DFI Recommended Power Supplies (UPDATED October 20th, 2006!)

[General Septem]

How about any of these, can you get these in Argentina?

 

Enermax

Sparkle

Fortron

 

There are others I think but you can look for yourself. Basically anything on the first article is a good PSU if you're not using SLI or expert.

[Raiker]

fortron i haven't seen, sparkle neither. enermax i think some1 carries the 535W, don't know if the new or old. also the same some1 has the liberty series but i heard about some problems with nf4 ultra-d... in any case, i buyed the tagan and had my first DOA, so i hope the guy sticks with the warranty and replaces it (for example i only get 6 months of warranty and even then the warranty only works if the psu starts (if the psu is inestable, not if it is dead))...

[DKAKINGCE]

I have a SS500 S12 with the 80%eff. I"m having problems with my computer, its in the intell 925x-t2 thread i'll keep it that but the sucker looks solid and is wraped up verry nice. After i fix the problem i'll come back and post on the SS,PSU.

[Nil Einne]

David, a little advise here:

 

If choosing between these PSUs which AFAIK can be considered resonably good brands, which one would you go for (price about same)

 

Enermax FMA 460W http://www.enermax.com/english/product_Display1.asp?PrID=28 the low end Enermax, seems similar in specs to Noisetaker 485W but only single rail and not as fancy...

 

Acbel Intelligent Power 450W http://www.acbel.com/html/products_/pc_1_main.htm (click on the Go not the more info which is a rather poor review) seems okay but info rather poor, dual rail and doesn't specify maximimum combined for +12V. I've e-mailed don't know if I'd get a response

 

Coolermaster Real Power 450W http://www.coolermaster-europe.com/index.p...l%20Power%20450 AFAIK OEM is Acbel according to you and it looks the same. Comes with a consumption display gimmick and may be a little cheaper then either of the above

 

BTW, you would consider Enlight worse brand then any of the above right?

 

Oh this is not for my system but for a DFI Nforce3 250Gb with a Sempron64 2700+ (probably to be OCed significantly) and Radeon 9700 Pro OCed to about a 9800 Pro (uses external floppy connector) and 2 HDs + 2 optical drives and 1 gb of RAM

[NEOAethyr]

I got a ?

 

If ocz is gonna keep pushing them and selling these things, are they willing to pay for the damage?

 

 

I'm am a very pissed off camper...

My entier rig seems to be dieing.

 

My 2 hardrives.

both new.

a wd1200jb and a wd1600jb.

The 120 has a smart error now.

The 160 keeps getting it's ntfs indexes checked and redone.

The perfomance for the 2 of them is getting worse and worse.

 

My dvdrw.

A benq ew162I.

Does'nt work all the time anymore, I find myself rebooting from time to time to get it working.

I mean, I can freshly bootup, cold boot, and get into windows and it won't see a dvd+rw.

Then reboot and it works, this happens warm boots too from time t time.

 

My floppy drive, same prob as the dvdrw but with floppys.

 

3 fans so far.

the 1st to go was a 120mm uh panaflo.

Which got so bad in one day on this psu that it made my ears ring for over 3-4 weeks!

The other 2, a generic 80mm which is making crackling noises.

The 80mm tornado, which I can now hear the bearings from time to time, now I'm constanly re-oiling it.

 

A brand new video card.

A msi 6600gt sp model, vivo, agp, 128m.

^^, very pissed.

Got stuck at agp 1x with this psu.

Once in a while it would yell about missing aux power.

When doing that, and yes the cable was hooked up, it would clock to 800mhz core in 2d!

 

Memory or motherboard.

This adata, which should get 266 100% stable, 275 stable as well, 288 suicide.

Can't even get 250 100% right anymore, in single chan.

 

Cpu, can't clock for nothing anymore, no matter how many time I redo the gel on it or mounting.

I can't even boot 2500 on it anymore.

I'm so stuck on 2.1ghz now it's not even funny.

 

 

My health.

The constant hearing loss I'm getting now since I got this thing is . me off so much.

This is the 2nd time it's happened now in 3 months.

Freaking ringing in my ears...

 

My time, like I said, been around 3 month.

I got the psu at the end of december.

Distress caused from it, hearing, time, and fustration.

My rig, rig is now going to crap fast.

A video card which I ripped out and broke in pieces.

At least 2 good fans.

2 brand new hd's, which 100's of 1000's of $ worth of stuff on it.

I've managed to back off most of my stuff.

But the most expensive stuff, I have not backed up yet.

I take sha1's of every single file in those folders, and if one is'nt matching, man I will be freaked to say the least...

My hd's should'nt be having any issues of any sort, abiet perfromance or stability, but yet they are suffering both, big time.

 

 

The cpu, a mobile mp socket a.

How would ocz replace that?

By me another one?

They don't make them anymore.

Same with the mobo, a nf2 lanparty ultra b.

 

The memory, granted I did not pay much.

$58 total I think it was for my tcc5 and adata but still.

If I don't have memory working at 500mhz min, at 100% stability.

Then I have nothing.

I mean that, I'm not being stupid, 250 is as low as I go here.

This should be child's play for me, but it's fast becoming more then I can handle.

 

 

Why would I post this?

 

I'm posting this because this ocz screwed my rig up.

If you're gonna buy an ocz, by all means do so.

Keep a close eye on things though.

If something seems not right...

 

Rails by default are way high btw as well, I mean way up there.

You have to void your warrenty to make them normal :

Is'nt that a crock?

 

$200 ocz psu

$200 video card

$58 memory

$130 mobo

$98 cpu

$200+ hd's

$45? dvdrw

$15 floppy

$25? in fans

 

Total cost of parts damaged incurred from buying an $200 ocz:

$771+ ^^

That's around $1000 spent for a rig that does'nt work for crap anymore when it did just fine for years.

 

Don't get me wrong, I don't mean to slander or whatever.

Just to warn ^^, I am a wee bit mad, that's all.

 

 

Edit:

, just found out that the high pitched sound is coming from the psu.

If I would of..., damn .

I could of tossed a $200 psu in the garbage 3 months ago instead of a $200 video card today.

 

This psu I'm gonna have to trash as soon as I can....

Freakin a....

What a great christmas present..., screwed my entier system.

I can't even afford to eat everyday freaking day let along buy another rig.

 

Heh..., well I guess that's it for me.

No more rig for quite some time.

I know what I have to do, but it's choosing...

Choosing either to tear out the psu, smash it to hell, or live with my hearing getting worse and worse until I can manage to get a cheapy $30 psu or something.

 

I know this is uncalled for but...

I hope the peeps at ocz goto heaven, because like I know I'm going to hell.

I don't wanna see them after I die lol...

 

Sigh...

Later dudes :.

Kinda bumed here that my rig is screwed because of this expensive @ss psu that was trash to begin with.

 

 

My quadro 900xgl is now acting up too, great...

[Nil Einne]

If you are certain the PSU has caused failured of components you might want to look into what, if any protections you are offered under law. You live in the USA so this probably doesn't apply but for people who live elsewhere, determine if your are protected against consequential losses under law. For example, in NZ you are covered from forseable consequential losses. I would say the death of components due to a faulty powersupply is a foreseable consequential lose but AFAIK, this has never been tested here in NZ or anywhere else.

 

There are several difficulties with such cases. Number 1 is that most likely, even if you approach the retailer or distributor with your claim, they will not agree and you will have to take them to court or a disputes tribunal as is appropriate in your case/country. Number 2 is that you will have to prove is very likely the component failure was due the power supply which might be difficult. If you can show for example that the power output is not within specifications and is likely to cause damage I expect you may have a case. On the other hand if the power output is always within specifications, I doubt your claim will be accepted.

 

BTW, of course if you have overloaded or modified the power supply, you will have no case/claim. Finally, as I have stated, since you live in the US, I guess you'll probably have no hope unless you can launch some sort of class action lawsuit that appears to be the rage in the US... Of course, since the number of people having the problems you have mentioned is small, you probably don't have much hope for such a course of action.

 

BTW you say there is a video card you ripped out and broke into pieces? I hope you mean you did this on purpose, because if not... In any case, although I understand how frustrating computer problems can be, if you purposely break a video card in to pieces in your frustration, I wonder if it's possible you might have an anger issue and might be a little careless or fooldhardy when frustrated possibly damaging components.

 

Also you say the rails are high by default? How high? If they were not within published specs, you should have sent it back. If they are high but within published specs, it would have been okay even if not optimal. I would not have recommended modifying the PSU. If you did modify it, it's quite easily possible you made a mistake and are the reason for the PSU being damaging to components. I'm assuming the voltmeter/multimeter you used to test the rails was properly calibrated and you tested the PSU under load, correct?

[NEOAethyr]

No I can't even afford to send anything back, nor a $15 dmm to check the levels.

 

And yeah I did the video card thing on purpose, though now I remember I need ps2.0 but ohwell.

My old quadro here is doing allright for the most part.

 

I got the psu as a christmas present^^.

I got the video card from a buddy for free, call it sponsership.

 

 

Yes I have anger issues ^^.

If something does'nt work after 3 straight months trying allmost every day...

 

 

As for the voltage lev's.

The only way I know they were bad:

Tornado rpm(before I modded the fan) was over 6100 rpm on this psu.

Stock rpm I think was 5700 or something.

 

My enermax 350 ran it from 5800 to 6000 at times, from a dusty heatsink to a clean one.

6000 max though and only for that day I cleaned the heatsink.

 

 

High pitched noises started the 1st day.

I left it be for the time.

Then after I while I decided it was time to fix the issue, I coud'nt figuer it out before it had effected me big time.

I could shut the rig off completely, sleep for say 12hrs, wake up and still could'nt hear much but ringing.

So much so I would have to ask my mom and such to speak up.

The was other misc things I could barely hear.

 

I had a buddy over that tried helping me with that.

He could'nt figuer it out either.

He got a monster headache from it.

My mom was able to figuer it out of all things.

It was one of my new panaflo's.

The one on my chipset and memory.

I took that off and had to trash it, keeping the outer case for it for someday later if I wanted to use it...

To tel the truth I could not tell if it was the panaflo or not.

But since I could have my mom verify it, then again my buddy verified it as well, I trashed it.

 

Last night I looked up some stuff and found that some people return the psu because of that issue...

Like it's a psu issue or something.

Maybe it's caps...

Anyways I'm having that issue again with the high pitched noise.

 

 

As for voltages.

I've tried to keep them around .1 or .2 under using mbm and the bios.

It's the only thing I could do.

I modded my tornado a while back and the rpm skyrocketed from that so I can't truly tell anymore what is what for the +12v rail.

 

I know what levels seem to perform the best, 11.89 or whatever - 11.97.

5.97 - 5.99.

The 3.3 confuses the heck out of me....

Turn it one way, unstability, the other way, again problems.

I have no idea.

 

 

I recently made a bios with 2 seperate romsip in it.

With hellfire's 3eg for aggresssive setting in my bios.

My old alpha omsip for the optimal setting in the bios.

I've only been running 250.

I am totally unstable with hellfires at this speed when I should'nt be, only single chan.

Again unstable, to the point where programs exit on me and etc with my alpha bios.

 

I lowered the +5v and +3.3v rail last night since I removed the 6600gt.

To that .1v under.

This made it so whenever I startup with my quadro, it like hicups :.

Hd speed got way slower again...

Which sucks to say the least.

 

 

I can say re-adjust the rails for optimal stability if I wanted to.

Which is around 5.87-89, 12.2x I forgot.

Rpms at that speed are off the charts though.

 

Stock rails read out were like this, if I can find them...:

Hmm must have it wrote on paper...

 

+12v: 12.11 - 12.29

+5v: 5.08 - 5.11

+3v 3.25 - 3.26

 

Regulation, software wise has gotten worse over time.

At one time it was +12v: 11.98 - 12.73!

 

3.3v rail is worse, but not by much, as well as the 5v rail.

The worst one is of course that +12v rail.

Fluctuating around .2+ avg.

 

I know software is nothign to go by, and that perhaps it's leading me in the wrong direction.

 

I still don't think 6300 or just below that in rpm is right for a little 80mm tornado.

I can also check by device sync by watching a ideo and looking at the stats in realtime as I adjust the rails.

I can as well, adjust the rails while checking for stability in memtest or s&m.

 

 

The video card acted up as soon as I aded the psu, it just took me some time to figuer out that actual problem.

I even got the problem when I swapped psu's for my old one, which was really screwed up.

 

My dvdrw and floppy don't even work on the old psu anymore.

As well as I cut off it's atx conenctor, I wish I did'nt do that now.

I was using it for a 2nd psu a wile there untill I decided the enermax was crappign out.

Now it seems that the ocz has really backed me into a corner.

 

To the point where it's less stable then a fake ibm(acer) k62 350mhz rig that has never seen memtest or prime ^^.

No joke here.

It's that unstable.

I would'nt want to burn anything like that.

 

The only thing I can think of is to make a special bios just for my rig.

Just so I can keep it stable 100% so I can burn stuff off it, take out my 160gig and stick with the 120gig with smart errors.

 

I'm so tempted to build another rig out of k62 parts or somehting.

The only prob is I don't have another psu at all.

The only other psu I have has 11a for the +12v rail, which is my mom's.

 

I wanted to kick my rig last night so bad but I just can't.

The mobo and cpu may be fine after I get something going.

That and it's the highest clocker I have straight out.

 

 

I know part my prob is the rails.

What I would need to do is up the +5v and +12v a bit until it's stable again.

And maybe up the+3.3v but I'm not sure abot that one, strange results with that.

Oil'ing my tornado everyday though just for that I dn.

That and I never got 2500mhz cpu stable with this sucker so I'm just clueless.

 

 

I've tried so many diffrent levels, never changing the rail led colors on the ocz.

One thing I do know, the least amount of device sync problems happen when the actuall software voltage readouts are most stable.

There's spots in where it's most stable at, upp'ing the rails or lowering them results in more fluctuation.

Just touching them does that...

 

 

I had my hd's performing well on dual psu's.

Then right afterwards on the ocz when I re-adjusted the rails to inlcude the load for everything, as a single psu.

Then afterwards it went down hill without me touching them.

I could never get my 6600gt out of agp 1x mode.

Over 20 fps device sync offset when capturing .

200megs a sec at times, up to a max of 1600megs a sec of agp bandwith at 100mhz 8x agp .

Instead of around 2300megs a sec.

 

I've been going through stability isues, video card issues and hd issues for over 3 months now.

Fixing one but not the other, and etc, never fixing that video card one nor the cpu one.

Actually I did have the cpu stable when I 1st got it though.

When the rail was around 12.30v.

I tried that though and it would'nt fix it...

 

And like I said, I just can't afford a dmm for checking the levels.

I could of last christmas when I had $80 but I spent that on food.

I wish I still had my analog voltage meter but I can't find it for crap, I must of tossed it long ago since I broke the needle on it.

 

I will have another go at it in a few mins here.

I will try getting my rig stable enough so I can at least continue burning with this rig.

Thats' the only thign I really need it for, that and continued work on my mods and tweaks, etc.

I don't think I'm gonna beabel to capture though.

Not even with this quadro.

Sucks watching tv and wathing someone's lips totally not match there voice, being allmost a sec off.

 

Oh yeah, I used to hear my tornado vary quite a bit during load with it's rpm, by ear.

But that stopped after moved my psu away from the cpu in to the front bays.

[ayeayre]

Just a quick question. Im going to get the OCZ powerstream for the dfi sli-dr expert but in the first post theres only 2 or so psu's that have the 8-pin (+12V) EPS connector REQUIRED for the expert.... ~>

 

PC P&C 510 ASL (nVidia SLI Certified)

Only the ASL model has the 8-pin (+12V) EPS connector REQUIRED for the Expert!

 

Enermax 535W (New) FMA 535W (nVidia SLI Certified)

Only the current model has the 8-pin (+12V) EPS connector REQUIRED for the Expert!

~>

 

Does this mean that there will be a problem/disadvantage with the other psu's and the expert? Seeing that the Powerstream is one of the many that dont have the REQUIRED 8-pin (+12V) EPS connector...

 

Thanks..

[NEOAethyr]

ayeayre, the ocz 600w powerstream does have te 8pin connector.

It's in 2 parts, splitable and joinable, stuff like that.

If you get one, get a dmm to check the levels that are setup stock.

 

I was recently told that using the adjustable rails does not void the warrenty I guess.

Not totally sure though...

But whatever.

 

 

 

As for my end, a status update.

I got 275x8 stable on my lpb board with a custom bios I put together meant specificly for my problems.

 

 

Still had some odd issues with media player classic with it exiting on me.

This may be a win2k type issue however, or other, more on this :

 

Primary Disgard Timer

Secondary Disgard Timer

 

Both setup to 1024 clocks instead of the default 32768 clocks.

Fixes the issues I have with port timeout.

I can't paste the function name because I can no longer reproduce the issue...

It now produces the same error as below.

 

Still have this issue:

!WaitForMultipleObjects

I open up a dvd image and play it with mediaplayer classic.

I then click repeatedly, fullscreening it and then windowing it.

After a few times it'll exit automaticly and I'll have a drwatson thingy in my logs.

It says the above function most., Like it did before with some sort of port timeout function.

 

Fixing the port timeout had the same exact issue.

Except that it would exit the 1st time I did it every time.

 

Setting up the video card's latency from 40h to 00h fixes the device sync issues.

Sync issues of over 200ms.

 

 

Ok, still have an average sync offset when capturing of 999ms!

That is obscene yo...

 

Anyways, I found a way around this , most excelent too.

I allways used vmr9 as a preview output, or a direct output for previewing.

The reason being is because I found it be the fastest renderer availible back when I tested it.

 

However.

I checked the default renderers, a vmr and non vmr ouput version(ie new ver's of windows default and pre sp1? of 2k's default renderer).

I ended up using the vmr ver because I prefer it from experience....

Anyways I got like .5 fps with it :.

0 sync latency though...

I hooked it up with the tee filter, then redid the entier graph so 720x480 was working correctly.., then bam.

Instant 29.x-29.7 fps .

Good quality too.

 

So I have captures out of the way .

Video playback has inproved probably 10 fold in performance.

 

 

However, I still have issue with my hd's.

The video card obviusly has some issues, which I allready knew of because I screwed that up, bad caps on it.

But...

 

There's defently hd's issues.

There might be agp bus issues still.

There might be sound driver issues with a lagging buffer?

 

 

Ok...

When I was 1st plaing around a while ago, a dvd image on my hd was getting alot of cdrom errors in the logs, using daemontools.

But after I started priming(after s&m and superpi), they never appeared again.

I have a fealing, and I was thinking this before but I was'nt sure, that my hd stability issues actually stem from memory errors.

 

However I know that my hd's are not performing like they should.

That and that one smart error bothers me, but that happened on my enermax when using dual psu's.

 

 

I have to figuer these out:

Do I have a sound buffer issue to fix, again(yes this would be the 2nd time, the last was with an overflow or underflow making major static during ac3 playback)?

 

Hd speed, I had this fixxed once.

When I had dual psu's, on the enermax the speed was good until I started having problems.

Then when I 1st moved them back to the ocz.

Probably a voltage level issue, may need breakin(ie may need time to take hold after set correctly).

 

Agp bus issue.

I can run agp 4x on this military ver of a quadro 900xgl, quantum3d ventana 5121...

I could only run agp 1x on my old 6600gt, which is gone now.

I assume this is because the 6600gt had an external power connector for the majority of it's power.

? is if I still have this issue or not though..., I can't really tell with a card clocked at 200mhz for memory :.

Though I have been inproving it with the above latency and time out tweaks.

Around 20megs a sec more bus bandwith from the above.

Maxxing around just under 900megs a sec now.

 

Dvdrw and floppy.

Floppy seems fine now.

Dvdrw, needs more testing, I need to reboot a buynch of times just to make sure.

I up'ed the votlages a tad on all 3 rails when using s&m at 275, because I noticed it was falling below the .1 line thing(ie 11.9 and etc).

 

The video card hiccup when starting up is still there though.

Maybe that's a windows issue I don't know, I may of messed up windows when I was running 250 and was unstable.

Hd's are no longer hitching video during startup, a tiny bit better.

 

 

So I'm close.

Just that I have a few issue to work out now.

I can say that I now have a workable rig ^^.

Just needed a nice kick in the butt to get there, and I needed some motivation...

I fealt alot better this morning to tell the truth.

That's probably why I sat down and actually fixxed some of the issues, glad I did.

 

 

Anyways sorry for the long and probably kinda freaky posts.

I reccomend that if you cannot get a dmm, that you don't buy the ocz.

It's just to hard to manage without one.

It may be a good choice if you're willing to work alot with it.

I would reccomend staying the closest to normal levels as possible.

Then adjusting as you go, for latency.

Voltage defently effects latency.

As well as stability.

That's what makes the whole thing hard to deal with.

 

 

One thing I will have to track down though is this high pitched noise :.

I would hope it's another fan, even though that's not cool.

 

 

Oh and going from 250 unstable to 275 stable...

I was only able to do so with a custom bios I setup just for this.

With this bios I can't even run 266 stable without hardlocks in memtest ^^, that's how crazy of a mod it is.

 

Laters.

[ayeayre]

get a dmm to check the levels that are setup stock<~ sorry wats a dmm? and how do i do that? n wats it for? lol

ok so theres 2x4pin connecters on the powerstream to use the 8 pin connector on the board..correct?

 

so instead of having 1x8 pin it has 2x4 pins instead..

 

thanks

[NEOAethyr]

Yep a 2x4 connector .

It says you can use any of the 2 on a 4pin mobo, but that's not the case, as onyl one of them will actually fit in the socket thing.

Anyways yeah it's a 2x4 that'll work fine in a 8 pin mobo .

 

A dmm, it means a digital multi meter.

Use it to check what voltages you have setup stock from ocz or any other psu that has adjustable rails.

If you don't have the time to fool with such a psu, and it does take time(in my eye I believe it does), then buy a psu that does'nt have them adjustable.

Like an enermax or other.

 

 

I say to check them because I've had so much trouble with mine.

I have a fealing that at one time when my software says I got one voltage.

Then at another time with a slightly diffrent setup, it says somehtign totally diffrent...

The fealing I have is that both are incorrect in some way.

 

Software:

Say that setup # 1:

11.80v

Adjust to 11.90(min, max will be higher)

 

Setup #2:

12.10v

Adjust to 11.90 min again.

 

Both results are diffrent.

This is what I'm not sure of.

Though on load it does seem to hold allright, but only if I adjust during load.

Blah lol.

 

 

Anyways get a dmm and you wold really know what is what and all ^^.

It's best that way, so you are'nt getting all confused like I am... lol

It may save you alot of grief too, because you'll have an idea of what votlage level does allright here and there, etc.

 

[Nil Einne]

NEO- If you tried to adjust your rails based on MBM, you made a big mistake. It's a well known fact that motherboards sensors are not always well calibrated and really should be taken as a guide only. Case in point, with some Asus NF2 mobo I had, reported readings for 5V and 12V were about 0.3 above or below (can't remember which) correct levels indeed I believe entered 'danger' zone on MBM on some occasions but when I tested with multimeter, readings were less then 0.1 different from the correct values (5V or 12V)... The fact that your fan MIGHT have been going faster then it used to didn't really prove anything. I suspect you didn't measure the speed of the fan in some way that is known to be accurate. The readings given by a computer are again generally not that accurate as it depends on the sensor among other things. Also, since I guess you did not test the voltages of your old PSU, there was really no reason to assume your old PSU was not undervolted assuming there was a voltage difference between your new PSU and your OCZ PSU.

 

If you really have ringing in your eyes several hours after you turn your computer off and you've even been asleep YOU MAY HAVE A VERY BIG PROBLEM. It may be Tinnitus or something, see a doctor ASAP. It's possible but quite unlikely IMHO to because by the PSU fan. Fans can be very annoying especially the high pitched ones but it is unlikely IMHO for any normal computer fan, even a tornado for example to be loud enough that they're likely to cause tinnitus or any hearing problems in such a short space of time. I don't know what your job is but if it involved a lot of loud noises you might want to make sure your employer is providing adequete protection. I guess you don't visit nightclubs or anything since it sounds as if you don't have the money.

 

Also, I'm not sure but I believe 12.73V for the 12V lines would be outside of the specs and if you really had such readings on an accurate source (e.g. a multimeter your motherboard using MBM is not reliable), you should have returned your PSU. If you cannot afford to send it back, you should have bought it from a local retailer instead who deals with RMAs. In any case, really when you buy something, you should assume you may have to send it back and should make sure you have enough money to cover it if you have to. If you won't you should reduce your budget until you do...

 

Really what you should have done if your computer was so unstable since you appear to have access to multiple computers is your should have tried the computer with only changing one thing, the PSU. If you have many problems with one PSU but not the other, the PSU is very likely the source of the problem and assuming it's not being overloaded which in your case I guess it wasn't you can assume it's faulty. Most retailers would accept a product for RMA/replacement if you have carried out such a test and if yours didn't you should have used a better retailer. (In any case, I think OCZ deals directly with customers for RMAs as well and I'm sure they would have accepted a return in such a case). This would have saved you a LOT of time and a lot of frustration and from the sounds of it a lot of money.

 

I'm still a bit confused about how you adjusted the rails. Assuming the OCZ has adjustable rails as you seem to be suggesting, do you really void your warranty by adjusting them? I don't know much about OCZ psus...

 

In conclusion, your PSU may or may not have been faulty. However given that you have adjusted the rails without properly testings them, you may have been partially the cause of the problem. (And as I have pointed out earlier, if they were within specs, there is absolutely no reason to adjust them unless your some insane overcloaker who wants things perfect which you clearly aren't and don't have the money to be). In any case, you've made numerous other mistakes in diagnosing the fault and really given your apparent financial difficulties, I would recommend you seriously reconsider the amount of money you spend on your PC. A 6600GT and OCZ PSU may sound cool, but you can't afford them. I would try to salvage anything you can and write off anything you can't salvage and sell anything you don't need that works and put it all down to an unfortunate learning experience.

 

BTW, afaik, most adjustable rail PSUs of reputable brands should on average be just as close to the correct voltages as non-adjustable rail PSUs of reputable brands. Generally, adjustable rails are included for the benefit of those who want to undervolt or overvolt and also for those who are very fussy about their voltages (with or without reason). Personally, with reputable brands, I wouldn't bother checking the voltages unless your a fussy person... Of course, I'm not saying they will never have problems, but it's so rare it's not really necessary unless you have reason to suspect there's a problem