NF4 Expert will not cold boot- psu issue?

[Limnos]

Here's some interesting info I found on the Antec site.

 

http://www.antec.com/pdf/manuals/powertester_Manual.pdf

------

The Power Supply Tester works with the newest ATX12V and EPS12V power supplies that come with a 24-pin power connector. It is also backwards compatible with previous form factor power supplies that come with a 20-pin power connector. Simply plug in the power supply's main power connector to test the five individual outputs (+5V, +3.3V, +12V, -12V and +5V Standby) and the Power Good signal at the same time.

------

The most interesting bit of information comes in section 4 of the instructions.

------

Output Indicators in Green + Power Fault light up: Faulty power supply. Your power supply Power Good signal is bad. A power supply with a "failed" Power Good signal won't boot up your system but will probably turn on your system fans.

------

 

Hey ExRoadie, you are going to like this. I got my new board (and I'm glad I did RMA with my e-tailer since I had some suspicions that I did not actually get a NEW board the first time, and the new board they sent basically confirmed this.).

 

Anyway, prior to even hooking up my board to the PSU I tried out my new little tester and I got the exact fault that you noted above (in bold). All the rails test perfectly (all green and all bang on with the DMM 3.37, 5.08, & 12.1), but I get the fault indicating a bad Power Good Signal. Now I am not sure if all Liberty PSUs would show this (for whatever reason) but mine does, so I am now in the process of trying to deal with Enermax on it since my e-tailer RMA period is past.

 

What I don't get is why the PSU apparently tested out fine at the local shop...unless they were just blowing smoke up my butt.

 

In the meantime, has anyone had any issues with the FSP PSUs, in particular the Fortron Epsilon? They look like decent PSUs and I have not seen any issues with it and the Expert.

 

Edit...I see ExRoadie is running the 600W Epsilon...good enough fer me.

[Limnos]

Been reading up on the Power Good (Power OK) signal since mine appears bad and have a few interesting notes.

 

http://www.pcguide.com/ref/power/sup/funcPowerGood-c.html

 

"To prevent the computer from starting up prematurely, the power supply puts out a signal to the motherboard called "Power Good" (or "PowerGood", or "Power OK", or "PWR OK" and so on) after it completes its internal tests and determines that the power is ready for use. Until this signal is sent, the motherboard will refuse to start up the computer"...

 

"Sometimes a power supply may shut down and seem "blown" after a power problem but will reset itself if the power is turned off for 15 seconds and then turned back on."

 

http://www.fonerbooks.com/power.htm

"A PC that boots on the second or third try is most likely suffering from a quick power_ok (or power_good) signal, coming on before the power supply has stabilized. The presence of the power_ok signal tells the motherboard that the power supply is stable, while its absence tells the motherboard to stay off to protect itself. It's possible the power supply isn't quite up to the current ATX standard or the motherboard is a little too demanding about timing."

 

http://web2.murraystate.edu/andy.batts/ps/POWER_OK.HTM

"This signal is a +5V active (nominal) high, usually present within 100ms to 500ms after applying A/C to the powersupply. Active high means that as long as the PS is functioning properly (active), the signal can be measured. There may be some variation in the voltage level on this line, but ranges of +2.4V to +6.0V are generally considered to be sufficient to force the processor out of RESET."

 

Perhaps a lower voltage on this line is not enough on the Expert board. Also, it seems possible that your rail voltages could actually be fine (seems to be my situation) but that somehow the actual Power Good signal is not present (or is intermittent), or not getting through to the MoBo, or its timing is mismatched (as above) with certain boards and certain PSUs?

 

I'm starting to think the answer to *some* of these issues lies within the Power Good signal somewhere. Of course I'm a biologist, so what the hell do I know about electrical engineering :shake:

 

Hmmm, PWR OK is the gray wire (8) on the 24 pin atx layout and should be @ 0v or somwhere between 2.4-6.0v when sending the signal. I think I'll have a look with my DMM and see what I get.

[Limnos]

Ok, sorry for all the posts today, but I have confirmed that on my PSU the PWR OK line is only supplying 1.25v when it should be supplying 5v or at least somewhere between 2.4-6.0v. I'm not sure if that is because it is not really loaded well enough on my little tester, but I would suggest others with a cold boot issue try their DMM to test the PWR OK line on their PSU (pin 8 gray wire) and see if a) the voltage is high enough and B) whether it fluctuates a lot. I suspect either would give 'cold boot' symptoms. Not saying this will be the cause of all the issues, but it is worth checking. Note that my normal 3.3, 5, 5VSB, 12, -12 all test fine with my tester and my DMM.

 

Edit, tried out my PSU on my new board...same boot issue so I definitely think it is the PSU in my case. It will boot once in a while and when it does the PWR OK line shows at 5.11v (as opposed to 1.25 when on my little tester). Still think my issue has something to do with that PWR OK signal though. Going to spring for an FSP Epsilon 600 and find out once & for all. If I get a new Liberty back on exchange, so be it, but I don't want to wait around and I can always use the spare PSU for another build.

[tekert]

i had the problem:

1. cold boot (psu on): would hang at loading ram screen

If i reset here, would boot normally

2. warm restart in windows, would randomly hang at loadup

If i reset when it hung, would boot normally

 

this would happen 100% of the time. I tried changing ram to some older corsair value

WORKED FINE

 

so it got me thinking (i was using 2*1gb mushkin samsung xp4000 uccc chips when i had the problems and running latest bios from www.dfi.com)

 

what if i try new ram+ old ram together = WORK FINE ( old ram in orange slots), but who the hell uses 3gb of ram! (+this error crap was giving me the sh1ts big time)

 

ive spent ~ 1 week trying to fix it

 

finally i flashed my bios to the 2nd latest one...23/6 (23rd july) version

and everything works fine now!..same ram settings...same everythig, jus diff bios...so WIERD!!

 

so i dunno..try the bios version

 

 

Same problem, ill tell my story.

 

After overcloking and all that, i began to suffer from that problem

Power up when the pc was "cold" (12 hours no use por example)

y spent 2 weeks, yes that much figuring out what was.

well, always, if i reset the mobo boots fine, but strange things happens some

times, some had to do with overclock, MANY with ram, and many more with psu.

one day i flashed the bios to a new one, the problems went away 2 days and come back later..

well, i did many many test, with very weird results, but one thing is consistent.

the problems went away when the PSU was warm... at least on mine, but not completely away.

the fully solution was to increase ram voltage to 2.8v.. and decrease cpu voltage. yes,

decrease, i think the cpu voltage has some influence on DRAM clock and vice-versa, at least

for the boot problems, i mean, its no entirely the dram or psu fault, its weird, but i

noticed after 450 reboots, cpu voltage and dram clock had some relation.

other EXTRMELY weird thing was. when i bought the mobo, stock settings in bios, etc.

i was unable to pass 290 HTT... after a week, same bios, same settings, exactly same everything.

i was able to reach 330 htt (of course lowering the cpu multiplier). strange.. its not

something i tested 2 times, many many times, 100 reboots per day. the thing i can think

off is mobo power asimilation hah.

i wonder whats happening.. well, al least im stable.

sorry my ugly english, i don't speak very often.

 

DFI lanparty NF4 Ulta-D

Opteron 148 at 250x11 1,408v (bios 1,425v), dram at 196mhz 2,8v

Some crap ram OCZ pc3200.

Topower 420P4, tagan version. (when is hot it works well)

[MidnightUT]

Hey Everyone. New to the board, I just built my computer a week or two ago and just set to OCing it this weekend. My experience is almost the same as Gandalcrap's.

 

My board initially had no problems with cold boot, but after my overclocking session this past weekend it has stopped cold booting without first turning off the rocker switch to drain all power from the mobo. However as Gene first mentioned and Gandalcrap repeated if I quickly tap the on switch on the case the computer boots up every time. I tried adjusting voltages to RAM etc with no luck before I found this fix, I haven't tried setting all settings back to default and return to stock speeds but I can if this will help anyone in any way.

 

I will continue to monitor this thread as I'm interested if there is a final solution or explaination.

 

A quick history.

 

Built thecomputer a week or two ago, used with no problems. Switched to OCZ Powerstream 520 on Thursday, initially saw no problems. Overclocked all weekend and startes seeing cold boot issues for the first time Sunday evening. However I don't shut the computer totally off very often its possible I've only turned it off 2 or 3 times since Thursday when I put the Powerstream 520 in.

 

Thanks for all the input from everyone, especially the quick tap solution that seems to be working for me!

[dow]

I don't have the expert board, but I think I'm having the same problem. I've had my board for about a month, and its been working fine. It was a little hard to get started in the first place, but running Linux and leaving it switched on all the time has meant I've had very few reboots, and probably no cold ones.

 

That was until this weekend just gone, when I received another 6800GT in the post. I had to power off to put the new card in, and flip the jumpers to sli mode. Then I couldn't switch it back on. All weekend I was trying a million things (I'd imagine most people on this thread have had a similar experience) but the board would not go beyond 4lit LED's after hitting the power button, on the case or the board.

 

I was wondering whether the bios could be screwed, and although I've never touched it, was thinking of hot flashing (glad I don't have to do that now, the idea makes me nervous), but this thread had the tip thats got me to one LED. One LED I'm not worried about because there is no drives connected at the moment, but the thing that worked was pressing the power button whilest the psu switch was off, then hitting it a couple of times after flicking the psu to on. Doesn't work all the time, haven't narrowed it down to "one click 1second before switching on and 0.5 seconds after" or anything like that, but this if the first time I've seen any LEDs go out in days so now I'm happy, and am going to try booting an OS... or am I barking up the wrong tree when I don't have an Expert, just a Sli DR UT... I did buy it only a month ago, anything to do with the new batch?

[scottR]

Update on my cold boot issue- I contacted DFI about the rma of my board. They were able to reproduce the cold boot problem with an OCZ power supply.

 

They mentioned that several power supplies were tested: OCZ, Enermax and two different Antecs. In my case the OCZ made power supplies seem to be the issue- they got the board to boot perfectly with the Enermax and Antec units.

 

Recommendations for me are to switch power supplies, at least temporarily, while DFI and OCZ work on this problem.

 

Perhaps this information may help others.

[Limnos]

Yes my situation has been resolved with a new PSU as well. Hats of to ExRoadie who initially said my problem sounded like the PSU. I did RMA the first board for two reasons 1) the apparent boot issue and to rule out the board as the PSU tested fine on my DMM and at a local shop, and 2) I was almost certain that the board I bought 'new' was a returned product as opposed to new.

 

I now have a new board and the same issue occured with it and my Enermax Liberty. I took ExRoadie's advice about the Antec tester and found that the Liberty had a faulty Power Good signal...not something you automatically test for. Anyway, I bought a new FSP Epsilon 600 and low and behold it tested fine on the tester and is now running my new Expert board with nary a cold boot issue to be seen (fingers crossed).

 

I liked the Liberty, but I like the Epsilon even more as it just seems to have an extra bit of quality (connectors, sleeving, etc) and of course it boots my bloody board!!

 

Lesson of the story is that all the rails on my Liberty tested fine on my digital multi-meter and it would boot fine lots of the time. I had to actually test the Power Good signal on the Antec tester and my DMM (gray wire pin #8 on 24 pin connector) to determine that it was faulty. If you are having similar issues, especially with the Liberty, you may want to invest in the Antec PSU tester and see if you have the same problem.

 

Anyway, props to those on here (esp ExRoadie) that were throwing out ideas which helped steer me in the right direction at least (even if not officially).

 

Let the fun begin!! :nod:

[Sachs]

Well, if you guys were to try changing the startup values and found that it fixed the problem, DFI would then know where the issue lies as it may be working on the rest of their boards but maybe right on the border.

 

There is no need to shoot down an idea of someone who is merely trying to help find a solution to the problem. If it worked, great. Is it right that it needs to be done for a brand new board, NO. But atleast you can be part of the solution and not the continuing whining that it's not working.

 

 

DFI , the king of cold boot......NF2, NF3, NF4...MY GOD.

[scottR]

Update on my cold boot status:

 

I actually sent my board back to RMA for a second round of analysis regarding the cold boot issue. The staff wanted to add a couple of capacitors to the board as a possible fix. When the board came back I examined it very closely and could not detect any additional capacitors anywhere on the board, I suppose it is possible that some were replaced with caps of different rating or manufacture. Anyhow, same result with my OCZ 600w powerstream- no go on cold boot!

 

Finally I broke down and bought an Antec Neo 550 HE psu and the board now works perfectly. For some reason certain Expert boards just will not work properly with the OCZ power supplies.

 

Now I have a spare $150 psu that will not work properly with my board. Should have followed my instincts on this build and gone with Antec from the beginning instead of reading too much into the psu recommendations for the Expert.

 

Live and learn...

[red930]

Update on my cold boot status:

 

I actually sent my board back to RMA for a second round of analysis regarding the cold boot issue. The staff wanted to add a couple of capacitors to the board as a possible fix. When the board came back I examined it very closely and could not detect any additional capacitors anywhere on the board, I suppose it is possible that some were replaced with caps of different rating or manufacture. Anyhow, same result with my OCZ 600w powerstream- no go on cold boot!

 

Finally I broke down and bought an Antec Neo 550 HE psu and the board now works perfectly. For some reason certain Expert boards just will not work properly with the OCZ power supplies.

 

Now I have a spare $150 psu that will not work properly with my board. Should have followed my instincts on this build and gone with Antec from the beginning instead of reading too much into the psu recommendations for the Expert.

 

Live and learn...

Sorry scottR but that's simply not true!

 

There are some instances where replacing an Antec with an OCZ will cure the "cold boot" issue.

 

If you had paid attention to the numreous threads on this and other forums regarding this issue, you would have seen that the greater issue involves the newer high power units.

 

There seems to be some sensitivity to the "PWR_OK" signal in many situations.

 

The mere fact that you replaced the affected unit with another that works is indicitive of the whole issue. You could have just as well replaced one OCZ with another and resolved the issue.

 

To make a blanket statement that an OCZ unit won't work with the DFI nF4 Expert board is shortsighted and frankly ignorant of the subject. Get your facts straight before posting.

 

ExRoadie

[red930]

Update on my cold boot status:

 

I actually sent my board back to RMA for a second round of analysis regarding the cold boot issue. The staff wanted to add a couple of capacitors to the board as a possible fix. When the board came back I examined it very closely and could not detect any additional capacitors anywhere on the board, I suppose it is possible that some were replaced with caps of different rating or manufacture. Anyhow, same result with my OCZ 600w powerstream- no go on cold boot!

 

Finally I broke down and bought an Antec Neo 550 HE psu and the board now works perfectly. For some reason certain Expert boards just will not work properly with the OCZ power supplies.

 

Now I have a spare $150 psu that will not work properly with my board. Should have followed my instincts on this build and gone with Antec from the beginning instead of reading too much into the psu recommendations for the Expert.

 

Live and learn...

Sorry scottR but that's simply not true!

 

There are some instances where replacing an Antec with an OCZ will cure the "cold boot" issue.

 

If you had paid attention to the numreous threads on this and other forums regarding this issue, you would have seen that the greater issue involves the newer high power units.

 

There seems to be some sensitivity to the "PWR_OK" signal in many situations.

 

The mere fact that you replaced the affected unit with another that works is indicative of the whole issue. You could have just as well replaced one OCZ with another and resolved the issue.

 

To make a blanket statement that an OCZ unit won't work with the DFI nF4 Expert board is shortsighted and frankly ignorant of the subject. Get your facts straight before posting.

 

ExRoadie