’re one of the most expensive units to feature in one of these round-ups, although $60 is still not a lot for an 850 Watter. Let’s see if this one has some potential or if it’s just very overpriced.
The logo just removed the last bit of doubt in my mind that this really is supposed to be a knock off of the Cooler Master GX series. The label says that this power supply has four 12V rails. However, this is not true. It is a single rail unit. It is semi-gloss black in colour, and uses a 140mm fan. It also (seemingly) has passive PFC, but I do think that it is a very bad idea to mount it to the rear grille, as it blocks so much of it off. In fact, there’s only about two square inches of the grille that air can actually get out of. The blocked off parts of the fan grille are shaded orange in the above right picture. As you can see, between the fan and coil, the majority of the grille is blocked off.
Load Testing
Test 1 (116.27W Load)
Rail | Load | Voltage | Ripple |
12V | 4.7A | 12.27V | 15.6mV |
5V | 4.98A | 4.98V | 16.8mV |
3.3V | 10.12A | 3.34V | 19.2mV |
−12V | 0A | −12.24V | 14.4mV |
5Vsb | 0A | 5.1V | 17.2mV |
AC Power | 142.5W | ||
Efficiency | 81.60% | ||
Power Factor | 0.65 |
.
Test 2 (201.39W Load)
Rail | Load | Voltage | Ripple |
12V | 9.4A | 12.18V | 15.8mV |
5V | 9.78A | 4.89V | 16.2mV |
3.3V | 9.97A | 3.29V | 19.0mV |
−12V | 0.1A | −12.48V | 14.6mV |
5Vsb | 1.0A | 4.99V | 24.8mV |
AC Power | 241.2W | ||
Efficiency | 83.50% | ||
Power Factor | 0.66 |
.
Test 3 (256.78W Load)
Rail | Load | Voltage | Ripple |
12V | 14.0A | 12.12V | 16.2mV |
5V | 9.8A | 4.9V | 16.6mV |
3.3V | 9.97A | 3.29V | 19.2mV |
−12V | 9.1A | −12.48V | 15.8mV |
5Vsb | 1.0A | 4.99V | 25.6mV |
AC Power | 308.5W | ||
Efficiency | 83.23% | ||
Power Factor | 0.66 |
.
Test 4 (306.3W Load)
Rail | Load | Voltage | Ripple |
12V | 18.4A | 11.97V | 16.8mV |
5V | 9.76A | 4.88V | 17.6mV |
3.3V | 9.88A | 3.26V | 20.2mV |
−12V | 0.1A | −12.5V | 17.2mV |
5Vsb | 0.99A | 4.96V | 25.2mV |
AC Power | 371.2W | ||
Efficiency | 82.52% | ||
Power Factor | 0.64 |
.
Test 5 (360.52W Load)
Rail | Load | Voltage | Ripple |
12V | 23.1A | 11.9V | 18.2mV |
5V | 9.76A | 4.88V | 17.8mV |
3.3V | 9.82A | 3.24V | 21.6mV |
−12V | 0.1A | −12.52V | 18.4mV |
5Vsb | 0.99A | 4.94V | 25.2mV |
AC Power | 439.1W | ||
Efficiency | 82.10% | ||
Power Factor | 0.64 |
.
Test 6 (410.27W Load)
Rail | Load | Voltage | Ripple |
12V | 27.5A | 11.82V | 19.2mV |
5V | 9.74A | 4.87V | 19.0mV |
3.3V | 9.79A | 3.23V | 23.2mV |
−12V | 0.1A | −12.52V | 21.4mV |
5Vsb | 0.99A | 4.93V | 25.6mV |
AC Power | 505.9 | ||
Efficiency | 81.10% | ||
Power Factor | 0.62 |
.
Test 7 (457.63W Load)
Rail | Load | Voltage | Ripple |
12V | 31.8A | 11.73V | 20.6mV |
5V | 9.72A | 4.86V | 19.6mV |
3.3V | 9.73A | 3.21V | 25.0mV |
−12V | 0.1A | −12.54V | 24.8mV |
5Vsb | 0.98A | 4.92V | 25.0mV |
AC Power | 571.3W | ||
Efficiency | 80.10% | ||
Power Factor | 0.62 |
.
Test 8 (498.47W Load)
Rail | Load | Voltage | Ripple |
12V | 35.8A | 11.6V | 22.0mV |
5V | 9.66A | 4.83V | 20.4mV |
3.3V | 9.61A | 3.17V | 27.2mV |
−12V | 0.1A | −12.59V | 27.6mV |
5Vsb | 0.98A | 4.88V | 26.8mV |
AC Power | 635.3W | ||
Efficiency | 78.46% | ||
Power Factor | 0.62 |
.
Test 9 (552.68W Load)
Rail | Load | Voltage | Ripple |
12V | 40.3A | 11.51V | ? |
5V | 9.66A | 4.83V | ? |
3.3V | 9.58A | 3.16V | ? |
−12V | 0.21A | −12.53V | ? |
5Vsb | 1.93A | 4.82V | ? |
AC Power | ? | ||
Efficiency | ? | ||
Power Factor | ? |
.
Edit: I’m quite sure that passive PFC coil is a fake, judging by the power factor results (most PPFC PSUs achieve a power factor of 0.75 or so). —LongRunner
The 12V rail started at 12.27V and dropped to 11.51V. This equates to 0.49V (4.1%) Regulation, and a drop of 0.76V, or 6.3%. The 5V rail started at 4.98V, and dropped to 4.83V, which gives us 0.17V (3.4%) regulation, and a 0.15V, or 3.0% drop. Finally, the 3.3V rail started at 3.34V and dropped to 3.16V, which equates to 0.14V (4.2%) regulation a drop of 0.18V, or 5.5%. While this result is still in spec, it’s not great, especially on the 12V side. In fact, this is one of the worst voltage regulation results I’ve seen in a while.
The efficiency peaked at 83.5%, which is fairly average on 230V. It dropped below 80% with the power supply loaded to 500W. When I attempted to load the GX850 to 550W, it ran long enough to read the voltages, but it failed shortly after, which is why there are question marks in the table for Test 9. The part that failed was a 12V rectifier, which means it is quite likely that it would have damaged attached hardware if it had been in a PC.
Rail | Test 7 (457.63W) | Test 8 (498.47W) |
12V | ||
5V | ||
3.3V | ||
−12V | ||
5Vsb |
.
The ripple suppression, on the other hand, was excellent, staying under half of the maximum allowed for the most part.
A Look Inside
The GX850’s input filtering consists of two common-mode chokes, two X capacitors, and three Y capacitors (including the one after the rectifier), which is enough components. The two primary capacitors are rated at 12ooµF and are supplied by HEC. Unlike the other power supplies in this roundup, the GX850 uses the more modern two transistor forward topology, and not the dated half bridge. The switching transistors are two Toshiba 2SK4108 MOSFETs rated at 20A.
Moving on the the secondary side, two STPS30H100CT schottky rectifiers are used on the 12V rail. They are rated at 30A each, so the 12V rail should theoretically be capable of up to 60A (assuming perfect current sharing, which however won’t happen). In practice, however, even 40A was too much to ask. This means that the rectifier must have been getting extremely hot. In fact, when I opened the power supply, the secondary heat sink remained too hot to touch long after the rest of the power supply had cooled down. The other rails use STPS3045CW rectifiers. The 5V rail uses two, and the 3.3V rail uses one, so those two rails would be capable of 45A and 22.5A respectively. The capacitors on the secondary side are supplied by GoldLink and BH. Both of those are no-name Chinese brands.
The fan is a ball bearing part made by Yate Loon. This model has maximum speed, airflow and noise ratings of 2800RPM, 140CFM and 48.5dB respectively. It remained quiet until about 250W load. The heat sinks are very thick and have lots of surface to air contact, but this does little good when the exhaust vent behind the secondary heat sink is blocked by the (fake) PFC Choke. Maybe this explains why the secondary heat sink got as hot as it did.
Specifications and Conclusions
Real Wattage | 500W |
OEM | Golden Tiger |
PFC | Fake |
Price | $60 AUD |
ATX Connector type | 20+4 pin |
Worst-case voltage regulation (12v, 5v, 3.3v) | 4.1%, 3.4%, 4.2% |
Worst-case ripple (12v, 5v, 3.3v) | 22.0mV, 20.4mV, 27.2mV |
Worst-case efficiency | 78.46% |
Input filtering | Adequate |
CPU Connector | ATX12V (4 pin) |
PCIe Connectors | 1x 6 pin, 1x 6+2 pin |
Molex (Peripheral) Connectors | 4 |
FDD Power connectors | 1 |
SATA Power connectors | 5 |
.
Pros: Good ripple suppression, Quiet, Sleeved cables, 140mm Ball Bearing Fan
Cons: Can’t deliver labelled rating (−2), Low quality capacitors (−2), Badly placed PFC Coil/runs too hot (−1), Mediocre Voltage Regulation (−1)
Score: 4/10