How to Inspect New Power Supply?

[genosmm]

Am trying to do my homework concerning new power supply(PSU) purchases.

Among things I have learned so far is that you can Not just buy a "Brand Name PSU" but must do your homework so now check the reviews and a "LifeTime PSU Warranty" is only as good as the company that gives it.

Have found a number of inexpensive PSUs (on sale) that seem more than adequate from the mfrs specs but no independent or even buyers reviews.

As a Newbie in this area, any tips on inspecting a new PSU so not to void the store return policy or mfr warranty?

Any comments or suggestions welcomed!

Thanks

Gene

[370forlife]

Re: How to Inspect New Power Supply?

About the only thing you can do is go by brand, which one is known for quality and which isn't.

Or you can peek through the back of the psu, typically, if its somewhat crapmed, and heavy with nice big heatsinks it is probalby ok. Try to see what brand capacitors it has in it.

[zandrax]

Re: How to Inspect New Power Supply?

First check reviews of those psus on reputable sites (Jonnyguru, X-bit labs, Hardware Secrets and others). Reputable sites are the one with real reviews: they test the psu on a load bench at various loads, chech for complaincy of voltage and noise/ripple with ATX specs, dissect the components and such; fake reviews are just connecting the psu to a computer and watching it.

Then check it: look inside as 370 suggested, then check cable thickness: if most calbes are AWG18 then the manifacturer didn't skimp too much on them and this usually is a good signal; majority of AWG20 or, worse, 22 on a 400+ W psu means cheapness.
Check if it has active, passive or no pfc: as a rule fo thumb active pfc ones are recent projects while passive and no pfc ones are old psus barely adapted; if the manifacturer advertise a passive pfc, then it should have a coil weighting at least 300 grams (0.8 lbs). Since a good 400 W psu with no pfc old design weights about 2 kgs (4 lbs) or more, a good psu with a passive pfc is supposed to weight about 2.3-2.4 kgs (4.8-5 lbs).
Lastly check the label: the maximum power for each rail should be equal to maximum current multiplied that voltage. A 5 V line with i.e. 30 A should allow up to 5 * 30 = 150 VA; if it doesn't, then the label is deceptive (if not completely fake) and the product suspicious. Labels with a lot of power on the 3.3. and 5 V rails usually smell of old as those psus were designed with Pentium 2 - 3 or Athlons in mind; since Athlon 64 and Pentium 4 the most important and loaded line is 12 V and most recent psu deliver most power on that line.
Look for combined power too: this is the maximum draw for a brunch of lines. This should give some hints about components' size: if the maximum load on the 5 V alone is i.e. 30 A (150 VA) and the one on the 3.3 V alone is i.e. 33 A (so 100 VA), you should expect the combined 3.3. and 5 V max load to be 250 VA, am I right? No: the combined load is something near i.e. 180 VA : this means the transformer shared between the 5 and 3.3 lines cannot deliver more than that power. Same for dual or triple 12 V rails: a single rail alone may be allowed to draw up to 20 A (so 12 * 20 = 240 VA, enough for most power hungry videocards) but the combined load can be way lower than 2 or 3 * 20 A each: i.e. only 35 A. At the end there is total power: it should be the sum of all combined power values, but even good brand psus show slightly different values; some reviewers joke on label writers and their math abilities :P
Lastly check for the Underwriters Labs number (a circle with an UL and a "E + 6 digit number" inside or a mirrored UR with the Enumber): you can spot the real manifacturer by looking that number on the UL index site.

Sorry for the long post

Zandrax

[genosmm]

Re: How to Inspect New Power Supply?

370forlife and zandrax,

Thank you both for the tips!!!

zandrax As Newbie took a while concerning pfc. I read a very good summary at http://en.wikipedia.org/wiki/Power_factor_correction am guessing that it is best to have active pfc.

Gene

[shadow]

Re: How to Inspect New Power Supply?

Power Factor Correction should not be a high priority. Hypothetically, a good power supply that uses Passive PFC is preferable to a cheap/inferior power supply with Active PFC. In practice, most inferior power supplies won't have Active PFC because it costs more than a Passive PFC (from what I can gather) or have none at all (the very very cheap power supplies).

Having said this, in most countries, some sort of Power Factor Correction is required by law (be that Passive or Active). A power supply without any form of PFC is a sign of an extremely cheap power supply that should be avoided at all costs.

[zandrax]

Re: How to Inspect New Power Supply?

@ genosmm: shadow and I substantially agree. While you could theoretically buy a cheap and bad psu with active pfc, this doesn't happen in pratice because chinese producers find it's way cheaper and simpler to manifacture pfc-less psu or to retrofit old power supplies with a coil.
About pfc: Wikipedia shows the electrical aspects and engineering challenges, but nothing else: putting away modesty, I link my own post about pfc

Summary of my post:
For electrical companies pfc is necessary, for the common man pfc doesn't bring any immediate benefit: any electrical appliance draw the power necessary for a job (e.g. the spinning of an electrical motor) and some additional power which flows from the mains to the device, does nothing and flows into mains. The former is called active power and it's necessary for the job, the latter is reactive power and, since it's unneccesary for the job, it's a sort of "waste": residential contracts usually charge only for active power nevertheless reactive one must be supplied to the device so it overloads mains cables and power stations for free and for electrical companies and the environment this is a Bad Thing (bigger expenses for the former, more pollution for the latter); purpose of power factor correction (aka pfc) is reducing as much as possible reactive power, be it for the average load (passive pfc) or dinamycally under nearly all possible loads (active pfc). In some countries (e.g. european Union) pfc is mandatory.

For variable loads (a computer is really a variable load which changes over time) active pfc is more effective than passive one in cutting reactive power.

Zandrax