Computer wont boot (friend's)

[VirtualLarry]

Well, I helped a friend of mine put together a computer (more likfffe I did while he watched), a Phenom II X4 965 (3.4Ghz), 2x4GB DDR3, Asus 880G mobo, Microcenter 64GB SSD, two 1TB platter drives.

He told me last week he overclocked it, using some sort of of auto-overclocking feature, 10%. He said it was running at 3.8. I didn't really say anything to him about it, perhaps I should have. I guess I figured that that was within the ballpark of what those chips were capable of, so I didn't think that much about it.

He was using the stock retail heatsink, which for the Phenom II X4s, is a nice heatpipe job I think.

I'm not sure what to tell him. I really don't want to have to tear the computer apart, rebuild it, and possibly re-install Windows 7 (due to activation and a new motherboard, OEM version of Windows). At least not for free this second time.

The first thing to do, would be to clear the CMOS of course, and I'll try that the next time I'm over there. I'm thinking, possibly swap the CPU next. Thinking that it might have been damaged by running too hot. (I have some spare AM3 CPUs.)

I hope the mobo didn't burn out (the VRMs). The board, IIRC, proudly proclaimed overclocking support on the box though, and listed the power phases. I'm going to have to look up that particular mobo on their site and double-check.

Edit: This is the mobo:
http://usa.asus.com/Motherboards/AMD_AM3/M4A88TDV_EVOUSB3/
Says 8+2 power phases. Hopefully an overclocked Phenom II X4 965 won't overload 8 power phases for the CPU.

Hopefully it's just the CPU that overheated or something.

Edit: If we have to replace the mobo + CPU, would a SB Core i3-2100 and an H61 mobo, be better or worse than the existing Phenom rig?

 

[dave_the_nerd]

I think you should relax, and not worry about buying replacement hardware until you've reset the BIOS/CMOS back to defaults. A 10% OC isn't likely to have destroyed anything by itself, but it might have pissed the motherboard off @ you.

Once you determine which part is faulty, if any, you'll know of the OC may have been an issue or not.

If the system is fairly new, it's probably just a defective part that barely passed QC and was going to fail any minute now anyway.

Unless you want to scare the crap out of your friend so he doesn't ever try to OC again. In which case I'm fairly certain the power surge fried the CPU, then travelled through the disk buffer and probably killed both hard drives. (You might be able to get the data off of them, but I wouldn't count on them being reliable ever again.) Fortunately, SSDs are immune to buffer overloads. Unfortunately, i3 processors are incompatible with AMD-compatible computer cases, so your friend will need to foot the bill for at least an i5 rig.

(That was all sarcasm. This, however, isn't: I wouldn't want to go from a quad core to a dual core, even with hyperthreading. An i3 might even be faster at single-threaded stuff, but if he's a heavy multitasker, he'll probably notice a speed drop.)

 

[denis280]

Probably he used amd overclock.and the auto overclocking.so i would do a CMOS and pretty sure everything will be back normaly

 

[westom]

I'm not sure what to tell him. I really don't want to have to tear the computer apart, rebuild it, and possibly re-install Windows 7 (due to activation and a new motherboard, OEM version of Windows). At least not for free this second time.

Power is controlled by a power controller. That controller will not even let the CPU execute one machine insutruction until after certain hardware conditions are met.

So you must decided (post) whether the controller is even letting hardware execute. Or does the BIOS execute and fail.

Heat at those trivial temperatures does no hardware damage despite popular myths that say otherwise. Heat only caused timing and threshold changes. Cool the semiconductor and those changes go back. But heat is also a great tool for finding defective semiconductors.

That entire machine must be perfectly happy even inside a 100 degree F room. If not, a hardware defect exists. But you must first define what is meant by not booting. A machine does many tasks before it even considers booting Windows. Does it do any of them?

You have zero reasons to think the CPU failed. CPUs rarely fail. But many think otherwise when the power controller would not even let the CMOS read even one CMOS setting. First get facts before changing, disconnecting, swapping, or concluding anything. Only then can informed answers arrive. Currenly posted is insufficient information to recommend any action.

 

[VirtualLarry]

Well, it looks like my fears were unfounded. He didn't burn out the CPU. In fact, it boots the BIOS, it says "bootmgr is missing", so it looks like the fresh Windows 7 install we did is hosed.

He said that he left the computer running overnight, and that it went into a sleep mode.

Well, he had a Microcenter SSD in there, with the older firmware (since I forgot about updating it when I was putting his system together), and those Sandforce drives are known to have issues with sleep mode, sometimes borking the drive. So either that happened, or something else happened. (Maybe the overclock wasn't stable, and trashed the Windows install too?)