SMC 2008 - The IEEE International Conference
 

 

 

 
 
 
Testing Hiper SF-350TS ATX PSU

The PSU was tested in a CoolerMaster ATC-101 case fitted with the following system:

AMD XP2100+
ABIT KX7 - 333r
256MB Corsair DDR 3200
Gainward 650XP Ti4200
2 x Maxtor D740 HDD
Sanyo 32x CD ROM
3Com NIC
SB Live

In use, the supply was very quiet, almost silent. It was quieter than the Enermax but remember the Enermax has two fans against the Hiper's one.

Using the Hiper resulted in a CPU temperature about 2°C higher than when using the Enermax. This is obviously due to the Enermax's second fan exhausting the hot air coming off the CPU HSF. Here the trade off is noise against cooling.

As usual the voltages were measured at idle and under load. The load tests consisted of running 3DMark2001 and Stability test V6 , (apparently no longer available). Also a directory containing 2.5GB of files was copied between HDDs to see the effects of 12V loading.

The voltages were measured using a Fluke 87 DMM set to record the min/max voltages. The meter records any changes lasting at least 100mS. The readings were taken at the ATX connector so the voltages being measured were those actually being applied to the motherboard. I also decided to have a look at the +5/12/3.3V outputs on the oscilloscope to see if there was any excessive noise or ripple. I compared the findings to the Enermax EG365P-VE.

Looking at the results below, both supplies would appear to be evenly matched apart from the Enermax's 12V rail which is rather high. This is because the PSU has been modded to bring the 5V rail up to 5V. Were as previously it was about 4.6V. The knock on effect of doing this is the 12V rail is also increased.

The table below does not give the full picture. I noted that during HDD activity significant ripple was produced on the 12V rail, in the order of 350mV for the Enermax and 300mV for the Hiper. Due to the design of these supplies there is an interdependency between the 5V and 12V rails. Hence during HDD activity the 5V rails also experience an increase in ripple, rising to 50mV for the Enermax and 40mV in the Hiper. The ripple increase in the 3.3V rails was insignificant. The quoted ripple figures are peak-to-peak. Unfortunately my aged 'scope only has a 1MHz bandwidth so a more accurate analysis of the ripple was not possible.

The main point to realise here is that the Hiper supply is at least the equal of the Enermax.

Model
Voltage
Min
Max
Variation
Noise + Ripple
Hiper 350TS
+12V
12.32V
12.48V
0.16V
30mV
+5V
5.00V
5.08V
0.08V
10mV
+3.3V
3.42V
3.43V
0.01V
20mV
Enermax EG365P-VE
+12V
12.68V
12.84V
0.16V
100mV
+5V
5.04V
5.16V
0.12V
10mV
+3.3V
3.22V
3.23V
0.01V
30mV