Things are getting desperate. I am really at my wits end now.

A ray of light appeared when I removed the Creative Labs X-Fi soundcard, and other USB peripherals. The thinking was that if it was a power issue, that might help.

Low and behold, there was an improvement. The crashing continued, but was no longer resulting in permanent screen corruption as my previous pics show. Basicly in Mass Effect, the graphics freeze, as usual, but after an alt-ctrl-delete, the game and the graphics will pop back in a playable state.The crashes are also less frequent, and its sometimes possible to play for half an hour or more, without it crashing once.  However, when it does freeze, the freezing seems to come in waves, where the game will freeze every minute or so for period, before becoming more or less playable again for a while.   It doesn’t matter what i do in game though. I just leave it and it will crash eventually, with the characters standing around idely.

Grand Theft auto, on the other hand, still gives corruption, again, only after a while sometimes, but no longer the pixel mess. It will corrupt the textures in game, then freeze, then crash out to the desktop. GTA is more consistent in crashing that Mass Effect though, at this point.

Stuff I have tried since the last post:

– Took the PC appart and running on the table.  More or less excludes shorting with the case.
– Memtest86, turned up nothing (remember hte memory is suspect because its one of the few common components across the old and the new hardware)
– Removed 3 of the 4 RAM, no effect
– Replaced the Keyboard and Mouse with a different keyboard and mouse, no effect
– Reconnected the old PSU, no effect

I have been going over how the PSU stuff is suppose to work, still thinking its a power issue.
The new 800watt PSU I got (Gigabyte Odin Pro) is a very good PSU, good reviews, and is know to deliver well on the voltage and amps. It has 4 separate dedicated +12v “rails”, two of which can deliver 28amps, more than enough for this setup.   Now I have been running Speedfan, and besides measuring temperature, it can give a readout on Voltage.

Now here is the worrying part. It shows 11.35v on the 12v readout. Way too little. However, the BIOS itself, when I boot into it, is telling me the 12v is at 12.04, which is spot on. I did read that software indications of voltage are unreliable. Regardless, I reconnected the old 650watt no-name PSU, and its giving me the same readout in Speedfan, and in the BIOS slightly higher, 12.152v

The whole thing still smells of a power issue. The way people on forums seem to have described power-related issues seems spot on with what I am having here. But it may just be confirmation bias of course.

There are 2 approaches I can take at this point. The first is to reinstall Windows, in this case Vista, again. I cant be sure Windows 7 is not at least a part of the problem, even though I doubt it because the problems started on Vista originally. I need to get my hands on a version with a valid key, but I think I know where, my original key didnt seem to work anymore as you know. I think I know where.

The other path is to continue digging into the power issue. I have seen several forum examples of people that had similar instability, that they managed to solve by upping the voltage of components ever so slightly. The CPU and/or the RAM for example. I may try this, there are certain margins that we can play in safely. I am not convinced either PSU is not performing well, but to really confirm it, I would need a multi-meter, which I dont have. I may buy one, its generally usefull to have around anyway.

I still have not heard back from the shop about my 295 card. I am assuming they have not recieved a replacement yet. I will call them tomorrow.  I may explain this whole thing to them, but I suspect they will give me the usual shopping list of things I have already tried.

And the very last thing I want to do is bring the parts round to them to have a look at. Its a logistical nightmare.