[X4U] DRAM failure (was Hard drive or hard drive controller gone)

[Jim Robertson]

On 1/18/06 1:08 PM, "Jon Warms" <jwarms at mac.com> wrote:

> In my experience, random hair-pulling failures are often attributable to
> a faulty hard drive. Your experience doesn't go against this; when the
> MacOS isn't on the hard drive, things go well.
> 
> Have you run a verify disk with Disk Utility? What's the S.M.A.R.T.
> value?
> If this were my machine, with the results you describe, I would really
> check out the hard drive. (Disk Utility has worked for me.)

Thank you so much for weighing in. I need any/all help I can get.

The first thing I suspected was the hard drive. There were two in the
machine (both IBM Deskstars). One failed testing with Disk Utility and
couldn't be erased. So, I purchased a new Seagate 300 GB/16 MB cache 7200
rpm Barracuda (in other words, not cheap), installed it, put OSX Tiger on
it, did all the Software Update runs needed to get current, and told my son
he was good as new.

An hour later, he called me in to his room to look at a frozen cursor.

So, I downloaded and ran memtest from the command line in single user mode
with all permutations of the 3 DRAM sticks I had originally in the machine,
and with all three present, there were repetitive reports of errors at two
different offsets. I took out the Apple-supplied module (256 MB): still 2
errors. So, I took out the larger (unfortunately, generic) DRAM modules, put
the Apple-supplied 256 MB module in slot 1, and THE MACHINE EMITTED ITS
MONOTONE HARDWARE FAILURE sound when I powered up. (I should also mention
again that each of the two 512 MB DRAMs reports a single address error when
installed alone or with the Apple-supplied module).

I moved the 256 MB module to slot 2, and as long as all I do is run memtest
from the command line or play with Disk Utility booting from the Tiger
install DVD, everything is fine (including repetitive verifications of the
hard drive). However, if the machine boots from the hard drive to the full
Mac OS, it will freeze within an hour, often within minutes.

The last bit of data I have (coupled with my sense of what it FEELS like
inside the box) is that it's MUCH warmer than it used to be. The processor
temperature sensor reads 56 degrees celsius or higher, and of course the
processor sits right next to the 1st RAM slot, which seems to be kaput. The
default bay for two internal drives is also near there, and I've had the
hard drive failure mentioned earlier.

I've even tried moving the hard drive bracket over to the other bay, farther
away from the CPU, to see if it makes any difference (it doesn't).

However, it's my sense that with the subjective excess heat in the box, the
failure (in just a matter of days) of two hard drives, a RAM slot, and two
RAM modules, there's something seriously wrong with the hardware.

I cannot explain why or how the thing can keep running memtest in single
user mode, how it can run for hours from the installer DVD, or (most
puzzling), how it can go through everything it needs to do make a fresh
installation of the OS, the BOOT from that OS and run Software Update for
several iterations, but then freeze when it's doing nothing except painting
the screen from the screensaver.

Jim Robertson
--