PooBaa's minor OC project.

[lhgpoobaa]

What i have, for those that dont know or have forgotten:
XP1800+ 'AGOGA' brown core. (unfortunately locked, with no unlocking skill or kit present) :frown:
EPOXY 8K3A+ kt333ce mobo, revision 1.2, 2812 bios.
Leadtek geforce2pro 32mb, sblive, 56k netcom modem, Pioneer 106S dvd, ricoh 12x10x32x, W.D. 800JB, 80Gb Maxtor D740X.
PC3200 cas2.5 512mb corsair set to hardest mem timings (2,2,2,5,2T,4way,8something,1T) + synchronous operation.

Programs used to test stability... UD cure for cancer, Toast, Prime95, arj archiving for IDE stability.
Stable for me = extensive operation for 4+ hours running all the above applications for at least 30min a piece with no crashes OR errors of any kind.

<b>Results so far:</b>

<b>133x11.5=1533Mhz</b>, XP1800+ in boot screen, 1536Mhz in MBM.
1.625v - stable as a brick pooshed.

<b>137x11.5=1583Mhz</b>, XP1900+ in boot screen, 1585Mhz in MBM.
1.625v - BSOD 'page fault' while copying cd 30mins after bootup.
1.650v - stable for 4+ hours.

<b>140x11.5=1610Mhz</b>, XP1900+ in boot screen, 1620Mhz in MBM.
1.650v - reboot upon win2k startup.
1.675v - BSOD 'page fault' after 2 mins of prime95.
1.700v - BSOD 'ntoskrnl fault' after 3 hours of cure for cancer.
1.725v - BSOD 'ntoskrnl fault' as soon as Toast was run.
1.750v - stable for 4+ hours.

<b>143x11.5=1645Mhz</b>, XP2000+ in boot screen, 1653Mhz in MBM.
1.750v - BSOD 'irq not less than or equal in ntoskrnl' after 5min of toast.
1.775v - BSOD 'kmode exception in win32k.sys' after 30min of prime95.
1.800v - stable for 4+ hours.

<b>145x11.5=1668Mhz</b>, XP2000+ in boot screen, 1680Mhz in MBM.
1.800v - BSOD 'unexpected nernel mode trap' win2k startup.
1.825v - BSOD 'kmode exception not handled' during background task loading.
1.850v - BSOD 'kmode exception not handled' during background task loading.
1.900v - BSOD 'kernel mode trap' after a minute of operation.

at this stage i did some further investigation, and found that for some unknown reason (probably high FSB) Motherboard monitor was crapping out on me. it would run for a random amount of time, usually no longer than a few minutes, then spit out error messages. if i got the error messages quickly and turned off MBM or suspended then resumed it was ok for a while longer... but if i didnt intercept the error messages it typically BSOD'ed on me soon after.

as this happened from 1.825v all the way to 1.900v it must be a high FSB/AGP/PCI issue. never seen it before. without MBM running the system appeared stable 1.850v. but i like to have MBM running, so 145fsb is not a viable option.

CONCLUSION.
CPU is good at lowish speeds. very low voltages at stock speeds a pleasant bonus, as is 1.75v @ 140fsb. above 140fsb (1600mhz) the cpu/voltage doesnt scale that amazingly. min acceptable voltage seems to increase in a very linear fashion.
Should i unlock it in the future, the XP2000+ speed will be quite easy to obtain with 1.85v, and possibly XP2100 speeds with 1.9 or 1.95v.
shame MBM gets stroppy at 145fsb/72agp/36pci.

unfortunately, the cpu is only "average", thus the mystical 166fsb (1900Mhz) with the 1/5 divider kicking in is pretty much impossible, unless i go cryogenic or unlock :lol:

currently running at a sweet spot of 140fsb XP1900+ speed @ 1.75v.
that should do me until im able to pick up a 2000Mhz+, 166fsb barton later this year.

<b>P4 'Wilty' Celleron 1.7, 128Mb PC100 Cas3 SDRAM, 5400rpm HDD, Integrated everything. YUM!<b>

 

[phsstpok]

How are your temps, room, case, and CPU? What are you using for an HSF?

<b>I have so many cookies I now have a FAT problem!</b>

 

[lhgpoobaa]

Opps sowwie

Room temp was between 16 and 20C for all testing.
mobo temps were typically 3-4C above room temp.
CPU temps varied from 50 to 65C depending on room temp, task and fanspeed.

HSF is a MCX-462 with arctic silver2 and a 3600rpm 42cfm pabst fan on a rheostat
2900rpm -> 3600rpm.

cpu temps look high, but thats normal for the internal diode. even with the CPU @ 1533mhz & 1.625v the full load temp was around 50-54C.

at no time during testing did i get a overheating related lockup. with the higher overclocks the fan was set to full.

allthough at one stage when running toast i reached 65C and found out that the overheat shutoff mechanism A. works well and B. responds to the internal XP temp sensor.
it was fluctuating between 63 and 64C, then suddenly zing. within a split second it was completley powered down. as my cpu is naturally hot at full load i increased the overheat shutoff temp to 75C.




<b>MegaHertz Matters! ... But you need a chipset with BALLS to back it up!!! </b>

 

[Scotty35]

That’s a very good effort and a well-written report!

No way am I trying to steal ya thunda dude, but just for interests sake I have shown what my system is at.

My systems sweet spot with the Vcore setting at the lowest the A7v266-E allows in jumper free mode:

Cpu XP2100+ @ 1.78Vcore (overvolting)
Volcano7+ on lowest fan speed 3500Rpm.
141FSB x 13 = 1836Mhz = XP????
Corsair 512MB PC3200(2.6v)
Hardest mem timings(checking)

Temps: room 24C
CPU: 57C @ 100%-UD
System: 27C.

I have not tried going any further, mainly because I do not want to or need to right now, I am extremely happy with it. But I will, I can't help myself, is there a support group for over-clockers who are relapsing ?


<A HREF="http://www.btvillarin.com/phpBB/index.php" target="_new">A better place to be</A>

 

[lhgpoobaa]

got me scratchin my as to why MBM is being so contankerous though! never done that before. quite strange, specially when i can run prime95 flawlessly.
145fsb isnt THAT excessive...

<b>MegaHertz Matters! ... But you need a chipset with BALLS to back it up!!! </b>

 

[Scotty35]

Could be a voltage setting in MBM set to shutdown, or dare I say it, a weird Epoxy voltage thing.

<A HREF="http://www.btvillarin.com/phpBB/index.php" target="_new">A better place to be</A>

 

[lhgpoobaa]

dunno. get sharing violations and error strings. tonite ill bootup and see if i can record some of the info popping up.
bit hard though when the system does a BSOD within 5 seconds of these messages appearing lol

<b>MegaHertz Matters! ... But you need a chipset with BALLS to back it up!!! </b>

 

[Scotty35]

What about in your W2K event viewer? It should list a few possibilities.

<A HREF="http://www.btvillarin.com/phpBB/index.php" target="_new">A better place to be</A>

 

[lhgpoobaa]

errr... yeah. usually it just says rebooted from bugcheck in blahblahblah garbage.
but im goin home now. will get the bugcheck info in a few hours.

<b>MegaHertz Matters! ... But you need a chipset with BALLS to back it up!!! </b>

 

[lhgpoobaa]

here we go.
another MBM error i get is 'MBM core - Privalleged instruction'

and the following dump message

"The computer has rebooted from a bugcheck. The bugcheck was: 0x0000000a (0x554005cc, 0x00000002, 0x00000000, 0x8053589b). Microsoft Windows 2000 [v15.2195]. A dump was saved in: C:\WINNT\Minidump\Mini081802-06.dmp. "


<b>MegaHertz Matters! ... But you need a chipset with BALLS to back it up!!! </b>

 

[Scotty35]

whats your ram voltage?

<A HREF="http://www.btvillarin.com/phpBB/index.php" target="_new">A better place to be</A>

 

[lhgpoobaa]

ram voltage was fine.
for all these OC attempts i was running synchronous, and my mem can do 166fsb with hardest timings @ 2.6v... so 145fsb with the same timings should be a walk in the park by comparison.


<b>MegaHertz Matters! ... But not without Cache our a decent chipset!!! </b>