Is this PSU a piece of junk? OCZ "StealthXStream 2"

M

Michael

I was in a pinch last week and needed a new 600w+ PSU fast. The only
store I could go to had a $80 option, or a $170 option. Guess which I
chose.... It's an OCZ "StealthXStream2" 700w "80plus" rated
(#OCZ700SXS2). Since installing it my system has been freezing once
every 24hours or so.

Suspecting the new PSU I installed Asus Probe (my mobo is an Asus
M4A87TD EVO, the CPU is an AMD X4 630). I notice the vcore readings
for the 1.3V are jumping all over the place, from 1.06v to 1.36v - it
is changing every second. A second program, CPU-Z, confirms this.

I've read that 5% or 10% is the maximum variance you should see with a
good PSU but would like to confirm that. Any thoughts guys? Are these
readings proof my new PSU is a piece of junk and the cause of my
freezes?

(ps - it's got a 3 year warrenty, will they let me do an RMA if the
thing works "mostly"? should I break it some more first? :)
 
P

Paul

Michael said:
I was in a pinch last week and needed a new 600w+ PSU fast. The only
store I could go to had a $80 option, or a $170 option. Guess which I
chose.... It's an OCZ "StealthXStream2" 700w "80plus" rated
(#OCZ700SXS2). Since installing it my system has been freezing once
every 24hours or so.

Suspecting the new PSU I installed Asus Probe (my mobo is an Asus
M4A87TD EVO, the CPU is an AMD X4 630). I notice the vcore readings
for the 1.3V are jumping all over the place, from 1.06v to 1.36v - it
is changing every second. A second program, CPU-Z, confirms this.

I've read that 5% or 10% is the maximum variance you should see with a
good PSU but would like to confirm that. Any thoughts guys? Are these
readings proof my new PSU is a piece of junk and the cause of my
freezes?

(ps - it's got a 3 year warrenty, will they let me do an RMA if the
thing works "mostly"? should I break it some more first? :)

120V --- ATX PSU ----- 12VDC ------ Vcore_regulator ------ 1.06v to 1.36v
(Cool N' Quiet!)

The Vcore variation would have been present with your old power supply
too. The most likely cause, is the features of Cool N' Quiet. Like
Intel SpeedStep, Cool N' Quiet cranks down the VID setting, when the
processor is not being used. The voltage changes moment to moment.
So what you're seeing, is FID/VID multiplier and voltage
setting changes, on purpose, on the processor.

Go into the BIOS and disable Cool N' Quiet, then go back to measuring,
and it should stop changing. (You can also do this from the Power control
panel, but I wouldn't bother.) It should then stay at the "high" setting,
which in your case seems to be around 1.36v. You should then
see a smaller variation, as the voltage still changes a bit with
processor level of activity. (If running CPUBurn etc, the processor
could draw 100 amps from Vcore, which will cause a slight dip
in voltage.) Even that variation is OK, as the processor manufacturer,
in the datasheet, shows "load lines" in the voltage versus current
graph, for what the processor can tolerate in terms of variation.
The tolerance is quite large (0.15V on my old processor).

On its input side, the Vcore regulator is tolerant of changes in the
12V level. It might not be immune to all forms of ripple however,
depending on the frequency of the ripple.

For review comments on power supplies and their quality, check Newegg.

http://www.newegg.com/Product/Product.aspx?Item=N82E16817341040

Click the feedback tab on the page.

There are only three reviews, but the top-most review right now, notes
an inability of that supply to work properly when "underloaded". That
seems to match your symptoms, but doesn't explain exactly what is
happening to the power supply itself. We would need many, many more
reports of that nature, before branding the supply as "bad". Seeing
one report, is just a warning.

Some behaviors on power supplies, will be transient in nature, which
can't necessarily be seen with a multimeter, but could be seen
with a storage oscilloscope. The ripple could be out of spec,
the thing might have poor transient (step) response. There
are lots of things we can't see with a home multimeter. So
even if the BIOS hardware monitor shows all the voltages
as "normal", there could be millisecond or microsecond wide
events, which are enough to upset the hardware, which
won't get caught by a voltage measurement.

Switching to another supply, is the fastest way to remedy that.

Sites like Jonnyguru.com , take digital storage oscilloscope
pictures of the DC outputs of the supplies they review, at various
load settings. I'm not going to look right now, but if I needed
detailed info, I'd look for the model of supply over there.

There are plenty of other power supply review sites, that
only note things like "it sure is shiny" :) It's hard
to get anything remotely resembling a technical review.
Only a couple sites come close, to proper testing. A lot
of guys, review power supplies so they can get a supply
for free from the manufacturer, which is why there are
so many "it sure is shiny" type reviews.

In the case of OCZ, they probably have a web forum, and
you can check their forum and see if anyone else is
having your problems.

http://www.ocztechnologyforum.com/forum/forumdisplay.php?56-Power-Supply-Unit-Support&s=

Paul
 
R

RayLopez99

I was in a pinch last week and needed a new 600w+ PSU fast. The only
store I could go to had a $80 option, or a $170 option. Guess which I
chose.... It's an OCZ "StealthXStream2" 700w "80plus" rated
(#OCZ700SXS2). Since installing it my system has been freezing once
every 24hours or so.

You bought somebody's lemon. Just trash it, and buy the more
expensive power supply.

RL
 
Top