PC sometimes not powering up

[bytheway_r]

First of all, I wasn't sure where to write about this. Hope here's ok. I've been having this issue since I moved ( studies in a different city ). Every now and then when I turn on the power strip, then click the power button on the PC it does nothing. Only after a while and several "resets" of the power strip does it work. The PSU is receiving power ( or enough to light up the diode in it's on/off switch ) but the PC is not responding to "power" and "reset" buttons. It's behaving like it's dead, no reaction whatsoever.

Now, I have some ideas myself as to what could be causing this but maybe there's something that's more likely to be the culprit than the rest.

First of all, there's a fuse in the power strip. It could've gone bad. Other than that, I'll be the first to admit that my PSU is a little low on power for what I'm running. Still, the fact that it's running fine when it's actually working is kind of baffling me. Not sure about the fuse but the PSU would likely cause some problems when the PC is stressed, right?

I'm back home for the next few days, so I won't be able to check anything but let me know if you have any other ideas. Motherboard or CPU problem?

I'll probably start with plugging the PC directly into the socket to rule out the fuse.

System:

MB: Gigabyte GA-MA770T-UD3P rev. 1.0, F10 bios
CPU: Phenom II x3 720BE -> x4 3.5Ghz
GPU: Powercolor HD 4870 512MB PCS+
RAM: 2x2GB OCZ Intel XMP 1600Mhz CL8 DDR3 -> 1333Mhz 6-6-6
PSU: OCZ StealthXStream 500W
OS: Windows 7 64bit

sound card: Asus Xonar D1

I've decided to mention the sound card as it has a few quirks, like a slight flicker of the screen when it's turning on during Windows boot.

[Dark Mantis]

HI
Well the most likely thing as you have shipped the computer to another place is a bad connection somewhere, so check them all by unplugging them and then reseating them. As you have already suggested take the power strip out of the equation and if possible even try another mains socket in case the socket itself is faulty which is quite common. It is far more likely to be something like this because it has been moved so check these first.

[bytheway_r]

Thanks for your insight. Yeah, I've started wondering today about the connections being loose or something. Guess I have a few things to check once I get back .

[bytheway_r]

Update:

1. Fuses checked and as far as I can tell, they're ok. I could try replacing them with new ones to be 100% sure.
2. Connections checked.

I haven't had this problem since checking the fuses so can't say if maybe taking them out and putting them back in helped. I have another problem now and it could be connected to the other one.

I've started playing Dragon Age and it's the first time this PC is under that kind of stress for up to ~2-3hrs. Now, I've had several crashes, some more, some less serious. I tried pulling back CPU voltage and frequency, RAM overclock and my modification of GPU fan speed. Nothing worked so far. At the moment I've taken off the side panel of my case to see if this will help.

What I figure could be the reason:

- overheating of RAM or CPU power section or NB or PSU
- power delivery problems with PSU, socket or power strip ( like, the PSU cable is twice as big as the power strip cable and I have a 24" LCD and a printer connected to it, too )
- damaged GPU?

Now, I'll add a description of these crashes, maybe there's a clue there:

- most common: game stops and I get the window that it stopped working, blah blah. I can start it again and it'll crash in ~15mins again.
- less common: game stops with a graphics and sound problem. Window pops up, etc. I have to restart the PC as I'm having rather serious display problems and I think ( will check if it happens again ) the sound's not working either.
- happened once: got a BSOD

I'm not too sure this could be caused by overheating, though air inside gets rather warm. Hard drive temp goes from 33 degrees in idle to 40 degrees after an hour of playing.

Any ideas?

All I can come up with is either overheating somewhere or power delivery problems or something dying on me.

[Dark Mantis]

The fiorst and easiest test is to remove the power strip. These are mainly cheap and nasty and a problem waiting to happen so get shot and plug directly into the mains source.

You haven't said what CPU heatsink/fan assembly you are using but make sure it is firmly fixed all round.

Try removing/refitting the graphics card and power cables.

[bytheway_r]

I'll try it, though it may be tricky to reach the socket .

CPU heatsink problem is out as it's a Mugen II that's mounted rock solid onto the MB. More importantly, CPU temps didn't go over 45 degrees during playing.

I've taken out the graphics card when checking that previous problem so that's covered, too.

Now that I think about it, I'm leaning towards power problems somewhere. I can't really reach the power section with the CPU heatsink in place but the NB doesn't get HOT. It's warm, sure, almost uncomfortably so but not burning.

Considering that at home I used a different and from the looks of it, better power strip without powering up issues... Well, this could be it, indeed.

Anyway, with these temps during gaming I'm thinking about an Antec 900 case. I reckon it should be quite a bit better than this cheap-o Cooler Master. There's one used that I have my eyes on.

[Dark Mantis]

CPU heatsink problem is out as it's a Mugen II that's mounted rock solid onto the MB

Make sure that the Mugen II isn't too tight as we have had problems with motherboard distortion when these  larger heatsinks are tightened too much. The tracks can crack and all sorts of nasty things!

... the NB doesn't get HOT. It's warm, sure, almost uncomfortably so but not burning.

You could always just fit an auxilliary fan in place to blow onto the northbridge to cool it a little.

Personally my money would be on the power though.

[CanUK]

Hey,

Just saying that I currently own a Antec 900 case, so if you have any questions about fitment/space issues.

[bytheway_r]

I seem to have interesting problems with my PC. Never something that can be fixed easily .

The PSU's pluged into the socket now and the problem's still there. Granted, it's the same socket as before so that could be the reason.

What else... with case open RAM, CPU, PSU, GPU overheating is out of question. They're all running cool. I can run it through Intel BurnTest without any problems. At least short term.

From what I could feel around, the power section is also only slightly warm. On the other hand, after some more playing around with the NB it can get to the point of being close to burning ( if you touch it for 3sec+ ).

Except for some really unusual reasons I can guess at:

- wall socket issues with high power draw ( though this seems unlikely to me );
- NB overheating;
- PSU failing;
- GPU failing;

I'm actually surprised that the NB gets this hot. It's running at stock volts, too... I'm seriously starting to consider a better heatsink for it. I'm a little worried about what Gigabyte was thinking when designing this board.

I'll probably try remounting the GPU later on. To be 100% sure it's not that.

P.S. Don't worry about Mugen. I did some reading before installing it and I really doubt that I screwed it on too tight. Let me just tell you that it was a royal pain in the behind to mount it on the CPU. Glad I had someone to hold the heatsink still. I have no idea how you're supposed to do it by yourself.

Thanks, CanUK! I doubt I'll have any problems with space, though. My current case is smaller ( or, at least, it's not bigger ) so it should be a breeze if I pull the trigger on it.

EDIT:

One last thing. The PC/game seems to crash sooner with higher CPU frequency applied. Generally, these crashes are highly inconsistent and with same settings can happen after 10-20mins just as well as after 2hrs+.

Guess I'll also check with CPU running 100% stock settings to rule it out for sure.

[Dark Mantis]

I would say the easiest thing would be to get  a desk fan or similar small fan and just point it towards the northbridge at a good speed and see if the problem still occurs. If it doesn't then you have your culprit. It does sound like it is overheating if it is getting that hot.

[bytheway_r]

Ok, so I'd say that the problem is with either power or CPU. I'm afraid it's the CPU but we'll see for sure by the end of weekend. As I've said in a different thread, I only have half of the pins covered on the CPU power phases.

Anyway, there's Cooler Master HAF 932 and Corsair TX650W that I'll be picking up in a day or 2, so cooling and power wise I should be set.

You know, I don't get that "not really there" reading that says something's running ~80 degrees. I'm not saying it's accurate but after replacing NB heatsink that temp went down from 81-82 degrees ( since changing CPU heatsink ) back to 78-79 degrees. Other than that, I hope I won't crush my NB with this new heatsink. It was hard to gauge how strongly it should be screwed on and that "cushion" on the sides is way too big compared to the one on the original heatsink. Good thing - now the heatsink is about as warm as the mosfets which is a good deal better than before.

[Dark Mantis]

At least the northbridge sound like it is a bit cooler when running now. Let us know how you get on with the new components when they arrive.

[bytheway_r]

I have to say I can't wait to see that monster of case under my desk . I've noticed you have it, too. What's your opinion on the stock fans? On the net some people say they're crap.

Considering that this case was only ~35£ more expensive than that used Antec 900 I think it was well worth the money. That and I've been reacquainted with how bothersome it is to take out an ATX MB out of a small case. Especially if you have some bigger heatsinks on it.

[Dark Mantis]

It is certainly a nice room case and you may be suprised how long it is too. About the fans, to be honest I only have one of the three original 200 mm fans still running(the front) as I water cool and so I have just a 120 mm at the rear and an 80 mm over the power phases as there is no cooling from a CPU fan in that area. The good thing is that unless I want to crank them up (which is never necessary) it is almost silent. I would say the supplied fans are fine but of course it depends on the rest of your system.

[bytheway_r]

Guess I'll continue in this thread.

Took way longer than it was supposed to but I've put the PC together today.

In the end, I've ordered a few more things and will be moving the old case+MB+CPU+PSU back home to replace the old Pentium IV there.

My rig now:

case: HAF 932
MB: Gigabyte GA-MA790XTA-UD4
CPU: Phenom II x2 555 @ x4 B55 running at 3.8Ghz as advertised by the seller ( I'll see if I can push it further later on ) + Mugen II
RAM: OCZ 2x2GB 1600Mhz CL8 1.65v @ 1333Mhz 6-6-6-19-25 1T 1.73v
GPU: HD 4870 512MB PCS+
HDD: Samsung 500GB F3
audio: Asus Xonar D1
PSU: Corsair TX650W

Now I have to figure out how to put the cables in the case as atm I can't put the side panel back on.

All in all I love this case and that sweet little CPU. On the other hand, I've been tempted by a 1055T as it was about just as nice bang per buck.
This HAF 932 is great. HW Monitor reports idle temps:

System - 30C
CPU - 22-23C ( clocked down do 800-1400Mhz )
TMPIN2 ( Northbridge? ) - 30-32C

GPU - 41-43C
HDD - 28C

That's quite a bit better than the previous case. CPU temp is not only lower but also the CPU fan is running 200-300RPM slower ( 400-500RPM ).

Now, I have 3 questions:

1) What's TMPIN2 on this board? It doesn't seem to be a bogus sensor like on the 770t-ud3p. Could it be the NB?
2) Isn't 28 degrees a little too cold for the HDD? I have no idea whether this still holds true but I remember that HDDs used to like temps ~35C best? And below 30 or above 40 they used to fail far more often?
3) RAM voltages above 1.625v are colored red on this board. Is it really any problem for the MB or CPU to run RAM at 1.73v? It's not all that high in the end.