Introduction
We would like to thank
FSP Group for supplying the Aurum Gold review sample.
FSP Group is one of the few PSU OEMs (Original Equipment Manufacturers). Besides lending their platforms to other companies they have also a retail presence in the market since 2003. The company was established in 1993 and its headquarters are located in Taiwan. Their major products, besides PC power supplies, are industrial and open frame PSUs, adapters, medical and LCD TV power products, Uninterruptible Power Supplies (UPS) etc. Currently FSP Group is the 5th largest power supply vendor in the world, which means that they are a pretty large company.
Recently FSP enriched their Aurum series with three units of 550W, 650W and 750W capacity. The main difference between the new units and the older ones, with nearly similar capacity, besides the small power increase is the addition of the modular cabling system. So internally the modular Aurums share the same design/platform with the non-modular Aurums.
In today's review we will take a look at the FSP Aurum CM Gold 750W (AU-750M here after) and find out if the addition of the modular cabling system justifies the price difference with the Aurum CM Gold 700W. In a quick look the AU-750M features Gold efficiency, four +12V rails, low profile flat modular cables for easier cable management, four PCIe cables and a 120mm FDP (Fluid Dynamic Bearing) fan.
Specifications
| FSP AU-750M Features & Specs |
| Max. DC Output |
750W |
PFC |
Active PFC |
|---|
| Efficiency |
80 Plus Gold |
Operating temperature |
10°C - 40°C |
|---|
| Protections |
Over Voltage Protection
Under Voltage Protection
Over Current Protection
Over Power Protection
Short Circuit Protection |
Cooling |
120 mm FDP (Fluid Dynamic Bearing) Fan |
|---|
| Dimensions |
150 mm (W) x 86 mm (H) x 160 mm (D) |
Weight |
2.0 kg |
|---|
| Compliance |
ATX12V v2.3, EPS 2.92 |
Warranty |
5 years |
|---|
| Price at time of review |
$179.99 |
The PSU is equipped with a quality FDP fan so we don't expect it to output much noise even at high loads/temperatures. Also all possible protections are present, something that ensures that the unit is fully protected and thus you will have your piece of mind. The latter is further ensured by the five year warranty that comes with the AU-750M.
What we didn't like in above characteristics were the high price and the fact that FSP states that 750W total continous output power is guaranteed only up to 40°C and not 50°C. Nevertheless we will crank up the temperature inside our hot box up to 50°C and see if the PSU can handle the heat or OTP will be triggered.
| FSP AU-750M Power Specs |
| Rail |
3.3V |
5V |
12V1 |
12V2 |
12V3 |
12V4 |
5VSB |
-12V |
Max. Power |
30A |
30A |
18A |
18A |
18A |
18A |
3.5A |
0.5A |
|---|
| 160W |
720W |
17.5W |
6W |
Total Max. Power |
750W |
|---|
The unit features four +12V rails with 30W lower combined power than the total max power, two quite strong minor rails and a 5VSB rail capable to give up to 3.5A. Since the PSU uses DC-DC converters for the minor rails generation I expected max power at +12V to be closer to 750W.
Cables & Connectors, Power Distribution
| Native Cables |
| ATX connector (550 mm) |
20+4 pin |
4+4 pin EPS12V (550mm) / 8 pin EPS12V (+100 mm) |
1 |
|---|
| 6+2 pin PCIe (550 mm+100 mm) |
2 |
| Modular Cables |
6+2 pin PCIe (550 mm+100 mm) |
2 |
|---|
| SATA (550 mm+150 mm+150 mm) |
3 |
SATA (550 mm+150 mm) / 4 pin Molex (+150 mm+150 mm) |
2 / 2 |
|---|
| SATA (550 mm) / 4 pin Molex (+150 mm) |
1 / 1 |
SATA (550 mm+150 mm) / 4 pin Molex (+150 mm+150 mm) / FDD (150mm) |
2 / 2 / 1 |
|---|
The number of connectors is sufficient for the PSU's capacity. There are three hardwired cables which are fully sleeved and all modular cables are flat sleeved but not with normal sleeving net. Instead a thick plastic sleeving is used which it may protect the wires well, but you can't call it eye candy. On the hardwired cables the distance among the connectors is too small, only 100 mm. Thankfully on the modular ones the distance is increased to 150 mm. Finally all connectors use 18 AWG wires, the recommended diameter by ATX spec.
| Power Distribution |
| 12V1 |
ATX, SATA, Peripheral |
12V2 |
8 pin EPS, 4+4 pin EPS |
|---|
| 12V3 |
PCIe1, PCIe2 |
12V4 |
PCIe3, PCIe4 |
|---|
As you can see power distribution is optimal since the EPS connectors use a separate +12V rail. Also the PCIe connectors, in pairs, utilize two +12V rails.
Packaging
The package is quite large and the used black and the gold colors make a nice mix. On the front there is an image of the Aurum unit, which is surrounded by the arrow's of the Arrow Flow Technology. As is usual, on the rear of the packaging we can find much more useful information about the product. In this case we are informed about the power specifications, the available cables and connectors along with their length, efficiency and acoustic noise curves and further characteristics.
Contents
Once we find the box we find a very neat packaging. The PSU sits in the middle while on its three sides there are equal compartments storing its hardwired cables and the bundled accessories. The only thing we disliked here was that not even a plastic wrap protected the unit itself and to be frank at such a price tag we expected a cloth bag. The bundle includes a set of thumbscrews for chassis mounting, a user's manual, several Velcro ties, the modular flat modular cables (no storing pouch for them either) and the necessary AC power cord.
Exterior
The black heavy matte finish is fairly scratch resistant and the peculiar design of the fan grill is eye catchy. Also on the front we don't meet the classic honeycomb design grill but a unique ventilation (Arrow Flow Technology) with arrow shape holes, which according to FSP improve air extraction. On the rear side we find five modular sockets and a grommet around the cable exit hole. While the hardwired cables are round, the modular ones are flat and have an industrial look sleeving (meaning that although it does its job quite well its looks aren't so great).
A Look Inside
Before reading this page we strongly suggest to take a look at
this article, which will help you understand the internal components of a PSU much better.
The main PCB is quite small, underpopulated and the primary and secondary heatsinks are too small, so we don't expect a significant delta difference between the input and output temperatures in the voltage regulation and efficiency measurements. FSP also uses two proprietary ICs, made by themselves, to control the primary and secondary side respectively. We will refer to them below.
The transient filter starts at the AC receptacle with two Y caps and the power wires are wrapped around a ferrite core. It continues on the main PCB with three coils, two X and two Y caps. We didn't find an MOV, something that FSP tends to make a tradition in their platforms. However in this case FSP claims that the MIA IC that this platform uses offers over voltage protection and it can absorb excess surges coming from the power grid. Still we would be more assured if we saw a MOV in the transient filter, since its cost is insignificant.
The single bridge rectifier is bolted on a heatsink. After it we meet the large APFC choke and the primary heatsink which holds the APFC mosfets, two
IPB60R165CP, and the boost diode. Right behind the boost diode, attached also on the primary heatsink, there is the 5VSB regulation mosfet, a GE03N70T. We found it strange that FSP claims 3.5A max at 5VSB since this FET supports only up to 3.3A continuous drain current at 25°C and only 2.1A at 100°C.
The smoothing/reservoir capacitor is provided by Rubycon (390μF, 450V, 105°C, MXG series). The topology that FSP uses in Aurum is called Active Clamp Reset Forward (ACRF). It offers high efficiency and good power handling capability but up to a certain power level it cannot really compete with other topologies, like double forward, full phase shift etc. In ACRF two mosfets are used, ones plays the role of the main switcher (Q1) and the other is the reset switch (Q2), which disconnects the main capacitor while Q1 is active. Also while Q2 is open, power is transferred from the primary to the secondary side. The main advantage of ACRF is the almost lossless switch of Q1, because while it is turned off the drain voltage is very low. In AU-750M in the role of Q1 is a
SPA17N80C3 and as Q2 we find a
FQPF3N80C. The APFC/PWM controller is a FSP 6600 IC, for which we didn't find any documentation on the net.
In the secondary side synchronous design is used so the generation of +12V handle mosfets (two
IRLB3036). The minor rails are generated from +12V with the help of a DC-DC converter. The mosfets responsible for the regulation of the minor rails reside on the solder side of the main PCB and consist of two pairs of
IPD031N03L and
IPD050N03L. The PWM controller for the DC-DC converter is the proprietary FSP 6601 IC and all filtering caps of the secondary side, on the main PCB, are provided by Nippon Chemi-Con and rated at 105°C.
We would like to see some heatshrink around these wires.
On the solder side of the modular PCB there are two CapXon caps for current filtering. A wise move from FSP, although we would prefer to see two Japan made caps here.
Soldering quality in general is good enough and we were very pleased with the fact that we didn't find any long component leads. On the secondary side we spotted four current shunts under the +12V wires area, so it seems the PSU indeed has four +12V rails.
The cooling fan is provided by
Protechnic Electric and its model number is MGA12012HF-A25 (12V, 0.45A, 2400 RPM, 84.8 CFM, 37 dBA, 155g).
Test Setup
All measurements are performed utilizing ten electronic loads (seven Array 3711A, 300W each, and three Array 3710A, 150W each), which are able to deliver over 2500W of load and are controlled by a custom made software. We also use a DS1M12 (Stingray) oscilloscope, a CHY 502 thermometer, a Fluke 175 multimeter and an Instek GPM-8212 power meter. Furthermore, in our setup we have included a wooden box, which along with a heating element is used as a Hot Box. Finally, we have at our disposal two more oscilloscopes (Rigol 1052E and VS5042) and a CEM DT-8852 sound level meter. In
this article you will find more details about our equipment and the review methodology we follow.
Voltage Regulation Charts
The following charts show the voltage values of the main rails, recorded over a range from 60W to the maximum specified load, and the deviation (in percent), when compared with the voltage values at 60W load.
Efficiency Chart
In this chart you will find the efficiency of AU-750M at low loads and at loads equal to 20-100% of PSU’s maximum rated load.
Voltage Regulation and Efficiency Measurements
The first set of tests reveals the stability of voltage rails and the efficiency of AU-750M. The applied load equals to (approximately) 20%, 40%, 50%, 60%, 80% and 100%, of the maximum load that the PSU can handle. In addition, we conduct two more tests. In the first we stress the two minor rails (5V & 3.3V) with the maximum load that our tester can apply to these rails, while the load at +12V is only 2A and in the second test we dial the maximum load that +12V can handle while load at minor rails is minimum.
Voltage Regulation & Efficiency Testing Data
FSP AU-750M |
Test |
12 V |
5 V |
3.3 V |
Power
(DC/AC) |
Efficiency |
Temp
(In/Out) |
PF/AC
Volts |
| 20% Load |
11.034A |
1.984A |
2.002A |
150.00W |
90.50% |
41.4°C |
0.952 |
| 12.090V |
5.038V |
3.297V |
165.75W |
42.5°C |
232.6V |
| 40% Load |
22.233A |
3.999A |
4.053A |
300.00W |
92.03% |
42.4°C |
0.981 |
| 12.000V |
5.002V |
3.257V |
326.00W |
44.6°C |
230.6V |
| 50% Load |
27.896A |
5.007A |
5.097A |
375.00W |
91.89% |
44.1°C |
0.987 |
| 11.955V |
4.993V |
3.237V |
408.10W |
46.9°C |
228.3V |
| 60% Load |
33.602A |
6.030A |
6.156A |
450.00W |
91.27% |
45.0°C |
0.988 |
| 11.910V |
4.975V |
3.216V |
493.05W |
48.8°C |
228.3V |
| 80% Load |
45.212A |
8.085A |
8.323A |
600.00W |
89.69% |
47.5°C |
0.993 |
| 11.802V |
4.948V |
3.172V |
669.00W |
52.5°C |
226.1V |
| 100% Load |
56.915A |
10.177A |
10.652A |
750.00W |
88.18% |
51.0°C |
0.996 |
| 11.714V |
4.913V |
3.126V |
850.50W |
58.6°C |
225.3V |
| Crossload 1 |
2.010A |
19.000A |
19.000A |
170.40W |
78.71% |
49.2°C |
0.968 |
| 12.531V |
4.582V |
3.061V |
216.50W |
52.8°C |
230.9V |
| Crossload 2 |
59.999A |
1.000A |
1.000A |
696.65W |
88.52% |
50.3°C |
0.995 |
| 11.471V |
5.145V |
3.260V |
787.00W |
56.4°C |
227.5V |
Although efficiency is pretty high, we were not satisfied wit the results. Only the 5V rail managed to stay within 3% while +12V exceeded this limit and 3.3V went even beyond 5%. Voltage regulation on the last two rails definitely needs tuning. Also the results at full load and Cross Load tests weren't the ones we expected. Judging from the voltage readings on all rails and especially at +12V it seems that the PSU struggles to deliver 750W of power. Especially at CL2 the above-mentioned rail dropped its voltage to very low levels, under 11.5V, something simply unacceptable for a PSU of this price/category.
Efficiency at Low Loads
In the next tests, we measure the efficiency of AU-750M at loads much lower than 20% of its maximum rated load (the lowest load that the 80 Plus Standard measures). The loads that we dial are (approximately) 40, 65 and 90W. This is important for scenarios in which a typical office PC is in idle with power saving turned on.
Efficiency at Low Loads
FSP AU-750M |
| Test # |
12 V |
5 V |
3.3 V |
Power
(DC/AC) |
Efficiency |
PF/AC
Volts |
| 1 |
1.920A |
1.992A |
1.998A |
40.00W |
72.27% |
0.718 |
| 12.189V |
5.020V |
3.303V |
55.35W |
232.6V |
| 2 |
4.388A |
1.992A |
1.998A |
70.00W |
82.45% |
0.740 |
| 12.171V |
5.020V |
3.301V |
84.90W |
232.4V |
| 3 |
6.874A |
1.992A |
1.997A |
100.00W |
86.43% |
0.928 |
| 12.135V |
5.020V |
3.299V |
115.70W |
232.6V |
Efficiency at low loads is the strength of the AU-750M as it seems, since in two out of the three tests the 80% mark was easily surpassed.
5VSB Efficiency
ATX spec states that the 5VSB standby supply's efficiency should be as high as possible and recommends 50% or higher efficiency with 100mA load, 60% or higher with 250mA load and 70% or higher with 1A or more load.
We will take four measurements, at 100 / 250 / 1000 mA and 3000 mA.
5VSB Efficiency
FSP AU-750M |
| Test # |
5VSB |
Power (DC/AC) |
Efficiency |
PF/AC Volts |
| 1 |
0.100A |
0.51W |
60.00% |
0.071 |
| 5.064V |
0.85W |
233.5V |
| 2 |
0.250A |
1.27W |
72.16% |
0.139 |
| 5.064V |
1.76W |
233.5V |
| 3 |
1.000A |
5.04W |
78.14% |
0.332 |
| 5.038V |
6.45W |
233.4V |
| 4 |
3.000A |
14.84W |
88.81% |
0.456 |
| 4.948V |
16.71W |
233.3V |
The 5VSB exhibited an astonishing efficiency. With 0.1A load, efficiency is 20% above the ATX recommended level and with 3A we measured very close to 90% efficiency!
Power Consumption in Idle & Standby
In the table below you will find the power consumption and the voltage values of all rails (except -12V), when the PSU is in idle mode (On but without any load at its rails) and the power consumption when the PSU is in standby (without any load at 5VSB).
Idle / Standby
FSP AU-750M |
| Mode |
12 V |
5 V |
3.3 V |
5VSB |
Power (AC) |
PF/AC Volts |
| Idle |
12.178V |
5.064V |
3.334V |
5.056V |
12.25W |
0.348 |
| 233.0V |
| Standby |
0.08W |
0.007 |
| 233.7V |
Phantom power is really low, among the lowest we have ever measured, so the AU-750M will not have any problem meeting not only the current ErP Lot 6 2010 requirements, but all future ones too.
Cross Load Tests
For the generation of the following charts we set our loaders in auto mode, through our custom software, and try over a thousand possible load combinations with +12V, 5V and 3.3V rails. The voltage regulation deviations in each of the below charts are calculated taking the nominal values of the rails (12V, 5V and 3.3V) as point zero. We should note here that we will run this test only with PSUs that have capacity equal or lower than 800W since it takes way too long and as the capacity increases the completion time increases exponentially.
+12V Voltage Regulation Chart
5V Voltage Regulation Chart
3.3V Voltage Regulation Chart
Efficiency Chart
Advanced Transient Response Tests
In these tests we monitor the response of the PSU in two different scenarios. First a transient load (10A at +12V, 5A at 5V and 6A at 3.3V) is applied for 50 ms to the PSU, while the latter is working at a 20% load state. In the second scenario the PSU, while working with 50% load, is hit by the same transient load (with the exception now that load at 3.3V is increased by 4A). In both tests, we measure the voltage drops that the transient load causes, using a Labjack that is attached to our loader and the Stingray oscilloscope. In any case voltages should remain within the regulation limits specified by the ATX specification. We must stress here, that the above tests are crucial, since they simulate transient loads that a PSU is very likely to handle (e.g. starting of a RAID array, an instant 100% load of CPU/VGAs etc.) We call these tests “Advanced Transient Response Tests” and they are designed to be very tough to master, especially for PSUs with capacities lower than 500W.
| Advanced Transient Response 20% |
| Voltage |
Before |
After |
Change |
Pass/Fail |
| 12 V |
12.091V |
11.950V |
1.17% |
Pass |
| 5 V |
5.038V |
4.982V |
1.11% |
Pass |
| 3.3 V |
3.298V |
3.214V |
2.55% |
Pass |
| 5VSB |
5.038V |
5.011V |
0.54% |
Pass |
| Advanced Transient Response 50% |
| Voltage |
Before |
After |
Change |
Pass/Fail |
| 12 V |
11.957V |
11.822V |
1.13% |
Pass |
| 5 V |
4.993V |
4.943V |
1.00% |
Pass |
| 3.3 V |
3.238V |
3.154V |
2.59% |
Pass |
| 5VSB |
5.011V |
4.984V |
1.08% |
Pass |
The results here are clearly acceptable but with 50% load the 3.3V rail dropped significantly, mainly because of its low initial voltage before the transient load. However with such a lousy voltage regulation at 3.3V we were positively surprised to see the AU-750M finishing these tests with success.
Below you will find the oscilloscope screenshots that we took during Advanced Transient Response Testing.
Transient Response at 20% Load
Transient Response at 50% Load
Turn-On Transient Tests
In the next set of tests we measure the response of the PSU in simpler scenarios of transient loads, during the turn on phase of the PSU. In the first test we turn off the PSU, dial 2A load at 5VSB and then switch on the PSU. In the second test, while the PSU is in standby, we dial the maximum load that +12V can handle and we start the PSU. In the last test, while the PSU is completely switched off (we cut off power or switch off the PSU's On/Off switch), we dial the maximum load that +12V can handle and then we switch on the PSU from the loader and we restore power. The ATX specification states that recorded spikes on all rails should not exceed 10% of their nominal values (e.g. +10% for 12V is 13.2V and for 5V is 5.5V).
The 5VSB registered a tiny spike here while +12V performed excellent with no voltage overshoots. Overall pretty good performance here.
Ripple Measurements
In the following table you will find the ripple levels that we measured on the main rails of AU-750M. According to ATX specification the limits are 120 mV (+12V) and 50 mV (5V & 3.3V).
| Ripple Measurements |
| Test |
12 V |
5 V |
3.3 V |
Pass/Fail |
| 20% Load |
26.2 mV |
7.8 mV |
19.2 mV |
Pass |
| 40% Load |
20.8 mV |
10.6 mV |
15.4 mV |
Pass |
| 50% Load |
23.6 mV |
10.8 mV |
18.2 mV |
Pass |
| 60% Load |
26.8 mV |
11.2 mV |
20.8 mV |
Pass |
| 80% Load |
32.2 mV |
13.4 mV |
24.4 mV |
Pass |
| 100% Load |
39.6 mV |
14.6 mV |
29.8 mV |
Pass |
| Crossload 1 |
46.8 mV |
28.6 mV |
30.8 mV |
Pass |
| Crossload 2 |
34.6 mV |
26.6 mV |
32.4 mV |
Pass |
Ripple at +12V is low, although not among the lowest we have ever measured. At 5V, ripple suppression is very good, except the CL tests where we clearly see that the minor rails are heavily stressed. Finally, ripple at 3.3V is in the normal range, although in most cases it is twice as much compared to the 5V ripple readings.
Ripple at Full Load
In the following oscilloscope screenshots you can see the AC ripple and noise that the main rails registered (+12V, 5V, 3.3V). The bigger the fluctuations on the oscilloscope's screen the bigger the ripple/noise. We set 0.01 V/Div (each vertical division/box equals to 0.01V) as standard but sometimes we are forced to use 0.02 V/Div, meaning that the fluctuations will look smaller but actually this wont be the case. For the first screenshot we used 0.02 V/Div, so actually the registered ripple is much bigger than it seems (compared to the other screenshots where 0.01 V/Div was used).
Ripple at Crossload 1
For the first screenshot we used again 0.02 V/Div. The order of images is +12V, 5V and 3.3V.
Ripple at Crossload 2
For the first screenshot we used again 0.02 V/Div. As above the order of images is +12V, 5V and 3.3V.
Value and Conclusion
 |
- The FSP Aurum CM Gold 750W retails for $179.99
|
|---|
 |
- High efficiency
- Delivered full power at 50°C flawlessly
- Silent operation even at high loads/temperatures
- Better than 3% voltage regulation at 5V
- Good efficiency levels below 100W and very high efficiency at 5VSB
- Fairly good ripple/noise suppression
|
 |
- High price
- Loose voltage regulation at +12V and 3.3V
- 3.3V dropped below ATX limits with full load
- Lousy performance at Crossload tests
|
| 7.8 |
The FSP Aurum's 750W competitiveness is crippled by its high price which puts it in the same price group with very strong opponents, like the Corsair AX750, NZXT HALE90 750W, and Seasonic X-750. With such competition you can imagine things are pretty tough for the AU-750M, especially since it didn't set new performance standards during our tests. It may have high efficiency and fairly good ripple/noise suppression but voltage regulation at +12V and 3.3V is loose and its performance at Crossload tests was terrible. When we pulled 60A from +12V during the CL2 test, this rail's voltage dropped dangerously low. In my opinion FSP should take a look at this matter and find a solution for the poor performance at Cross Load tests. Load on the rails won't be equally balanced in any system, so this improvement would have a positive effect. Also I think that the PSU needs an urgent price revision, in order to make it much more competitive.
With the current price tag and the performance we saw at full load and Cross Load tests I simply cannot recommend AU-750M as a good buy. If FSP decides to offer it at a much lower price, around $110-$120, then things would be much more different and I would have no problem recommending it to any user who seeks for a reliable middle-wattage PSU to cover his needs. |