Motherboard and broadwell-e dead again

[Geforce man]

Just went through a failed PSU of my own. Thought it was CPU or board. Was a corsair hx 850, high quality expensive PSU. Bought an EVGA 650 platinum, no issues since. Very bizarre as the hx powered my 2700k for a long time with no issues.

 

[2blzd]

I bought a 24pin motherboard extension cable. I don't know how that couldn't be compatible. Same with the cpu and gpu ext. cables. . I'll take them out though.

I meant clean slate in regards to the problem parts. Replacing everything would be crazy. Same ram, GPu, case and drives.

 

[MongGrel]

I don't know if this is relevant or not, but PSU fans blow in to cool their contents. The PSU exhaust goes out the back.

Older PSU's have fans in front. They blow in. Newer PSU's have their fans on the top or bottom, depending on how you orient your PSU. The fan blows in -- up if you have the PSU with fan side down; down if you have your PSU fanside up. You can test this; you don't have to take my word for it.

Resting a computer on a carpet is OK, but in such a case you need to have the PSU resting fanside up. Air will go into your PSU and out the back. Again, test this.

I don't know if this is relevant to your case, but the fan in a PSU should be an intake fan, not an exhaust fan. If it actually blows air out, the fan was installed incorrectly. Lord knows what else is wrong with it if that is true.

A thought: could your PSU be overheating and turning itself off? Could your PSU be overheating and killing your stuff?

Personally, I have a small HAF cases were the intake on the PSU is on the bottom upside down with a filter and output is on the rear for my HTPC in the master bedroom.

It is on a laminate floor though.

I have no idea how high the legs are on the OP's case, but it sounds like the carpet is pretty thick and might be allowing no air to circulate through the PSU more or less.

If the intake is blocked that PSU is just going to overheat no matter how hard the fan works.

Basically, what ehume said, I'm just seconding that.

 

[2blzd]

Up and running on the new system. Everything is running fine and nice and cool (just like the last 2 times I built this).

Removed all extension cables, including ones that weren't effected by the problem, (video card pcie extension cable - what a waste of $30 for all those cables)

Installed PSU fan up this time, pointing inside the case. Benchmarked my drives so far, everything looks good. What kind of tests can I run to make sure everything is ok? I mean I did this the last two times as well and nothing rear-ed its head, until it died.

Here's a screenshot of AIDA64 Extreme

Let me know if anything looks abnormal or of any tests I can run besides standard benchmarks.


Thanks for your help everyone, I hope this is the last of this...(knock on wood).

 

[fastamdman]

I read through the start of this thread, but didn't finish all of it. Glad you got everything back up and running. Have you tried running something like OCCT's PSU test? I don't recall if you just rma'd the board and cpu, or if you rma'd the evga power supply as well. I was thinking it was a PSU issue from the get go as well. Put some strain on the system and watch your 3.3v 5v and 12v rails and see what happens. Also I would run aida64 for awhile and let it log, even at idle. Because the last time it failed was while it was at idle, maybe the 3.3V standby is fluctuating during idle, I don't know just a thought. Just monitor temps and voltages and keep a close eye on it for awhile. Hope it lasts for you

 

[2blzd]

I read through the start of this thread, but didn't finish all of it. Glad you got everything back up and running. Have you tried running something like OCCT's PSU test? I don't recall if you just rma'd the board and cpu, or if you rma'd the evga power supply as well. I was thinking it was a PSU issue from the get go as well. Put some strain on the system and watch your 3.3v 5v and 12v rails and see what happens. Also I would run aida64 for awhile and let it log, even at idle. Because the last time it failed was while it was at idle, maybe the 3.3V standby is fluctuating during idle, I don't know just a thought. Just monitor temps and voltages and keep a close eye on it for awhile. Hope it lasts for you

Thanks for posting and caring enough to read through this and offering some insight. I appreciate it and everyone else's time and responses. Very grateful.

I RMA'd the PSU and the replacement EVGA sent me looks brand new, so thats cool. Intel sent me a new 6800k and I got a new motherboard. As far as logging within AIDA64, I assume this is something I have to turn on? I have not run OCCT's PSU test. I will look into that, thanks for the recommendation!

Edit: I just gamed in DOTA 2 for 50 minutes, at 2560x1600p @ 240 fps - ultra settings, dx12 not enabled. Cpu temps topped out at 45-47C. I noticed my 12v rail went from 12.288 to 12.199 or so during game. So it drop .09ish..is this ok?

 

[Ranulf]

At this point I wouldn't trust anything. Personally I wouldn't trust an 840 evo for anything other than mad science bit rot experiments.

That reminds me. I need to dig out my first 840evo and test it. I never updated its firmware and I think its been sitting with an old image file for over a year now.

Glad things seem to be working well 2blzd.

 

[ClockHound]

Thanks for posting and caring enough to read through this and offering some insight. I appreciate it and everyone else's time and responses. Very grateful.

I RMA'd the PSU and the replacement EVGA sent me looks brand new, so thats cool. Intel sent me a new 6800k and I got a new motherboard. As far as logging within AIDA64, I assume this is something I have to turn on? I have not run OCCT's PSU test. I will look into that, thanks for the recommendation!

Edit: I just gamed in DOTA 2 for 50 minutes, at 2560x1600p @ 240 fps - ultra settings, dx12 not enabled. Cpu temps topped out at 45-47C. I noticed my 12v rail went from 12.288 to 12.199 or so during game. So it drop .09ish..is this ok?

Good news. 3rd time lucky!

Yes, you need to turn on logging. As seen here, logging is selected from preferences.

And since you edit/render video - a good test that includes Handbrake in a batch is Asus RealBench

The 12V rail voltage drop is well within EVGA specs. That 0.73% variance is under the JohnnyGuru 1.4% excellent rating.

 

[BonzaiDuck]

This all reminds me of a flaw in my building of my 2700K. In the 2600K, the PSU was fed by interior air with the fan facing up in the bottom of the case. Yet the case-bottom has an intake vent for the PSU fan. The PSU fan has never spun up to my observance, because the PSU only barely gets warm when pulling ~ 500+ watts. With the nearly identical 2700K but with SLI, I had done what would be the first inclination -- use the vent.

But it's never been filtered. The box sits on 2" double-caster wheels, maybe 2.5 to 3" off the floor. On the other hand, it's a wool carpet, in a dry and dusty part of the country. Now I'll have to decide just how soon I flip the box over and have a close look. I should be able to simply unplug the modular wires, flip the PSU, and then twist each wire in an appropriate direction and plug it back in.

Somebody mentioned the deviation of a 12V rail from the 12V spec. The 12V ATX spec is really a variance or deviation from the 12V target, or +/- 2% as an acceptable range. If the 12V rail is slightly in excess but within the 2% spec, that's a good thing. If it is less but above 11.8V, this electronical illiterate is only guessing that it may not run to spec wattage, or maybe has a barely-acceptable hold-up time. somebody correct me or elaborate if you and your fingers have the time. But that's my intuitive conclusion about it.

Some of the EVGA PSUs of recent vintage are rebranded Seasonics. A review of a 750W "gold" or "platinum" model showed that it could sustain above 900W without shutting itself down, so the model performs beyond its spec.

 

[2blzd]

Real Bench gives me a "instability error" when I chose "up to 32gig" of ram on the stress test settings ( I have 32gig). But runs fine on 16Gb. I search google and noticed a lot of people have had this issue.

Is this an actual error, or the applications inability to use all 32gig (Im assuming because Windows is using some) ?


edit:

I just noticed there have been 3 bios updates to my board in the past two months of me owning it. The last two bios updates were on Oct 6th and Oct 11th, 5 days apart. And they both have the same notes:

"Version 1201 - 2016/10/11

STRIX X99 GAMING BIOS 1201
1. Optimize power management
2. Improve PSU compatibility



Version 1101 - 2016/10/06

STRIX-X99-GAMING-ASUS-1101
1.Optimize power management
2.Improve PSU compatibility
"

My computer died on the 11th. I know the dates are a coincidence, but are the issues a coincidence? Maybe its been the motherboard all along and not the PSU. All the motherboards Ive had died have been Asus X99 boards (x99 A-II and x99 strix gaming) and I've found a half dozen or so people on the internet with my same exact issue. (red pled1 by cpu socket lights up, no power to board)

Maybe this is an Asus x99 board issue. In my general research into x99 boards around the Haswell-E launch, I read there was a larger than normal defective rate, compared to the 115x chipsets.

I'm updating my bios now...this can't be a coincidence! ( aslo wtf is psu compatibility? If it has enough watts and had the right connectors, what else needs to be compatible ? )

 

[BonzaiDuck]

There's always the prospect of multiple causation -- as a possibility among discrete possibilities. There's a possibility of two causes interacting to create an event or result. Sometimes, and event is so singular that you can rule it out as something random -- it has an assignable cause. Howsoever inconclusive, any rational person might look at the odds of such a thing being random when it appears not to be, and accept it as an explanation.

It's interesting that the bios update change logs highlight "power management."

Usually, if I build a new machine, I'll pick a PSU that was touted for the very last chip generation and chipset, if not for my current processor and chipset choice.

There had even emerged a problem of compatibility between PSUs and certain models of CyberPower UPS systems -- I think it was the Active PFC feature. No problem for me: but I had to redeploy the UPS and find another for the system. What's going to take precedence? A Seasonic PSU just purchased? Or a 3 year-old UPS that can be used elsewhere?

 

[MongGrel]

There's always the prospect of multiple causation -- as a possibility among discrete possibilities. There's a possibility of two causes interacting to create an event or result. Sometimes, and event is so singular that you can rule it out as something random -- it has an assignable cause. Howsoever inconclusive, any rational person might look at the odds of such a thing being random when it appears not to be, and accept it as an explanation.

It's interesting that the bios update change logs highlight "power management."

Usually, if I build a new machine, I'll pick a PSU that was touted for the very last chip generation and chipset, if not for my current processor and chipset choice.

There had even emerged a problem of compatibility between PSUs and certain models of CyberPower UPS systems -- I think it was the Active PFC feature. No problem for me: but I had to redeploy the UPS and find another for the system. What's going to take precedence? A Seasonic PSU just purchased? Or a 3 year-old UPS that can be used elsewhere?

There is always the possibility that someone is just doing it wrong, after multiple fails.

Occam's Razor, etc.

 

[Torn Mind]

This all reminds me of a flaw in my building of my 2700K. In the 2600K, the PSU was fed by interior air with the fan facing up in the bottom of the case. Yet the case-bottom has an intake vent for the PSU fan. The PSU fan has never spun up to my observance, because the PSU only barely gets warm when pulling ~ 500+ watts. With the nearly identical 2700K but with SLI, I had done what would be the first inclination -- use the vent.

But it's never been filtered. The box sits on 2" double-caster wheels, maybe 2.5 to 3" off the floor. On the other hand, it's a wool carpet, in a dry and dusty part of the country. Now I'll have to decide just how soon I flip the box over and have a close look. I should be able to simply unplug the modular wires, flip the PSU, and then twist each wire in an appropriate direction and plug it back in.

Somebody mentioned the deviation of a 12V rail from the 12V spec. The 12V ATX spec is really a variance or deviation from the 12V target, or +/- 2% as an acceptable range. If the 12V rail is slightly in excess but within the 2% spec, that's a good thing. If it is less but above 11.8V, this electronical illiterate is only guessing that it may not run to spec wattage, or maybe has a barely-acceptable hold-up time. somebody correct me or elaborate if you and your fingers have the time. But that's my intuitive conclusion about it.

Some of the EVGA PSUs of recent vintage are rebranded Seasonics. A review of a 750W "gold" or "platinum" model showed that it could sustain above 900W without shutting itself down, so the model performs beyond its spec.

It is a 5% spec. 11.4 volts and 12.6 volts are the lower and upper limits of the range in which voltage is acceptable for the 12 volt rail.

 

[LTC8K6]

Yes, the tolerances are rather large for the mobo rails. Much of the time when people are worried about their voltages, they are actually well within the standard ATX design specs.
People see 11.8 volts under load and want a new power supply, when that's actually good.

But if you suspect your power supply, then it's the mfgs tolerances for that power supply that matter, if they are published. If it does not meet the mfgs tolerances, then you have a complaint.

The ATX design guides on the Wiki page are very informative.

 

[2blzd]

I talked with the EVGA tech that I've been corresponding with since this happened and he thinks it's the extension cables I got, especially since it worked for a little bit and died after a few weeks. They haven't got the test results back from the PSU yet but I never brought up I used extension cables until today, so he was like 'oh, well that was probably it' - And went on to explain in detail why..And it made sense. Well hopefully that nips things in the bud! Either way, still going to keep an eye on things.

 

[Aratahu]

I just noticed there have been 3 bios updates to my board in the past two months of me owning it. The last two bios updates were on Oct 6th and Oct 11th, 5 days apart. And they both have the same notes:

"Version 1201 - 2016/10/11

STRIX X99 GAMING BIOS 1201
1. Optimize power management
2. Improve PSU compatibility

Hi,
New user here, found this thread when looking for info on my own setup's issues. So, I've got the Strix X99 myself, bought the new system in June, and am on my third 6850k. Last one died when trying to update to bios N-2 where N=1201.
I really have no idea what I was doing wrong, so can't help but think there's something about the power management updates that relates to it. First one died just doing basic surfing or something, nothing out of the ordinary in Win 10.

I have an APC Voltage regulator which is actively bringing it down from 250v to 220v, as well as a UPS in front of the EVGA SuperNOVA 1000 G1 powersupply.

I noticed that voltages would go into the high 1.3x ranges on auto, and adaptive would always give me 1.31 or so even when asking for less. So I've locked it down to 1.26 manual now, with a 42x multi. This gives me about 42c on the Package.

Also, I dialled the screws on my H115i cooler out by a third of a turn as well, in case the pressure was somehow too high. Hopefully this will prevent a third failure. I've never had so many problems with a PC build. Old 2600k system is still ticking along nicely at 4.4ghz.

Other components include 980 SLI, Soundblaster ZXR, 32gb (8x4) G-Skill @ 3200mhz, Samsung Pro 950 / 512GB.

Edit: Adding a picture of my build.

 

[beginner99]

Edit: this doesn't make any frigging sense. How could the power supply kill my system ? One day after working perfectly for weeks it just decides to zap my cpu?

I don't understand

Stuff breaks. Was gaming, browsing on my PC. Shut it down for the night. Next day I wanted to turn on my PC and nothing happened. No LED, no fan. Nothing. Checked if cable is plugged in, PSU switch on, that the socket has power. Everything fine. googled with my laptop, tried paper-clip method and got 0 voltage. no power. Decided PSU is broke, sent it for RMA, got it returned couple hard weeks later and everything worked fine again since then. PSU was 4 years old at the time and had 5 year warranty. They also sent back my unit, i could tell due to cable bend.No idea what actually was broken.

 

[Chicken76]

They also sent back my unit, i could tell due to cable bend.No idea what actually was broken.

A fuse probably.

 

[Aratahu]

I noticed now checking voltages that the CPU System Agent Voltage on my X99 Strix is 1.328v, for my 6850k . I left this on Auto in the motherboard. Seems a bit high? I am simply using the XMP setting on my G-Skills. http://www.gskill.com/en/product/f4-3200c16q-32gtzb . Thanks.

Edit. Quick Google, found some people saying no more than 1.25 for 24/7. Damned... So, changed the voltage offset from Auto to + 0.20, to pick a more conservative number. This gives me a runtime value of 1.160V, which seems a lot more sane. Have only done 5 minute tests (cpu, memory) in XTU, but held up ok for them. I'm using my PC for Visual Studio and basically just don't want it to fry up again...

How does this look? Is that DRAM voltage of 1.34 another thing to worry about? Thanks, I'm a bit of a noob at this. (Re-edit: Ram spec page says it's tested at 1.35 so probably ok, still I'm a paranoid now).

 

[2blzd]

Stuff breaks. Was gaming, browsing on my PC. Shut it down for the night. Next day I wanted to turn on my PC and nothing happened. No LED, no fan. Nothing. Checked if cable is plugged in, PSU switch on, that the socket has power. Everything fine. googled with my laptop, tried paper-clip method and got 0 voltage. no power. Decided PSU is broke, sent it for RMA, got it returned couple hard weeks later and everything worked fine again since then. PSU was 4 years old at the time and had 5 year warranty. They also sent back my unit, i could tell due to cable bend.No idea what actually was broken.

Except everything in my scenario was brand new. A 4 year old PSU doesn't really relate to this.

 

[richaron]

Except everything in my scenario was brand new. A 4 year old PSU doesn't really relate to this.

Well at least we agree on this.

A brand new PSU is much more likely to be broken than one which has been reliable for years.

 

[2blzd]

Well at least we agree on this.

A brand new PSU is much more likely to be broken than one which has been reliable for years.

Yea no, I disagree.

 

[wilds]

Yea no, I disagree.

He has a point. There are always brand new defective PSU's out there. I also had a brand new EVGA one come DOA a few years back.

 

[richaron]

A new PSU is not like a new house mate. And it's not like a hand made apple pie.

Like all electronics they're flying off the production line by the thousands, and the result of failure is simply an inconvenience. This failure rate is balanced with costs, but it's significant compared to many things you're used too.

There have been many experienced people in this thread trying to share knowledge with this dude and they seem to be sticking their head in the sand.

Again: A brand new PSU has a much higher chance of failure than one which has been reliable for years.

 

[Jackie60]

I lost two cpus and three Asus x99 mobos. All chips fried with no apparent reason but I suspect Asus's X99 socket, not a single problem with Asrock X99 all other components including psu are the same. Asus even refused 3rd rma claiming bent pins then 6 months later sent replacement mobo after I had been running the Asrock system without issue. So now I have an Asus X99 mobo I wouldn't dare risk a chip in sitting in its anti static bag completely useless that I paid £230 for. Never again Asus!