300w generic psu caused hard drives failures?

[CleanSlate]

I have had 4 or 5 hard drives die on me, few lasted 6 months. In all fairness these drives were random drives I had laying around since I haven't had the cash to upgrade lately.

Now what I'd like to know is is it likely my psu?

at idle the my rails are (sisandra pro, source):

CPU core voltage: 1.68v
Aux voltage: 3.09v
+3.3v: 2.91v
+5v: 5v
+12v: 10.11v
-12v: -8.82v
-5v: -3.95v
+5 standby voltage: 1.45

At load my rails are:

CPU core voltage: 1.68v
Aux voltage: 3.09v
+3.3v: 2.85v
+5v: 5v
+12v: 10.11v
-12v: -8.82v
-5v: -3.95v
+5 standby voltage: 3.68

I'm not sure what it means but I gather the voltages should match their rail, like +5v should = +5v -/+ 5%.

I read through the FAQ and I didn't get any real info on how to tell if your psu is not enough for your system or not...

Thanks for any help you give me

~Adam

[Geeky1]

It's entirely possible that the power supply is killing the drives, yes.

A generic 300w power supply is really not something that I would trust to power an Athlon or P4 system. Generic power supplies are generally not nearly as high quality as Antec, Enermax, PC Power & Cooling, etc.

I've seen a generic "550w" power supply die when I asked it to deliver <300w. It was rated 550w, but could only put out that much power for 17ms at 25*C according to the manufacturer's website. Spend $40 on a decent 350w+ power supply.

[Ice Czar]

there is quite a large section in the Corruption 101 (http://hardforum.com/showthread.php?t=761272) FAQ devoted to Power

yes its a good chance
worth a read > Power supplies become increasingly expensive (http://www.theinquirer.net/?article=11279) June 13, 2004
specifically the ATX standard bruhaha

He explained that the later P4 CPUs take their power from a 12 Volt feed and, using the onboard voltage regulators, generate the high current, low-voltages they need to operate (anywhere from ~ 1.6V to around 2.7V). So the first requirement is a PSU that has plenty of power available from the 12V supply. Since the older ATX compatible PSUs didn't supply much current from the 12V section you have to ensure that your new PS delivers enough current (or power in Watts) from the 12V section. That's why using your old P3 ATX PS is a big "no-no".

He then described the history of the post-P3 power supply mess:

The earlier Athlon motherboard manufacturers decided to solve the 12v problem by using an additional 4-pin square connector to get the extra 12V those CPUs required. However many older PSUs didn't provide the 12V 4-pin square connector.

The motherboard makers then wised up and decided that there was nothing magical about a square 4-pin connector, so they put a normal hard disk type socket on the motherboard and provided the 12v power via a standard hard disk (4-pin inline) connector. Problem solved - you may now use your older PSUs provided they supply sufficient 12V current (many did not).

Further confusion came from PSU manufacturers not specifying the capabilities of the PSU in a way that allowed the end user to verify it's 12V power output rating.

In the meantime the ATX spec was saying "use the new 6-pin" inline connector - and very few motherboard makers implemented it.

Just to be sure, some motherboard makers, implemented *both* the hard disk style 4-pin inline connector and the square 4-pin connector. They said "use either or both in any combination you like".

The spec then evolved to the 24-pin main connector. Again, most motherboard manufacturers did not wish to make their customers mad by mandating that they replace their power supplies. This might cause their customers to avoid motherboard upgrades. Some used a "special" 24-pin connector with the extra 4-pin connection blocked off, or colored so that the user could plugin a 20-pin plug into the correct end of the 24-pin socket on the motherboard. Many others simply ignored the 24-pin requirement in the specs. Again - problem solved - use your older PSU.

[Elledan]

At load my rails are:

CPU core voltage: 1.68v
Aux voltage: 3.09v
+3.3v: 2.85v
+5v: 5v
+12v: 10.11v
-12v: -8.82v
-5v: -3.95v
+5 standby voltage: 3.68

I'm not sure what it means but I gather the voltages should match their rail, like +5v should = +5v -/+ 5%.
With those voltages I'm surprised this system runs at all.

Tolerances:

+3.3V - 3-5%
+5V - 5%
+12V - 5%
-12V - 10%
-5 - 10%
+5VSB - 5%

Either the readings are completely wrong, or this PSU shouldn't even be allowed near any system.

First thing you should do is disconnect this PSU from the rest of the system, put a load on it using an old HDD, fan(s) and/or resistors, turn it on using the 'paperclip'-technique, and test the voltages using a DMM.

[Vertigo Acid]

Either the readings are completely wrong, or this PSU shouldn't even be allowed near any system.

First thing you should do is disconnect this PSU from the rest of the system, put a load on it using an old HDD, fan(s) and/or resistors, turn it on using the 'paperclip'-technique, and test the voltages using a DMM.

First off, there is no doubt that those readings are off. It wouldn't boot if they were correct.

However, you shouldn't use Elledan's technique to test the voltages, because it does not represent actually conditions for you system. Get a multimeter or voltohmmeter, and test the voltages, but with everything still attached. That way you can see how things are during actually conditions with your setup

[CleanSlate]

I would test it but I'm afraid I might hurt something in my system with the voltmeter, I saw how to test it in the faq.. but yet I'm still sorta worried.. I guess I will


Actually, I'm just going to replace it because I've always been worried about this psu, it's generic and the voltage is the lowest recommended for this motherboard.. I want to get a psu that will last me through at least my next motherboard, an NF2 400 ultra, suggestions?

~Adam

[Ice Czar]

you can see how important I consider voltage regulation (transient response)
as posted in the Corruption 101 link

to wit

PC Power & Cooling TurboCool Line 1% Voltage Regulation
Antec True Power Series 3% Voltage Regulation
ATX Standard (dated) 5% Voltage Regulation

[CleanSlate]

Well I'm probably just going to wait until I get paid, in 2 weeks.. :|

~Adam

[Ice Czar]

as far as Im concerned
the amount you pay for a PSU should be proportional to the amount you spent on the system, if you note my sig, youll see that $200 for a PCP&C was cheap in comparision

but I dont run PCP&C in my other rigs, for starters they have a lower FSB
and are more tolerant mobos, all are on PSUs with 5% regulation
and of various quality, from an antique (at this stage) Antec PP-403X
to a Zippy MR3-6460P 650Watt N+1 Redundant (which Id hope has a fairly tight regulation but is still listed in the 5% spec)

another prime consideration is the amount of draw on the 12V rail required in the newer boards compared to the older ones (see quote above)

there are alot of good power supplys out there, and then there are a few excellent ones
as youll note in Vertigo Acid's sig
"This is not a court of law, when it comes to a "cheap" PSU they are presumed guilty until proven innocent - Ice Czar" :p
Ive more or less extended that to even mid range power supplies until I see transient response figures or a multimeter readout under a dynamic load

so
My list is pretty short for recommendations at this time for a "modern high draw high FSB mobo"

[CleanSlate]

Yes I seem to agree with what I could fully understand in what you said but with that said, what's your short list of recommendations for a 2600+ on a nforce 2 ultra 400 with a gig of ram an 80 gig and a ~9600/9700 pro/XT(of course I know there are no XT's for 9700 series)

That's about what my rig will be when I'm done with it, hopefully.. I was thinking a 420 or 450watt Antec/thermaltake, the thermaltakes seem to be much less expensive than the Antecs, and they are on the FAQ list of recommendations.. what do you guys think of this Thermaltake 420w:
http://www.newegg.com/app/ViewProductDesc.asp?description=17-153-006&depa=0

~Adam

[Ice Czar]

Antec TruePower

because Ive seen the transient response under a dynamic load (http://terasan.okiraku-pc.net/dengen/tester/index.html) & twice (http://www.silentpcreview.com/article58-page3.html)

note the second review lists the published transient response as
+5V,+12V outputs return to within 5% in less than 1ms for 20% load change.
not sure if that is an oversight or if it changed
its published at Antec as
+5V,+12V and 3.3V outputs return to within 3% in less than 1ms for 20% load change.
http://www.antec.com/specs/true550_spe.html

but the actual review had it at 1% anyway

[CleanSlate]

So what you're saying is that you say Antec is the best because it has 1. an extremely low fluctuation level, 2. very reliable and high quality and then 3. it has a very low response time in response to load changes which is very good?

#3 is what I got from your last post.. could be wrong, that's just what I could understand.. am I right?

~Adam

[Ice Czar]

1 and 3 are basically the same (voltage regulation \ transient response)

I say its the best choice also because of the cost,
half or less than a PCP&C like in my sig

and most importantly, because I can see published independent tests
that confirm 1-3

there may be other supplies for less that are close, but without the specs
or published tests...

[CleanSlate]

Thermaltake seem to be about 60% of the cost of the Antecs.. what do you think of them? Especially that one link I put up a few posts ago.. thanks

~Adam

[Ice Czar]

what do you think of them?
that other people with various mobos, with various levels of unknown onboard voltage regulation have had no issues with them :p

Id say they fall within the likely good category
but as I said before, where are the Load regulation figures? ......±5%
you could likely save even more and get similar performance from a Fortron
another likely good

we are talking probabilities here
how important is it to reduce them as low as you can?


but this is just a link in a chain
the quality of the voltage regulation on the motherboard is another
as is the power conditioning or UPS you employ
or what it has to deal with in the form of brownouts, spikes and blackouts

[CleanSlate]

Ok thank you, I will try to find an Antec for about the same price as that 420w thermaltake, if I can i will buy that as that seems to be the best around. If not I will just have to stick to Thermaltake because I am on a budget =/. Thanks :)

~Adam

[Vertigo Acid]

FS/T is a good source for Antecs cheap; my deal for a PCP&C fell thru, and now i'm looking at an Antec TrueBlue 480w for $50, from which i'll immediatly rip the obnoxious blue LED Fans from, to make it a straight True 480.

[Ice Czar]

my deal for a PCP&C fell thru,

sorry M8
I know how excited you where about that :(

[Vertigo Acid]

LOL @ how quickly things change around here w/ FS/T and e-bay

Going to get both now, as my RAM just sold on e-bay, and keep one as a backup. The only bummer is that it isn't the deluxe, so i've got some sleeving ahead of me (4 PSUs :eek: total )
.
.
.
.
.
And everyone lived happily ever after.

[Ice Czar]

count yourself lucky
not only do I need to resleeve 4 power supplies
but lop them down from 20" wiring harnesses to 3.5" (for the main power connectors)
from full tower length to rackmount distance :p

[Vertigo Acid]

Ouch, have only had to do that once, and that was on a LC&C (deer) PSU that is powering my DHCP server right now. No worries about it being a deer, its powering a Pentium Pro 200 in an pizza-box style mac case, even dremmeled the heatsinks down to make it fit in that form factor. I should post some pics, it's rather entertaining.

/thread hijack

[Elledan]

However, you shouldn't use Elledan's technique to test the voltages, because it does not represent actually conditions for you system. Get a multimeter or voltohmmeter, and test the voltages, but with everything still attached. That way you can see how things are during actually conditions with your setup
It's still not a bad idea to test the PSU first under less extreme conditions, though. If the PSU's voltages are already way off with such a light load, you can be certain that it's worthless.

[Vertigo Acid]

It's still not a bad idea to test the PSU first under less extreme conditions, though. If the PSU's voltages are already way off with such a light load, you can be certain that it's worthless.

Testing a power supply under real conditions *always* beats testing in an artificial condition. The loads might be light enough to be below the minimum load draw for that line to be in regulation. Also, how would you suggest easily, using computer parts, loading the 3.3v line short of pluging in a mobo?

[Elledan]

Testing a power supply under real conditions *always* beats testing in an artificial condition.
Not if the PSU is the worst POS you've ever come across. I would never want to hook up a PSU to any system without at least checking its voltages while it's still disconnected from the system's components.
Also, how would you suggest easily, using computer parts, loading the 3.3v line short of pluging in a mobo?
I would say that a resistor would do nicely. You could even create a circuit that would create a dynamic resistance, similar to that of a functioning mainboard.