VAIO RAM Problems
Here's an interesting solution to SODIMM problems:
For about five days now, the copy of Windows 2000, I've been having problems at Random times with Windows 2000 failing to restore properly from hibernation on Wildfire, my little Sony VAIO PCG-FX105K. Today, whilst trying to vapourise the hibernation file to see if that would help, it was made obvious that I was missing about 2/3 of the ram I should have had available to me.
Needless to say, I considered this to be something of a problem. After a bit of fiddling with the SODIMMs in there, it became obvious that any RAM put into the top SODIMM socket was not being recognised by the system.
Much frantic searching for potential problems then ensued. Nothing seemed to work, and I couldn't spot any obvious problems with the electronics or their physical interfaces, until I noticed that the locking lever on the left hand side of the SODIMM was slightly more wobbly than the oone at the right hand side.
This shouldn't have been a problem, as the SODIMM was held firmly in place, but I can only assume that there is some weird mechanism for detecting physical presence of the SODIMM to do with this, because when I powered up with this held down everything started to magically work again.
So, should this problem happen to you, here's my solution:
- Open the RAM access panel
- Light a match
- Blow the match out
- Break off the end of the match that has been (semi-)burned
- Break the remaining match in half
- Put the two halves of the match together side-by-side
- Wedge this between the locking lever and the case
- Rotate through 90 degrees so it is just below the lip of the RAM access panel
- Relplace the access panel
Okay, so it might not work for you, but it's been a lifesaver for me. Especially since I found out that my complete stupidity insurance on Wildfire doesn't cover this sort of thing. ;-)
In case you're interested, here's a low quality photo of the result, which is completely invisible when the panel is replaced:
Additional
This doesn't seem to have permanently fixed the problem, it still regularly "forgets" about the RAM in the top SODIMM socket until it's given a bit of a poking. This is really annoying, but at least I've got my larger SODIMM in the bottom socket now. If I find any other potential problems, I'll be sure to post them here.
For the moment, I've removed the smaller SODIMM because I'm investigating software suspend under Linux and I don't really want varying RAM sizes to be adding an extra level of complexity to the provess.
More progress
I had another poke around trying to get the top RAM socket working today, and found that it would consistently recognise the presence of a SODIMM in that socket if you gripped the edges of the socket together quite tightly during the boot process. As a result, I modified my matchstick method slightly to reproduce this result.
I now have two carefully cut lengths of matchstick pressing on the SODIMM socket, one on either side. These are positioned perpendicular to the socket and the side of the case, rather than parallel to it. It was quite a bit tricker to fit than the parallel arrangement, but provides a greater force against the socket and has less change of becoming noticably squashed during use.
The result so far is very promising - Wildfire has consistently detected both SODIMMs, and life is continuing well. The matchstick mod is shown in the picture below, which incidentally was created by placing Wildfire on my shiny new (well, second hand) HP ScanJet 6100C and turning the gamma up a bit:
Wildfire has since been retired due to a problem with the power management subsystem. As such, I will not be able to report on any future developments with the SODIMM slots on this VAIO.