CPU fan broke; now no POST after replacing both fan and CPU

[lefenzy]

Our old 5+ year computer had some sort of CPU fan malfunction the other day that resulted in the computer shutting down automatically. After the sudden shut down, we tried to start the computer without success, and it was only after opening the case that we saw the CPU fan wasn't spinning. It had in fact seized up; normally fans are easy to rotate, but spinning this broken fan was not very easy.

Ok, so the fan was obviously broken, and the computer wasn't POSTing either. Maybe the computer can't boot without a working fan? So we replaced the fan, but still no POST. So next we bought a used Q6700. I replaced the old CPU, not noticing anything obviously burnt on the processor or the socket. After all that effort, still no POST.

What's up. It seems the possibilities are that the new (used) CPU is DOA or the motherboard is also somehow messed up.

The replacement Q6700 by the way should be compatible, as the previous processor is a Q6600.

Thanks!

 

[VirtualLarry]

First of all, it is highly unlikely that the Q6600 was damaged.

But I supposed it's possible that the CPU fan header may have been damaged if the fan was seized. However, it's probably equally likely that it wasn't.

Did you remember to re-connect both the 24-pin ATX power connector, as well as the 4-pin ATX12V or 8-pin EPS12V secondary power connector to the motherboard?

Did you connect the new CPU fan connector to the proper CPU fan header on the motherboard?

Did you remember to use the motherboard's CLEAR CMOS jumper, after installing the new CPU?

Edit: It's entirely possible, that the shutdown was NOT due to the seized fan, and that it was limping along (and probably throttling), and that something else went on the system. Possibly the motherboard, or the PSU.

If it was on the same PSU for 5+ years, it's due for replacement.

Unless your motherboard implemented solid caps, then it could be the mobo too.

I had a fan fail and seize up due to dust. It was an aftermarket OCZ Vendetta, on an AM3 Thuban Phenom II X6 1045T, overclocked from 2.7 to 3.5Ghz.

I thought something was funny, when CoreTemp reported Tjunction temps of like 80-90C max recorded.

But amazingly, the machine was running DC, and didn't crash.

That's why I think that the problem is not your CPU.

 

[lefenzy]

Thanks Larry. I'll check on connections, but I doubt that the fan header is broken. The new fan spins up just fine. The motherboard has a CMOS reset button that I pressed, but I also later tried taking the battery out for a period of time.

 

[VirtualLarry]

If that motherboard is "Deluxe" enough to have a clear CMOS button, does it also have a POST LED diag readout (two digits)?

Do you have a speaker connected to the mobo? If you remove all of the RAM, and have a speaker connected, does it beep several times at you? If it does, then that means that the CPU is working OK. If it doesn't give any beeps, that could be a bad sign.

Interesting, though, that the CPU fan spins up. Does it keep spinning, and not stop after like 2sec? As long as the ATX12V 4-pin aux connector is connected properly, then that's actually a good sign.

If this rig has a GPU, have you tried re-seating it?

Give complete specs, please.

 

[lefenzy]

Saw your edit on your first post just now.
I have a system that has been upgraded a bit over the years.

MSI P6N SLI-Fi (version 1) http://www.newegg.com/Product/Produc...2E16813130082; has solid caps, but not all caps.
Q6600/Q6700 (both in my possession now)
Memory: 6 GB ram DDR2-800 (2x1GB, 2x2GB)
Graphics: nVidia (galaxy) GT430 (a replacement from a year or so ago)
PSU: Corsair CX430 (also a replacement)
Three SATA hard drives

Unfortunately it doesn't have a LED readout. There is no beeping at all from the computer speaker. After I removed all memory sticks, it's the same situation: fans remain on but no sound from the speaker. When the computer was working, it would usually beep thrice upon POST. This doesn't change whether or not the GPU is installed in the system.

The CPU fan does remain spinning in all cases. Bought a nice new fan (http://www.newegg.com/Product/Produc...82E16835709037).

I thought that the CPU should be fine since I figured the heatsink is beefy enough (cooler master geminii s524) to allow the CPU to safety shut itself off from overheating, but it's a family computer, and I wasn't there when it first crashed and the several attempts to turn it on.

There's no difference with or without the CPU.

One previously omitted detail: when removing the heatsink to change the CPU, I knocked my hand into the north bridge cooler, and the two plastic, spring-loaded pin clips holding it in place snapped. I had to go to a computer hardware store to find replacements, and I also had to remove the stock TIM and replace it with arctic silver. The pin clips had become brittle with age I believe; they were slightly discolored (brown..). The north bridge of this computer (nforce 650) has always ran hot (heatsink hot to touch).

I don't want to think that the motherboard is broken, since it would be a pain to replace (from buying a new one to reinstalling the OS), but process of elimination seems to suggest it. I should note that all the LED lights on the motherboard as well as the case fan (connected to mobo) are operational. The LAN port LEDs even flash a bit.

It's entirely possible, that the shutdown was NOT due to the seized fan, and that it was limping along (and probably throttling), and that something else went on the system. Possibly the motherboard, or the PSU.

Very plausible; the family had complained of slowness.

 

[lefenzy]

The only other thing to do is to run it in an open box situation: CPU, memory, power, maybe GPU.

 

[VirtualLarry]

I hate to say it, but it's start to sound strongly like the motherboard. I am also curious about applying the AS5 to the Northbridge heatsink. Did you put an excessive amount on? AS5 is capacitive, so they say, and most Northbridge chips are bare die, which is OK, but then they have supporting capacitors around the die. If you bridged one of those caps with AS5, I'm not sure what would happen.

If you were careful, and not excessive with the AS5, then that may not be an issue.

Edit: At this point, this is what I would do:
Take the board out of the case, put it on cardboard.
Plug in the motherboard speaker.
Take out the RAM.
Put in the Q6600, that was working before.
Put on the new replacement heatsink.
Connect the GPU, and the PSU.
Short the power pins with a screwdriver.

If you don't hear beeps, then the board may well be dead.

I would, as a last-ditch effort, also try a new PSU. But with fans spinning and LEDs lighting up, it seems most likely that the PSU is alright.

I would also carefully inspect the LGA775 socket, with a light and magnifying glass if necessary, to see if it has bent pins.

 

[lefenzy]

Yes; I do have that fear of AS5. It was hard to clean off the old TIM (isopropyl
alcohol). I accidentally got some AS5 outside the chip in the first attempt, and wiped it off; I hope I cleaned it all. In the second attempt, I removed the heatsink right after installing it, just to see how much of the surface would be covered. There seemed to be no overflow. in the end (third application), I put very little.

I'll work up the energy to run things in an open box situation.

 

[lefenzy]

Update:
In the process of taking the computer apart in preparation for a motherboard replacement, we saw that there was a single burst capacitor (1000 uF, 6.3 V) previously unnoticed and hidden by the heatsink.

After some effort, we desoldered and replaced that capacitor. The computer's now operational again!

Who knows how long the computer was running with that broken fan until the failed capacitor finally knocked it out?

Back in action: http://einstein.phys.uwm.edu/hosts_user.php?sort=rpc_time&rev=0&show_all=0&userid=586495

Thanks for all your help Larry!

 

[VirtualLarry]

Wow, great job diagnosing and repairing the motherboard! (Didn't know that you were a DC-er too. That's cool. )

Edit: You're quite welcome.

 

[tweakboy]

Sorry you went thru this.