Sandy Bridge 2500k Just Died

[DrSteve]

Now I'm a bit worried.
Just ordered the 2500k with the MSI board and XMS3 that was on offer which is rated at 1.65v. Going to be running with water cooling but I'm now pretty nervous about clocking it!

 

[SimonMaltby]

Am I going to be OK using standard Corsair XMS stuff then?

OcUK shows the voltage for this memory as 1.6-1.7v but I've seen it listed as 1.65v elsewhere. 1.5v rated stuff seems to cost a load more.

Well I just don't know. Both my kits are rated at 1.6v - 1.7v on the Ocuk site, but SPD returns 1.65v. I was running mine at 1.65v in the Bios using the SPD values. Asus say, in the manual, that 1.65v is ok, but not to go any higher.

Remember, my issue looks quite different to the one this thread is about. I have an actual CPU reported error and that could have been caused my many things including just having a bad CPU. So far mine looks to be unique.

 

[Shetlander]

Im a bit worried now. I ordered the Patriot 2 Sector 5 2400mhz mem for my build with a voltage of 1.65v. Hopefully it will run slower with less volts as its going in an Asus p8p67 pro board and after reading this thread it doesnt inspire confidence in the setup.

 

[Vertigo]

Well since the vast majority of the 1600Mhz stuff is rated at 1.65v and the basic XMS3 kit is on the Asus compatibility list, surely it must be ok at this voltage?

I really don't need any more than 4GB so the cheap XMS3 kit was perfect but I'm now worried that running at 1.65v is "too close to the wire" and maybe I should get some 1.5v, even though it costs a load more.

EDIT: Just seen this on the Corsair site for their Vengeance stuff:

"1.5V VDIMM spec ensures compatibility with 2nd generation (Sandy Bridge) Intel Core processor family"

That's a bit worrying - suggests anything more than 1.5v doesn't "ensure compatibility"

 

[Skyrocket]

Calm down guys.
There is nothing to suggest that 1.65 on memory damages the chip.
Damage is more likely when ppl push the vcore too much.

Memory controller is exactly the same as 1156 CPUs.
Which are 1.5 up to 1.65

 

[Vertigo]

Well I don't know too much about this but isn't the problem something to do with the on-die memory controller voltage not being more than a certain amount (0.5v) above the main "core" voltage of the processor, or else damage can result?

As the SB processors are 32nm as opposed to the 45nm of Lynnfield, wouldn't this core voltage be lower?

 

[exotic444]

I was thinking of buying a i5 2500 and giving it a small overclock since its not the K series, would I be better off getting the i5 and just overclocking it. or are the SB cpu's reliable?

 

[Gibbo]

Hi there

I've spoke to Intel this morning and got the following answers from them:-

1. Anything over 1.40v on air is risky, upto 1.50v on water should be OK but risky.
2. Anything over 1.50v on memory is not recommended but should not kill a CPU.
3. Clocking on BCLK is highly un-recommended due to onboard graphics.

So the biggest factory they say is vcore, pushing above 1.40v is a risk, therefor I'd say don't go beyond 1.35v. BCLK clocking is a no no for sure.

I've asked for documents regarding 1.50v as motherboards manufacturers seem to think 1.65v is the safe limit, though Intel are saying 1.50v is more advisable.

I am going to update all our CPU descriptions with this information.

Anyone with 1.6-1.7v memory should run the memory at 1.50v as in theory it will run fine at those voltage as all memory manufacturers are now on new IC's.

 

[lollyhayes]

This thread is getting linked from various places, but is there actually other reports of this? Only thread I can find on "2500k died" or variations is this one.

Also "was running a Vcore within the spec, well within the spec at 4.8ghz" sounds a bit clandestine. What vCore were you actually running?

 

[DrSteve]

Hi there

I've spoke to Intel this morning and got the following answers from them:-

1. Anything over 1.40v on air is risky, upto 1.50v on water should be OK but risky.
2. Anything over 1.50v on memory is not recommended but should not kill a CPU.
3. Clocking on BCLK is highly un-recommended due to onboard graphics.

So the biggest factory they say is vcore, pushing above 1.40v is a risk, therefor I'd say don't go beyond 1.35v. BCLK clocking is a no no for sure.

I've asked for documents regarding 1.50v as motherboards manufacturers seem to think 1.65v is the safe limit, though Intel are saying 1.50v is more advisable.

I am going to update all our CPU descriptions with this information.

Anyone with 1.6-1.7v memory should run the memory at 1.50v as in theory it will run fine at those voltage as all memory manufacturers are now on new IC's.

Thanks for that.
What about you pre overclocked bundles? Were they set at 1.65v with the corasir RAM?

 

[Apem]

There is another thread about a SB chip dying on here this is rather worrying glad i held off on my order now.

 

[Rofflay]

Gibbo there was a new ripjawx memory range launched this morning that runs at 1.5v 2133mhz.Its a 4GB kit and its the only one of the range that runs at 1.5v 2133mhz.the rest all run at 1.65v which i find very odd considering intel dont reccomend that.

The only difference is timings and i hope you stock these in and add them as an option to the 4.6GHZ bundles.

Impressed with the service i know ive moaned at price before and i will in future but at least someone is getting us some info on the subject as i wont touch SB until we find out why 2 cpu's 2 days old have died.

 

[Gibbo]

Hi there

In our own testing we have found the only thing killing the CPU's is core voltage. We've found setting memory voltage at 1.65v seems to be doing no harm, but were now gonna setup 3 rigs, one with 1.45v core volts, one with BCLK overclocking and one with 1.65 memory clock to see which one if any dies first. We shall do this on 2500k chips.

 

[DrSteve]

Hi there

In our own testing we have found the only thing killing the CPU's is core voltage. We've found setting memory voltage at 1.65v seems to be doing no harm, but were now gonna setup 3 rigs, one with 1.45v core volts, one with BCLK overclocking and one with 1.65 memory clock to see which one if any dies first. We shall do this on 2500k chips.

Sounds like a fun experiment actuallly.
Are you going to provide live coverage?

 

[Gibbo]

Hi there

Another update:-

Intel recommend 1.50v for all their I5 and I7 processors, not just Sandybridge, which makes us further believe the culprit for dead CPU's exceeding 1.35-1.40v range on vcore.

 

[Vertigo]

Thanks for that Gibbo - does clarify things.

Might be worth putting a note against all your 1.6v+ memory saying it's not compatibly with SB at that voltage.

I'm now trying to decide whether to leave my order for the XMS3 stuff and just run it at 1.5v/1333Mhz or splash out on an 8GB 1.5v/1600Mhz kit.

Also, given the comment on the Corsair site I mentioned above, I've posted on their forums asking for their take on this issue.

 

[the wopster]

Hi there

In our own testing we have found the only thing killing the CPU's is core voltage. We've found setting memory voltage at 1.65v seems to be doing no harm, but were now gonna setup 3 rigs, one with 1.45v core volts, one with BCLK overclocking and one with 1.65 memory clock to see which one if any dies first. We shall do this on 2500k chips.

now thats a company trying to help its customers out,not many of the so called other net companys would do that

 

[DrSteve]

Thanks for that Gibbo - does clarify things.

Might be worth putting a note against all your 1.6v+ memory saying it's not compatibly with SB at that voltage.

I'm now trying to decide whether to leave my order for the XMS3 stuff and just run it at 1.5v/1333Mhz or splash out on an 8GB 1.5v/1600Mhz kit.

Also, given the comment on the Corsair site I mentioned above, I've posted on their forums asking for their take on this issue.

Same here. I have the XMS3 on order and may swap in for something else just in case.

The GeIL stuff looks like a good choice now

 

[lollyhayes]

guys, I have got the cpu back again albeit at 1.38vcore. I think the advice Gibbo is giving is good, and I have booted my ram (rated 1.65) at 1.5v with no ill effects... I am also running it at 1.6v now but only because it rules out the ram as a source of the problem whilst I prime95 the CPU...
I think maybe this thread should be re labelled as it now gives a false interpretation of the current status of the CPU.
So far there has only been one failure (in another thread of this forum) which was the IMC, my problems are liekly simply to be due to my overvolting @ 4.8ghz. Which whilst they lasted for a short duration may have imapacted on the subsequent performance of the processor... Who knows...
I will post a screenie of Vcore, temps and prime95 results in a few hours.

 

[bulldog147]

I just been contact Intel customer support and they say both i5 and i7 2500 and 2600 chip must NOT go over 1.380v in the bios and the memory volt must NOT go over 1.65v and that's the safer one!