Tom's Power Supply Article re: Antec TP 2.0 550w

[Venomous21]

This post is referencing the "Stress Test: Power Supplies Under Full Load" article on tom's, specifically the following link:

http://www.tomshardware.com/2005/07/11/stress_test/page10.html

I happened to buy this power supply prior to this article. I want to make a couple quick assumptions, I assume the article references the non-EPS version, and secondly, I assume that even though the "Nvidia SLI Ready" logo is a new thing, that versions of this power supply bought before this label existed, would also have received the same Nvidia logo if it had existed previously.

Anyways, I wanted a clarification on a point in the article, specifically:

Important Update - Antec technicians from Taiwan and the USA visited the Munich THG laboratory for just this purpose. The reason for the 3-hour visit was the divergences we discovered in the ripple test, which were not within spec - although the power supplies did not fail. The technicians from Antec performed new ripple tests on their power supplies together with the THG laboratory engineers. It turned out that the tolerances and high-frequency vibrations with their PSU can lead to varying results in the ripple tests. However, we didn't observe this phenomenon in any of the other candidates.

Because we determined that the ripple tests can remain entirely within spec in the present models and that the other values were flawless, the Antec devices still earned positive test grades.

Could someone seek to clarify the bold/italized part of the above quote in layman's terms? I've had extremely bad experiences with power supplies in the past, and I was hoping this power supply would be very stable as I plan to make another go at overclocking my system soon and I am a stability freak. I expect prime95 and 3dmark 2k1 loops to be able to run at least for 48 hours with no errors, as well as the same from memtest86.

Thanks again. Oh and first post!

 

[Guest]

Ripple is an artifact of switching power supplies and is impossible to avoid in this type of supply which is driving a load. It can however be filtered to within acceptable limits. DC sources with ripple actually put out a time varying signals, not a flatline like a battery. Typically this "ripple" will vary around the nominal value and will ideally be small as compared to the DC level the signal would represent in a more perfect world. Depending on switcher frequency it will take special equipment like an oscilloscope to see this ripple, and it wont show up on a DMM.

The peak variation can be measured as a percentge of nominal or in a peak to peak value. This value can be compared to manufacturer tolerance to tell whether ripple is within or outside of manufacturer ratings.

I'm not sure exactly what happened here that they made the comments they did, but here is my interpretation: they saw some weird effects caused by vibrations which didn't occur with other manufacturers supplies. Either the Antec guys showed them the flaws in their test setup, or told them that the supplies they tested are outdated. I can't tell which or if it somethin else entirely.

It seems like there is missing pieces to the story or maybe I am just misunderstanding what I read.

 

[emogoch]

Here's my interpretation:

When the Antec guys came in to see the Tom's setup, they could not consistantly get the ripple below the ATX spec (pass), nor could they consistantly get it above the spec either (fail).

Now, given that condition, and that the ATX spec allows for a +/- variation on of 5% (+/- 165 mV on the 3.3V rail), they decided that because the observed ripple wasn't causing the PSU to fail, nor causing it to go outside the rail-voltage spec, AND that a excessive ripple may even occur when the unit it being used, it decided that overall the minor problem was not worth failing an otherwise perfect PSU.

As for the SLI Ready sticker: it may or may not have received it. Typical GPUs now have an additional 6-prong external power connector to supply additional power to them, that you attach directly from your PSU. A PSU may come with 0, or 1 of these connectors, and you use a molex converting cable to get an (additional) one if you need it. My Enermax NoiseTaker Is "SLI-Ready" as well, but has 2 of those 6-prong connectors on it. Thus, the pre-SLI models of the PSU may have only had 1 connector, while the other ones have 2 (but both having the same PSU unit, jsut a matter of different cable ends).

 

[RichPLS]

I got the Antec TruePower 550 and it has two SLI connectors.