Booting problems with ABIT AN7 (post code 9F)

[tortilio]

Hello everybody, I'm pretty new to the forum, and I'm 'desperately' looking for an answer before going definitively mad!!
I also have to confess I'm not so expert about hardware, but I believe I could read and do whatever I might hear!

Well... I'm trying tu run a once working Abit AN7 (Socket A), with a once working AMD Athlon XP 2600+, but whenever I try to turn the system on, it automatically shuts down after a very few seconds (before booting or even loading bios) with the motherboard uttering an alternate sound similar to an emergency siren (although sometimes it not even reaches the time to utter such sound, self shutting down earlier).
The last post codes shown by the motherboard are the following:

9a
98
99 (that's what it shows when it shuts down)
9E (and after a handful of seconds from the shutting down, the 99 turns for a very short while in what appears to be a 9E...)
9F (...than shifting to a permanent 9F)

The story is pretty long, but these are the clue facts: the motherboard once correctly worked with an Athlon XP 3000+, then I tried to put such CPU on a different (previously checked for compatibility) ASRock K7VT4-4X motherboard without success: power correctly reaching the mobo, but no video signal, nor sounds... Therefore I put the CPU back on the Abit AN7, but the system did not work any more (even here no video signal nor motherboard sounds when powering on)!?
I then tried putting on the AN7 another CPU (an Athlon XP 2600+, previously finely used on the ASRock mobo), with it working just a 'little better': the system couldn't work with default bios settings loaded - usually self resetting when updating the DMI pool or earlier - but could go with fail safe configuration, and could correctly load Windows in safe mode. After several hardware and bios tests and attempts, I eventually opted for upgrading the bios to the latest version (719), and everything started working correctly.
At this point (damn myself!!!) I tried to put back on the motherboard its previous CPU (the Athlon Xp 3000+), with it still not working (no video signal). I put the Athlon XP 2600+ having correctly run just a few minutes before back on, and the system started behaving like described at the top!?!

I think I tried about everything: different graphic cards (with some, the system shuts down even more quickly), different RAM blocks (no difference), disconnecting evry drive, disconnecting and then reconnecting the power supply, clearing the CMOS... the CPU and cooling fan are correctly mounted and running... And I currently have no way to make any attempt to downgrade the bios...

Another 'funny' fact is that after unsuccesfully trying the 3000+ CPU on the ASRock mobo, not even the previous 2600+ worked any more on it (though behaving a little more ordinarily: video signal and usual starting beeps, the self power down)!

My questions are: what might ever be wrong? The fact that the system was working with an upgraded bios, before trying to mount the 3000+ CPU makes me suspect it cannot be a matter of hardware... But how to go back??

I feel pretty discouraged and perplexed, and - as a matter of fact - I now have two previously woking systems currently out of order!

Should you need any additional information, just let me know, and please forgive my inexperience with forums, the long post (and moderate experience with hardwrae too...)!

THANX

[Atomic Rooster]

Have you made sure that you plugged in the AUX ATX12V plug to the motherboard? It's the four pin connector near the CPU.

[tortilio]

Yep, it's plugged in (and I also tried two different power suppliers).

[Jarhed7276]

with the motherboard uttering an alternate sound similar to an emergency siren

This sounds like a CPU overheating warning.
You did apply a fresh coat of thermal paste between the CPU and heatsink didn't you?
Also check that the heatsink is seated properly.

[tortilio]

This sounds like a CPU overheating warning.
You did apply a fresh coat of thermal paste between the CPU and heatsink didn't you?
Also check that the heatsink is seated properly.

'Unfortunately' the CPU fan is correctly plugged and spinning, and the heatsink is mounted on the right side.
The only think I do not have is the thermal paste, but the system always worked correctly without it, and I tend to exclude that a just turned on system might raise to allarming temperatures in a very few seconds, just because of the absence of the paste (I'm not an expert, though).

[stroyal]

You need thermal paste, the heat sink has probably warped.
It is a pretty rare system that can run without thermal paste.
Even a good heat sink, can't make 100% contact.
My NF7S makes an alarm like a siren, immediately on start up, if overheating.

You are wasting your time, diagnosing with such an obvious problem.
As far as your not heating right away theory, many years ago when I was young and stupider, I had my finger on a CPU that I thought was bad.
Within a half a second of starting, I got 3rd degree burns on my finger, and the CPU really was dead then.
That was a P3, your CPUs are much hotter.

Welcome to TechIMO!

[tortilio]

I'll then look for some thermal paste, and try using it, although I don't feel very optimistic about such option (the system started up with no problems previously, and it did not have the paste).
I'll let you know as soon as I can get my hands on some paste...

tortilio

[RacerX]

Any computer store, even radio shack has some.
not the best, but better than none.

Stroyal is right on the money with his advice.

Even thought it "may have run before", you were still cooking the CPU over time.
Even too thin of a layer of paste can raise temps.

Those CPU's don't have as good of thermal management that newer CPU's have, so you were probably running it hot the entire time, which can cause things to break down internally.
You may have killed it already, but we'll see what the paste does.

[tortilio]

Well, I bought some thermal paste (Nexus TMP-1000 silver), applied it (I think I put a little more than needed), and then the system started up, loading BIOS in fail-safe state, and showing a CPU speed of 1250 (while it's actually a 2500+ AMD Athlon).
Nevertheless, shortly afterwards the siren alarm started again, with the system self shutting down again (although this time its turned on time reached out to the Winodws start up...).
I entered the BIOS setup and gave a look at the CPU temperature (with CPU running at fail safe 1250 Mhz speed) and it was around 70C (=158 Fahrenheit), which appears to be still pretty high.
After trying to load BIOS defaults, and a new self shut down, I looked athe temperature again witnessing a 97C (=206 Farenheit).
An easy conclusion seems to be that my CPU overheats excessively, and extremely quickly. At this point my question would be: do you think I have to look for a new CPU?

[stroyal]

Not yet!

You can put too much paste on. Especially the silver tyre.
The paste is only for the parts of the heat sink that don't touch the CPU directly.
This is usually on a microscopic basis, unless the heat sink is warped.
With a good flat heat sink, the paste should look almost opaque, and even when the heat sink is removed again.

Is this an all aluminum sink, or a copper/aluminum one?
Did you clean it and the CPU with isopropyl alcohol? It needs to be 100% clean.
Is the hear sink as smooth as glass? It should be.

Does the heat sink feel warm, or hot to the touch,right after shutdown?
It should with those CPU temps.
If not the paste isn't installed correctly, so it is not transferring the heat to the heat sink.

[electric]

and CPU fan plugged into the MOBO [ easy tio forget ]

if you do get your system working, when in the BIOS select " softmenu setup"
to set up all parameters regarding CPU FSB speed,multiplier factor etc.

[stroyal]

What is your room temp?
My barton 2500, in the basement Idles at 26-30C, in a cool 60F room.
If the temp in the basement was 70F that would go up to 36-40c.

My cooler is a little better that the stock AL/CU one.

I run an old Vantec CU/AL hearsink, with a 3200 RPM case fan, and a homemade duct.

[tortilio]

I eventually found out the heatsink being the cause of the problem.
I'm pretty astonished by the facts, but if the previous heatsink (been working efficently for years, and with the fan corrrectly spinning...) had the CPU temperature raising to allarming temperatures in a handful of seconds (with littly shorter times in case of absence of thermal paste), a different one leaves the CPU at an ordinary average temperature of 40 C, without any operating problems.
I find it hard to believe that a piece of metal 'stopped working', and I tried to inspect it: it features a well cleaned (though not with alcohol) copper plate, with a 'natural' slighltly ruled appearance but with no visible scratches or wrappings (the rest of the heatsink being in alluminium). I really don't know what to think about, and how an appareant absence of defects might result in drastically different heating behaviours, but let's say that everything's fine as long as it ends finely...

A BIG thanks to everyone for your help!

[stroyal]

Either by warping or by pitting, the old heat sink was not making 100% contact. If it is really bad, paste can't make up for that.
The only other thing it could be is, if the whole heat sink was tipped somehow, making almost no contact.
The only way the human eye can see small imperfections is by lapping. (sanding the bottom with a very fine grit, on a perfectly flat surface, until it makes 100% contact)

In the olden days, when heat sinks where just a crude chunk of aluminum, overclockers use to lap the heat sink directly to the CPU. This was far more efficient than paste, although paste was still used to fill in the microscopic scratches.

The bottom line though is if the old one did not work and the new one did, the old one was not contacting the CPU enough.

I know, more than you wanted to know, but glad you fixed it.