Voltage too high? (Poll)

[Abzstrak]

Have increased the voltage on my Barton core to 1.825 to take it to 2500MHz... running at 36C Idle, 43C under load... Temps look pretty good to me for air cooling, whatchya guys think?

does anyone have any GOOD evidence (links are good) as to the effects of running a higher voltage over time? I see people running 2+V on these and it just seems like somehow the gates have to be strained by this over time.



EDIT-
btw, this isn't a new barton, I've had it for some time now (7 months or so) running 2300MHz at 1.7V.
the more I think about it, this voltage bump is only 10.6% above default, I'm gonna leave it and hope it all works out good.

 

[myocardia]

Well, I don't ever run mine over 1.775v, but I've seen people who say they run theirs as high as 1.85v, but I wouldn't, personally. Most of us "old-timers" don't recommend going over 1.75v with a Barton (mine is a T-bred B). That seems to be a very safe voltage, but hey, it's your chip man! If you don't plan on keeping it more than a year or so, I'm sure it'll run at 1.80v for that long.

 

[jjyiz28]

most i'd do is 1.75. but since you getting some monster oc results, and temps arnt that high, thats kool. i wouldn't push it any more though

 

[JustStarting]

1.85 is fine for me. just watch temps.

 

[Corporate Thug]

i dont run anything higher than 1.75V for bartons and 1.65V for northwoods

 

[myocardia]

Originally posted by: JustStarting
1.85 is fine for me. just watch temps.

You don't have a Barton, though. Supposedly, the Bartons are a bit more susceptible to overvolting than the T-breds.

 

[Abzstrak]

just saw I miss wrote it, its 1.825, not 1.8... edited

 

[Abzstrak]

Originally posted by: JustStarting
1.85 is fine for me. just watch temps.

yea, kinda what I'm figuring, but it only gets up to 43C after an 1+ hours of prime 95...

 

[cowdog]

I run my oc'd Barton 2500+ at 1.75 most of the time (for over 9 months) but have been running it at 1.8 lately to squeeze out a little extra (11.5 x 200). I don't really notice that much temp difference from 1.75 to 1.8 - 1-2 deg C. I have a "spare" 2100+ just in case though.

 

[Abzstrak]

Originally posted by: myocardia
Well, I don't ever run mine over 1.775v, but I've seen people who say they run theirs as high as 1.85v, but I wouldn't, personally. Most of us "old-timers" don't recommend going over 1.75v with a Barton (mine is a T-bred B). That seems to be a very safe voltage, but hey, it's your chip man! If you don't plan on keeping it more than a year or so, I'm sure it'll run at 1.80v for that long.

Any reason you think it will decrease its lifespan? as long as its stable and cool electron migration won't occur which should prevent alternate pathway's from being burned into the chip right?

 

[slaves123]

running that high effectibly decrease lifespan and has nothing to do with the temperature, but if it working and you like to take the risk no problem

 

[jjyiz28]

Originally posted by: Abzstrak

Originally posted by: myocardia
Well, I don't ever run mine over 1.775v, but I've seen people who say they run theirs as high as 1.85v, but I wouldn't, personally. Most of us "old-timers" don't recommend going over 1.75v with a Barton (mine is a T-bred B). That seems to be a very safe voltage, but hey, it's your chip man! If you don't plan on keeping it more than a year or so, I'm sure it'll run at 1.80v for that long.

Any reason you think it will decrease its lifespan? as long as its stable and cool electron migration won't occur which should prevent alternate pathway's from being burned into the chip right?

i may be wrong on this, but runnign higher voltage introduces "micro hot spots" from within the cpu core even though the cpu is running very cool as a whole. i think i read about this for northwood sudden death syndrome, same may apply here.

 

[Abzstrak]

Originally posted by: slaves123
running that high effectibly decrease lifespan and has nothing to do with the temperature, but if it working and you like to take the risk no problem

but why? I haven't had a logical desription as to why the lifespan would be decreased... I'm really wanting to know, but just telling me it does isn't helping... I've had a number of people tell me this I want to know if its true, but I'm skeptical by nature and require a reason.

 

[Abzstrak]

Originally posted by: jjyiz28

Originally posted by: Abzstrak

Originally posted by: myocardia
Well, I don't ever run mine over 1.775v, but I've seen people who say they run theirs as high as 1.85v, but I wouldn't, personally. Most of us "old-timers" don't recommend going over 1.75v with a Barton (mine is a T-bred B). That seems to be a very safe voltage, but hey, it's your chip man! If you don't plan on keeping it more than a year or so, I'm sure it'll run at 1.80v for that long.

Any reason you think it will decrease its lifespan? as long as its stable and cool electron migration won't occur which should prevent alternate pathway's from being burned into the chip right?

i may be wrong on this, but runnign higher voltage introduces "micro hot spots" from within the cpu core even though the cpu is running very cool as a whole. i think i read about this for northwood sudden death syndrome, same may apply here.

Interesting, going to go look for more info on these "micro hot spots"

 

[Duvie]

Anytime you run the chip out of the manufacturers suggested vcore range you are shortening the lifespan of the chip versus a chip at manufacturers spec....For the most part those manufacturer's specs likely is 5-10 times the length you would even own the chip.......

I haven't heard it called micro hot spts, but search for a post by PM and engineer in the field of chip design.....

From my understand the higher voltage starts leaving electron deposits on the tiny gates and over time can burn the gates out...If this happens any thread trying to access registries in those certain areas may results in errors....


Callin it Sudden Death Syndrome has always been incorrect with the northwood chips...Most experience loss of overclocking prowess that may eventually lead to errors even at default speeds..however I have seen some backed down to spec levels and the chip continued to run fine to this day......

True sudden death is like the guys who pushed the tbreds over 2v and they upped and flat out died in weeks.....

 

[Abzstrak]

Originally posted by: Duvie
Anytime you run the chip out of the manufacturers suggested vcore range you are shortening the lifespan of the chip versus a chip at manufacturers spec....For the most part those manufacturer's specs likely is 5-10 times the length you would even own the chip.......

I haven't heard it called micro hot spts, but search for a post by PM and engineer in the field of chip design.....

From my understand the higher voltage starts leaving electron deposits on the tiny gates and over time can burn the gates out...If this happens any thread trying to access registries in those certain areas may results in errors....


Callin it Sudden Death Syndrome has always been incorrect with the northwood chips...Most experience loss of overclocking prowess that may eventually lead to errors even at default speeds..however I have seen some backed down to spec levels and the chip continued to run fine to this day......

True sudden death is like the guys who pushed the tbreds over 2v and they upped and flat out died in weeks.....

Been looking for a bit and haven't found anything about micro hot spots...
also, electron's can't deposit on a gate... thats bad physics man, it has to go somewhere. The primary problem as I understand it is due to electron migration from one gate/path to another. This would happen due to a higher potential between two paths rather than down the proper path, which should only happen due to an increase in the resistance of the path. Since resistance doesn't spontaneuosly change in transistors the only thing changing it is the temperature since higher heat = higher resistance, thus my thought that keeping it cool should prevent the migration.

I'm gonna post something over in HT and see what they think... that is if they think its highly technical enough ;-)

 

[Duvie]

Originally posted by: Abzstrak

Originally posted by: Duvie
Anytime you run the chip out of the manufacturers suggested vcore range you are shortening the lifespan of the chip versus a chip at manufacturers spec....For the most part those manufacturer's specs likely is 5-10 times the length you would even own the chip.......

I haven't heard it called micro hot spts, but search for a post by PM and engineer in the field of chip design.....

From my understand the higher voltage starts leaving electron deposits on the tiny gates and over time can burn the gates out...If this happens any thread trying to access registries in those certain areas may results in errors....


Callin it Sudden Death Syndrome has always been incorrect with the northwood chips...Most experience loss of overclocking prowess that may eventually lead to errors even at default speeds..however I have seen some backed down to spec levels and the chip continued to run fine to this day......

True sudden death is like the guys who pushed the tbreds over 2v and they upped and flat out died in weeks.....

Been looking for a bit and haven't found anything about micro hot spots...
also, electron's can't deposit on a gate... thats bad physics man, it has to go somewhere. The primary problem as I understand it is due to electron migration from one gate/path to another. This would happen due to a higher potential between two paths rather than down the proper path, which should only happen due to an increase in the resistance of the path. Since resistance doesn't spontaneuosly change in transistors the only thing changing it is the temperature since higher heat = higher resistance, thus my thought that keeping it cool should prevent the migration.

I'm gonna post something over in HT and see what they think... that is if they think its highly technical enough ;-)

Actually electron migration is what I was thinking about...sorry I am not an electrical engineer....

Try this link...It is about the myth of Burn in but was written by PM....



NMOS hot-electron gate-impact ionization

I will try to find one on the electron migation....


Also I know the heat theory has to be BS cause some of the big losers in northwoods dieing on them were water cooling and more exoctic cooling users who pushed the vcore up to 1.85-1.9v and had load temps below my idle cpu temps....So heat wasn't the issue there unless there is a case of spots reaching much higher temps in very micro areas....


Also soulkeeper who killed many an AMD chip had decent cooling with temps reported i the normal range and he still killed them in rather short time....

 

[Duvie]

Try this one as well as it mentions hot spots...Also why as we go down in fabrication size the fragileness of the tinier components may come even in more play....

Xbit labs on SNDS

I wonder if the prescott has so much leakage does that mean OCing may be much out of the question and any vcore much above default may really stress it like using 1.8v on the northwoods??


Edit: the above may be a good reason why the temps being reported are not all that reliable since they are some many micro areas in a chips design where heat may build up or react differently....Then the question come to what area is the thermal diode reading from...

 

[Gravity]

Dude, keep pushing the envelope. Perhaps the core's are slightly more stable now. Run it that way for a while then give it a nudget. Let's see where the envelope REALLY is.

 

[Ionizer86]

I say it's a bit high considering how I run my Tbred B at 1.525 and my bud has his Barton at 2200 with 1.60V. But it's your chip, and if you upgrade a lot and don't plan on passing the chip down or selling it, then go for it.

 

[CarpeDiem99]

that voltage for a barton is fine by my standards
Unless your running 1.9 or higher, it should last you quite a while since its only .2V more than stock

 

[joe2004]

Up to 1.85V is fine for Barton on air, assuming the cooling is adequate, say Thermalright 700 or 800 with a decent fan.
Still I doubt you'll see 2500 MHz.

 

[User1001]

I believe increasing the voltage causes the circuit pathways to come toghether faster and this destroys the board. I don't own a Barton but 1.9 seems way to high. Adding .2v is a lot.

 

[hytek369]

just fine, but you need a vantec tornado, lol

 

[joe2004]

If I would have warranty to run it 6 months on 2.0V on air I would do that, no question asked. Who cares about $80 processor now when 64s are coming down in price?