Random Stop Errors at random times

Well, I'm officially stumped...

I started having a problem with my RAID O array behind a Promise 20276 chip running the fasttrak utility right after I loaded

SP2. I'd recieve 8E, 50, 7F errors (and others that I can't remember) at random times and the the error returned seemed to

be random as well.

Having looked through this forum and other web hints, I checked the memory with the Windows Diagnostic. That turned out

okay. Oddly, the routine stopped after a single pass. Should I be concerned about this?

I've updated the MB bios. I've downloaded all of the latest drivers for the hardware in my machine, except the Promise Raid

driver. It is not XP certified and I have not been able to find one-even having checked with the MB manufacturer, Gigabyte.

Still no luck; random errors at random times

I decided that while I was on the long holiday break, I would re-load XP Professional. I've now tried so many times that I

remember the Product Code.

During the latest series of installations, I've now started getting alerts that several files (again random) cannot be copied

initially from the CD. (It is an official Windows XP Pro CD and not a pirated copy.) I tried another original XP CD with

similar results. I haven't resorted to replacing the CD drive, but I'm real close.

I gave up after I received two STOP 0x50 errors during the set-up. The Microsoft KB Article 171003 claims that this stop

error is a result of faulty RAM so I am on the hunt. The MB uses RAMBUS, so this might be an expensive fault to solve.

However, the symptoms seem to point to bad memory (random files not copying correctly from CD seems to imply that the memory

looses a bit or three when transferring the file from the CD to the HDD--sound reasonable?)

The other recommendations from MS are to disable cache in the bios (not an option for my board); or, replace the video

adapter. I'd try this but I do not have an old 1.5 V AGP card sitting on the shelf.

Is it worth shelling out the 100 bucks for a new RAMBUS stick to test with? Since RAMBUS must be installed in pairs, I'm

stuck with buying only one that would sit on the shelf if both sticks turn out to be okay.

Grateful for any advice...

Dan

 

I agree with your conclusion. RAM first

Can you buy it from a vendor with a reasonable return policy? Most brick and mortar shops will let you bring it back.

 

RAMBUS Availability

Ben,

Will give that a thought... living in Belgium as an American, my first thought is to buy online. The RAMBUS stick is a couple of years old --PC1066 and somewhat scarce. I'll check with some of the computer companies that I deal with at work to see what they can assist with.

It's my son's machine and he's getting anxious to have the box back to normal.

 

As a general rule, you really can't get good testing of installed memory from a software routine. If it does get errors, there likely is a problem. If it does not get errors, that simply means it didn't get errors and says nothing about how good or bad the memory is.

 

Just an update to the random STOP errors

ALL,

I ran the Memtest86+ program. The results
were as I suspected: in 2 passes over a span of 32 minutes, the test
returned 3833 errors. Oddly, running through two more passes returned
no further errors.

I stopped the test to get the details off the modules and popped them
back into the other RIMM Slots (and put the CIMM where the RIMMS were).
The results are that in this configuration there were no immediate
errors. Given no further errors after the first two passes as outlined
above, I guessed that the problem must be heat related. I left the
machine overnight and ran the test again in the morning. Result errors-not as
many, but more than one.
In summary, then

Memory in slots 1 &2: Cold-errors
Memory in slots 3&4: Cold-errors
Memory in slots 1&2: Warm-no errors
Memory in slots 3&4: Warm-no errors