OT: BSOD & PC PSU

[Endulini]

Hi All,

Slightly off-topic but I am looking to fix this myself so I guess it is DIY at a stretch.....

My ten year old PC has been running Windows 10 mostly quite happily for a while now but since the upgrade it's developed a consistent habit of crashing with a BSOD citing the following error "Video_TDR_Failure" and driver nvlddmkm.sys. Updating the drivers and/or reinstalling the Geforce GT630 graphics card didn't make any difference. Changing to a Geforce 8400 GS seemed to reduce the frequency but not eliminate it and it seems to be creeping back up again. There's no common thread to what's happening at the time of the crash and I don't run any particularly challenging software so I can't see why it should be crashing.

A colleague has mentioned looking at the PSU as this is around 10 years old and I understand that they do deteriorate with age but I'm not sure what symptoms to expect if this was an issue.

Googling the problems shows a number of people with similar issues albeit with no resolutions. Running Whocrashed sheds no light on what it happening and there's nothing obvious (to me!) leaping out of Event Viewer.

Any thoughts/ideas/suggestions welcome, if it helps it is on an Asus P5N-E mobo with an Intel Core2 Q6600 processor.

Thanks

[Lee]

The P5N-E is one of those known to suffer from bad caps. Do a google image search for what to look for and then see if any of the caps on the board look domed or leaking.

Of course it could also be PSU, so usual advice, see if you can borrow one to test the system with.

[Bill Wright]

I'd be replacing a ten year old PSU as a matter of course. A little bit of ripple on the outputs can cause all sorts of funnies. Also, get a good blower and clear out all the fluff from the insides of the machine. Have you tried running it with the case off?

Bill

[NY]

I wouldn't advise running a PC with the case off. The case performs a vital job of routing the airflow through the intake ducts, across the hot components and out through the fans. Remove a side and the air that is sucked in by the case fans is not being channelled across the motherboard components and so these can only cool by convection and not also by forced-air cooling.

With my PC I can see the CPU temperature and hard disk temperature increase by about 20 deg C within about ten seconds of removing the side. The CPU fan also starts spinning like crazy, rather than at its normal speed, in an attempt to cool the CPU down.

For cleaning out dust, I'd use a vacuum cleaner rather than a blower so the dust is sucked away and removed, rather than being blown all over the room. Preferably a mini PC vacuum, but failing that, the tapered end attachment on the hose of a normal vac.

[Lee]

Oh dear worms->can :) :) I use a vacuum cleaner though :)

[The Natural Philosopher]

Try removing banks of RAM. Or running memtest.exe on it

It does sound like a bad bit of RAM that just happens to be being used post upgrade that wasn't before.

when you work out which bit is bad, replace it

[Harry Bloomfield]

NY a écrit :

I take mine out to a bench in the garden and blow the dust out with compressed air from my garage supply. Having blown it out, I will often give it a hose soaking with a garden hose, then leave it in the sun to dry out.

[Peter Parry]

Odd crashes for no obvious reason on an old motherboard would certainly point to failing capacitors.

[Bob Eager]

And whether using air or vacuum, hold the fans still. Easy to overspeed them.

[T i m]

I think the trick is to do it outside . I have a portable oilless compressor for such jobs and remember to stop any fans spinning as you clean round / though them (you can poke a cable tie into them to keep them still, especially the one in the PSU).

The only thing I would say with that is it's good for removing the general buildup, say in the bottom of the case but I'm not sure how effective it is at getting the dust out of the heatsinks or PSU etc.

Plus, you may have to be careful of a static buildup (especially with many cleaners having plastic bodies and hoses etc). Maybe a natural bristle brush (with a metal surround and hold the metal) would help move the dust into the path of a cleaner?

My mate runs a PC shop and you would be surprised (or maybe not) just how many machines come in with everything completely bunged up.

The worst looking was one of those clear / Perspex cases and it looked like the inside of a Dyson! ;-)

Cheers, T i m

[Bob Eager]

My son said he was getting a new laptop because the old one had got very slow. Actually, he did so and sold the old one to his mum at a knockdown price.

It needed a new DC jack, so I stripped it down. Removed the felt air filter (about 0.5cm thick) from the air intake. Then realised it was actually just solid fluff.

It went a lot faster after it stopped overheating.

[T i m]

Funny you should say that, here is one I did previously:

They generally do. ;-)

I have sometimes used 'SpeedFan' to see how the temperatures are, before and after cleanout.

Cheers, T i m

[The Natural Philosopher]

Or failing RAM

[Dave Liquorice]

What is the issue with "overspeeding"? I keep seeing it mentioned but no explantion of what the consequencies are.

[Mike Clarke]

Probably that it will act as a generator and apply undesirable volts back onto the motherboard.

[Andy Burns]

If it was a brushed motor it would act directly as a dynamo and overspeed would give overvolts, but most (all?) PC fans nowadays are brushless.

I didn't think much power would find its way backwards through the driver circuitry, but a few minutes later noticed a multimeter within reach of a 14cm fan, hooked it up and spun it by hand, got 2V out of it (obviously it has a capacitor inside as it decayed away slowly).

Tried again with a 4cm fan powered by the vacuum cleaner, managed to get

8.5V and 25mA out of it, seemingly no capacitor in the smaller fan, so I wondered how spiky it was and hooked it to the scope. The waveform is basically a half-wave rectified sine wave with a bit of chopping, but it was 24V peak, so yes I'd say wedge the fans when cleaning dust out of them.

Impressively inefficient, less than 1 watt from a 1600 watt vacuum cleaner, but I've stopped now before the neighbours wonder if I've taken up home dentistry :-)

[The Natural Philosopher]

Welcome to the wonderful world of wind turbines

[Bob Eager]

And possibly not do much good to the bearing.

[Dave Liquorice]

mentioned but

I guess so, some one mentioned 24 V but IIRC that was measured by a scope and thus effectively open circuit. What happens with say a 100 R load resistor?

If the bearing can't take 10 seconds a bit quick it's probably on the way out anyway. Most fans are simple sleeves with a fair bit of metal to take away any heat. Balls are a little different as things can move and bash into each other.