The Problem.
Reduced ability to accomodate memory AND inability to accept memory- seems to be the number of GB rather than number of sticks but could be the number of full slots. Sympton is boot failure fixed by removing memory sticks.
How it developed
Had boot problems. Over a period of months it took 2 -3 rebooots, then 10 then 20 then infinity. Had 4x512MB installed. Solved problem first time by removing 1 512Mb RAM stick. Ran OK for a while with 3x512 then same problem again solved by removing another 512mb stick. Was working OK but 1G not enough and as there seemed to be damage on the gold contacts I bought 2x1GB thinking the RAM sticks were the problem and I could insert the 2x1GB to get total of 3GB. Didn't happen.
Removed 1 of the two 512mb located side by side - now would not boot first time even though I had removed not added. Tried various combinations of 512mb and 1GB sticks and found that there seemed to be no one stick or no one slot that was identifiable as a problem. The board will accept 2x512mb or 1x1GB but not 2x 1GB which is the minimum I really need. Did not try 512mb + 1GB.
Usually it got as far as mouse and keyboard lights before stalling (stalls before monitor backlight come on) before I removed the offending sticks. Now it gets this far with 2x1GB and then stalls before monitor backlight, but does NOT get this far with all slots full i.e 2x1GB and 2x512mb. boot hangs earlier it seems.
Observations that may help someone to isolate the problem:
Main board is Intel D865PERL. Both old and new memory is the correct spec (i.e listed as tested/compatible) and proven to work as individual sticks
I considered installing a LATER version of the Bios but on looking through fixes in release notes I could see references to hanging at post issues and memory issues but could not see anythoing that specifically related to my problem. Right now the PC is funtioning so I am reluctant to install a new BIOS and risk losing that functionality if I do not know for sure I am specifically addressing the problem.
As I am using correct spec memory and the problem developed after having worked fine for a couple of years I assume it is not an incorrect voltage setting as there have been no changes and the problem happens with the old memory also.
There may be two seperate issues here.
Existing in place memory fails because it is damaged by a power surge, BUT there is also a separate issue with recognition of new memory.
Getting it to boot after changing sticks and slots is pot luck - but once a configuration is accepted and boots it will boot from cold (at least until the next powerr surge maybe). This suggests some issue with recognition?
When I took out 1 of 2 working 512MB it would not boot even though i had just removed one stick that at that time was known to be working (was clearing slots for upgrade). had to move to different slots from the previously working configuration to get it to boot with memory known to work.
Boot problem was not present with warm start ie. green restart button pressed unless I set CHKDSK in which I had the non boot problem.
If during a power cut I kept the UPS on with PC switched off it would usually (but not infallibly) boot first time. If I switched off UPS there was always a problem rebooting and usually also a problem with Samsung monitors claiming they were not connected to PC. Had problems with a previous quad card not running all four monitors after a switch off. Have tried without UPS and it makes no difference so not UPS. Already tried a new Power Supply.
We went through a process of elimination of everything eg removed video cards etc etc one by one with someone who makes a living with computers - it seems after going through that process we isolated it as memory, further confirmed by the fact that the problem was solved a second time by removing a second stick of memory. I tried different keyboard and mouse and different monitor. Mouse and Keyboard recognition seems to happen before the stall, although I would have though RAM is checked first so mouse and keyboard circuitry/controllers/drivers etc cant be eliminated. I don't think it is memory sticks themselves I think it is an issue with the components and circuitry on the MB that is immediately connected to the RAM manifesting through memory. I base this on the way it took 2 then 10 then 20 then infinity times to boot up - this along with the sporadic nature of the problem seems to indicate component failure.
The PC runs 24/7 so I assume if the CPU fan was an issue this would manifest somehow e.g overheat causing a power off.
I am aware of the replies in this thread describing my problem as evidenced by OLD RAM not working when put back in and my memory is correct spec/tested by intel, works as one stick in situ.
http://www.tomshardware.com/forum/229874-30-d865perl-boot-installing-memory
Bought a new CMOS battery but my friend who knows more than me said he did not think it was the CMOS battery (gave an explanation as to why he felt this could be eliminated). Accordingly I held back from changinmg the battery as I understand it, if I take out the CMOS battery the BIOS settings will revert to default, and as I did not build the PC I do not know if there are non default settings used which I do not know how to restore. Right now I have a working computer.
Often there is no clearly discernable logic with these problems (although there is always a reason) and trial and error is the only way.
Suggestions:
I think this is some kind of dying mainboard/component problem - which means it is going to be hard to isolate/fix - but could it be something simple like bad earthing? The problem is linked to mainboard power off rather than PC power off it seems. As I understand it there is power flowing to the main board when PC is switched off.
I think I have eliminated total memory+ total page file being over the (4GB?) limit as a possible cause but not entirely.
Any suggestions?
Reduced ability to accomodate memory AND inability to accept memory- seems to be the number of GB rather than number of sticks but could be the number of full slots. Sympton is boot failure fixed by removing memory sticks.
How it developed
Had boot problems. Over a period of months it took 2 -3 rebooots, then 10 then 20 then infinity. Had 4x512MB installed. Solved problem first time by removing 1 512Mb RAM stick. Ran OK for a while with 3x512 then same problem again solved by removing another 512mb stick. Was working OK but 1G not enough and as there seemed to be damage on the gold contacts I bought 2x1GB thinking the RAM sticks were the problem and I could insert the 2x1GB to get total of 3GB. Didn't happen.
Removed 1 of the two 512mb located side by side - now would not boot first time even though I had removed not added. Tried various combinations of 512mb and 1GB sticks and found that there seemed to be no one stick or no one slot that was identifiable as a problem. The board will accept 2x512mb or 1x1GB but not 2x 1GB which is the minimum I really need. Did not try 512mb + 1GB.
Usually it got as far as mouse and keyboard lights before stalling (stalls before monitor backlight come on) before I removed the offending sticks. Now it gets this far with 2x1GB and then stalls before monitor backlight, but does NOT get this far with all slots full i.e 2x1GB and 2x512mb. boot hangs earlier it seems.
Observations that may help someone to isolate the problem:
Main board is Intel D865PERL. Both old and new memory is the correct spec (i.e listed as tested/compatible) and proven to work as individual sticks
I considered installing a LATER version of the Bios but on looking through fixes in release notes I could see references to hanging at post issues and memory issues but could not see anythoing that specifically related to my problem. Right now the PC is funtioning so I am reluctant to install a new BIOS and risk losing that functionality if I do not know for sure I am specifically addressing the problem.
As I am using correct spec memory and the problem developed after having worked fine for a couple of years I assume it is not an incorrect voltage setting as there have been no changes and the problem happens with the old memory also.
There may be two seperate issues here.
Existing in place memory fails because it is damaged by a power surge, BUT there is also a separate issue with recognition of new memory.
Getting it to boot after changing sticks and slots is pot luck - but once a configuration is accepted and boots it will boot from cold (at least until the next powerr surge maybe). This suggests some issue with recognition?
When I took out 1 of 2 working 512MB it would not boot even though i had just removed one stick that at that time was known to be working (was clearing slots for upgrade). had to move to different slots from the previously working configuration to get it to boot with memory known to work.
Boot problem was not present with warm start ie. green restart button pressed unless I set CHKDSK in which I had the non boot problem.
If during a power cut I kept the UPS on with PC switched off it would usually (but not infallibly) boot first time. If I switched off UPS there was always a problem rebooting and usually also a problem with Samsung monitors claiming they were not connected to PC. Had problems with a previous quad card not running all four monitors after a switch off. Have tried without UPS and it makes no difference so not UPS. Already tried a new Power Supply.
We went through a process of elimination of everything eg removed video cards etc etc one by one with someone who makes a living with computers - it seems after going through that process we isolated it as memory, further confirmed by the fact that the problem was solved a second time by removing a second stick of memory. I tried different keyboard and mouse and different monitor. Mouse and Keyboard recognition seems to happen before the stall, although I would have though RAM is checked first so mouse and keyboard circuitry/controllers/drivers etc cant be eliminated. I don't think it is memory sticks themselves I think it is an issue with the components and circuitry on the MB that is immediately connected to the RAM manifesting through memory. I base this on the way it took 2 then 10 then 20 then infinity times to boot up - this along with the sporadic nature of the problem seems to indicate component failure.
The PC runs 24/7 so I assume if the CPU fan was an issue this would manifest somehow e.g overheat causing a power off.
I am aware of the replies in this thread describing my problem as evidenced by OLD RAM not working when put back in and my memory is correct spec/tested by intel, works as one stick in situ.
http://www.tomshardware.com/forum/229874-30-d865perl-boot-installing-memory
Bought a new CMOS battery but my friend who knows more than me said he did not think it was the CMOS battery (gave an explanation as to why he felt this could be eliminated). Accordingly I held back from changinmg the battery as I understand it, if I take out the CMOS battery the BIOS settings will revert to default, and as I did not build the PC I do not know if there are non default settings used which I do not know how to restore. Right now I have a working computer.
Often there is no clearly discernable logic with these problems (although there is always a reason) and trial and error is the only way.
Suggestions:
I think this is some kind of dying mainboard/component problem - which means it is going to be hard to isolate/fix - but could it be something simple like bad earthing? The problem is linked to mainboard power off rather than PC power off it seems. As I understand it there is power flowing to the main board when PC is switched off.
I think I have eliminated total memory+ total page file being over the (4GB?) limit as a possible cause but not entirely.
Any suggestions?
Facebook
Twitter
Instagram