• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

P4 Overclocking Worsening?

X

xcelr8

Member
I built a computer over the summer with a P4 1.6a overclocked to 2.4ghz on an epox 4g4a+ with the voltage set at 1.55V. This system was perfectly stable and running fine from summer through December 20th. This is the day that my school let out for winter break. The computer stayed, shutoff in the apartment, and when we went back a month later there were BSOD left and right with WinXP. After formatting and running Prime95, I found that the problem wasn't a corruption in windows, but the chip became unstable at that voltage. Now it must be run at 1.60V to be stable. This isn't horrible, but how could it happen, especially when the computer wasn't being run, does anybody have ideas?

Thanks.
 
Well telling you in experience the epox mobos tend to undervolt so what you have set in bios may not be what is being delivered at idle and significantly less under load. The mobo regulators could be degrading and the power supply could be degrading...

IN my experience when my northwood started to go it was and became more sensitive to higher vcore and I had to drastically lower it to near default levels to make it stable...


The epox may undervolt .05-.03v at idle meaning...

1.55v set = 1.52-1.5v actual at idle and under load could fall to 1.46-1.44v.....

1.6v set = 1.57-1.55v actual at idle and under load could fall to 1.51-1.49v.....

Anyway you look at it I would be highly shocked if it was the cpu worsening considering with known epox undervolting at 1.55v you were running this thing at default vcore and lower then most ppl with asus mobos at default bios settngs....

I think the problem may lay elsewhere....
 
Thanks for the replies. I doubt that it was a lightning strike since it is winter time here in central Illinois and we don't get that this time of year. I guess that it could be the thermal paste. To Duvie, what makes no sense is that nothing changed and the computer wasn't on at all for the month, then on coming back the chip needs an extra .05volts more than what was required for the past 6 months it was up and running. Now it's been running prime95 for about 24hours with no errors and appears stable, I just don't like the temperature increase and wonder if something like this will happen again in the future some how.
 
What do your temps and voltages look like????

What kind of stability testing did you do before and how long???


My 1.6a has been running 2.71ghz now for 7-1/2 months at 1.71v and I have not seen actibity like that....


Remember when ocing the fsb with a p4 you are stressing the memory and the motherboard chipsets....What is the memory set to in terms of speed and timings and how that compares to default specs...

Sometimes stressing prime95 for an hour may have not been long enough to get that part to fail or hiccup....


Run memtest86 and see if the ram is fine...Post temps (sys and cpu) and vcore as well as 3.3v,5v, and 12v lines....
 
i dont think its the cpu either ~ something else went wacky.
feel free to send the chip to me or Duvie ~ well test it out for ya 😉

ill also guess its the board, mem or PSU ~ not cpu.
 
When the computer was first setup the ram, samsung 2700 256mb, was set at SPD in the bios and at a speed of 374mhz, the voltage was at 2.5V. The chip was at 2.4ghz and 1.55V in the bios, I don't know what the actual voltages are/were in windows exactly, but they were very close to this. I ran prime95 for about 30 hours when the computer was first setup and it went without error. I then looped 3dmark2002 for a 24+ hour period and there were still no problems so I deemed it stable and it ran fine until it was left off for the month of break. The typical idle temperature before, and on coming back, at 1.55V and 2.4ghz was 37ºC both before and after this problem started. Now it's up to ~40ºC when idle at 1.60V. I also tried running prime95 after the problem started with the RAM voltage boosted to 2.6V, but it failed, and I lowered the RAM speed to 225mhz and it still failed. Upon raising the cpu voltage to 1.60V and setting the ram back to 2.5V @374mhz it has run prime for 24 hours now fine, so I doubt it's the RAM, is there any simple way to check if it would be the power supply?

Thanks again
 
Just report:

vcore
+3.3v
+5v
+12v

report them as you get into windows environment..ie idle

then run prime95 and wait until about at least test 16 and report them again...

We are looking for big swings in the vcore and the +12v range...

This itself will not tell us for sure as a mobo can go wacky and start varying the voltages it is regulating.


prime95 itself is not a good test for memory anyways...it is a basic test but it does not check the whole module like memtest86 does. I recommend running it at several diff settings and see if you get errors.

 
is the power supply fan still spinning at a constant speed. Some times if a power supply's fan will start to noticably change speeds then you know the PSU is on its last legs.
 
Originally posted by: Duvie
Well telling you in experience the epox mobos tend to undervolt so what you have set in bios may not be what is being delivered at idle and significantly less under load. The mobo regulators could be degrading and the power supply could be degrading...

IN my experience when my northwood started to go it was and became more sensitive to higher vcore and I had to drastically lower it to near default levels to make it stable...


The epox may undervolt .05-.03v at idle meaning...

1.55v set = 1.52-1.5v actual at idle and under load could fall to 1.46-1.44v.....

1.6v set = 1.57-1.55v actual at idle and under load could fall to 1.51-1.49v.....

Anyway you look at it I would be highly shocked if it was the cpu worsening considering with known epox undervolting at 1.55v you were running this thing at default vcore and lower then most ppl with asus mobos at default bios settngs....

I think the problem may lay elsewhere....
Click to expand...

duvies right my epox 4g4a+ posts at 1.47v but is set to 1.55v

i blew a cheap power supply early on. now i have the truepower 430. i can overclock higher but i have to go to 1.7v (setting in bios) 1.62v actual for 3200 mhz on my 2.8 processor

but when i bought it it read 1.52 in bios set at 1.55 ymmv
 
*shameless plug*This is another object lesson in why all computers should be on 24/7 crunchin' DC projects for TeAm Anandtech 😉
 
I had an Epox 4G4A+ for a few months, but finally sold it because the voltage fluctuated too much, and, like you, I could have sworn it overclocked better when I bought it than when I sold it.

Epox usually makes great overclocking boards, but I feel they kinda missed the mark this time around. I had it at 1.72 vcore in Windows, and couldn't even get 2.88 GHz 100% stable. I bought an Abit BD7-II RAID and I am now stable at 2.9 GHz with Motherboard Monitor reporting 1.62 volts on the core. I attribute this to the more stable voltage regulation of the Abit.

Anyway, I wouldn't worry about it too much, unless it keeps going, and you have to step up to 1.65. Then I would start looking for a new board.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top