AX1200 after power failure

[coma200s]

Hello to everyone. I have encountered an issue with a Corsair AX1200 PSU powering intel DX58SO motherboards in the company I work for.

 

I will not get into more details if needed, but I really am fighting to figure out this problem for more than a week now to no avail.

 

To simple down things:

 

Corsair PSU + DX58SO (tried 3 different revisions of the motherboard, 404, 504, 703) plus different bios updates up to the latest.

 

Motherboard setup:

CPU Intel Core i7 Extreme 980X (Gulftown 32nm)

RAM Kingston KHX1600C9D3K3 (12GB kit of 3x4GB)

VGA Palit NVidia GeForce GTX 550 Ti

HDD Typically 6x WD Velociraptor 1TB WD1000DHTZ (3x Raid1 arrays setup)

 

I know the first thing to do in order to figure out a problem is try to recreate it. So in my case, I stripped out all the HDDs, even tried moving around (adding or subtracting)

the ram modules but the problem remains, and it is the following:

 

In case of a power failure, the machine does not really seem to be able to recover to its previous (powered on) state.

 

Scenario 1: Simulated power failure for <5sec. After power resume, the machine boots up normally. (the green mobo led next to the 24pin remains on during this scenario)

 

Scenario 2: Simulated power failure until the green mobo led turns off, which means the mobo is not getting any power. After power resume, the machine boots up normally.

 

Scenario 3: Simulated power failure for >5sec but BEFORE the green mobo led turns off. After power resume, the machine fails to boot. Pressing the case power button

does nothing. (although the green mobo led remains ON)

 

At scenario 3, it seems there is a time period between the power failure and the complete loss of power to the motherboard - after about

10sec - when the PSU will once do a "clicking" sound and then start an indistinct "hissing" sound, that I needed to stick my ear on

the PSU to hear it. And now the machine is in a "stuck" state. Seemingly getting power but not really doing anything. No fans, no nothing,

just the led of the motherboard and the led of the PSU on, and the "hissing" sound as well. Case power button press does nothing.

 

To work around the issue and have a successful boot, I either need to drain the power completely (make the green motherboard led go OFF)

or, interestingly, to just hold the case power button for around 5sec (when at that time the PSU's indistinct "hissing" stops and an ever

more silent continuous sound comes off the PSU) release the button and then press it normally. The green mobo led remains ON during that time,

but by the difference in the sounds of the PSU I can understand when it gets "unstuck" and is ready to power on normally again.

 

I have read about the "latch-off" issue and incompatibilities with hashwell cpus, but I do not have a hashwell and even if I did, the AX1200

is officially posted as fully compatible. So what is wrong?

 

-Remember I have tried two AX1200 PSUs with several DX58SO mobos of different revisions and the result is the same. Needless to say that other

PSUs do not cause this "stuck" state.

 

Forgive my long post, but I'm trying to be as clear as possible with this subject. I am running out of options and believe the only cause is a

spec incompatibility or something.

 

Of course I will be happy to provide you with more info that I may have omitted but really I think I have reached a wall in my self-research.

This issue is not only a testing scenario, but a real thing, because in my region the power fluctuations sometimes do cause short-term power failures.

If a power failure is between this "critical" time period of ~10sec, the machine will not switch back on, even if the motherboard is set to

"After power failure= power on". (and yes, we have UPS device but this has nothing to do with the problem I'm dealing with atm)

 

Thanks in advance,

 

Spyros

[peanutz94]

If a power failure is between this "critical" time period of ~10sec, the machine will not switch back on, even if the motherboard is set to

"After power failure= power on". (and yes, we have UPS device but this has nothing to do with the problem I'm dealing with atm)

Doers the issue still remain if you are not plugged in to the UPS? If so, use the link on the left to request an RMA.

[coma200s]

Doers the issue still remain if you are not plugged in to the UPS? If so, use the link on the left to request an RMA.

 

Not only does it do it with either the UPS in between or directly to a power socket from the electricity provider, but (as i mentioned before) it is not just one AX1200 that does that. Even newly unboxed AX1200 does the same on any DX58SO revision of motherboard i gave it a shot at. (may I suggest if someone can test the same setup, to post results as well?)

 

So this is a matter of specifications or something similar, but surely NOT a PSU that once was working "this way" and now "something has broken"...

 

What good will a RMA do in this case? I have 3 PSUs that are all under warranty (7years). So I will get 3 new of the same model? Same model=same issue... That will not work..

 

Please, do you have any other indication/advice/solution? I'm at a loss.

[jonnyguru]

I was going to say "bad motherboard", but the fact that it "does the same on any DX58SO revision of motherboard i gave it a shot at" is really puzzling.

 

I Googled your problem and it seems a lot of people have this same issue with this motherboard and several different make and model PSUs. Some people are having the problem due to certain memory modules that are used while others are having the problem depending on the particular graphics card you have. But nobody has narrowed it down to PSU because people have tried different power supplies and have been able to replicate the problem regardless of PSU.

 

If you can't use different RAM or graphics card to narrow down your problem, I'm not sure how else to help. Several of the other users were able to solve the problem with newer BIOS's, but you said you've tried different BIOS revisions.

 

The problem wouldn't be related to the "known Haswell issue" because the problem with Haswell CPUs and certain "older" topology PSUs involves going into sleep mode; not initial power up. Besides, you're not even using a Haswell CPU. ;)

 

If there was any reason this would be a PSU related problem, I'd have to point to an over sensitivity the motherboard may have to power good signal. Either PG voltage or delay. But usually the only way to address that is with a BIOS revision and you've already tried that.

 

Best thing I can do is put in a trouble ticket with engineering, have them duplicate the issue and then escalate it to Intel. On your end, like I said before, try other RAM or graphics card.

[jonnyguru]

Just to make sure... this is the older AX1200 Gold and not the AX1200i, right?

 

FYI: I talked to engineering and they said it could be because the hold up time for the AX1200 is so long and the +5VSB is still energized on the motherboard that the protection circuit in the PSU never really knows that there a brown out, so it doesn't actually realize the PC is off. The protection circuitry of the PSU prevents it from turning back on again.

[coma200s]

Just to make sure... this is the older AX1200 Gold and not the AX1200i, right?

 

FYI: I talked to engineering and they said it could be because the hold up time for the AX1200 is so long and the +5VSB is still energized on the motherboard that the protection circuit in the PSU never really knows that there a brown out, so it doesn't actually realize the PC is off. The protection circuitry of the PSU prevents it from turning back on again.

 

This seems like a reasonable reply/explanation to my issue (and initially the same conclusion I came at, although not really put that way since i lack the technical knowledge regarding psus). So this PSU never really "sees" that the PC is off that's why it doesn't really do anything when power comes back up.

 

I confirm I'm using AX1200Gold Units and not the "digital" (i) ones. I've tested different ram modules as well as VGA cards (from standard low-spec ones to power-hungry monsters) and still get the same results.

 

Again a RMA that would result in me getting back a brand new boxed AX1200Gold (or actually 3x of them) would not solve my problem.

 

I've also seen people sending AX1200Gold units and getting back AX1200i ones, but again how can I be sure the same issue is not present also with this combination? (ax1200i & dx58so).

 

Thanks for your help so far. I'm really looking forward to an answer from you and/or the engineering dept when they duplicate the issue, to work out what can be done with my case. If it is of any assistance, I can also post the three serial numbers of the units i'm currently testing, in case someone needs to track back the build of each one etc..

 

Spyros

[jonnyguru]

Again a RMA that would result in me getting back a brand new boxed AX1200Gold (or actually 3x of them) would not solve my problem.

 

I've also seen people sending AX1200Gold units and getting back AX1200i ones, but again how can I be sure the same issue is not present also with this combination? (ax1200i & dx58so).

 

There is no guarantee.

 

And engineering was able to duplicate the issue last night.

 

Question: What is the real world circumstance that's causing this issue? Are you really somewhere that experiences so many brown outs and are you not using these with a UPS?

[coma200s]

It took me some time to respond, but I hope this case is still "open"?

 

To answer your question, that particular machine is planned to be used as a remote-backup server put off-site in a location where physical access to it will not be frequent. It will, of course, be behind a more-than-adequate Apc smart ups 1500 (which will only power the main unit and its little NAS companion;) ).

 

At that location brown outs occur often, and even if the procedures of this machine do come to a halt (i'm not a big fan of most of the UPS companies' "reliable battery uptime"), it is imperative that the machine itself will always be online without human interaction.

 

As i understand you can contact Corsair's engineering dept. So, since they have duplicated the issue, have they come to any solutions/workarounds?

 

Thanks again,

 

Spyros

[jonnyguru]

If you're going to use the PSU with an appropriate sized UPS, your own testing has shown that you wouldn't have a problem. Only if the PSU was subject to brown out would you see this issue. So you should be fine.

 

And this PSU has been discontinued for quite some time, so I'm afraid there would be no workaround.

[kwisatzhaderach]

What if he connected a device that was powered from the USB port even when the PC is off, that would draw down the hold up power faster. A fan controller? I know something like an Aquaero would do that, but I don't know of any other devices but that would be an option.