PC keeps crashing

[Fargo]

It keeps cutting out after about 10 minutes and then trying to turn itself back on, seems to stop if I remove some of the ram thou. I'm guessing its the psu which is 500w. I'm running a q6600, 6 gigs of ram, and an ati 4870 1 gig. Please help

[briareos_kerensky]

Or an overheating CPU. The BIOS cuts off the power from the PC and turns itself back on when the temperature has dropped at an acceptable level. However, for such a drastic measure on a Q6600 the heat sink should be undersized or assembled in an incredibly bad way (including the lack of thermal paste)...or some sort of failure of the motherboard's thermal diode.
If the PSU is branded, it must be defective to be unable to power such a PC when idling, if it's a no-brand PSU it might be true that it won't be able to support the machine once something (a CPU usage spike, HDD accessing or whatever) kicks in.
You might also want to rule out the RAM by running Memtest on a single module at a time, just to be sure.

[VR46]

Got to be lack of power or overheating.

There is a site that calculated how much power you needed (you entered your set up via a series of dropdowns) but I can't remember what the URL is sorry.

EDIT: I type too slow

[Fargo]

Would the ram require much more power in dual channel mode than it would otherwise ? I'm starting to think it must the PSU . Thanks for the quick responses guys

[briareos_kerensky]

Nope, dual or triple channel don't affect power requirements.

[Matt]

If it's overheating you won't be able to turn it back on straight away. If it's power it's more likely to only be cutting out in games or something that draws a lot of power.

If neither of those apply, and removing a bit of RAM helps, I'd say it's possible the RAM isn't getting enough power. Go to the BIOS and give the RAM a LITTLE more power. Not a lot, start with the smaller increment and see if that helps. Read up on your RAM and mobo first to see how high you can safely go. To keep my overclocked system stable (though it needs more RAM voltage even when not overclocked) I've gone up .15 volts I THINK. I may have a decimal place wrong somewhere so read up first

[abigsmurf]

my first guess would be the PSU. Bad memory causes blue screens more often than random rebooting.

Is the PSU unbranded or old? Both of those reduce the power it can put out. Run Speedfan, look at your temps and look at the voltages.

Voltages should typically be 12v: 11.9v, -12v: -12.1v etc. if you're getting stuff like 11.3v for 12v that's a sign it's not delivering enough power.

May also be worth checking the wiring. Your PSU may be capable of delivering 500W but not all of it down one 12V cable. If you've more than one cable, try giving the gfx card it's own cable with nothing else on it (it'll be by far the biggest draw). 20A down one of those cables is a big ask.

Usually for temp issues you get blue screens or artifacts on the screen. Look at Speed fan. Ideally your CPU should be running at 65c or so. They can safely run all the time at up to around 85c. Anything above that indicates cooling problems. GPUs tend to run hotter but I'd still say anything above 90 is problematic.

[Fargo]

The PSU is a brand new antec. It wont boot windows or even the bios now it just cuts out after a few seconds. The funny thing is when I had first put it together on monday night it was on for 2 hours or more, long enough to install both windows and crysis I even played crysis for a few minutes to see how it ran. The problems only started when I opened the case again because I had placed the ram in the wrong order and it wasn't running at full speed. Seems strange that if it was the psu it seemed fine at first. I think that over the weekend I'll try a 1000w psu and just rebuild the whole thing maybe add another fan to rule out overheating but I doubt thats it anyway. Thanks for the tips guys

[Matt]

Give your RAM a little boost first. It won't hurt it, and you may be surprised. RAM uses very little voltage, so removing a stick and the system then working isn't a sign you've not got enough total voltage.

I've had this happen to me; 1 stick fine, 2 sticks rebooting, sometimes just after the BIOS. Voltage boost and it was fine. If it doesn't work, you've lost no more than a few minutes.

[Fargo]

Ram voltage was set to auto so I changed it to 1.8v which doesn't seem to help . I'll see if the graphics card cable is also connected to something else next. BTW has anyone here heard of a PSU brand called antrix ? at least thats what I think it was called. I saw a 750w PSU for 56 euro today and was tempted to buy but everyone says to stear clear of cheap brand names.

[Matt]

Antec? Good brand. Don't buy cheap, look around the £70 or so for a decent PSU given how much power you seem to want. I'm not sure it is the PSU though.

Daft question, but have you ran a RAM test?

I think I had my RAM up to 1.85 to be stable BTW.

[Fargo]

Installed a new psu this morning, 720w up from 500w. It still freezes but the system doesn't restart instead it just locks up and the mouse keyboard etc go dead. In the bios I get a message stating the crash is due to overclocking or voltage changes. Any more ideas ? Ram seems fine btw

[NemesiS]

Most of the advice here pretty much covers your issues, but one thing i would do now you have replaced the PSU is, remove the heatsink and CPU. Clean the CPU and reapply some Arctic Silver thermal paste and re-seat the heatsink/fan. It could be that simple.

I would download a CPU overclocker software and reset it to the deafult if it has been overclocked.
Use Systool form here: http://www.techpowerup.com/systool/

A system hang with an overclocked CPU can also be slow memory.

In Bios as well, make sure your fans are operating at max to keep the system cool.

[Fargo]

It was a bad stick of ram . Did a second ram test last night I must have mixed up the sticks the first time I ran the test.