Contents
Bliss ATX-350
My experience with Bliss Products is fairly limited. The few I have run into over the years have been mostly low end products which didn’t impress me with their build quality. Given that this model is available for only $30 from some retailers, my prediction is that it will be anything but blissful.
The label indicates that this unit uses a dated 5V-heavy design, which hasn’t been practical for the last decade. The power supply itself is a plain grey colour. Unfortunately, the ventilation is poor, with just a row of slots in the front as the intake.
Load Testing
Test 1 (119.55W Load)
Rail | Load | Voltage | Ripple |
12V | 4.89A | 12.22V | 58.0mV |
5V | 5.06A | 5.06V | 18.8mV |
3.3V | 10.18A | 3.36V | 19.8mV |
−12V | 0A | −12.82V | 18.8mV |
5Vsb | 0A | 5.03V | 10.4mV |
AC Power | 146.88W | ||
Efficiency | 81.39% | ||
Power Factor | 0.61 |
.
Test 2 (209.49W Load)
Rail | Load | Voltage | Ripple |
12V | 9.78A | 12.22V | 76.4mV |
5V | 9.98A | 4.99V | 20.8mV |
3.3V | 10.15A | 3.35V | 18.6mV |
−12V | 0.1A | −12.3V | 56.8mV |
5Vsb | 1A | 4.98V | 15.6mV |
AC Power | 247.72W | ||
Efficiency | 84.57% | ||
Power Factor | 0.6 |
.
Test 3 (267.61W Load)
Rail | Load | Voltage | Ripple |
12V | 14.58A | 12.15V | 93.0mV |
5V | 10.02A | 5.01V | 22.6mV |
3.3V | 10.15A | 3.35V | 19.09mV |
−12V | 0.1A | −12.49V | 74.0mV |
5Vsb | 1A | 4.98V | 17.2mV |
AC Power | 318.27W | ||
Efficiency | 84.08% | ||
Power Factor | 0.6 |
.
Test 4 (323.23W Load)
Rail | Load | Voltage | Ripple |
12V | 19.28A | 12.05V | 112.2mV |
5V | 10.08A | 5.04V | 28.6mV |
3.3V | 10.12A | 3.34V | 19.0mV |
−12V | 0.11A | −12.65V | 88.2mV |
5Vsb | 1A | 4.98V | 17.4mV |
AC Power | 391.4W | ||
Efficiency | 82.58% | ||
Power Factor | 0.62 |
.
The 12V rail varied between 12.22V in Tests 1 and 2 and 12.05V in Test 4. This equates to 1.83% regulation, and 1.42% variation. The 5V rail managed 1.2% regulation and 1.4% variation, while the 3.3V rail managed 1.81% regulation with 0.61% variation. That is good regulation, as ATX standards allow for up to 5%. However, many power supplies can do better in terms of variation, keeping the voltages more constant.
The efficiency wasn’t too bad for a low end product, peaking at 84.57% during Test 2. The power factor never made it past 0.62. This is a fairly poor result, but it’s typical of power supplies that lack PFC. I did attempt a fifth test, with the power supply loaded to 350W (it’s rating), but the switching transistors exploded during that test.
Rail | Test 3 (267.61W) | Test 4 (323.23W) |
12V | ||
5V | ||
3.3V | ||
−12V | ||
5Vsb |
.
The ripple was well suppressed on 5V, 3.3V and 5Vsb rails, staying well within limits at all times. The 12V and −12V rails were a different matter. While they were both in spec, the ripple was above half the maximum limit. On the 12V rail, it was coming very close to the 120mV limit, with less than 10mV of headroom.
Disassembly
The input filtering consists of two X capacitors, two common-mode chokes and five Y capacitors (including the one after the rectifier), but there are no MOVs for surge protection. The bridge diodes are rated at 3A, and the primary capacitors are 470µF parts made by 12Kuang Jin. The switching transistors are (now rather charred) Huashan KSH13007s rated at 8A in the dated half-bridge configuration. Unlike most $30 power supplies, this one uses a switching IC for the 5vsb, rather than the more failure-prone 2-transistor circuits. The IC used is an ST Micro VIPer22A. The PWM controller is an SDC2921.
Moving on to the secondary side, the 12V rail employs a Mospec F20C20C Fast Recovery rectifier rated at 20A. This rectifier is good enough to deliver the 18A claimed by the label, but it would be better use use a Schottky rectifier as they offer better efficiency. The 5V and 3.3V rails both use a Mospec S20C45C schottky rectifier, which is also rated at 20A. This is just good enough for the 3.3V rail, as the label claims it to be capable of 20A, but a more powerful part should have been used on the 5V rail, as it is supposed to be capable of 35A. The capacitors on the secondary side are all made by Nicon. The quality of these parts is questionable, which raises some concern about the long term reliability of this product.
It looks like Mary is back with her little Wam. The fan was fairly quiet, but unfortunately, the lubricant in the bearings was insufficient. The heat sinks do have reasonable surface area, but are very thin, so they won’t conduct heat away from the silicon and up to the fins as well as thicker heat sinks would.
Specifications and Conclusions
Real Wattage | 320W |
OEM | Unknown, Possibly XHY (Xin Hui Yuan) Power |
PFC | None |
Price | Unknown |
ATX Connector type | 20+4 pin |
Worst-case voltage regulation (12v, 5v, 3.3v) | 1.8%, 1.2%, 1.8% |
Worst-case ripple (12v, 5v, 3.3v) | 112.2mV, 28.6mV, 19.8mV |
Worst-case efficiency | 81.39% |
Input filtering | Adequate |
CPU Connector | ATX/EPS12V (4+4 pin) |
PCIe Connectors | None |
Molex (Peripheral) Connectors | 4 |
FDD Power connectors | 2 |
SATA Power connectors | 2 |
.
Pros: Good input filtering, Good Voltage Regulation, Quiet
Cons: Can’t deliver labelled rating (−2), Low quality capacitors (−2), Low quality Fan (−2), Mediocre 12V ripple suppression (−0.5), Old 5V-Heavy design (−0.5)
Score: 3/10