Safe 24/7 Voltage for 2500K Overclocked

[Thebrave85]

Hey guys I contacted Intel Customer Support because I read that 1.52v is the MAX Safe voltage for 24/7 usage without degrading the chip and I wanted to ask them what the dela was with it. I currently have an i5 2500K @ 4.7Ghz overclocked at 1.45v stable with c3/c6 enabled. Idles at 30c and under 30c most of the time with a CM Hyper Master 212+ cpu cooler. I ran P95 for 10 hours last night and the highest temp I got was 69c on one of the clocks after all that time without any errors. I contacted Intel and what they said was at 1.52v or under it is safe for 24/7 useage. Here is some of the conversation I had with them. I talked to two different Intel Representatives:


talked to two representatives via Intel's Onine Chat Support Service. I am going to do it again now actually lol.

Adolfo: Hello. Thank you for using the Intel Customer Chat Support service. We are glad to be of service. How may I help you?
Adolfo: Keep in mind that 72.6 degrees Celsius is the maximum recommended temperature for the CPU, not per core, it is totally normal to see the cores running a lot higher temperatures
Adolfo: I would recommend using thermal monitoring software that allows you to check the CPU temperature and not the cores temperature
mf2385@gmail.com: Ok so that is pretty normal then because I idle at pretty good temperatures, I think atleast, in the high 20c and low 30c.
Adolfo: that is perfect temperature
Adolfo: since 77 degrees Celsius is the temperature per core that you are reporting, it is OK

Richard: Hello. Thank you for using the Intel Customer Chat Support service. We are glad to be of service. How may I help you?
mf2385@gmail.com: Just wanted to find out since I do have a 2500K series i5 processor, the max safe voltage for 24/7 use? Is that 1.52v?
Richard: Thank you for the information, just read the short description you provided us.
Richard: Sounds perfectly fine.
Richard: The memory controller is embedded on the processor and is very sensitive to high voltage. anything higher that that could dame the processor.

mf2385@gmail.com: But where I'm at now at around 1.45v now is safe for 24/7 usage?
mf2385@gmail.com: I am trying to stay well below that max
mf2385@gmail.com: not touching 1.5v
Richard: Great practice. While you maintain the voltages under those numbers there should be no problem at all.

I actually recently lowered the vcore to 1.425v level 5 load line calibration and perfectly stable after 2 around 10 hour tests on the latest Prime95 without any errors and highest temp on one of the cores was 69c.

 

[Lex Luger]

Even though intel says 1.52 is safe max, the real true safe is more like 1.425

If you are using more voltage than that, you shouldnt run the cpu any hotter than 70c.

The truth is no one knows for sure what the safe max for these chip is, but Ill tell you this,

if you have cpu at 72.6 degrees and were running 1.52 under load, it would degrade fairly quickly.

 

[Thebrave85]

yeah I hear you but I don't even get 70c under 9 hours of Prime95, and I am actually at what you had said 1.425v. Level 5 LLC. 69c max prime95 almost 10 hours of testing with no oerrors.

 

[Thebrave85]

Well I've talked to a bunch of Intel Customer Support and the only thing for a fact that is known is that you should stay under or at 1.52v for 24/7 use.

 

[cmdrdredd]

You might be able to step down the voltage some. Worth a try.

 

[E6700]

alright....now i can run 4.8 at 1.42v with confident

 

[RussianSensation]

Why would Intel's safe 24/7 100% load CPU voltage for 32nm Sandy Bridge be higher than any of their chips in the last 5 years?

1) 32nm Gulftown CPUs = 1.375V
2) 45nm Nehalem CPUs = 1.375V
3) 45nm Lynnfield CPUs = 1.400V
4) 45nm Wolfdale CPUs = 1.3625V
5) 65nm Kentsfield CPUs = 1.500V
6) 65nm Conroe CPUs = 1.500V

It sounds to me like what I think they meant that you can run your memory at around 1.5-1.52V with your Sandy Bridge processor.

Plenty of people have reported failed Sandy Bridge CPUs at 1.5V, and even at much lower 1.45-1.50V voltages. You can PM aigomorla.

I am pretty sure most people use Gulftown as the most accurate yardstick, in which case going above 1.400V is no longer safe.

 

[996GT2]

Why would Intel's safe 24/7 100% load CPU voltage for 32nm Sandy Bridge be higher than any of their chips in the last 5 years?

1) 32nm Gulftown CPUs = 1.375V
2) 45nm Nehalem CPUs = 1.375V
3) 45nm Lynnfield CPUs = 1.400V
4) 45nm Wolfdale CPUs = 1.3625V
5) 65nm Kentsfield CPUs = 1.500V
6) 65nm Conroe CPUs = 1.500V

It sounds to me like what I think they meant that you can run your memory at around 1.5-1.52V with your Sandy Bridge processor.

Plenty of people have reported failed Sandy Bridge CPUs at 1.5V, and even at much lower 1.45-1.50V voltages. You can PM aigomorla.

I am pretty sure most people use Gulftown as the most accurate yardstick, in which case going above 1.45V is no longer safe.

~1.4V is generally fine for a 24/7 overclock on Sandy Bridge, as long as you have proper CPU/VRM cooling and a good motherboard with beefy power circuitry. Most people (enthusiasts included) won't even run their CPUs at full load more than 10-20% of the time, so with Speedstep and C1E enabled the CPU is usually sitting at ~1 GHz and ~1.00V anyway.

I would definitely not run 1.52V for 24/7 use, though.

Btw, any links to your claim that people have had SB chips die on them at just 1.45V? I can understand damage at >1.5V, but I've yet to see someone have a K series SB chip be damaged from 1.45V. Perhaps the person was going by the voltage reported in Windows, when in reality the motherboard was actually delivering much more voltage to the chip?

 

[MrTransistorm]

I really wish people would stop looking at the max VID and assume that it is the max Vcore.

I also take online chats with Intel CS with a load of salt. Someone did the same thing back when SB was released and got different answers from different reps. Until I see an official statement in writing that >1.4V is safe, I'm not believing it.

I am pretty sure most people use Gulftown as the most accurate yardstick, in which case going above 1.400V is no longer safe.

This. Considering that the official max Vcore (Vcc) for Gulftown is 1.4V, I think it's pretty reasonable to assume that it holds true for SB as well.

 

[Acanthus]

Why would Intel's safe 24/7 100% load CPU voltage for 32nm Sandy Bridge be higher than any of their chips in the last 5 years?

1) 32nm Gulftown CPUs = 1.375V
2) 45nm Nehalem CPUs = 1.375V
3) 45nm Lynnfield CPUs = 1.400V
4) 45nm Wolfdale CPUs = 1.3625V
5) 65nm Kentsfield CPUs = 1.500V
6) 65nm Conroe CPUs = 1.500V

It sounds to me like what I think they meant that you can run your memory at around 1.5-1.52V with your Sandy Bridge processor.

Plenty of people have reported failed Sandy Bridge CPUs at 1.5V, and even at much lower 1.45-1.50V voltages. You can PM aigomorla.

I am pretty sure most people use Gulftown as the most accurate yardstick, in which case going above 1.400V is no longer safe.

I think most of the dead CPUs were due to high PLL voltages. (IIRC)

 

[Binky]

Something seems fishy with your idle temps. A "high 20s low 30s" idle temp at 1.45v and 4.7ghz doesn't sound right (unless your room is less than 20c).

 

[MrTransistorm]

Something seems fishy with your idle temps. A "high 20s low 30s" idle temp at 1.45v and 4.7ghz doesn't sound right (unless your room is less than 20c).

It doesn't sound fishy to me. My 2600K usually idles less than 10 degrees above ambient. Right now with the room temp around 20C, it idles at less than 30C. Power saving features work wonders.

 

[Binky]

It doesn't sound fishy to me. My 2600K usually idles less than 10 degrees above ambient. Right now with the room temp around 20C, it idles at less than 30C. Power saving features work wonders.

Your comment doesn't necessarily disagree with mine. Also, the system in your sig appears to be running at less than 1.45v and it's watercooled.

 

[MrTransistorm]

Your comment doesn't necessarily disagree with mine. Also, the system in your sig appears to be running at less than 1.45v and it's watercooled.

It's due to the power saving features (C1e, EIST, etc.). When it's at idle, it downclocks to 1.6 GHz w/Vcore @ ~1 V. It's consuming very little power at that point, so it could easily be only a little above ambient.

The same story held when I went on a 5 GHz hell run. The load temps obviously increased, but the idle temps stayed the same because it still downclocked to 1.6 GHz.

The watercooling makes very little difference at idle. It had pretty much the same idle temp delta when it was on air (Thermalright HR-02).

EDIT: I hope you don't think that I was trying to argue with you. You just seemed perplexed that such low idle temps were possible at higher OC's. There's no doubt that the idle temps would be higher if I did disable power saving features and used a fixed Vcore, but there is little reason to do so on SB. IME stability is not much better than with power saving enabled.

 

[MadScientist]

Hey guys I contacted Intel Customer Support because I read that 1.52v is the MAX Safe voltage for 24/7 usage without degrading the chip

Where did you read this?

An excerpt from an article at TH: http://www.tomshardware.com/reviews/automatic-overclock-motherboard-cpu,3048.html

"We've learned through trial, error, and dead processors that voltage levels beyond 1.45 V at above-ambient temperatures can kill an Intel CPU etched at 32 nm (Sandy Bridge-based parts included) very quickly. Those same processors die a fairly slow death at voltage levels between 1.40 V and 1.45 V (somewhere between weeks and months on our test benches). And we're expecting more than a year of reliable service from the parts we've dutifully kept below 1.40 V. Not all motherboards are perfect however. Voltage instability on a particularly cheap motherboard fried one of our processors when it was set to only1.38 V. Subsequently, you've seen us use 1.35 V for the overclocking tests in older motherboard round-ups, embracing 1.38 V to 1.40 V in more recent pieces covering higher-end platforms."