GA-EP45-UD3L NB heat question

[kick]

maximum nb v is supposed to be 1.25
However lots of people use 1.3-1.4v to facilitate memory/cpu communication

I can get a useable 490 cpu ref (1960 fsb) working with 1.3v mch and it seems stable for awhile,but it isnt reliable(for me).
I have a seperate temperature sensor drilled directly into the mch.nb heatsink.All i can tell u on that is ,it doesnt like to be over 42C with my setup.
As I type this its at 36.6C.nb has a 120 mm fan blowing on it(angled at 45 degrees).

http://forums.tweaktown.com/f69/q8400-q8200-overclock-ep-45-ud3-high-fsb-37179/

are settings I use and have used for a long time,however 4 core communication can put more strain on than 2 cores

[Triton05]

Re: GA-EP45-UD3L NB heat question

When I get home, Ill drop it down to 1.28v and see if it holds. If so, Ill just go with that.

I understand its over the 1.25v max, but hopefully not far enough over to result in perm. damage. Seeing as some have thiers at 1.35 - 1.4+ Ill reamin optimistic!

[profJim]

Re: GA-EP45-UD3L NB heat question

You'll make it easier for us to help you if you add your hardware setup, bios & mobo version, o/s version, etc. to your forum signature.
Most EP45 setups with 2x2GB of memory use MCH Core = 1.24 - 1.28 with moderately high overclocks. Some GSkill setups need only 1.05 volts for stability.
MCH Core that is too high or too low will cause boot loop or stability problems.
With monitoring software (ET6, HWmonitor, etc) the "System" temperature sensor is located near the North Bridge chip.
Without knowing your computer case brand/model, cooling fan setup, and room temperatures, it's hard to say if you have a problem.
MCH Core should not be set higher than 1.40 volts. Naturally, lower is better.

[Triton05]

Re: GA-EP45-UD3L NB heat question

Thank you kindly for the reply. Just updated my signature. I created the account today while at work

When I got home, I lowerd my MCH core to 1.28v. I then ran IBT for 10 passes, then ran prime small FFT's for the last hour. No issues at all.

CPU temp peaked at 65C during IBT and 59C during prime. I feel a little more comfortable having my MCH to 1.28 .. when I first clocked to 4.0Ghz I started at 1.32, so Ive backed it down a bit.

EDIT: Just booted my system again and in the BIOS for system temp its going between 40C-41C. I assume thats the temp probe near the NB sensor?

[Triton05]

GA-EP45-UD3L NB heat question

Just built this system this past weekend.

GA-EP45-UD3L rev1.1
E8400 E0 3.0Ghz @ 4.0Ghz
G.SKILL PC8500 2x2GB 5-5-5-15 memory ( 4GB total )
Arctic 7 Freezer Pro Rev2 cooler
500Gb Seagate Barracuda
GTX285 OC 1Gb video card
OCZ 600watt SXS PSU
Windows 7 Ultimate 64

Vcore: 1.35 bios with LLC enabled. 1.32 windows
MCH core: 1.30
VTT: 1.30
PLL: 1.55
DRAM: 2.04v
Strap: 2.4B ( effective memory 1066Mhz ) w/444FSB

Rock solid system. Couldn't be happier. Idle CPU temps 29C - 30C. Full load with prime is 59C and IBT 5 passes peaks around 65C.

Heres my only concern. That NB heatsink gets HOT. As in I cannot hold my finger on it longer then 3sec. I assume this is normal. But my concern is how close my memory is to it.

I can def. feel heat from the NB heatsink warming up my slot 1 memory dimm. I just want to make sure this isn't going to create any issues long term. I *would* move it to slot 2-4 instead of 1-3 but that would require me removing my heatsink and fan from my CPU which hovers directly over it. So Id rather not as long as this isn't going to be a concern. The memory is fully shielded, so Im sure that helps.


Should I be concerned, or am I good to go? Thank you kindly for any advice.

-Cheerz!

[yuu]

Re: GA-EP45-UD3L NB heat question

You are overvolting the chipset too much, My chipset runs 500 Mhz FSB with only 1.16V on MCH Core, runs less than 500 with 1.14V both at 4.00Ghz, CPU 1.325V bios no LLC.

Perhaps 1.30 on MCH could be optimised to 1.20 .

Even on 400FSB with 1.05V the chipset is way too hot, 1-2 seconds burns. no problem so far with memory in 1+3

There were long periods of time, 2-3 months actually this last year i was running intelburntest nonstop and it survived. 5 passes is like nothing.

Shielded memory is likely to overheat more, because it does exactly the opposite of what you think, it shields the heat in the chips not letting it out.

[Triton05]

Re: GA-EP45-UD3L NB heat question

Well Im not having any issues preformance wise. 3Dmark06 / Vantage / Ive run Prime95 several times, for periods of 30min doing small FFT's. No issues at all.

Your running your MCH core below 1.2v ? Everywhere Ive read, people with E8400s' giong to 4Ghz or higher had to use 1.3-1.4v. I figured I was on the low end. Ive already backed it down from 1.32 intially to its current 1.30.

I guess my main question is just the Heat from the NB saturating the memory in slot 1. The system is working better then I could ask, so Id rather not tweak anything else from this point. Unless of course Im *not* safe running my [email protected] .

Regards,

EDIT: Just to clarify, I was under the impression, running my memory with a 2.4B strap and effectivly at 1066Mhz. The 1.3v MCH was needed. As I understood, running the 2.0D strap with a lower effective memory speed, required less NB voltage. *correct me if Im wrong*

Thanks :)

[profJim]

Re: GA-EP45-UD3L NB heat question

Thank you kindly for the reply. Just updated my signature. I created the account today while at work

When I got home, I lowerd my MCH core to 1.28v. I then ran IBT for 10 passes, then ran prime small FFT's for the last hour. No issues at all.

CPU temp peaked at 65C during IBT and 59C during prime. I feel a little more comfortable having my MCH to 1.28 .. when I first clocked to 4.0Ghz I started at 1.32, so Ive backed it down a bit.

EDIT: Just booted my system again and in the BIOS for system temp its going between 40C-41C. I assume thats the temp probe near the NB sensor?

The forum time stamps have fouled up the post order in this and other threads, so it's impossible to follow the order of the posts.

Yes, the SYSTEM temperature reading is from a sensor near the North Bridge chip, but this reading is not the true NB temperature.

It's best to run any of the Prime95 tests for at least 8 - 12 hours or longer. When stress testing, I run IBT or LinX first for a relatively quick test and then follow with extended Prime95 or OCCT tests. I always run 10 loops of IBT or LinX using the maximum amount of available memory and THREADS = 2 for dual core processors with IBT. Many times IBT or LinX pass their tests only to have Prime95 or OCCT fail, sometimes within minutes or possible after 6 - 10 or more hours.

[Triton05]

Re: GA-EP45-UD3L NB heat question

*bump*

[Lsdmeasap]

Re: GA-EP45-UD3L NB heat question

For 4Ghz with 2x2Gb of memory you should only need 1.2-1.24 MCH, maybe 1.26 if you are running 500 FSB or super tight memory timings

Using D memory multi vs B memory multi is not better, that was only in the past. You should use D since you are over 400 FSB

Yes though, when using B multi's past 400 FSB you may need more MCH voltage, but since you are already concerned about heat I suggest you switch to the D strap multi's

http://forums.tweaktown.com/f58/2-00b-vs-2-00d-399-500-fsb-simple-strap-comparision-31656/

[Triton05]

Re: GA-EP45-UD3L NB heat question

Well, Im only concerned *if* 1.28v for 24/7 use is bad. I have ample air flow in my case.<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" /><o:p></o:p>
<o:p></o:p>
Im using the 2.4B strap to effectively run my memory at its rated speed. (the point of using 444 FSB) 444 FSB * 2.4B = 1065 Mhz <o:p></o:p>
<o:p></o:p>

<o:p></o:p>

<o:p></o:p>

Thanks :)

[Triton05]

Re: GA-EP45-UD3L NB heat question

Got home today from work, booted up, dropped MCH to 1.26v.

Kicked off IBT Max Stress 10 passes : PASS
Kicked off Prime95 Blend : 2 hours : PASS


Looks like Im keeping my NB @ 1.26v

[kick]

Re: GA-EP45-UD3L NB heat question

very strange.Im down as the thread starter

Nothing bad has happened to my system on 18 months of mch voltage of 1.38-1.42v . I mean allways at that voltage.
I dont leave it that high for fun, I just find it necessary.
I think i could run ok at 1.2-1.3v on a lower fsb .
Not that much lower ,say 1850-1900 but at anything over 1800 settings are very individual.

Anything over 450 cpu ref(1800 fsb) seems component specific.

[Lsdmeasap]

Re: GA-EP45-UD3L NB heat question

Nice to hear you were able to lower the voltage a bit, how's the temps now?

[Triton05]

Re: GA-EP45-UD3L NB heat question

Everything seems great.

CPU idle is around 30C, max load is 61C.

Normal operating temps for the NB is up to 1.25v so having it at 1.26v is pretty much within spec. Intel rates the NB at 1.2v for a 1600FSB.

Haven't had ONE BSOD this entire rebuild. I have the temps now to my liking. CPU at 1.32v on full load in windows. Memory running at 1068Mhz with 2.08v. Its rated for 2.0-2.1v so that too is within spec.

:) just wish my 285 didnt get so hot!! but not much I can do about that. The card in general just runs hot. Taps 89C now and again in BC2.