CPU Voltage basics.

Mozno1

Mozno1

Mozno1

Associate
Joined
16 Apr 2010
Posts
17
My apologies to Admin just realised i posted in the wrong place :rolleyes: please delete the post with the same title in the hardware general forum.

Hi all,

I was wondering if anyone could spare me a minute to answer a quick couple questions about CPU voltages. I am new to overclocking so you might have to stick with me.
The system I have is listed in my sig.
I have a lot of different voltage settings in my BIOS:
VCore
CPU VTT
CPU PLL
NB and SB
+1.2v HT
and of course memory.
As I overclock the CPU do I need to raise these all these voltages slightly (as in if I raise Vcore all others must be raised to keep up), or is there only certain voltages I need to concern myself with as I raise the CPU speed? I am also presuming that raising the memory voltage is not required when talking about CPU overclocking (I have my memory unlinked form the CPU) is this correct?
Please note that I am not really asking "what should I set my voltage settings to" but more "what is the basic principal behind raising my voltages".
I hope you guys can help me out because I'm loosing the will to live.

Cheers guys.
 
Thank you for your response.
That's pretty much what I have done, raised my FSB to 1533 which is giving me 3.4 but if I go up another level (say FSB up to 1573) things start to go unstable.
When I am thinking of raising the voltage to stable things out should I raise everything a little or just certain settings? Also when I do raise voltages how much should I raise them by 0.1v or maybe 0.01v I'm really not sure.
Any more help out there guys?

Thanks again.
 
Here lies your problem imo. Asus Striker 2 Formula

Thats a 780i chipset. NVIDIA chipset (especially x80i series) + Quad (especially 45nm) = FAIL. :(

A nice P45 based board would probably net you 400 x 9 for 3.6 GHz by just upping CPU voltage a bit.

A Q9650 should do 400 x 9 with ease, IF you can get the board to do 400 FSB, you shouldnt need to consider any of the other CPU voltages.

Try concentrating on just the CPU and NB voltages, your best bet, would be to lock the multi at its lowest (6 i think?) and put cpu voltage on stock, then keep upping the FSB and bumping the NB volts to try and get the FSB stable, then when you do get the FSB stable where you want it, slowly up the multi, and bump the cpu voltage as you go, so you get an overall stable system.

That said, you could have already reached the limit for the board. I couldnt get an NVIDIA chipset to handle a Q6600 at stock stable :(
 
OK I will try that.
So I'm going to try setting my multiplier in at x6 then set my FSB setting to 1600 and up the north bridge voltage slowly until I manage to get it stable. Once I have that turn my multiplier back up one notch at a time also gradually up-ing the VCore until stable again.
That's a totally different way of doing things than I was looking at, hope I get some joy. I had a Q6600 in this board and managed 3.0GHz without even touching any voltages!? Strange.
Thanks for your help alexp999.
Anyone out there had / heard about / got any info on the Asus Striker Formula / 780i chipset?

Thanks again.
 
Mozno1 said:
OK I will try that.
So I'm going to try setting my multiplier in at x6 then set my FSB setting to 1600 and up the north bridge voltage slowly until I manage to get it stable. Once I have that turn my multiplier back up one notch at a time also gradually up-ing the VCore until stable again.
That's a totally different way of doing things than I was looking at, hope I get some joy. I had a Q6600 in this board and managed 3.0GHz without even touching any voltages!? Strange.
Thanks for your help alexp999.
Anyone out there had / heard about / got any info on the Asus Striker Formula / 780i chipset?

Thanks again.
Click to expand...

You got it. Alternatively, what I would try is raising the fsb incrementally (with a low cpu multi) to find what the limits of your board are before trying to overclock it. That way you know just how hard you have to push it and can avoid frying your NB. Once you've established your max fsb, you can then set about clocking your cpu knowing that instability is likely to be caused by the cpu (and remedy this with more vcore). Something to consider here, is a very useful post by Big Wayne in this thread explaining the relationship between multipliers and the Northbridge clock.

http://forums.overclockers.co.uk/showthread.php?t=18143931

his posts begin at #15 and he explains a bit more later in the thread. Very interesting stuff for understanding how the northbridge works.
 
Hi again,

Thanks for the update i will take a look at the thread you suggested later.
I'm busy at the moment trying to get the FSB stable at 400, not having a lot of luck at the moment though.
I have raised the voltage on the north bridge in 0.02 increments form 1.32v to 1.58v currently. I have tested with Prime95 (Blend test) after each increment and had either failure on one or two workers or in some cases a nice BSOD.
One other thing I would mention is I have disabled things like Intel Speed Step Tech and enabled Load line Calibration (reduces VDroop).
I will let you know later if I ever get it stable.
How high do you think I should / can take the NB voltage safely? This board has little LED indicators which show green / amber / red as you raise the voltages, as of yet the NB voltage is still green, which I presume means that it is well in the safe limits!?

*edit* - OK so I have now tested in increments of 0.02v all the way up to 1.62v at which point I got errors at POST and loading in to windows. Does this mean that this board truly is rubbish and will never even achieve a measly 400MHz bus? It seems odd however that just messing about I have achieved a FSB of 1533 giving me CPU speed of 3.45Ghz without even touching the NB voltage all I did was raise the Vcore form 1.3v to 1.34v. Help guys I'm stuck.

PS - What you guys think to this? I was thinking of maybe investing in the Blitz Formula, only trouble is its a P35 chipset instead of P45.

http://hothardware.com/Articles/Asus_Blitz_Formula_and_Extreme_P35_Motherboards/
 
Last edited:
I'd suggest the P5Q SE if you're on a budget, or the deluxe if you want a really good board (easy 500MHz fsb here :D).
http://www.overclockers.co.uk/productlist.php?groupid=701&catid=5&subid=1155

I'm wondering, leading on from Big Wayne's posts, if the lower cpu multi is causing the NB to work harder than you need it to. Might be worth dropping him a line as he seems to know what he's talking about (beyond the basic principles). Personally I wouldn't want to push much more voltage through that NB, sometimes you just hit a wall. Just as an experiment to test Big Wayne's info, make sure your RAM is on a 1:1 divider, and then crack your cpu multi right up to its highest (native) value, and start raising fsb and see if you get any results that way. In theory, the lower cpu multi raises the actual NB clock proportionately, increasing bandwidth, but also causes it to run hotter and faster which may be limiting your overclock. I'm not sure if that relates to your highest fsb or if I'm interpreting it wrong, but read through that thread I posted and see if you think it's worth trying.
 
Hmmm OK.
I don't understand why I would want my memory linked to the FSB in a 1:1 ratio? Surely this would lead to the memory being overclocked and then I would be unsure if it was a memory or cpu overclock causing instability?

Cheers.
 
Well after all this time of trying I am still only at a FSB of 1573. I have this running nice and stable now but as soon as I try to step up to 1600 *BOOM* the wall of overclocker death. :eek:
Anyone any idea at all what might be the issue?
I have raised both NB and Vcore voltages and no joy. From the voltages listed above anyone think there may be a voltage I am missing.
 
You must log in or register to reply here.
Back
Top Bottom