cancel
Showing results for 
Search instead for 
Did you mean: 

Dead CPU Rampage V Extreme Unidentified LED

Arne_Saknussemm
Level 40
Anyone help on this?

I have just had a CPU die on me. It was a rather fine 5960X. It was a good OCer...tested but not benched since it's too hot here at the moment...

So...it was running at stock when it died. Switched off fine one evening...dreaded code 00 the next morning.

When I switched on that morning a very bright red LED lit up next to the 4pin ATX connector on the RVE...it is not listed in the manual....the PCIe lanes also lit up but not the correct ones for the occupied lanes. And indeed one of them was flasing?

Any clue from ROG Illuminati (pun intended) to help me understand what happened?

The LED next to 4pin ATX only lit up on first fatal boot attempt...has not come on since....strange PCIe light behavior has been same.

Not the board...have a new CPU in it now and all working great....

:confused:

EDIT: 29/Oct

Have replaced motherboard with 3.1 version of same; night and day difference in behaviour; there is definitely something wrong with VRM of affected board and I think it probable given the silence in this and other similar threads that the board killed the CPU by overvolting on startup

EDIT:31 August 2016

In the interests of truth/knowledge...lack of either...

Board worked OK with other CPU, though had slightly strange startup still...now not sure if board VRM was killer or FIVR on CPU was killer or combination or none of the above...

Have seen multitude of similar complaints...but all different hardware, OCing, behaviours etc etc. very difficult to pin down.

ASUS, no comment, which is not on....
40,340 Views
84 REPLIES 84

SB_NL_
Level 9
The strange thing is, I also noticed a higher CPU temperature after trying out the XMP settings. If it's related or not I can't tell.
Even when I put it back to manual, it stays higher than before XMP (same roomtemp etc).

But when I flash the BIOS with the previous version, upgrade again and apply the exact same manual setting,
the CPU temp goes back to the previous "normal" temp range.

On previous BIOS'es I also had the overvoltage warning after applying XMP but a clear BIOS helped.
I never saw the overvoltage warning again. And because I don't use XMP anymore perhaps they are somehow one and the same thing.


ps
Especially after a Failed overclock reboot, I can't do anything more than clear the BIOS boot completely to Windows.
If I go directly into the BIOS and try to select a saved OC profile the BIOS freezes.
I stopped experimenting with this board to be honest. I put it on my max save OC with acceptable temps
and left it alone.

A damn shame, because I could have bought a cheaper "scarypants" edition MOBO.

ps
Arno : Hamster eating MOBO's ROFLOL 😄 😄

'SB[NL wrote:
;547235']The strange thing is, I also noticed a higher CPU temperature after trying out the XMP settings. If it's related or not I can't tell.
Even when I put it back to manual, it stays higher than before XMP (same roomtemp etc).

But when I flash the BIOS with the previous version, upgrade again and apply the exact same manual setting,
the CPU temp goes back to the previous "normal" temp range.

On previous BIOS'es I also had the overvoltage warning after applying XMP but a clear BIOS helped.
I never saw the overvoltage warning again. And because I don't use XMP anymore perhaps they are somehow one and the same thing.


ps
Especially after a Failed overclock reboot, I can't do anything more than clear the BIOS boot completely to Windows.
If I go directly into the BIOS and try to select a saved OC profile the BIOS freezes.
I stopped experimenting with this board to be honest. I put it on my max save OC with acceptable temps
and left it alone.

A damn shame, because I could have bought a cheaper "scarypants" edition MOBO.

ps
Arno : Hamster eating MOBO's ROFLOL 😄 😄


Crumbs!...weird behaviour...any board with the overvolatge error needs replacing I reckon....so you can sleep at night

gene-iv-ocz-rev
Level 7
I thought it was just me doing something wrong but:

"Especially after a Failed overclock reboot, I can't do anything more than clear the BIOS boot completely to Windows.
If I go directly into the BIOS and try to select a saved OC profile the BIOS freezes."

After any failure I have to push the reset to clear bios or I can't get anything but black screen and then trying any previous saved OC profile hangs on.

OK once I click it.

I thought that it may take time to go thru the settings and modify all the changes I made but I had it sit for an hour while I did something else and it was still stuck on:

OK

The most annoying thing is having a raid array defined when doing overclocking because after the clear of bios it often shows a drive is not included in the array and it's failed so I have to delete and recreate the array then boot up and use testdisk to recover the partition which is more than annoying. At least I have not lost data yet from there and it is backed up anyhow.

If I learned one thing with this board. Make sure the case is close to the seat so you can reach the buttons because it's going to be a LOT of presses if you plan on doing any changes at all in the BIOS.

I do like to burn things up but it's frustrating to spend 30mins to an hour trying to get back to where the damn thing will do more than a black screen.

Anyone else here have that besides us 2?

Vlada011
Level 10
I never notice so much dead CPU as Haswell E. Mostly i7-5960X but now I see even i7-5820K.
That's very bad because I can't remember so good chance to people get really good CPU/Memory performance for less money than usually.
Almost perfect platform. Perfect will be when Skylake E show up with 48-60 PCI-E lanes Gen 3 together Chipset and CPU.
I expect Rampage Gene than... 28 lanes CPU, 20 Chipset. Layout need to be as Rampage 4 Gene, same only with M.2. 😄

gene-iv-ocz-rev
Level 7

gene-iv-ocz-revo3-240 wrote:
I'm sure this is unrelated but worth a read:

http://www.legitreviews.com/intel-x99-motherboard-goes-up-in-smoke-for-reasons-unknown_150008


I read it a long time ago 🙂 thx for sharing because it also started when he used the XMP profile although they suspected the combo of a Corsair PSU with older firmware and the VRM design of the X99 DeLuxe.

The things that remain the same in this case is changing to the XMP profile built in the G-Skill memory banks and the 5960X. (correct me if I'm wrong).
Perhaps it shows more often with a Corsair PSU but it should be worth investigating by all involved.

It would be really scary if there would be a design fault in the X99 CPU's architecture.
And seeing the design change they made removing the integrated voltage regulator from Skylake...I just wonder...

'SB[NL wrote:
;547301']I read it a long time ago 🙂 thx for sharing because it also started when he used the XMP profile although they suspected the combo of a Corsair PSU with older firmware and the VRM design of the X99 DeLuxe.

The things that remain the same in this case is changing to the XMP profile built in the G-Skill memory banks and the 5960X. (correct me if I'm wrong).
Perhaps it shows more often with a Corsair PSU but it should be worth investigating by all involved.

It would be really scary if their would be a design fault in the X99 CPU's architecture.
And seeing the design change they made removing the integrated voltage regulator from Skylake...I just wonder...


I'm far from an electrical engineer or expert in CPU micro-architecture, but I've been building and OC'ing systems for decades and I think you're on the right track. If FIVR was such a grand idea, why have they apparently abandoned it already??

Azusa
Level 7
Interesting read, but not related at least to my experience.

When my first proc failed, I first sent the RVE back to ASUS for RMA. When they returned it unchanged, I knew it was the proc that caused the failure. At that point my system had been down for 2 weeks, so I purchased a separate 5820K to use while my original was out for RMA to Intel. I swapped the new one in, system booted up fine. put the old one back in, still dead.

I happed to still have the RMA processor Intel sent me when the second one died, so I had a proc on hand. Swapped the RMA part in, same results. Two dead processors, RVE itself seems to work fine, but at this point I'm sure it's killing these $400 parts, despite ASUS returning it to me as though nothing was wrong.

Frustrating thing is I'm basically at a dead end. My board has already been through the RMA process and they said nothing was wrong, so what now? I guess I either replace the board or keep flushing $400 down the drain every 4 months, because Intel is sure to stop RMA'ing the processors at some point.

Korth
Level 14
On older Asus mobos, PLED1 is just a power LED. I would expect it to be in series with the ancient chassis PLED1 pin, they should both light up whenever the mobo is powered on. (Actual power LED, not standby power LED.)

But on the R5E it's physically adjacent to the VRM bank which services the LGA2011-3 socket. I suspect it's an onboard VRM status indicator meant to light up when the mobo is attached to Asus factory diagnostic gear. A conspicuous lack of any public documentation suggests it's not intended to assist end-users with troubleshooting. The SMT LED traces attach to a zener on the VRM side and ultimately feed (through some passives) to the magical black box ROG chip - they clearly won't light up unless a certain voltage threshold is met and a secret ROG signal is sent, so methinks PLED1 is probably meant to indicate (to a qualified Asus technician, of course) that one of the VRM components has flagged a fail condition of some sort. "DIGI+ VRM" is just fancy marketing of PWM-controlled variable voltage regulators, it all boils down to whatever firmware and PWM capabilities are running in the ROG chip (fancy Asus marketing name for embedded ASIC microcontroller).

I've yanked my ROG chips out and plugged my JTAG in, the ROG parts are of course unmarked proprietary codelocked stuff without datasheets but it's just a matter of time before I can duplicate their functions with a more robust programmable microcontroller part and, incidentally, figure out what undocumented components like PLED1 are meant to really do.

First guess is that it's showing a main VRM failure, the 8-phase DIGI+ VRM is really one digital controller running a gang of eight discrete voltage regulators wired in parallel, and any imbalance or failure in one or more of those VRM parts will put a heavier load on the other VRM parts (which can really yank them all around if one part is oscillating or has intermittent thermal failure or overcurrent runaway or something) - which throws the entire output voltage (as seen by the controller) entirely out of skew in all sorts of bad ways. Meaning that your mobo is not sensing correct voltages and is working with incorrect voltages, meaning that even if it is working within correct voltages they might momentarily surge and waver out of tolerance and bombard some or all of your poor little CPU pins with all sorts of half-uncontrolled power spikes. Hope you can RMA that mobo before you burn another CPU, cooked silicon is an ugly smell. Everything *everything* on this mobo is digitally controlled except for things like that little zener diode which is physically wired to sense an analog voltage threshold, so if PLED1 is lit then either that zener is bad or all the onboard digital information is bad - I'm inclined to say the zener is properly calibrated because if it wasn't then it could falsely light the LED but no way could it cook your processor socket.

And the sad thing here is that you've lost a killer $1000 CPU and $500 motherboard most probably because a $0.25 electrolytic failed to halt a $3 VRM from oscillating out of calibration.
"All opinions are not equal. Some are a very great deal more robust, sophisticated and well supported in logic and argument than others." - Douglas Adams

[/Korth]

Arne_Saknussemm
Level 40
Cheers Korth...nice reasoned thinking...sounds about right and fits with what happened...and with my surmising...

I have recently swapped out the board...new 3.1 version. Night and day...new board just boots from cold or restart with no fuss...old board was starting to flash 00 and pcie lane lights and then shut down before restarting...so yeah it's going back

That was a great CPU and is one of the hardest things to take....I am astronomically lucky that it's replacement is even better so makes it all easier to take...

Lots of sad stories in this thread....CPUs die it is a fact and OCing makes them die faster sometimes a lot faster...user has to take responsibility if it's thiers...plenty of stories of overvoltage that look like bad boards...bad VRMs.

Maybe bad FIVRs?....FIVR definitely a bad idea....good riddance.

Azusa wrote:
Intel is sure to stop RMA'ing the processors at some point.


For sure they will....sounds like you definitely need a new board :eek: