美商艾澎的AG-650M能買嗎？價格很香耶 - 3C

http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story4&reid=554
But that's not why I'm bringing this up. This was my second sample. The first
sample did this too... to the tune of 300mV. It was also out of spec most of
the time on the standby rail, something this unit didn't have an issue with.
Efficiency at the time was atrocious, ruling out newer designs that could
have produced this effect in order to comply with some of Intel's latest low
load efficiency demands.

I am not going to sugarcoat this... this anomaly concerns me. I've seen two
units do it back to back, and one of those units had "hardware killer ripple"
written all over it. Because of the rules, I'm going to have to score this
unit fairly and give it full marks for ripple suppression. But you absolutely
should keep this anomaly in mind if you are buying one.

300mV -> hardware killer ripple

12V ripple 不能高於120mV

超能课堂(93)：这个高了肯定不行，电源纹波电压的二三事
http://www.expreview.com/54999.html

EVGA G2 550W 架構同 Leadex Gold 550W
http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story4&reid=440
We have less than 10mV on all rails

300/10=30倍
兩者 ripple 能差30倍
再便宜又有什麼用?
現在回過頭來看看
ripple 可能比電壓穩定度更重要
Intel ATX規範
12V -> 11.4V~12.6V
12V ripple 不能高過120mV
ripple 要超標很簡單.看看這顆首利代工的PSU就知道了...
振華南波萬 >_*

--
http://www.coolaler.com/threads/psu.342054/

--