What is my oc limitation...

[Duvie]

Here are the specs...the rest are in my rig setup...

1.8a northwood w/ 4000rpm AVC fan from www.pcpowercooling.com
P4s333 sis645 w/ 2 stcks of kingmax pc2700ddr
sparkle 350watt PS

1.8a@2.4ghz 133fsb w/ 4:5 ratio for 333mhz ddr...rock stable at 1.65v prime tested and app tested...
temps:
idle (bios after 10min) 36c
full load (hours into prime) 45c
voltage
idle 12v = 11.91v
full 12v = 11.77v
vcore set in bios 1.65v actual from asus probe 1.67v
load vcore =- 1.62-1.65v


1.8a@2.43ghz 135fsb w/ 4:5 ratio for 240mhz ddr...rock stable at 1.675v prime tested
temps:
idle (bios) 37c
full load 47c
voltage
idle 12v = 11.91-11.84 fluctuates...
full 12v = 11.71v
vcore set in bios to 1.675v actual from asus probe 1.7-1.68v
load 1.63v-1.67v

1.8@2.45ghz 136fsb w/ 4:5 ratio...so far stable at 1.7v...not at 1.675 (fails in 1 hour in prime)
temps:
idle 37-38c
full load 49c
volatge 11.91-11.77 fluctuates
full vcore = 11.71v didn't see any lower but fluctuated more


can't get 2.5ghz stable in prime at 1.725v (high temp 51c)...fails 1.7v in minutes, though it never gets up to 48c...does run apps and bm's fine...though in sandra cpu i can get a fluctuation in scores in terms of 150-200 pts...

at highest voltage i am able to set 1.8v it will post all the way up to 143fsb...haven't run test though...


questions....

Does it look like my limitation is:

A) power supply do to the 12v rail dropping and fluctuating wildly at higher voltages??

B)Heat???

C) mobo chipset...chipset temps have never been higher then 26c....


I have never had a freeze up or a BSOD...just it will get an error in prime testing...

 

[WarCon]

What is your ddr voltage at? Have you tried playing with memory settings? It might very well be the 12volt rail on your supply. I also have noticed increased fluctuation in the 12v rail, when I am pushing it. But mine is running stable all the way down to 11.43 on the 12v rail, so its really hard to blame that for sure.

I know that the supply I am going to order better have a great 12v rail rating. With Athlon's it was the 5v rail and with these its the 12volt rail.

I don't think any of the above is an answer, just thoughts and more data to accumulate into a theory.

 

[SSGTi]

Well I'm at the magical 2.5ghz mark. Using the stock hsf combo w/ artic silver 3. I'm using a 300w enermax ps. I think the problem sounds like your power supply. On my MSi 845 ARU (great mobo btw) according to mbm my voltage never drops below 12.0v . Usally it's around 12.1v.

I havn't tried any higher yet (still breaking in the artic silver), but I think I should be able to reach 2.6ghz with some more voltage and cooling. My temps at full load though are kind high at 60c with prime 95. But it's totally stable (tested w/ prime for 3 hours)

my ddr and agp voltages are stock. anyone know advantages to increasing them ?.

 

[Duvie]

I am sure it is not the memory as:

1) I have ran it at lower fsb with ratio that made the memory at 380mhz ddr and it ran fine...

2) I have also ran it at 1:1 settings so the memory at 133 would be underclocked to pc2100 standards

3) I have not tweaked memory at all...left it to cas 2.5t


I may also look at the 300watt enermax...

I have a sparkle and I hear about adjustable "pots"...What is this?? How do I know if I have this capability???

 

[Duvie]

bump...any more thoughts....

 

[SSGTi]

Well try with 1 stick of ram... or have you tried that already ?. Generally more sticks leads to less stability. If that and a new ps don't help, then it's your chip .

 

[Duvie]

I have my ram in 2 sticks...1 in the 1st slot and the other in the 3rd slot...could this be issue??? I did it for cooling reasons and the manual said I could occupy any of the 2 slots...

I have also ran it at a 1:1 ratio and still had errors so I don't think it is related to ddr voltage, settings, or number of sticks....

I thik it is either power supply (12v rail in particular), chipset stabilty (wheter heat or what not), or basic limitations of the chip....


I can boot this at 2.52ghz at 1.65v and run all sort of apps....but I can't find a stabile voltage up to 1.725v just to get it to pass prime95 stability test...At 1.725v it will hit 52c and fail within 45-60 minutes....

 

[o1die]

If your board permits it, try increasing the memory voltage by +.2. It works better than raising the cpu vcore. I also use heat spreaders. Fry's had them for only $5.95 this weekend.

 

[exp]

<< I thik it is either power supply (12v rail in particular), chipset stabilty (wheter heat or what not), or basic limitations of the chip....et >>

I agree. It is almost assuredly not your chipset or RAM. If you can't get any higher with a new power supply then it's probably your chip. 1.8A's often top out at 133MHz FSB so those kind of results are pretty typical.