Umm yet another PSU thread. Interesting However..

[craig588]

-- Craig --

I take it from your reference to adding a cap and the ATX spec that you were checking the voltages without the supply hooked up to a m/b?

Regarding the OVP, at what point did it kick in for the 510 and did you try adjusting the OC pots?

It was hooked to a mobo, they all were. Almost all of them were tested on a p4c800e-d with a 3GHz cpu priming at stock speeds and volts. (The 350 watt PSUs I tested couldn't handle more than that, the 180 watt needed to be run on a PIII system) I never knew you needed to use a cap before you measured the voltage so I was just measuring off of the mobo header.

I sold the PCPNP a long time ago, but I belive the OVP came in at 3.5V on the 3.3 rail. I didn't push other rails back then, so I didn't test their limits.

[craig588]

I really dont know what craig588 is talking about when he says he has to put a cap before he measures....

I belive it was you that said I was supposed to do it according to the ATX specs. I had never heard of doing such a thing before though.

The PCP&C 510 has two 12 volt rails inside both rated at 17 amps

Both of mine contained only a single rail rated for 38 amps.

I've owned just about all revisions of the 510 and they are all the same design and all have the same exact measurements.

Conveniently not the ones I tested.

What you see on the scope can be totally off if you dont know the proper place to put your ground lead

I have a grounding rod right outside of my room because the normal ground in my house is so crappy. The only circut that grounding rod is connected to is my room

If your looking at the ripple voltage with the power supply connected to the computer, you'll get tons of induced noise from the computer and a large 60, 72, 75, or 85 hz ripple due to the refresh of your video card which is inducing noise on your voltage rails at the refresh rate of your monitor. If your measuring ripple when benching, your computer will be dumping so much noise on the lines it is laughable to even think it comes from the supply.

So why didn't any other PSU display these symptoms to anywere near the degree of the PCPNP?

It's pretty frickin shady to publish inaccurate findings based on poor measurements and poor variable control situations.

Publish?


If you show my why my results are wrong I will stop complaining about PCPNPs 510 unit. (I know more/am smarter than you is not enough of a reason for me) If you show me how I can get any of my other PSUs to show fluxuations like the PCPNP showed, or provide an "approved" set of testing instructions where I can find I made an ommision.

I would prefer the former because I want to know what I might have done wrong, not just the right way if I am in fact using the scope incorrectly.

[SPL15]

I belive it was you that said I was supposed to do it according to the ATX specs. I had never heard of doing such a thing before though.



Both of mine contained only a single rail rated for 38 amps.


Conveniently not the ones I tested.


I have a grounding rod right outside of my room because the normal ground in my house is so crappy. The only circut that grounding rod is connected to is my room


So why didn't any other PSU display these symptoms to anywere near the degree of the PCPNP?


Publish?


If you show my why my results are wrong I will stop complaining about PCPNPs 510 unit. (I know more/am smarter than you is not enough of a reason for me) If you show me how I can get any of my other PSUs to show fluxuations like the PCPNP showed, or provide an "approved" set of testing instructions where I can find I made an ommision.

I would prefer the former because I want to know what I might have done wrong, not just the right way if I am in fact using the scope incorrectly.

Sorry, I re-read my post and I didnt mean to sound like such an aXX. I apologize man.

I did not say that the ATX standard calls for a cap before the scope?!?! That just plain does not make sense. If I did say it, I was tripping on the liquid crystals from an LCD panel. I do remember saying something to the effect of that when you put your scope on the "AC" coupling setting, there is an internal high quality capacitor (Inside the scope) wired in series with the scope probe which is used to block the DC component so you can measure and amplify just the AC component of whatever it is your measuring. If your trying to meausure ripple voltage you need to be using the AC coupling setting of the scope. It has nothing to do with the ATX standard.

How did you determine that your PCP&C 510 only had one 38 amp 12 volt rail?? The design is clearly a dual 12 volt rail desin indicated by the two dual inductor coils used on 12volt lines in 99.99% of computer power supplies. These are located behind the fan, you can see one easily, the other is right behind it, kinda hidden from view. I bet if you measure all of the 12 volt lines, you'll see ones that measure .1 or .2 volts higher than the other 12 volt lines. Depending on what rail your measuring and the load on that particular rail, the apparent measured ripple/noise will not be accurate.

I wont go into the proper use of a scope because it takes judgement and knowledge that I can't convey to anyone no matter how smart without being right there to show them how to do it correctly. Even I'm guilty of sometimes not taking into account PCB layout and design and current paths when using a scope. Getting an accurate picture on a scope is much more than sticking the ground on anything remotely grounded and probing away. I have designed circuits where moving the ground point 1 cm on the ground trace made the difference of measuring 1.2Vp-p ripple or 12mVp-p ripple.

PCP&C uses a much tighter regulation than any other PC power supply which means that it's transient response is much higher which means it's regulation circuitry is constantly adjusting the PWM IC to adjust the voltage. This does translate into a more visible ripple on a scope, but this is the trade off you pay for ultra tight regulation. It is not possible to get ultra tight regulation AND totally pretty picture perfect looking voltage on a scope. Just because things on a scope dont look like how we want them to, doesnt mean the circuit it is supplying will give a flying fk about it. When I measured the ripple and regulation on my PCP&C power supplies, they all measured WELL within tolerance. Less than 0.75% variation. Your computer cares more about tight regulation and high transient response than ripple voltage, but even the ripple I measured was below tolerance levels.

I dont know the specific details on how you did your test, where you took measurements, etc so I cant tell you specific things that were bad. Did you take into account the dual voltage rails on most supplies? what rail did you measure? If the device causing all the ripple is on the line your measuring, you'll see it on a scope plain as day, but if you measure the other rail that has minimal load from less noisy devices, you'll get great looking results (Although it will not be an accurate picture of the power supply). The only load that does not induce noise into the circuit is a resistive load. This is how your supposed to measure power output of any power outputting device. If you're using a load that induces noise, your measuring the noise of your load, not the power supply.

I've done many tests on these types of things. I've tested multithousand dollar power supplies used MRI scanners that must have ridiculous power output and low noise. An MRI is essentially a massive high resolution amplifer driving a huge voice coil. A low noise power supply is a must. The PCP&C has better ripple performance than these units and so do 100% of the computer power supplies I've measured so ripple voltage is not an issue here. The thing that makes PCP&C sooo good is the reliability, NO BS ratings, rock solid regulation, Solid design. This is what makes a power supply good. The 510 delivers 510 watts at 50C ambient. I would like to see the OCZ put out 600 watts into a resistive load for more than an hour in 40C ambient. I doubt it would last.

[craig588]

How did you determine that your PCP&C 510 only had one 38 amp 12 volt rail?? The design is clearly a dual 12 volt rail desin indicated by the two dual inductor coils used on 12volt lines in 99.99% of computer power supplies. These are located behind the fan, you can see one easily, the other is right behind it, kinda hidden from view.

I was just going by the ratings provided in the documentation with the PSU. I just thought of something though, is there any PCPNP PSU counterfitting going on?

The only load that does not induce noise into the circuit is a resistive load. This is how your supposed to measure power output of any power outputting device. If you're using a load that induces noise, your measuring the noise of your load, not the power supply.

Unfortunately I have already sold all of my PSUs I used to do my tests, but I really wish I knew that before hand. I figured I would use a computer because people would want to know real world results rather than results based on something I built. (I thought about copying the unit that Toms Hardware uses to test PSUs, but decided against it because of the "non real world" testing environment it created.) If I ever test PSUs again I will build a unit like that.

Did you take into account the dual voltage rails on most supplies? what rail did you measure?

I measured the 12V, 5V and 3.3V rails from the ATX header. I didn't bother with the drive rails because we really don't need clean voltage going into the HDs because we arn't OCing them.

[blinky]

the sales rep is telling you BS to get you to guy the PCP&C

tons of people on this forum have ran OCZ600s with SLI 6800s, so it will work
if you talk to someone like Mdzcpa hell tell you that he got better clocks with his PCP 510 ... so its a tossup

id go with OCZ they have good warranty, and nice reps

[antipop]

PCP&C has a 5years warranty on the 510!!! I think that's pretty amazing and tells you how confident they are in their products

i checked ocz has the same warranty, however i'd go with the PCP&C as some serious oc'er only swear by this brand and if i have the money i'd buy one too

[computerpro3]

Everything the OCZ can do, the pcp&c can do better.

Including warranty.

Seriously, think of something that the OCZ can do better? I honestly can't...

[SPL15]

the sales rep is telling you BS to get you to guy the PCP&C

tons of people on this forum have ran OCZ600s with SLI 6800s, so it will work
if you talk to someone like Mdzcpa hell tell you that he got better clocks with his PCP 510 ... so its a tossup

id go with OCZ they have good warranty, and nice reps

Ha HA sales rep? If your talking about me, I'm an independant contractor that deals with companies much much MUCH larger than PCP&C. And I work at a local multi Billiion corporation that has no stake in the computer industry. I have no affiliation with PCP&C except that I use four of their 510's which have improved my productivity unmeasureably.

I rely on PCP&C for my test stands and my workstation computers. If I have constant failures, which I did with other brands, I lose money and lots of it. I've had one failure with one PCP&C on a workstation that has 4 gigs of ECC ram, raid 5 using five 15,000 SCSI drives, and two zeon processors OC'd to 3.4Ghz. Other units powered it to the same overclock and what not, but after being in use for several hours full CPU load, they would randomly reset the computer needless to say the PCP&C 510 solved this problem for all my workstations.

I have no stock in the company, honest. But my lively hood does depend on the reliability of their Power supplies so that is why I support them so much. If OCZ writes an E-mail to a customer saying that their 600 will not do the job reliably, and shows them a link that has their competitor's unit gaurenteeing reliability, I think it says a lot of the 510, and a lot about OCZ's great ethics. I'm not saying other power supplies are crap you should only get PCP&C, I'm saying that for the best of the best in computer power supplies, the PCP&C 510 is really the only choice according to my own experiance, Nvidia's testing, and OCZ's opinion.

[EMC2]

*tries to avoid the minefield*

--- Craig ----

If you were hooked up to a M/B, then there shouldn't have been a need for added capacitance. I thought you might not have been because of your reference to the spec in your cap comment (see comment to SPL15 below).

FYI... the PCP&C's are UL certified, so they *have to* at least partially split rails for the 12V. There is a UL safety requirement that puts a 240VA limit on any single rail (20A for 12V). (any individual OC circuit is limited to the 240VA).

Regarding the OVP limit, you didn't comment, so I guess you didn't adjust the OC pots to see if they had an affect?

As far as measurement... were you making differential measurements?


--- SPL15 ---

Just an FYI... the ATX spec does specify the use of 2 caps when doing ripple measurements *if you are testing with resistive loads* to simulate minimum load filtering (see section 3.2.6 of the spec).

Questions for you - were you using realtime or equivalent time modes on your scope and what probe set? Single ended or differential?

Noticed you've modded your PCP&C. From your look see and mods, what is the limiting factor/weak point in the 12V rail design in respect to max current?

Side comment - amen on the grounding issue.

[I]*exits minefield, stage right*[/I}

[SPL15]

*tries to avoid the minefield*

--- Craig ----

If you were hooked up to a M/B, then there shouldn't have been a need for added capacitance. I thought you might not have been because of your reference to the spec in your cap comment (see comment to SPL15 below).

FYI... the PCP&C's are UL certified, so they *have to* at least partially split rails for the 12V. There is a UL safety requirement that puts a 240VA limit on any single rail (20A for 12V). (any individual OC circuit is limited to the 240VA).

Regarding the OVP limit, you didn't comment, so I guess you didn't adjust the OC pots to see if they had an affect?

As far as measurement... were you making differential measurements?


--- SPL15 ---

Just an FYI... the ATX spec does specify the use of 2 caps when doing ripple measurements *if you are testing with resistive loads* to simulate minimum load filtering (see section 3.2.6 of the spec).

Questions for you - were you using realtime or equivalent time modes on your scope and what probe set? Single ended or differential?

Noticed you've modded your PCP&C. From your look see and mods, what is the limiting factor/weak point in the 12V rail design in respect to max current?

Side comment - amen on the grounding issue.

[I]*exits minefield, stage right*[/I}

Ha ya mind field!! I like living dangerous I guess .

Did not know ATX spec requires outside capacitance for ripple voltage measurement. This kinda is stupid in my opinion, but I guess it is like amplifier measurements that require an inductive load when measuring real output, but in the case of amplifiers an inductive load makes output measusurements lower than with resistive (Unless it is a switching ampm, ei Class D class T). I feel that Ripple voltage on DC power supplies should be rated with purely resistive loads because it shows exactly what the power supply is putting out. A good power supply should not rely on outside filtering for improved performance In My unhumble opinion. if the caps are being used for as Bypass caps for stability reasons then I can see why, but by measuring directly from the PCB and not at the end of an 18inch cable, you make this a non issue. I believe you should measure power supplies by what they are doing design wise internally, outside influences like cable resistance should not be factored in since this is variable with different power supplies.

I used realtime with two probes in differential mode to negate outside interference on my probe lines. I like this way the best and feel it gives me the best measurements.

I believe maximum current in today's Computer power supplies are limited by heat dissipation/inefficiencies. I've got 13.8 volt DC power supplies to bench test automotive amplifiers that put out 350 amps!!! They are massive and pull tons of current from two 480 volt phases phase input and put out tons of heat when used with the 3000 watt amps I put on it.

Also the typical 15 amp wall outlet is a limiting factor especially when there is not PFC in the design.

But really I think space/noise/heat limitations are holding back computer power supply power output. Until manufacturers start using full bridge techniques in power supplies to gain additional efficiency, I dont see the power rating increasing in the current ATX size power supply.

Also cable resistance and parasitic resistances play a huge role in voltage regulation when drawing massive currents. This is a major design issue in today's computer's because of the low voltages and ultra high currents. Until manufacturers start using larger cable diameters and more redundancy of cables, I dont see voltage regulation getting any tighter than it is right now.

Compared to highend industrial SMPS's, top of the line computer power supplies are a steal!!! Too bad Most computer power supply manufacturers use highly misleading ratings. <---- Now that is a mind field

[The Mofo]

Oh my, im at work little over 24 hrs and i got more information then rice at a chinese convention! Well thanks for all the info. Since PCP&C is near my work, im going to got get my PSU from them. I can have all of my questions answered by them and maybe they might be able to tell me what other mods would enhance my psu. Thanks for the info guys.

[antipop]

Good choice, don't forget to tell us how you like it and if it has help your oc

[SPL15]

Oh my, im at work little over 24 hrs and i got more information then rice at a chinese convention! Well thanks for all the info. Since PCP&C is near my work, im going to got get my PSU from them. I can have all of my questions answered by them and maybe they might be able to tell me what other mods would enhance my psu. Thanks for the info guys.

Good choice man! You never have to worry about an RMA if it ever does go bad being that close!! You wont be disappointed!

[enzoR]

could you tell us about your "Asus P4C800E-Dlx Rev. 1.xx Never Before Seen Mods&Tweaks" please?

[haPpydUde]

SPL15 nice post, Lots of usefull info in there thanks. My vote is that you get the pcp&c as well, i have owned both psu's in question and i like the pcp&c better for stablitly, Tho I do love the nice ajustable rails on the ocz, I just wouldnt trust it to power everything ie lots of fans couple drives and 2x 6800ultras ect

[blinky]

ship

lol, no i was talking about the email he got from the PCP&C guy

im sure u know what ur doin

[EMC2]

--- SPL15 ---

If you had said equivalent mode or SE, I would have had to throw you into the mine field head first

You misread one of my Q's though... I didn't ask your opinion on the general topic of p/s design constraints and issues I said I noticed you had modded your PCP&C supply and was inquiring specifically about their 510 in regards to what was their design's limiting factor/weak point (i.e. what is their weak link component in the design which sets the limit on their max current, ignoring the artificial UL safety rule limitation - subjective view based on your fiddling and modding of it)

Regarding specs of anything... kind of like a friend saying the blind date they're trying to set you up with is cute to sell you on taking her out, eh?

[SPL15]

--- SPL15 ---

If you had said equivalent mode or SE, I would have had to throw you into the mine field head first

You misread one of my Q's though... I didn't ask your opinion on the general topic of p/s design constraints and issues I said I noticed you had modded your PCP&C supply and was inquiring specifically about their 510 in regards to what was their design's limiting factor/weak point (i.e. what is their weak link component in the design which sets the limit on their max current, ignoring the artificial UL safety rule limitation - subjective view based on your fiddling and modding of it)

Regarding specs of anything... kind of like a friend saying the blind date they're trying to set you up with is cute to sell you on taking her out, eh?

Yea for noisy high EMI/RFI environments, differential mode is the ONLY way to get unmolested measurements on a scope, and if you have a scope that is decent enough to do differential mode.... why not use it? I haven't even scratched the surface of what my scope is able to do. It's a fine piece of lab equipment that is for sure!! When I do use it I set it to run in realtime but it records in the background in realtime also and if there is an anomaly with certain characteristics I can choose from or anything different from the normal periodic wave, it stores the event and I can play back the wave form and see what the wave looked like before and after the event. The main thing I like about my work is my test equipment.... pretty sad huh ?

The overcurrent protection variable resistors are the limiting factor for current output .

Seriously though, as for PCP&C 510 and my experiances, I would have to say the main constraint on more output current (ie Power output) is space and heat limits. That power supply is jam packed!! It is a real real REAL pain in the arse to take apart and put back togethor. Which brings up another point. From my experiances in general, the more difficult and complex the assembly process for the manufacturer, the higher the quality of design there seams to be in the actual PCB's themselves and overall design (Definitely not a hard/fast rule though).

Also, since they are using a universal input design, the 120-240 volt primary side input capacitor is not exactly beefy in capacitance since the voltage rating must be rated at least 400 volts, but this is common among all computer power supplies that can work on 240 VAC. Since I live in the states and this power supply will never see 240 VAC, I took out the stock cap and put a lower voltage one in (250WVDC) and more than doubled capacitance. Again this deals with space restraints on current power supplies.

I know for a fact that the current PCP&C could put out more current in stock form, but heat issues are the only thing to blame for the present factory limited current rating right now. Techinically it could be said that consumer culture limits the output power right now in all power supplies do to everyone being sooooooo incredibly stuck on a silent power supply (Even when they have 6 120mm Vantec tornados in their case). Until there is a huge breakthrough in efficiency (Fullbridge outputs) heat is going to be the main issues with any well designed power supplies regardless of manufacturer. Also Consumer culture has a lot to do with cost. Are customers willing to pay lots more for more output current when it really is not needed? Even when your current best is more than enough?

If you were to put a vantec tornado in the 510, you could safely boost the overcurrent protection limits with everything else being stock, but I doubt you'll ever draw the factory rated current limits with a half way normal Overclocking setup. Until Fullbridge is implemented, we'll be limited to about 70-75% real world efficiency which means 25 to 30% of the power your suckin from the wall is being dissipated in your power supply (Put that in your pipe and smoke it). A silent fan on a 500watt power supply under heavy load is like running a marathon in the desert in a rubber suit. People put fans and heatsinks on a bunch of things that are not even getting luke warm on their systems, but they leave their power supply to tuff it out in an enclosed steel box with a low RPM silent fan.... Go figure. I think a lot of people could upgrade their fan in the power supply and get higher OC's due to current derating in their power supplies from being hot.

Specifically design wise for the PCP&C 510, I would say the 12 volt output inductors would need another parrellel conductor in order to carry higher current with minimal DC resistance, but this is inside the feedback loop of the supply so DC voltage drops across it are negated by the regulation circuitry. I would still like to see them have one more conductor though. I really think the design is solid enough to go higher, but again heat is an issue. Also, on the newer revisions of the PCP&C 510, there is a PCB COVERING half of the heatsink fins in the supply!!! It's being used to give extra support to the mains line filtering inductor, I chopped off the excess on mine. It wasnt there on the first revisions.

Also, On the older revisions, they use a nydec Beta SL fan on the 510 rated for 44 CFM, but now they use a delta rated at 55 CFM. The funny thing is is that the Delta seams quieter than the Beta SL. It is a different noise, the Nidec Beta SL sounds more like rushing air, The Delta has that low frequency PWM RRRRRRRRRRR to it until you get a devent load and then it sounds like rushig air, but still softer than the Beta SL. If your system is totally silent I guess it could be annoying, but I doubt you have the need for a PCP&C 510 if your computer is totally silent running.

I changed my rectifiers out for International Rectifier's HFA50PA60C, which have a low forward conducting voltage drop and softerfast recovery which theoretically decreases heat,innefficiency, and noise but to tell you the truth I doubt it did anything worth the effort.

It's gettin late. I'm going to bed. Long day... Cold beer

[craig588]

Regarding the OVP limit, you didn't comment, so I guess you didn't adjust the OC pots to see if they had an affect?

Earlier in this thread

I sold the PCPNP a long time ago, but I belive the OVP came in at 3.5V on the 3.3 rail. I didn't push other rails back then, so I didn't test their limits.

were you making differential measurements?

My scope isn't good enough for that. (Maybe I should take a switch and flip it back and forth really fast ) I was acctually just looking for differences between PSUs, not so much the specific quality of one. The PCPNP showed fluxuations, the others didn't. I made sure to only change the PSU and keep every other variable as close as possible.

[jinu117]

May explain some death of memory controller over time on one of my friend whose PC&Ps (Believe he has 3) with direct vdimm mod and Enermax. He didn't experience this with OCZ 520W, Antec TruePower, and I sure didn't experience it with Zippy. (I did have PC&P but was too afraid to test on FX-55)
Believe he did burn through about 5-6 FX-55 to figure it out eventually.. -_-; (AMD do have killer RMA process after all, in fact, the FX-55s I had are from one of his spare ones :P)

[EMC2]

--- SPL15 ---

The main thing I like about my work is my test equipment

Toys, toys, toys

The overcurrent protection variable resistors

space and heat limits

*sigh* bypassing obvious well known constraints applicable to most things... :p

with a half way normal Overclocking setup

Normal? What's that, a setting on your washing machine?

Specifically design wise

Ahhhhh... finally Output inductors and possibly recovery diode ratings. Thank you


--- craig ---

Yes, I had noticed you had gotten rid of it long ago Was just wondering when you had it if you checked to see if the OC pots had any affect on the OVP point as well

Maybe I should take a switch and flip it back and forth really fast

That comment invokes a hilarious vision of a mad scientist scene

Or you could get a good differential probe


Simple Q... at approximately what amplitude were the fluctuations and what repetition rate or freq? Second part, were they short term transients or long term periodicals? (no, I'm not talking about your mag 'scripts :p)

[computerpro3]

I sold the PCPNP a long time ago, but I belive the OVP came in at 3.5V on the 3.3 rail. I didn't push other rails back then, so I didn't test their limits.

Sorry, not true, I ran mine at 4.1v for benching with my bh5 for 3.8v. IF you don't believe me I'll post pics of it hooked to a multimeter. My 3.3v came STOCK @ 3.6.

And once again, I'll ask the question that no OCZ guys answered:

What does the OCZ do that the pcp&c doesn't do better?

[SPL15]

Sorry, not true, I ran mine at 4.1v for benching with my bh5 for 3.8v. IF you don't believe me I'll post pics of it hooked to a multimeter. My 3.3v came STOCK @ 3.6.

And once again, I'll ask the question that no OCZ guys answered:

What does the OCZ do that the pcp&c doesn't do better?

You guys are both right. The older revisions had a much lower 3.3 volt limit. On my oldest PCP&C 510 from mid 2003, The 3.3 rail only go's to 3.43 when it maxes out, the newest revision I have which I got 2 weeks ago, has a limit AND adjustable range above 3.6 volts stock. I dont know how high it goes, I cranked it up expecting only about 3.45 and saw 3.63 volts and quickly backed it off. Could have gone higher too.

If your unit really did come stock at 3.6 volts, that is not a good thing. Extra voltage means extra work for regulation circuits. sure it can be a good thing for OC'ers but they should not be shipping out at 3.6 volts that is for sure. Everyone I've gotten so far is exactly .05 to .06 volts higher than the target 3.3 which is desired. A 0.3 overvolt is out of tolerance for manufacture and shouldnt have been shipped until adjusted correctly. Sure your meter doesnt just read .3 volts higher than actual? Lots of cheap to mid price meters dont read low voltage DC accurately. Especially analog meters not laying flat on its back. Low voltage AC is even more innacurate due to tolerances in Vdrop across the rectifying diodes.

[computerpro3]

You guys are both right. The older revisions had a much lower 3.3 volt limit. On my oldest PCP&C 510 from mid 2003, The 3.3 rail only go's to 3.43 when it maxes out, the newest revision I have which I got 2 weeks ago, has a limit AND adjustable range above 3.6 volts stock. I dont know how high it goes, I cranked it up expecting only about 3.45 and saw 3.63 volts and quickly backed it off. Could have gone higher too.

If your unit really did come stock at 3.6 volts, that is not a good thing. Extra voltage means extra work for regulation circuits. sure it can be a good thing for OC'ers but they should not be shipping out at 3.6 volts that is for sure. Everyone I've gotten so far is exactly .05 to .06 volts higher than the target 3.3 which is desired. A 0.3 overvolt is out of tolerance for manufacture and shouldnt have been shipped until adjusted correctly. Sure your meter doesnt just read .3 volts higher than actual? Lots of cheap to mid price meters dont read low voltage DC accurately. Especially analog meters not laying flat on its back. Low voltage AC is even more innacurate due to tolerances in Vdrop across the rectifying diodes.

I asked them to pre-adjust it lol. Forgot to mention that. Mine has the external pot option, but I wanted it to be ready to plug in immediatley without measuring or anything. Mines also from early 04, I just recently had it re-done for SLI.

[mcnbns]

Not to crap on this wonderful thread, but is the Enermax 600W or Topower 600W anywhere close to OCZ's 600W PSU? They're rated pretty well, and aside from the Enermax's under temp protection at 10 C they look good. Thank you!