I've successfully built four virtually identical machines for clients, using the ASRock A330ION Atom/ION motherboards. All run Windows 7 (32bit mostly), 2GB RAM, 500GB SATA HDD, Samsung DVD, InWinBM639 cases.

I've recently built a fifth, same specs, and it is causing me major headaches.

The RAM used is Corsair CMX2GX3M1A1333C9 2GB - don't know if this would cause a problem.

The problem is this:

The computer boots just fine, runs perfectly. It will go to sleep, and resume from sleep. The computer can even be asked to Hibernate, and will resume from Hibernate. BUT... if the computer is allowed to go to sleep by itself, and then is left for a prolonged period (where it can then go into 'hibernate') strange things happen on 'resume.' I have the blue light flashing on the front of the PC, the fans are shut down, the computer is in normal sleep. I press power to resume the PC, and sure enough the fans fire up, and the hard disk starts accessing, and accessing, and accessing... and you can hear the hard disk trying to spin up in 'cycles', then spin down again, repetitively. The display does not wake at all. I've had occasions where the computer even 'resumed' and booted LogMeIn, allowing me to log in remotely and do things on the computer even though the computer's own mouse/keybd and display were not working.

Then the really odd thing is this... if I power down or restart, the computer cannot boot and I find out that the CMOS checksum is invalid... Basically, something is causing the CMOS to reset WHILE the computer is in sleep/hibernate, which essentially knocks out all USB/display devices and changes the HDD mode from SATA IDE to SATA AHCI which, presumably, is why the HDD has such problem resuming - because its mode has been changed mid-usage...

I have to reset the CMOS configuration and reboot, and then the whole process begins again as soon as 'sleep' is used as a feature.

In the interest of troubleshooting I've tried the computer with 'suspend to ram' switched off (bloody noisy fan during 'sleep' and 'hibernate'). In Win7 power settings I have Hybrid Sleep switched off, sleep set to 20 mins and hibernate set to 120 minutes. The drivers are the latest ones. Windows Updates are the lates. BIOS is 1.60 by default, and yes, I've tried reflashing it too, just to be sure. I've replaced the CMOS battery. It still occurs.

It is an embarrasing proposal to prevent the computer from sleeping or hibernatng so as to 'ignore' the problem, when I've promoted it as a 'power saving' energy efficient computer, so I don't really want to concede defeat in this regard. I want to know why the CMOS appears to be resetting when the computer has been allowed to sleep for a certain period. I'd like to be able to fix it.

I have a hunch that the other machines were all built using firmware 1.40, which is no longer available on the website. I wonder if firmware 1.6 is faulty.

Does anyone have ideas, hunches, something to go on?