this conversation is being copied from the original forum of its posting, as the problem is broadening.
Erratic behavior with HX620W
I built my system about a year and a half ago, and recently rebuilt a lot of it, due to problems... after removing a device from the IDE channel (No, it wasn't plugged in), the memory controller and SATA controller on the old mainboard died, so I replaced the mainboard. The 1.5TB drives also started having random read delays, a precursor to failure, though they passed all other diagnostics. I also ran QuickTech and other mainboard diag tools on the new mobo (and memtest 24 hours), and it checks out fine. Now I'm having random Windows meltdowns, even after just reinstalling the system fresh... on a few occasions, the system has hard locked, and on others, it simply failed to terminate or launch applications until rebooted (by power cycling). Looking at MBM5 logs, it's showing (what it believes to be) the 12v rail experiencing as high as 80% spikes on occasion. I say what it believes to be, because the decimal is in the wrong place for the 12v rail, and I'm not sure that's the only thing the MBM5 developers screwed up.
There's nothing in the event viewer, no bluescreens or other logged events, but it's hanging or otherwise malfunctioning as much as twice a day, often when the GPU is active for video decoding or something like that (It's not the apps, it's happened in 3 different apps, playing 3 different videos from 3 different sources)
Can anyone think of anything I can try, apart from going out and buying expensive diagnostic hardware (that might not show anything anyway) to determine if the PSU is causing these failures?
The randomness of your symptoms may be an indicator of a memory issue. Have you run Memtest? Also, were all of the parts you are currently using plugged into the previous MOBO that failed?
Yes, and Yes. I did determine that two sticks of memory had gone bad with the mobo, and after removing those two, it ran and passed memtest for 24 hours on the remaining two sticks. All of the parts had been plugged into the bad mainboard, with the exception of the 1.5TB HDD that had also started failing diagnostics (it was plugged into a separate controller, though now is plugged directly into the new mainboard)
I am also now seeing DPC spikes of 20-30% CPU, no core affinity, often plateauing at 20% for a minute or longer at a time. I have no missing drivers, and my IRQs look solid. All hardware, excepting the backup drive, passes diagnostics. I have systematically killed unnecessary processes and seen non change so far. I have not yet installed an antivirus or active scanner, since Kaspersky would get in the way of me reinstalling a slew of programs I need after reloading the system.
My voltages on the +12V rail over the last 12 hours according to MBM5 are:
current: 14.2v; low: 11.5; high: 23.7; average: 12.8
Plateaus of DPCs seem to coincide with slight increases of voltage that hold steady for a period, such as the current.
I have Everest, Sandra and Belarc system profiles available if someone wants to see them, but this forum won't let me add attachments of that size.
$10 via paypal to anyone who can conclusively solve this one.
Motherboard Gbyte EP45c-DS3R
Processor Intel E6600
Memory DDR2 800 (D9GMH Micron)
Video Card # 1 Zotac AMP˛! GTX260-C216 OC
Hard Drive # 1 WDC250 (RAID0)
Hard Drive # 2 WDC250 (RAID0)
Hard Drive # 3 WDC250 (RAID0)
Hard Drive # 4 ST250 (RAID0)
Hard Drive # 5 ST1500 (AHCI/SATA)
Hard Drive # 6 ST1500 (in eSATA enclosure)
Optical Drive # 1 LG BR/HD-DVD/DVD-RAM
Power Supply Corsair HX620W
Case CM Stacker T-01
Sound Card Audigy SE
Monitor Samsung Syncmaster T240 + Viewsonic Optiquest Q20wb
Operating System Windows XP Professional x86