CPU or MB - what died?

[Mistral]

So, a few days ago I clicked "Sleep" in Vista by accident. The PC obediently complied and went to sleep mode. I clicked almost immediately after to wake it up. Instead, my PC kept restarting.

The PSU had died – so I got a new one. Now when I start the PC I get no POST screen. Fans and drives spin up, but it’s just the CPU (e8400) led on the MB (GB EP45S-DS4P) that flashes once, and then there’s a faint click noise (doesn’t sound like the normal beeps from the PC speaker). Then it repeats.

I inspected the components and there’s no visible damage. Can anyone please give me any suggestions? Is there a way to make certain if the MB or the CPU is the culprit, short of trying the processor in another machine?

 

[grunt_408]

Did you make sure that you plugged in the cpu power supply? You know the 4/8 pin plug that supplies power to the CPU
Hopefully the PSU did not take anything to grave with it.

 

[Mistral]

Aye, everything is plugged in the designated places.

 

[grunt_408]

Fair enough as it has happened to people before .

 

[chuck216]

Try clearing the cmos, then powering it up.

 

[Wile E]

Try clearing the cmos, then powering it up.

If that doesn't work, try a single stick of ram in the first slot. If that doesn't work, try the other stick. If that doesn't work, try every combination of a single stick in every slot.

 

[hat]

Clear the CMOS. After that I dunno. When a power supply goes, it's possible that it kills everything, so you're stuck testing everything out and seeing if it works or not.

 

[grunt_408]

If that doesn't work, try a single stick of ram in the first slot. If that doesn't work, try the other stick. If that doesn't work, try every combination of a single stick in every slot.

I was going to suggest that man

 

[Mistral]

Had tried the "clear cmos" thing already - didn't seem to change anything.

I will give the memory-move-around a try too, though I doubt it'll help. The freaking GB mobo is like a Christmas tree and the ram-problem-indication-leds don't light up. Besides, wouldn't it make it to the post screen with bad ram? And the manual also states "1 long, 1 short" beep for memory or motherboard error...

Can anyone who has had a similar experience share some impressions on Intel's RMA procedures? I guess in the end I'll have to just take the CPU, and waltz in some computer store and have them give it a spin...

 

[Wile E]

Had tried the "clear cmos" thing already - didn't seem to change anything.

I will give the memory-move-around a try too, though I doubt it'll help. The freaking GB mobo is like a Christmas tree and the ram-problem-indication-leds don't light up. Besides, wouldn't it make it to the post screen with bad ram? And the manual also states "1 long, 1 short" beep for memory or motherboard error...

Can anyone who has had a similar experience share some impressions on Intel's RMA procedures? I guess in the end I'll have to just take the CPU, and waltz in some computer store and have them give it a spin...

If the ram is completely unstable, sometimes the board won't even make it far enough in post to give error codes.

 

[Mistral]

Swapping RAM around brought no change.

Removing both sticks had the expected OMG-Where-is-my-RAM-BBQ effect on the MB, making it blink (the RAM error led, no CPU error led...) and beep as per the manual.

 

[kylzer]

Sounds like mobo as cpus just don't die than much theses days its usually always the motherboard.

 

[goldfries]

Can anyone who has had a similar experience share some impressions on Intel's RMA procedures? I guess in the end I'll have to just take the CPU, and waltz in some computer store and have them give it a spin...

about Intel's RMA, it was very easy. I went to their site, request for RMA on my E2140 that refused to work. they Q&A me and later they sent a parcel over to me to proceed with the RMA process. It was a DHL shipment container or whatever you call it.

I didn't send it immediately as I was busy with other things such as my business and my website. After a while (few weeks) Intel called me and asked me whether I've sent the item and that the RMA process will be invalid soon.

I'm impressed by how it was handled, anyway I shipped it out (DHL pickup) and after a few weeks, they sent a replacement E2140 to me. Met the country manager a little later and told him about my experience, he was happy too.

 

[Sir_Real]

Av u tryed disconnecting everything but the bare escentuals needed to get the mobo the post to the bios. Disconnect all drives inc hd's

Really you need to get your hands on another socket 775 cpu to test the mobo. Or another mobo to test the cpu !

As kylzer has mentioned i'd agree its more likely to be the mobo at fault rather than cpu or ram.

 

[95Viper]

Try turning off the PSU switch, un-plug it from the wall, wait 30 + seconds, plug it back in the wall, turn on the PSU switch. Push your Power-on button, cross your fingers.

Sounds like it might be the power good signal from the psu might not be registering with the MB for some reason. My GB MB did this a couple of times.

 

[Mistral]

OK...

Went to a PC store yesterday. CPU's good, RAM's fine, there goes the motherboard.
I kind of expected that but was hoping it wouldn't be the case, since freaking RMA can take well over a month on those things...

 

[Velvet Wafer]

they tested cpu and ram in another board`?

 

[Mistral]

they tested cpu and ram in another board`?

Aye, before my very eyes...

 

[overclocking101]

That's what I was gona say buy an asus m8 much better boards

 

[twilyth]

I've had this happen several times over the past couple of years and it has ALWAYS been the mobo. And this was even high end boards like Gigabyte P35 DQ6 and DS4.

The last couple times I pulled out the board and looked at the front - and especially back - sides. I both cases I was able to spot small spidery cracks in the coating/lacquer.

I think the reason this tends to happen during a transition, like being powered down, is because that is when the thermal stress changes rapidly and therefore amplifies any weakness that's already present. I'm not certain, but I don't think any of my crashes happened during normal use. It was always a power failure (even though plugged into a UPS) or a required system reboot, or something like that.

For some reason, we tend to look at the mobo last. I guess when you're shelling out $200-300 you figure that mofo should fuckin' bullet proof. But oc'ing takes it's toll and even the best mobo's well eventually succumb. At least that's been my experience.

 

[vbx]

how does putting a computer to "sleep" destroy the mobo?

 

[twilyth]

how does putting a computer to "sleep" destroy the mobo?

It doesn't make any sense. That's why I came up with the heat stress theory. It may not be right, but it does fit the facts.

 

[Sir_Real]

I do know a big killer of socket 775 mobo's is heatsinks !

Yeah apparently theirs alot of mobos being returned to manufacturers with damage due to excessive pcb bending around the cpu socket.

Most socket 775 sinks have zero bend in them & rely on the height of the socket & cpu being accurate enough for it to require only a slight flex in the mainboard to get the clips clicked in place.

A common problem occurs when either the height of the cpu socket or the clips on the heatsink are slightly out from norm. This can then mean excessive force is needed to flex the mobo enough to click the clips inplace. Thus resulting in a warped motherboard. Obviously this can cause damage to the circuity either strait away or after a period of time with the added pressure caused by expansion & contraction due to heat.

 

[grunt_408]

Sucks you have to get a new mobo man. On the bright side at least everything else is all good.

 

[hat]

Mobo bending can be a problem, but with backplates it shouldn't bend.