The AMD Phenom II OC'ers Club

[crazyeyesreaper]

you need to keep tweaking the system remember to set the NB and HT multis down 1 from there default to give you breathing room its about finding balance it took me a LONG time to find it in my system keep working with the voltage try to stay under 1.3 volt NB 1.3volts CPU- NB and 1.5volts CPU it will take time but its neccesary if you want the top speed you can get with less heat its all trial and error

 

[Thor2020]

you need to keep tweaking the system remember to set the NB and HT multis down 1 from there default

Could you explain a little bit more this part.. perhaps provide an example? Thanks

 

[Chicken Patty]

thor, I think we are getting a bit ahead of ourselves here.

To me the easiest way to find stability is set everything at their default clocks and then try raising the multiplier one notch at a time. If prime or LinX fails then you can always then you can go back into the BIOS and raise CPU vcore slightly until it passes. By doing it this way you are only overclocking the CPU and nothing else, not the RAM, not the NB.

This method to me works great because you overclock only one thing at a time and you know what to change when the stress test does not pass. Once you get the CPU where you want it then you can go back and start doing the same with the NB or the RAM.

 

[Chicken Patty]

Guys, I've been bothered by my temps lately. They are in the mid 50's and I'm only running 1.392v which is stock. I know the rig crunches non stop but still. I'm going to try and reseat the heatsink later using less TIM. Any particular way you guys apply the TIM? I just put some in the middle and spread it evenly across the surface. I think I'm just using too much of it.

 

[BlackOmega]

you can drop the voltage down to 1.425-1.45 1.5volts is way to high for that speed friend ive the Same CPU you do and 240FSB 1.425 volts to the cpu got it stable no issues even managed to get 2400mhz NB with a few voltage adjustments and i suggest using LinX much quicker to find system stability usually it seems in my travels if theres an error to be found it will be found either at LinX pass 2 pass 9 or pass 12 so do a 15 run LinX stress test at max ram to be sure

Remember every CPU is different. 2 identical CPUs' with the same steppings and everything will overclock differently. Batch number plays a big role in this as well.

 

[crazyeyesreaper]

depends on the TIM i use noctua NTH1 as its non conductive etc i use 2 small dots between the heat pipes and i mean small dots they fill in he creases then i use the usual method of TIM + card to apply a ting layer over the top that way i get maximum coverage with no air pockets

 

[BlackOmega]

Guys, I've been bothered by my temps lately. They are in the mid 50's and I'm only running 1.392v which is stock. I know the rig crunches non stop but still. I'm going to try and reseat the heatsink later using less TIM. Any particular way you guys apply the TIM? I just put some in the middle and spread it evenly across the surface. I think I'm just using too much of it.

For you, just a small pea sized drop in the very center should be enough.

Watch this

How TIM spreads.

 

[PaulieG]

Guys, I've been bothered by my temps lately. They are in the mid 50's and I'm only running 1.392v which is stock. I know the rig crunches non stop but still. I'm going to try and reseat the heatsink later using less TIM. Any particular way you guys apply the TIM? I just put some in the middle and spread it evenly across the surface. I think I'm just using too much of it.

Are you running push/pull with your Mega? As far as TIM goes, I put a small 1/2 pea size amount in the middle, then 4 tiny dots closer to the corners of the heatspreader.

 

[erocker]

I just cover the entire IHS with a thin layer of TIM and a very tiny dot in the middle. Never fails.

 

[BlackOmega]

Obviously you guys haven't watched that video I linked to.

 

[Chicken Patty]

Looks like the pea method works best IMO.

Paul, I am using just one pull fan. I use one pull in order to accomodate my RAM cooler.

Thanks for the vide omega

 

[erocker]

Obviously you guys haven't watched that video I linked to.

No, I'm going by what the instructions for my heatkiller says. That video is also useless for HDT coolers.

 

[crazyeyesreaper]

exactly HDT is difference it wont spread that way do to the gaps

 

[Thor2020]

thor, I think we are getting a bit ahead of ourselves here.

To me the easiest way to find stability is set everything at their default clocks and then try raising the multiplier one notch at a time. If prime or LinX fails then you can always then you can go back into the BIOS and raise CPU vcore slightly until it passes. By doing it this way you are only overclocking the CPU and nothing else, not the RAM, not the NB.

This method to me works great because you overclock only one thing at a time and you know what to change when the stress test does not pass. Once you get the CPU where you want it then you can go back and start doing the same with the NB or the RAM.

Thanks for the advise and will do that.. As you stated to start with increasing the multiplier first. If i understand correctly the concept once i maxmize cpu mhz w/ stable vcore, i can then start tweaking the RAM tighness through the FSB, then I can work on the NB and HT.

 

[Chicken Patty]

Thanks for the advise and will do that.. As you stated to start with increasing the multiplier first. If i understand correctly the concept once i maxmize cpu mhz w/ stable vcore, i can then start tweaking the RAM tighness through the FSB, then I can work on the NB and HT.

Yeah, the key here is trying to do one thing at a time so you can know what is causing the instability if any. The NB has it's own multiplier so you can raise that one by one and stress test after each increase just like you would do with the CPU, you get what I'm trying to say?

 

[Wile E]

No, I'm going by what the instructions for my heatkiller says. That video is also useless for HDT coolers.

Plus, the dot method only works on a mostly flat IHS. Both my 6400+ and my QX9650 have non-flat IHS's. The spreading method works best for me.

Not to mention, look how much tim they used in the spread method. The amount used has as much to do with it as the method used.

 

[crazyeyesreaper]

all i know is i hope i can come to a deal with theonedub on that 965 so i can try overclocking with a different CPU

 

[Thor2020]

Yeah, the key here is trying to do one thing at a time so you can know what is causing the instability if any. The NB has it's own multiplier so you can raise that one by one and stress test after each increase just like you would do with the CPU, you get what I'm trying to say?

Absolutely, one last question: i'm writing all of my default settings, speeds, multiplier and voltage so that I know where to start from. In cpu-z i'm thrown off my core voltage saying it's 1.152 but it's more like 1.38-1.39 cpuid hardware monitor.

 

[Chicken Patty]

Absolutely, one last question: i'm writing all of my default settings, speeds, multiplier and voltage so that I know where to start from. In cpu-z i'm thrown off my core voltage saying it's 1.152 but it's more like 1.38-1.39 cpuid hardware monitor.

I believe the 940's default voltage is 1.336v give or take. You can just set it at 1.35v to start if you want.

 

[jjFarking]

Most specs can be gleaned from AMD's website HERE.
Interestingly, regarding voltages, they simply specify the voltage ranges, not including the default voltage..

 

[Chicken Patty]

Most specs can be gleaned from AMD's website HERE.
Interestingly, regarding voltages, they simply specify the voltage ranges, not including the default voltage..

That's why I told him to just start off a 1.35v. Pretty close to default and not high at all. Seems like a good starting point to me.

 

[crazyeyesreaper]

default voltage for the 940BE is 1.35volts on both my asrock and gigabyte 790gx boards

 

[BlackOmega]

Looks like the pea method works best IMO.

Paul, I am using just one pull fan. I use one pull in order to accomodate my RAM cooler.

Thanks for the vide omega

No problem.

No, I'm going by what the instructions for my heatkiller says. That video is also useless for HDT coolers.

It's always best to follow manufacturers instructions (usually).

Regarding HDT coolers, it actually depends on the cooler. On a Xiggy S1283/DK the 2 line method on the HS works best. But on a Sunbeam Core Contact, the pea method still works best.

exactly HDT is difference it wont spread that way do to the gaps

Please see above with your crazy eyes

Plus, the dot method only works on a mostly flat IHS. Both my 6400+ and my QX9650 have non-flat IHS's. The spreading method works best for me.

Not to mention, look how much tim they used in the spread method. The amount used has as much to do with it as the method used.

If you're not using an HDT cooler the pea method usually works best. But the key is, you have to apply enough force to make it spread. This especially holds true for the thicker TIM's like ICD7, OCZ Freeze, AS5, etc.

I was part of the test bed for IC Diamond 7 at another forum and that was the most frequent problem when using it. I believe the guy from IC said that almost 80 lbs. of force is required to get it to spread when it's room temp. Putting in a cup of boiling hot water for 5 minutes helps it become more viscous.

Also, from my experience with AMD's, all of their IHS are concave.

 

[crazyeyesreaper]

well im not going to argue pea method i used when first setting up my system resulted in the Xigmatek posting WORSE temps then the stock cooler using my method i top at 50'c stock clocked in a room that hits 95'F + so ill stick to my method as it works for what i need and im not using the typical thermal compounds other ppl use i use Noctua NTH1 so i dont have to worry about force it spreads very easily and without issue

 

[BlackOmega]

well im not going to argue pea method i used when first setting up my system resulted in the Xigmatek posting WORSE temps then the stock cooler using my method i top at 50'c stock clocked in a room that hits 95'F + so ill stick to my method as it works for what i need and im not using the typical thermal compounds other ppl use i use Noctua NTH1 so i dont have to worry about force it spreads very easily and without issue

I wholeheartedly agree about the Xigmateks. On my Dark Knight I use the 2 line method. Like this:

I too tried the pea method with this cooler and it does not cool well. This method works the best.