Sign in with
Sign up | Sign in
Your question
Closed

Sandy bridge max voltage opinions

Last response: in Overclocking
Share
a b K Overclocking
January 6, 2012 6:33:40 PM

lets hear it. what do you think the max safe voltage for a two year life span.

if you have personal experience that be best. lets try to keep this none biased and factual. and dont like 12 month old data.

personally i draw the line at 1.4 as long as the temps are under 70 degrees.

also a topic is does the max temp under stress testing matter or is it max operating temp the one that matter. so should people aim for under 70(or max good temp that you believe) degree stress testing or under normal operations?
a b K Overclocking
January 7, 2012 7:23:29 AM

you may want to read this
http://www.tomshardware.com/forum/271980-29-z68x-ud3p-i...

HostileDonut quoted
"Intel recommends 1.52v maximum for that CPU, and that is with a great cooler. I wouldn't go past 1.43v myself even if my temps are in the 60s load. 1.39v is a bit high for that OC. I run mine @ 4.4Ghz with 1.32v maximum and some people would say that is still a little high. Use a offset voltage instead of auto. At least the CPU is stable. Running 1.39v shouldn't hurt the CPU with those temps, as they are pretty good running Prime95, but you could probably get the voltage lower, making your CPU last longer"
Score
0
January 7, 2012 10:12:52 PM

What i think and whats fact is probably not lined up exactly, but... i remember my older e8600 where everybody i met and talked to said anything over 1,4v would kill my cpu fast. So i tried to have 1,45-1,5v over 3 years without turning the computer off... had i off for like all-in-all 1 week on vacation. And i use it a lot, never was a problem at all.

My point being that i'm sure the sandy's can take much more than we think. I've only gotten mine this week, but been overclocking my mates computer. I think it depends a lot on the chip, but i would say with 1,45-1,5v you still have the chip working fine for a LONG time... Not necessarily a lifetime, but longer then you would have the chip before replacing it... But this is as long as the temps are withing reason and not like 70+all the time. I run a fairly big loop on my cpu (water ofc) and are seeing 60-65* @ 1,38-1,4v on load over time. They can probably take more than we think :)  my 2 cents...
Score
0
Related resources
a b K Overclocking
January 7, 2012 11:16:00 PM

ottesen said:
What i think and whats fact is probably not lined up exactly, but... i remember my older e8600 where everybody i met and talked to said anything over 1,4v would kill my cpu fast. So i tried to have 1,45-1,5v over 3 years without turning the computer off... had i off for like all-in-all 1 week on vacation. And i use it a lot, never was a problem at all.

My point being that i'm sure the sandy's can take much more than we think. I've only gotten mine this week, but been overclocking my mates computer. I think it depends a lot on the chip, but i would say with 1,45-1,5v you still have the chip working fine for a LONG time... Not necessarily a lifetime, but longer then you would have the chip before replacing it... But this is as long as the temps are withing reason and not like 70+all the time. I run a fairly big loop on my cpu (water ofc) and are seeing 60-65* @ 1,38-1,4v on load over time. They can probably take more than we think :)  my 2 cents...


judging by those voltages your at around a 4.7 overclock?
Score
0
January 8, 2012 1:00:27 AM

cbrunnem said:
judging by those voltages your at around a 4.7 overclock?


My mobo is faulty so i can't change voltage on mine now, so i'm running on "auto" which gives me 1,37-1,39v @ only 4,5ghz... My mobo is so faulty so i don't dare to restart my computer...haha. Will be talking to the guys where i bought it in a few days.

On the other 2500k i clocked we have 4,5ghz @ 1,29v rock stable. Did some easy benchmarking at 5ghz with 1,39v :)  Seems like a good chip :) 
Score
0
a b K Overclocking
January 8, 2012 2:30:27 AM

ottesen said:
My mobo is faulty so i can't change voltage on mine now, so i'm running on "auto" which gives me 1,37-1,39v @ only 4,5ghz... My mobo is so faulty so i don't dare to restart my computer...haha. Will be talking to the guys where i bought it in a few days.

On the other 2500k i clocked we have 4,5ghz @ 1,29v rock stable. Did some easy benchmarking at 5ghz with 1,39v :)  Seems like a good chip :) 


it seems that i5's overclock with a lower voltage cause of the lack of hyperthreading....
Score
0

Best solution

a c 224 K Overclocking
January 8, 2012 5:51:55 AM

If someone is actually expecting to get 2 years out of their Sandy using it 24/7 I'd say stay well below 1.40v preferably below 1.35v should be very safe, most 2500Ks will run 4500mhz in the 1.300v ~ 1.330v range.

Presently 4500mhz will do whatever you need it to do, so there's really not much reason to clock it any higher once you satisfy the curiosity of reaching the higher clocks.

Of course it also depends on how hot your CPU is running 24/7 at 4500mhz with the voltage it takes to achieve stability, the cooler the better so that's also an important factor.

It also depends on whether you're running stress tests all the time as some out here seem to think that's a daily duty, stress tests are to reach a stability point then stop running them as they are shortening the life of your CPU.

If CPUs had a MTBF rating in hours then what I just wrote would be easier to understand, if say for instance a cooling fan is rated MTBF 50,000hrs @ 12v but you decide to run yours at 14v, do you think the fan will still last 50,000hrs.

Well you could actually say that you have 26,280 hours before your warranty runs
out, but some on auto voltage are running higher voltages than those of us manually overclocking, hmmm, could that be a problem.

Curious that your M/B warranty runs out before your CPU warranty does and your M/B is what's actually regulating that extra auto voltage,and cooking up your CPU!

Well you did ask for opinions. :) 

Everything on this planet has an end, the key is how fast will you reach it?

Are voltages left on auto good or bad there's something to argue about, IMO when voltages are set manually they're fixed at a certain level, voltages on auto fluctuate sometimes going to extreme levels to keep the computer from crashing and catch up to what's going on.

It may seem the safest route and easiest to accept and be pleased with but sudden electrical surges are not good no matter how low the voltages are, especially when some of their peaks are way past what you'd like to safely use daily.

Some motherboards run auto voltages way higher than needed, but they never designed the things to run forever if they did you wouldn't need to buy another, and they'd go out of business.
Share
a b å Intel
a c 197 K Overclocking
January 11, 2012 12:07:19 PM

I agree with ryan. My reasoning:
Max recommended voltage for 65 nm CPU's: 1.50 volts.
Max recommended voltage for 45nm CPU's: 1.45 volts.
Max recommended voltage for 32 nm CPU's: I would not go higher than 1.40 volts.
Score
0
a b K Overclocking
January 11, 2012 1:34:06 PM

Best answer selected by cbrunnem.
Score
0
August 2, 2013 10:43:30 PM

I have been running my i7-2600k for about 1 1/3 years now at 4.6ghz and I recently am in the market to buy an Ivy Bridge Extreme edition chip and thought I would give it a bump up to 5ghz o say I can. I ran 4.6 @1.32v for over a year and then changed the digi vrm to a mediocre setting at first to see where that would take me. After bumping it up to 4.8ghz at 1.392v and a stable IntelBurnTest through 50 runs. I decided that I would not take it any higher. I can run 9 out of 10 runs on the Intel Burn Test at 5ghz at those voltages but its not rock solid stable. Plus the temps got into the eighties, SO I dialed it back to 4.8 again and am sitting stable at 1.392, So I would have to agree with the rest of the folks that 1.4 should be the limiting voltages for this chip. Im also using a z77 ASUS sabretooth mobo which I think may help in the stability aspect of things. But I have ran this benching technique at 1.45v as well as 1.48v @5.0ghz and have found that the chips starts to degredate after certain voltahes leaving some testing out of the question. PRIME runs were fine at these voltages running heat into low eighties for a 4 hour run with no BSOD's but were not stable on Intel Burn Tests runs. I also think that the RAM may have been the limiting factor in those benchmarks as well. I found my Kingston Hyper X Genesis 1600mhz modules did not like the stock 1.65v and preferred someplace in the middle. I settled on 1.575v. But Again. I would have to agree that for a 24/7 rock solid stable voltage on this chip is gonna have to be around 1.375-1.399v @ 4.8ghz and that is also using hyperthreading. Which if I turned it off. I dropped 10c off the Intel Burn Test temps as well as dropping the voltages down to a place where you could get 5.0ghz stable and still into the low to mid 70's. Message me if you would like some details on how I got there or if you have any comments about this post
Score
0
!