True limits of the increased voltage tolerance of Mobile AXPs..

MatthewF01

Senior member
Mar 1, 2002
728
0
71
I have the IQHYA or whatever stepping of 2400+... Im doing 213*12 and hitting a little bit of instability. I was fine at 11.5, but Ive maxed my FSB and wanna keep pushing the mult. til I hit 2.7Ghz hehe...

I recently purchased and installed an SLK-947u on top of which I installed a YS Tech 80-mm adjustable fan... at 2556mhz Im getting temps of 53c or so... which I think is pretty damn good considering my 1.9v setting.


My question is... am I safe to pick up voltage some more? The machine cant even complete 3dmark once its so unstable at the 12x mult... I dont remember what my BIOS allows past 1.9v but i have a lot to work with.
 

VisableAssassin

Senior member
Nov 12, 2001
767
0
0
Model 10 athlons (Bartons) are rated to a max of 2.15v MAX !
but that is really friggin high man ! and on air cooling that bad boy is gonna be roasting !
 

LeChuck59

Member
May 4, 2004
42
0
0
When I first put together my oc'ed mobile sytem a couple weeks ago, I too wanted to know how far I could safely push the voltage and did a bit of research. I found that AMD's recommended maximum voltage is 1.85v for the Barton line and that generally those with air cooling didn't feel comfortable above 1.9v. I suppose though that as long as your load temps stay at or around 55C, you don't have to worry about the cpu frying, but it will reduce its life expectancy.
 

SickBeast

Lifer
Jul 21, 2000
14,377
19
81
Your system will probably overheat and refuse to boot before you manage to fry your CPU, provided that you increase your voltage in small increments. I would echo what others have said and reccomend 1.90v as your max. Beyond that there's not much point; you might get 10mhz more but you'll be running 10C hotter. Whatever you do don't be stupid and try to run at 2700mhz at 2.1v or something silly. You will fry your chip.

Just for fun, I tried running my system at 2.0v at 2700mhz...it posted fine, but I went straight into the BIOS to look at the temps to be careful. I got in there and watched the temps slowly climb. After a minute or so it became apparent that my cooling couldn't keep up with the heat being generated by my chip; the temperature just kept climbing, and this was at IDLE! Then it took my computer like 15 minutes to cool down before it would even post again, even after resetting the CMOS.

I'm now happily settled at 2500mhz at 1.825v. :)
 

VisableAssassin

Senior member
Nov 12, 2001
767
0
0
Originally posted by: LeChuck59
When I first put together my oc'ed mobile sytem a couple weeks ago, I too wanted to know how far I could safely push the voltage and did a bit of research. I found that AMD's recommended maximum voltage is 1.85v for the Barton line and that generally those with air cooling didn't feel comfortable above 1.9v. I suppose though that as long as your load temps stay at or around 55C, you don't have to worry about the cpu frying, but it will reduce its life expectancy.

according to AMDs tech doc's their max is .5v OVER stock. desktop stock is 1.65v + .5v = 2.15v :)
1.85v is what most peopel feel comfortable with, especially on air cooling.
 

Avalon

Diamond Member
Jul 16, 2001
7,571
178
106
I brought a mobile up to 1.975v on air cooling once, but that was only to see if i could get it to boot into windows @ 2.7ghz :)
I would definitely not recommend trying any higher voltage, unless you just merely want to see what the processor can do.
Going above 1.8v is generally going to reduce life expectancy.
 

Zap

Elite Member
Oct 13, 1999
22,377
7
81
Originally posted by: SickBeastThen it took my computer like 15 minutes to cool down before it would even post again, even after resetting the CMOS.
That was probably your heatsink (probably all copper) retaining heat. Your CPU could have been cranked up right away if you had changed the HSF. Or, if you can (using a tester or shorting out whatever two pins on the ATX power connector) power up your power supply and run the fan for a minute, the heatsink will be cool enough for the CPU to work again.
 

SickBeast

Lifer
Jul 21, 2000
14,377
19
81
Interesting, thanks for the tips Zap. :)

And yes, I'm using an SLK-800A and it stays hot well after my system powers down.

I've given up the insane overclocking bit tho. I've killed two CPUs, and don't want to go for three.
 

MatthewF01

Senior member
Mar 1, 2002
728
0
71
It seems like with my cooling solutions temps have been pretty reasonable so far, and Im thinking it could probably handle another voltage bump... my question is just how safe it is. I know a lot of people dont feel comfortable with above 1.9v...but thats a desktop chip, and thats THEM not feeling COMFORTABLE... I can get past that little switch in my brain that tells me to slow down, stop, turn around...


im kinda itching for 2.7ghz, but it perplexes me how people are seeing stability around 1.85v at 2500mhz and im not stable at 2556mhz on 1.9....

I guess I'll go for it and see how things turn out, ill watch the temps... Ive had desktop bartons float around 65c and Im comfortable with that.... knowing that they die at 85-90... plus my BIOS will cut it at 85c anyway.

what do you guys think.


ill have to wait til the weekend though... im about to complete my last week of high school!
 

Falloutboy

Diamond Member
Jan 2, 2003
5,916
0
76
air max I'd do is 1.9 and thats with A/C with water I'd be comfortable pushing 2.05-2.1 which I am doing now
 

SickBeast

Lifer
Jul 21, 2000
14,377
19
81
Yeah but come on you're Falloutboy...you're the resident expert at dealing with core meltdowns! Give 'er 2.3v!!! ;)
:beer:
 

Soulkeeper

Diamond Member
Nov 23, 2001
6,732
155
106
amazingly i wouldn't go over 1.75v :)
.13u is not .18u
the components responsible for carrying the current are now smaller and thus more "fragile"
even with subzero cooling i've managed to fry a few .13u athlons by using too much voltage between about 1.75v and over 2v
some people do it, but assuming you use your computer everyday, voltages over 1.75 seem kinda high for longterm operation and enjoyment of your .13u athlon product.
 

Falloutboy

Diamond Member
Jan 2, 2003
5,916
0
76
Originally posted by: Soulkeeper
amazingly i wouldn't go over 1.75v :)
.13u is not .18u
the components responsible for carrying the current are now smaller and thus more "fragile"
even with subzero cooling i've managed to fry a few .13u athlons by using too much voltage between about 1.75v and over 2v
some people do it, but assuming you use your computer everyday, voltages over 1.75 seem kinda high for longterm operation and enjoyment of your .13u athlon product.

1.75 pfffff..... I bet some crappy motherboards give that all the time. hell. 1.8 is still within 10% which is the "safe" zone according to amd.
 

Soulkeeper

Diamond Member
Nov 23, 2001
6,732
155
106
Originally posted by: Falloutboy525
Originally posted by: Soulkeeper
amazingly i wouldn't go over 1.75v :)
.13u is not .18u
the components responsible for carrying the current are now smaller and thus more "fragile"
even with subzero cooling i've managed to fry a few .13u athlons by using too much voltage between about 1.75v and over 2v
some people do it, but assuming you use your computer everyday, voltages over 1.75 seem kinda high for longterm operation and enjoyment of your .13u athlon product.

1.75 pfffff..... I bet some crappy motherboards give that all the time. hell. 1.8 is still within 10% which is the "safe" zone according to amd.


yeah i'm talking about actual voltage
cause yeah even with boards i have sometimes i have to set them as much as .1v above what i actually want to get for the vcore reading.
1.8v might not sound too crazy, but even with good cooling has been know to cause damage with time and lotsa reboots :)
But people can run whatever voltages they want and see what happens, everything is different for everyone's situation and risks i suppose.

btw, do you got a link to amd saying anything about the "safe zone" or any % numbers dealing with overvolting it's processors ??
 

Jeff7181

Lifer
Aug 21, 2002
18,368
11
81
Telling someone to go ahead and run their Barton on 2.0 volts is irresponsible. You have uninformed users visiting the forums all the time and if they see someone say 2.0 volts is safe, they'll go ahead and run their's on 2.0 volts, which I'd think most of you would agree, is not the smartest thing to do.

Also, it's not surprising that your processor is stable at 2.5 GHz on 1.85 volts, but not stable at 2.55 Ghz on 1.9 volts. More voltage doesn't necessarily equate to higher clock speed. You'll reach a point at which it takes a larger and larger voltage increase to increase the clock speed the same amount. I bet your processor is probably stable at 2.4 ghz on 1.75 volts or less too isn't it? These things aren't directly porportional to eachother. For example... here's the results I got from mine...

1.575 volts - 1.866 - 2.2 GHz
1.650 volts - 2.2 - 2.3 GHz
1.700 volts - 2.3 - 2.4 GHz
1.760 volts - 2.4 - 2.48 GHz
 

Falloutboy

Diamond Member
Jan 2, 2003
5,916
0
76
so hell according to that 2.1 is within the max although I do agree anything over 1.9 on air is stupid. I was going on my situation running a good water cooling setup
 

VisableAssassin

Senior member
Nov 12, 2001
767
0
0
Originally posted by: Falloutboy525
so hell according to that 2.1 is within the max although I do agree anything over 1.9 on air is stupid. I was going on my situation running a good water cooling setup

yeah 2.1 is....but is crazy to try..........but ill try once my s-tdx is here and modded :)
 

Soulkeeper

Diamond Member
Nov 23, 2001
6,732
155
106
to tell you the truth it don't matter if yur on air or subzero cooling
unless yur temps are over 80c (approaching the thermal limit in other words) then damage is more likely to come from the effects of vcore (snd, electron migration, "burning out", or whatever you wanna call it).

also when comparing the athlon XP .13u to the mobile .13u versions there should be little if any difference between voltage limits before damage occurs.
 

VisableAssassin

Senior member
Nov 12, 2001
767
0
0
Originally posted by: Soulkeeper
to tell you the truth it don't matter if yur on air or subzero cooling
unless yur temps are over 80c (approaching the thermal limit in other words) then damage is more likely to come from the effects of vcore (snd, electron migration, "burning out", or whatever you wanna call it).

also when comparing the athlon XP .13u to the mobile .13u versions there should be little if any difference between voltage limits before damage occurs.


electron migration is the main killer.
the only difference that I can tell is the mobile silicon is rated to like 90 or 100c MAX.
but everything else should be 100% the same
 

Soulkeeper

Diamond Member
Nov 23, 2001
6,732
155
106
yeah yur right, i just get a little more "carefull" i suppose with my voltages everytime i fry a cpu at a higher voltage i suppose :)

anyways, good luck with all the overclocks everyone , and if i can recommend anything it would be: once you reach that limit of deminishing returns from vcore increases it usually isn't worth it to go another .1 or .2v for that extra 50 to 150mhz :)
 

VisableAssassin

Senior member
Nov 12, 2001
767
0
0
Originally posted by: Soulkeeper
yeah yur right, i just get a little more "carefull" i suppose with my voltages everytime i fry a cpu at a higher voltage i suppose :)

anyways, good luck with all the overclocks everyone , and if i can recommend anything it would be: once you reach that limit of deminishing returns from vcore increases it usually isn't worth it to go another .1 or .2v for that extra 50 to 150mhz :)

carefull is good :)
but by the time I kill this chip Ill be on a A64 :)
Once I get my S-TDX and mod it Ill see if I can get 2.7ghz or 2.65 or more :) ill stop at a max of 2.0v-2.05v TRUE vcore !
 

Falloutboy

Diamond Member
Jan 2, 2003
5,916
0
76
from what i've been told about vcore is yes it reduces the opterational life of the processor. at 2.1V if cooled properly proubly on 18mounths compared to 10years. now if I had high temps and a high vocre much much quicker. but my load temps even when i'm running at a 2.1 vcore only max out at 44c and I normaly run for everyday use 1.95v which only gets to 39C under load.

but even if my chip lasts a year I'll most likely be upgrading to the next hot Overclocking chip thats under 100 bucks. I'm of the theory I rather spend 100-200 dollars (CPU every year, vid every 1.5-2years, MB when needed). upgrading my computer yearly than spend 800 and upgrade every 3 years.
 

Soulkeeper

Diamond Member
Nov 23, 2001
6,732
155
106
i've burnt up .13u athlons in like 2 or 3 months before :)
i think it took like 3 or 4 tho before i learned my lesson
and i still don't know if i've learned any restraint hehe