|
Statistics
Visitors: 942203
Advertisement
Notebook Computers
For the best deals on laptop computers visit www.Geeks.com.
|
|
Written by Mikhail Ivanenkov
|
|
Monday, 18 June 2007 |
|
Page 2 of 3
Both fans are made by Ruilian Science (now Xinruilian Electronics), model RDM8025SA and rated at 12v, 0.13A for a power rating of 1.56watts. I couldn't find the exact specs for this particular blower but looking at the various fan on this page it most closely resembles the RBL8025S which is rated at 29.36dba. Regardless, these things are loud. They push a significant amount of air (so much that my PSU fans are the only exhausts currently running in my system), but they're loud nonetheless. These appear to be sleeve bearing parts so they probably won't last as long, but seeing as how the same fans have been used in Channel Well Tech PSUs (same company that made Antec's) and Koolance units, it shouldn't be much of a concern.
As I mentioned earlier the capacitors are a little on the small side for a 400watt PSU. The heatsinks aren't all that either, but still have more surface area than a lot of generic ones which are nothing more than a slab of aluminum with ribbed "teeth" up top. While I don't have industrial equipment for testing a power supply, that would be synthetic in its own right anyway. What's more important is how it performs in an average everyday system. And here's what I used:
- AXP 1700+ JIUHB DLT3C 1.5v (11x133 = 1466mhz) @ 11x166 (1833mhz) 1.5v and 11x200 (2200mhz) 1.7v
- Biostar M7NCG nForce2 motherboard
- 2x256mb PC3200 Crucial DDR400 @ DDR333 & DDR400 speeds
- Gainward GeForce3 Ti200
- Onboard sound, LAN
- Western Digital 1000JB SE 100GB HDD
- 16x Lite-On DVD drive
- 12x10x32 Plextor CD-RW drive
- Mitsumi FDD
- 2x 12" blue CCFLs
- 2x 80mm fans
- 1x 60mm fan
Motherboard Monitor 5 recorded the results in five second intervals, averaging the last 1000 readouts. And this is where things get interesting. The JIUHB stepping Athlon XP is known to be a high overclocker, especially the DLT3C with a stock 1.5v. I had no problem running it slightly above spec (166mhz FSB x 11 = 1833mhz). It would also post at 11x200 (2200mhz) at the same voltage, but then I ran into problems. Increasing the voltage in 0.025v increments yielded better results every time, but it's hard to believe that I had to reach 1.7v and even then stability was an issue. This was under minimal stress, i.e. running a couple Internet Explorer windows and Windows Media Player. The latter had problems rendering the visualizations and my screen would flicker constantly.
So I decided to run the first test as "idle" conditions, letting the PC sit and do absolutely nothing. Keep in mind this first test was run about a week after the PSU was installed, meaning it had plenty of burn-in time so any discrepancies are clear indications of the power supply's shortcomings. Later that night I figured out why I couldn't make my processor stable no matter how high the voltage went: the rails were too "loose" to keep sufficient power flowing. Let's recap:
CPU 1.7v line - average was 1.69v (less than 1% difference). The low was a shocking 1.54v (almost 9.5% difference). The recorded high of 1.84v (about 8.2% difference) was no better either. So we have a total range of about 17.7% which in my book is pretty unacceptable.
AGP 1.5v line - average was 1.48v (about 1.3% difference, slightly undervolted). The low wasn't very good at 1.41v (6% difference). The high was 1.58v (about 5.3% difference). Combined range was better than the CPU line at 11.3%, but still way too unstable.
3.3v line - average was 3.24v (1.8% undervolted). The low was 3.12v (about 5.5%). The high was 3.36v (1.8% difference). Total range here was 7.3%. Which is better than the two previously mentioned lines but it's still too broad for any kind of serious overclocking.
5v line - average was 4.8v (4% undervolted, not good at all). Low was 4.44v (an astounding 11.2%). The high was 4.92v (1.6% difference). This is the only line that doesn't go above the rated voltage. Instead it's constantly undervolted and needs to be adjusted accordingly. The range here was 12.8%, nothing to smile about.
12v line - average was 11.9v (less than 1% difference). The low was 11.67v (about 2.8% difference). The high was 12.4v (about 3.3% difference). Combined range was 6.1%, the best of all the lines, but still too much to compete with a high end model.
There are two things to keep in mind here. 1) This is idle testing, without much stress placed on the PSU. 2) The extreme high and low values were not one time occurrences; it wasn't uncommon for the rails to jump back and forth.
|
|
Last Updated ( Monday, 18 June 2007 )
|
|
|
