Failing hardware ? how to track down

BusError · 5 Jun 2011 at 13:54

I've had that Q9650 workstation since 2008 (see signature), it worked perfectly well since then, and I started having problems recently:
+ Failed to boot at some point, manage to recover by (very wierdly, but it seems now relevant) unplugging all the USB bits I had...
+ A few weeks later, machine failed to boot too (lots of beeps). I wrongly tought it was the RAM (OCZ 2.1V) but it wasn't, when I tried some spare.
+ It was the PSU packed up entirely. On a sunday of course !

I bought an "emergency" 680W PSU from ******, that looked quite overspeced ("G7" 680w, while my old one was 500W -- this system only has a small graphic card and 2 disks, no fancy gaming)

Since then, I've had /heaps/ of problem booting the board. I had to downgrade the overclock, but even 3.1 Ghz most of the time fails to boot; the machine beeps a few times (failed posts) then boot at the default 3.0Ghz (but if course with the ram at 1.8V, it won't even work properly!)

When it boots, it's quite stable, passes the memtest, passes CPU tests, is nice and cool etc etc etc. Basicaly, it works ! Apart from a crash every couple of weeks or so, totally unexplained. This morning it was the disk that had problems!

So I'm starting top suspect a faulty components somewhere (a cap?), that makes the voltages unstable. That could also affect the boot and thats why it mostly fails to boot, despite being well within specs!

Anyone has an idea ? I'm not really wishing to splash out more money on a system where I could possibly get a new mobo/cpu by stretching it a bit more the money, on the other hand, it's still a fantastic system at 4Ghz (could run at 4.2 too !) so I'd love to find the reason!

95thrifles · 5 Jun 2011 at 19:06

hmm, sounds like it could be mobo, but would be hard to pin down

honosuseri · 5 Jun 2011 at 19:21

It's not uncommon for some PSUs to take components out when they "pop". I've seen mobos fried first hand when called out to look at problem PCs.

Your best bet is really to replace the mobo and work from there. It appears the CPU is ok, although you now seem tohave doubts about the HDDs or DVD.

Bad mobos can display all sorts of symptoms, you seemed to have done the obvious and reset the CMOS and checked the RAM. Guess it comes down to cost of a replacement mobo as to wether you try to salvage it or just upgrade.

BusError · 5 Jun 2011 at 21:45

Well it's rather annoying, it's a monster mobo that I bought (on OcUK) for really cheap, and even my "upgrade" would feel a bit like a downgrade on a few points, for example, I have (and use) 2 gigabit ethernet ports... it has a 2 digits hex display for mobo codes, it comes with 1.5kg of copper heatsink!

Since my message I realised my problems always appear on sunday mornings. In fact, I remember the machine packing up on a sunday morning right as I walked in the office... I'm now suspecting maybe a neibours is screwing up the mains, somehow, and has damaged something

And yes, it seems that something around the SDRAM is not working properly... event at 1.8V it has problem booting with more than one stick. It's probable something is wrong on the regulator on that side of the board.

freddie64 · 6 Jun 2011 at 06:34

I would start with a better PSU, this Antec gets rave reviews.

Even if this doesn't solve the problem you will have a more reliable PSU than the one from "you know where".

What volts are you using (cpu, ram and northbridge) and have you overclocked the ram?

Admiral Huddy · 6 Jun 2011 at 08:15

Have you looked in the Event Log for any clues as to what might be going on? Google anything suspicious. It might also be worth re-installing or repairing Windows but before you do eliminate the HDDs first.

Firstly, identify the manufacturers of your hard drive(s) then visit the manufactures website for diagnostics tools. Most of these tools are executed from a live CD which means they’ll boot straight from the CD. This eliminates the need for booting Windows from the hard drive. Change your BIOS so the first boot disk priority is your optical drive. Insert the CD and boot. Run the diagnostics tool and verify to integrity of the drive.

If the drive looks ok, then try the windows repair/install. If still no luck then it's looking more like a faulty motherboard, possible popped caps or even faulty VRMs.

ps - might also be worth checking that there are no shorts on the motherboard or with the front panel connections. Try testing out of the case if you can, using a screw driver to short the power up pins.