4 Replies Latest reply on Oct 12, 2017 1:46 PM by paramountain

    When nothing else works, replace the RAM

    paramountain

      This is not a warranty question.

       

      I was reinstalling W-7 on a well-used DH67CL / i5-2500K system. On three different drives connected to the SATA III port, two SSDs and a Black Caviar HDD, the installation failed with a chkdisk after the first auto-reboot, though no errors were found. I thought: "OH NO! IT'S THE INFAMOUS DH67 CHIPSET BUG!" I tried once more with a laptop HDD -- and it was able to install.

       

      Okay, so the original DH67 bug involved the SATA II ports.

       

      I still had a working W-8.1 system on a different drive, so I ran memtest and IPDT, but neither showed any errors.

       

      After wondering if I should dump the motherboard and build a newer system, I pulled both RAM sticks (Crucial Tactical Ballistix DDR3-1333 4GB) and substituted a single SK hynix 2GB that I keep only for debugging. Eureka! I was able to install on the Black HDD. Then I replaced the SK hynix stick with two spare Crucial Tactical Ballistix DDR3-1333 4GB sticks and everything went back to normal.

       

      That's just weird. Has anyone else seen a chkdsk due to memory failure? Also, what's up with the laptop HDD working when nothing else did? Maybe since it is slower than the others it put less stress on the system, but it's a Black Scorpio with 16MB of cache so it's not that much slower than the Black Caviar with 32MB.

        • 1. Re: When nothing else works, replace the RAM
          N.Scott.Pearson

          Have you run the tool to verify whether or not your system is actually vulnerable to this issue?

          • 2. Re: When nothing else works, replace the RAM
            N.Scott.Pearson

            Oops, I clicked on Add before I was done.

             

            If you have bad/incompatible memory, the problem could rear its ugly head in all sorts of ways. If I remember, I had some early Ballistix memory (which I purchased specifically for the Thermal-Sensor-On-DIMM feature) that I had problems with. I discovered that the memory required 1.65V and didn't work reliably at 1.5V. Once I configured for this, the problems resolved themselves. Note that this was with a DZ68ZV; I don't think I ever got it to work correctly in a DZ68BC (it wouldn't even work long enough for me to get into BIOS and make the configuration change)...

             

            ...S

            • 3. Re: When nothing else works, replace the RAM
              paramountain

              I assume you are talking about a tool to verify if a board has one of the bad chipsets. I had no idea such a tool existed. I just did a search for it, but I found nothing. You wouldn't happen to have a link? I found a Chipset Identification Utility, but it only works on 32-bit OS. And I found another utility, Board ID Tool for Desktop Boards, but that just gives the information I can find on the board sticker.

               

              As for the memory, it's probably not incompatible, as it has been in the PC for five years or so. Bad is definitely possible, though. My theory, which changes by the hour, is that it was damaged slightly by my misadventure in OCing RAM which you helped me out of by pointing me to maintenance mode, though at the time you mentioned that it could also be the processor memory controller.

               

              I neglected to mention before that the BIOS is 0160, the latest available.

              • 4. Re: When nothing else works, replace the RAM
                paramountain

                It appears that my debugging was far too hasty. I stopped the W-7 installation at a certain point, but I later learned that the chkdsk could appear after either of the first two auto-reboots (or both). The memory actually made no difference.

                 

                So I tried installing on the Black HDD using the other SATA III port, which worked. Then I tried installing using a SATA II port (port 2), which worked. So it's port 0 which causes chkdsk, which makes sense. And I swapped cables to remove it from suspicion.

                 

                The only remaining question is, how closely related are the two SATA III ports, in other words, since one is unreliable, does it mean that the second one (port 1) may soon follow? I would imagine they are very related.