Contents
NSCom P4-500W
NSCom (Short for North Shore Computers) is actually the name of a computer parts retail shop based in Australia, and are not well known as a power supply manufacturer. I’ve seen one power supply from this brand some years ago, before Hardware Insights existed, and my opinion of it was anything but favourable. It was right down there with units like the Excel and A-Power units in last year’s roundup. Let’s see now if they’ve decided to get their act together, or if their power supplies are still complete junk like my last unit.
_
Yet another 5V-heavy design.
Load Testing
Test 1 (113.26W Load)
| Rail | Load | Voltage | Ripple |
| 12V | 4.7A | 12.01V | 117.2mV |
| 5V | 4.9A | 4.9V | 137.8mV |
| 3.3V | 9.97A | 3.29V | 62.8mV |
| −12V | 0A | −11.89V | 34.2mV |
| 5Vsb | 0A | 4.97V | 31.4mV |
| AC Power | 146.2W | ||
| Efficiency | 77.47% | ||
| Power Factor | 0.6 | ||
.
Test 2 (196.43W Load)
| Rail | Load | Voltage | Ripple |
| 12V | 9.4A | 12.01V | 160.2mV |
| 5V | 9.54A | 4.77V | 159.4mV |
| 3.3V | 9.82A | 3.24V | 71.8mV |
| −12V | 0.1A | −12.47V | 86.0mV |
| 5Vsb | 0.99A | 4.96V | 42.0mV |
| AC Power | 248.5W | ||
| Efficiency | 79.05% | ||
| Power Factor | 0.57 | ||
.
Test 3 (249.9W Load)
| Rail | Load | Voltage | Ripple |
| 12V | 14.1A | 11.84V | 183.0mV |
| 5V | 9.54A | 4.77V | 162.6mV |
| 3.3V | 9.73A | 3.21V | 81.8mV |
| −12V | 0.11A | −12.78V | 105.2mV |
| 5Vsb | 0.99A | 4.93V | 45.8mV |
| AC Power | 318.8W | ||
| Efficiency | 78.39% | ||
The 12V rail started off at 12.01V, and dropped to 11.84V during Test 3, which gives us a variation of 0.17V, or 1.41%. The 5V rail was at 4.9V during Test 1, but was near the minimum allowed in ATX specifications at 4.77V during Test 2. This equates to 4.6% regulation, with a drop of 0.13V, or 2.6%. The 3.3V rail stayed between 3.29V and 3.21V, which equates to 0.09V, or 2.73% regulation, and a drop of 0.08V, or 2.42%.
The efficiency was better than we saw with the Codegen and the PSI, but it still didn’t quite manage 80%. On 230V, many non 80plus certified power supplies can at least peak at 80% efficiency. Unfortunately, I couldn’t kill this unit with the load tester. It would shut down when I tried to ask for 300W load. I guess I’ll just have to kill it with my sledge hammer instead.
| Rail | Test 2 (196.43W) | Test 3 (249.9W) |
| 12V |  |
 |
| 5V |  |
 |
| 3.3V |  |
 |
| −12V |  |
 |
| 5Vsb |  |
 |
.
This power supply’s ripple suppression was absolutely horrible. The 5V rail started out at almost 138mV – more than double the maximum allowed limit. In the other tests, it was over three times the maximum limit. The 12V rail’s ripple was also very bad – off the scale during Test 3. The 3.3V rail fared somewhat better, but it was still too high. Only the two low current rails had acceptable ripple levels. Ripple this high will put considerable stress on attached hardware and shorten its life. It will also cost the unit all 10 points from the final score.
Disassembly
_
Normally, we start with the input filtering when discussing the internals of a power supply. This unit, however, lacks it entirely. There’s not a single coil or capacitor in sight. There is PCB space for two Y capacitors, but the manufacturer didn’t install them in order to cut costs. The two primary capacitors are rated at 560µF, but most 560µF capacitors are physically larger than these parts, so it wouldn’t surprise me if they are actually only 470 or even 330µF. The switching transistors are 13007s. I couldn’t identify their real manufacturer, but in all likelihood, they are rated at 8A, like all other 13007 transistors.
The 12V rectifier is from an obscure manufacturer, who I don’t recognize the logo of. Going by the part number (F16C20), it looks to be a 16A Fast recovery rectifier. The other rails use WuXi HuaJing 2CZ2045A8 schottky rectifiers, which are rated at 20A. Considering that the label claims 40A and 26A respectively on the 5V and 3.3V rails, these parts are very undersized. The capacitors are also from obscure manufacturers. Some are branded Huahong, and others are branded Telecon (and no, that’s not supposed to be Telecom). There is only one of them on each of the main rails, and there are no PI filter coils. This explains why the ripple was so high during the load tests. Interestingly, one of the 5V wires is soldered in with the toroid coil.
_
The fan is branded HeQun. There was some lubricant in the bearings, but there was no rubber plug to prevent it from leaking out. This power supply lacks a fan controller, and the fan was quite loud. The heat sinks have a good amount of surface to air contact, but are thin, so they won’t be as good at conducting heat away from the silicon as they should be.
Specifications and Conclusions
| Real Wattage | 0W |
| OEM | Unknown |
| PFC | None |
| Price | $15 AUD |
| ATX Connector type | 20+4 pin |
| Worst-case voltage regulation (12v, 5v, 3.3v) | 1.3%, 4.6%, 2.7% |
| Worst-case ripple (12v, 5v, 3.3v) | 183.0mV, 162.6mV, 81.8mV |
| Worst-case efficiency | 77.47% |
| Input filtering | None |
| CPU Connector | ATX12V (4 pin) |
| PCIe Connectors | 1x 6 pin |
| Molex (Peripheral) Connectors | 2 |
| FDD Power connectors | 1 |
| SATA Power connectors | 2 |
.
Pros: Working OPP
Cons: Can’t deliver anything with ripple in spec (−10), Mediocre voltage regulation, Old 5V-Heavy design, Low quality capacitors, Low quality Fan, Noisy, Inefficient
Score: 0/10
Loading ...