o/c'ing Mobile 2500+

Need advice regarding my cpu:

I'm trying to o/c my 2500+ barton mobile to as high as possible (2.4-2.6), but I keep getting errors on Prime95 and I don't know what to do.....

I've tried the following so far:

210x11.5 @ 1.7V - Large FFT <1 min ; Small FFT 9 min (i stopped it) ; Blend <2 min

220x11 @ 1.725V - Large <10 sec ; Small I stopped at 4 min ; Blend < 1 sec

200x11 @ 1.65V Very stable...ran all 3 tests and no errors within 10 min....nothing has crashed or anything in past 3 weeks...

My question is: since Large and Blend tests use Ram, would that be the problem? I have Corsair 2x512Mb 500Mhz TwinX on my Abit Nf7-S rev 2...running at 8-3-3-3 and at default 2.6V...

I've read that the higher the FSb, the more the memory could be at fault...but I have 500Mhz ram! shouldn't it be able to handle 220 fsb???!!

I haven't changed any other settings besides fsb, cpu multiplier or the VCore. Changing the memory timings to 7-4-4-3, 6-3-3-2.5 or 11-3-3-3 results in Blend and Large tests failing within 10 sec....

I someone recommends increasing ram voltage then: how much? and how do I know when it's too much (don't have anything to monitor ram temp or chipset...).


Here is some more info on my system:

1) using MBM:
a) Core 0 = 1.63V
b) Core 1 = 2.61V
c) +5 = 4.95V
d) +12 = 11.8V
e) -12 = -12.19V
f) -5 = -5.24V
g) 3.3 = 3.31V

cpu = 2500+ Mobile Barton @ 2.424 Ghz
Mb = Abit Nf7-s rev 2
ram: 2x512Mb Corsair 500 Mhz
HD : 2x 36Gb Raptors
1x 40 Gb
hs: sp-97 with 92mm at 48 cfm
video: Radeon 9800 Pro
ps: Enermax 550 W
case fans: 2 rear blowing out
1 top blowing out
1 side blowing in
(ps: 92mm blowing out)
*Case fans are still being set up - have post in cooling section

temp of cpu: idle = 35 celsius
load = 30.5 celsius
(AC is on - case temp is 25 celsius)

as for the cpu stepping......don't know.......can't check because it's installed and don't know how to install/remove cpu/heatsink and don't want to with these kind of parts installed - blew a ps already...........

Hope this helps some more to be able to help me some more....

To me looks like you need to give it more juice, however if you havent already I would start from the beginning:

Step 1. Set everything default, run prime.

If success: see step 2.
If fail: troubleshoot normally, check/replace power supply, ect..

Step 2. Keep Oc'ing the CPU (NOT FSB) with multi only in @100MHZ increments and running Prime after to determine stable OC.

(from my experience with prime, if your comp reboots itself while running its usally power problem and if it just errors usually heat/ram related)

Step 3. Once you've reached your highest CPU OC, start on the ram/FSB.

Step 4. Once you've deteremined highest FSB with DEFAULT ram timing, start tweakin the timings/voltage.

Step 5. Run benchies and marvel at all your hard work.

It seems this chip is stable at 200 fsb and any multiplier between 11 and 12....above 200 fsb...............

Highest i've gone: 22x11 @ 1.8v and ram at 2.7V with 8-3-3-3
No luck....Small FFT test passes...but Large FFT and Blend fail within 1-30 seconds....which makes me think it's something with the ram, but don't know what....it was suggested by someone i try cas 2.5 which i will do next....also to increase ram voltage to 2.8V (default is 2.6).

The best stable setting i could get so far:

200x12 @ 1.8V . ram @ 2.7V 8-3-3-3 Load=46 celsius

Also, I don't think it would be the power as i have the 550w Enermax......