"The RAM passed all Memtest86+ test running for 12 hours or so but as soon as Windows was booted things started to become quite unstable".
Even though you're likely not saddled with one or more of those "trick" FPGA'd modules, if you're so inclined, you may still care to give the old MS WinDiag memory diagnostic a try. In spite of it not being as fancy as Memtest86+, it often flags modules as bad that sail right past Memtest86+ as being A-OK.
If you do decide to play around with WinDiag, note that if you hit the *P* (for pause), a new option will appear at the top of the screen: The *M* option (for menu). You can then hit M -> 2 (advanced) -> 1 (change cache settings) -> 2 (turn off caching for all tests). This allows you isolate if there are any CPU cache related issues, which are usually non-obvious.
There's also a "change the test suite" option that will let you perform some additional/more rigorous tests.
One other thing I like to do while running memory diagnostics is take a hair dryer and deliberately heat up the DIMMs under test. (Though you have to be careful to not overheat them to excess). As they heat up, the CL-timings tend to shift, sometimes to an out-of-bounds value. This will usually show up on-screen for several test loops before you crash. <grin> This approach is far more likely to find a marginal DIMM than running a diagnostic by itself for days at a time.
Have fun!
I sent Scan a message on Sunday which they finally replied to yesterday. They asked if I'd ran memtest and said they won't change the RAM for something else, just replace it. Tried memtest again last night with just the dodgy RAM in and it said there was a problem straight away.
Waiting to hear back from Scan from yesterday now.
Oh FFS! I put the other RAM in the computer the night before last because I got fed up with how slow it was running with only 2GB (does Windows do anything special when using 4GB that would make it run like shit if 2GB is taken away? It's running so slow and thrashing the HDD constantly, the old computer wasn't like this with 2GB) but it started crashing straight away so turned it off on the PSU.
Tried last night to turn it back on, lights came on on the card reader but the power button did nothing. Turned it on and off but just couldn't get it to boot. Tried this morning and as soon as I turned the PSU on there was a lovely fop sound. I have a feeling it might have died :(
Rather than re-explain the basic diagnostic procedure for a box that appears dead, see this post of mine on the PCmag board. Let us know what you find.
I'll have a look through that later when I'm at home. Fairly sure the PSU has gone to heaven or maybe just a blown fuse. The PSU has a light on the power switch and that didn't come on after the pop something made.
I did have to replace the first PSU from Hiper which died after about two years and it looks like the replacement has lasted about the same time. I'll check it when I go home later.
I've been looking at PSUs and it seems I have to go above 500W to get two PCI-E connectors and a P8 connector so I'm thinking either of these:
-Thermaltake 575W Toughpower XT Modular PSU
-OCZ Fatal1ty Series 550W Modular PSU
The first is a little more expensive but mentions solid state capacitors and has a 5 year warranty instead of 3.
Because I was in a rush to make my previous post I skipped over your question, so I want to back up a little.
"...does Windows do anything special when using 4GB that would make it run like shit if 2GB is taken away? It's running so slow and thrashing the HDD constantly".
Not particularly. Obviously, it depends on what's loading at startup. If what's loaded consumes more than 2GB, the overflow is pushed out to the swapfile/pagefile on the HDD, which would account for all the drive thrashing.
To get a cursory view of what's running (and how much memory is consumed by each item) you can use the Task Manager and/or Resource Monitor. For how to interpret various indicators see Measuring memory usage in Windows 7. For a more granular view you can download the free Process Explorer from MS's SysInternals division. It's like Task Manager on steroids, and will let you identify hidden processes nested by either SvcHost.exe or RunDll32.exe.
You can use the MSconfig utility to prevent unneeded/unused programs from loading at startup. (AutoRuns is the free steroids version of MSconfig). You can do the same for Windows Services; have a look at the Black Viper's site. Be sure to see his Win7 SP1 Service Configurations page. By disabling unneeded/unused services and programs you can significantly shave your memory footprint.
You may also care to see RAM, Virtual Memory, Pagefile and all that stuff, which should give you a better idea of their relationship to each other. Various versions had been previously published by MS as KB2267427 and KB555223, both of which have been removed.
Moving on to your immediate problem, unless you know for sure that the "lovely fop sound" you heard came from the PSU, I would be cautious. If you can, try substituting another known-good supply before buying a new one. You want to make sure the supply you had been using didn't take anything else with it for the ride over the cliff.
I'm assuming you've run your requirements by a few of those PSU Calculators? (Thermaltake has their own calculator).
The PSU Stickies at the Overclockers forum are also a good resource.
The 575 Watt Thermaltake Toughpower XT (P/N TPX-575M) has good specs all around. It's a similar design to the Corsair TX650, so you may want to look at a few PSUs from Corsair.
I also took a look at the 550 Watt OCZ Fatal1ty Series.
Personally, I would go with the Thermaltake. Not only for the better warranty (which is a good measure of the vendor's confidence level and expected MTTF/MTBF), and the solid dielectric capacitors. But, according to Newegg, it's also EPS12V (v2.91) compliant. Because EPS12V supplies are often employed for high-reliability server use they use a failsafe design that minimizes the likelihood of follow-on (domino effect) damage of anything connected to them.
From my perspective, unless it's a throwaway box, the PSU is not the place to pinch pennies.
Something else to consider is what I call the "dust bunny margin". What many used to do is deliberately oversize the supply to obtain additional thermal margin before having to clean out the dust bunnies. Unfortunately, with many new, high-efficiency models (compared to older, lower wattage units) there is often no increase in heatsink size or fan CFM when going to a higher wattage model. The result being there's no increase in dust bunny margin before overheating occurs. So you wind up having to clean it out with greater frequency.
It's worth noting that excess heat can significantly reduce the life of most passive components. For instance, electrolytic capacitors are typically rated for use at a nominal temperature of 25°C. For every 10°C above that temperature the rated MTTF (i.e., expected life) will go down by half. So if a part is rated for 50K hours at 25°C, it's only going to get 25K hours at 35°C. Just something to keep in mind when premature failures are encountered.
Unfortunately I don't have a spare PSU with a P8 or any PCI-E connectors. I've had two previous PSUs blow up, one took out the motherboard with it and the other caused no collateral damage. The PSU calculator suggests 412W with some capacitor aging so shouldn't be overloading it.
Actually I had trouble starting the old computer recently with this PSU so wouldn't be surprised if it had died..
I do usually do a bit of optimisation to reduce memory usage as much as I can which I didn't do with the 4GB installed as it was running super smooth anyway so could just be it's using more memory than usual and using the swap file lots.
I've taken the PSU out and opened it up (without electrocuting myself this time (although that was about 10 years ago)) and it looks like the PSU suffered the same.fate as the old motherboard, leaky capacitors. Probably why it was playing up.now and again. Not sure what went pop though, nothing obvious.
So need to send the RAM back (should Scan pay the postage?) and order a new PSU and hopefully I will have a stable computer at last.
"Not sure what went pop though, nothing obvious".
Most electrolytic caps have a safety vent on their bottom. You will hear it blow, but there will usually be no visible signs on the PCB top side, other than possibly some expulsed electrolyte around the cap's sides.
"I was quite surprised how much dust was inside the PSU which certainly wouldn't have helped it's lifespan".
Hence, my mentioning the "dust bunny margin".
"And unlike the inside of the computer it's not something that can be removed easily with the little warranty stickers they put on the screws".
Which is why I either ignore those stickers (as more than a year is really too long to wait) or put a sticky note, having the warranty expiry date on the case's outside, reminding me to do a cleaning as soon as possible thereafter.
It also helps if you have a case that has a removable filter on the air-intake, like some Lian-Li models.
The vent is supposed to be the scored mark on the top and the plug at the bottom. Not sure how that works on the bigger ones though because they don't have the ridges on the top, or at least the ones in my PSU. There are some pictures on wiki showing exploded caps where the whole casing has come off and showing the insides.
Looks like I have 5 of them dead. 2 bulging, 2 leaking a little out the top and 1 where the bottom plug has been forced out. Definitely dead, but I felt a bit odd pulling the caps off it.
"The vent is supposed to be the scored mark on the top and the plug at the bottom".
Most older designs used the scored topside as a backup. A few vendors used it as the sole venting means. But most just have the bottomside vent plug.
Between 2001 and 2003 the market was flooded with bad caps. The problem had serious impact on both MoBo and PSU vendors, among others. A lot of products containing bad or marginal caps are still out there.
At the time, it was written up in the Electrical Engineering Times and other trade rags. Rather than go digging for those old, dry technical articles, have a look at PCmag's Hushed-Up Disaster Dept.
To better help identify bad caps, be they on the MoBo or in a PSU, see the following:
• Bad System Board Capacitors!
• How to identify bad capacitors
The above have plenty of good close-up photos.
"Looks like I have 5 of them dead. 2 bulging, 2 leaking a little out the top and 1 where the bottom plug has been forced out".
Sounds as though your PSU was long overdue for replacement.