PC switching off and rebooting randomly... all of a sudden

[manisandher]

Hi y'all! Sorry for my 'radio silence' these last few months - the summer has been crazy with lots and lots of work travel. But I'm now more settled and eager to finally find the time to upgrade to XX2.02. However, I have a problem with my PC that I need to solve first...

I've been using a Teradak ATX 275 linear power supply for over 6 months now. For most of this time, the PC has been simply left switched on (it sits in the basement and I've usually forgotten to switch it off even when there's been no one at home for a few weeks). And I've had zero problems... until a few days ago.

Playing music in Unattended, the PC now randomly switches off and reboots. This behaviour really does seem to be random. Sometimes, the PC will switch off during the reboot cycle, and sometimes not for 20-30 minutes of totally normal PC behaviour (including playing music). The PC is currently on right now as I type this post. It's sitting quite happily in Normal OS, doing nothing. It's been on for about an hour or so (just as a test). It seems totally happy. As I said, the switching/rebooting seems random.

When the PC switches off, one of the output LEDs on the Teradak PSU (I think it's the 3.3v output, but could be one of the 12v outputs - I'm not entirely sure) switches off. But I have no idea why the output is switching off. One hypothesis is that something in the PC is drawing too much current.

I have attached the output specs of the Teradak. Here is the hardware configuration of my PC:

- ASRock x79 Extreme4 mobo
- i7-3960X CPU, TDP 130W (running at default 3.3GHz) (water-cooled)
- 4x 4GB DDR3-2400 RAM modules
- 120GB SSD (OS and XX only - no music)
- LAN port (for NAS drive and RDC)
- PCIe Silverstone USB card
- *no GPU card* - RDC only (although PC has booted with GPU card installed no problem, when I've had to access the BIOS in the past)
- *no keyboard* - RDC only

The obvious reason would be that the Teradak is simply too weak to power the PC. If the PC had never ever managed to pass the boot up cycle, I would suspect this too. But the PC has booted up, many times with the Teradak. And everything has worked perfectly for over 6 months. The PC is working right now... but may switch off and reboot at any time.

What do you think might be happening? Let me know...

Cheers, Mani.

[PeterSt]

Hi there Mani,

All I can tell from personal experience is this :

At some stage my "XXHighEnd PC" (with very similar MoBo) started to reboot too. Normal switching PSU in there. However, this started to happen when I, after a year or so, set back the processor to a more normal speed (but still only 2GHz IIRC). Because I noticed this (the change as the cause) I set back the speed to my regular 430MHz and all was fine again (720MHz also worked fine).

Somehow I had the hunch that it was related to the XXHighEnd version of the time but nobody else complained and I couldn't really tell because I was using the 430MHz for a very long time. All I know is that prior to that, I used the normal 3.2GHz speed (I even used 4GHz in the beginning). This has worked for "months".

The processor speed can not be related to the 3.3V. Still you could try to slow down the speed (use the Ratio field in the BIOS).

Regards,
Peter

[manisandher]

Thanks Peter, really helpful. I'll try reducing the speed in the BIOS (will have to reinstall the GPU card).

Here's something else that I've lived with a a while (way before the Teradak arrived): Whenever I try any of the shortcut keys in Unattended (e.g. Alt-U, Alt-D, Alt-N, etc) the PC will switch off. (I'm sure I reported this when it first started happening, but it was so long ago that I just learned to live with it - simply coming out of Unattended whenever I want to do something.)

Could there be a connection?

Mani.

PS. PC still working happily.

[PeterSt]

Mani, hmm ... I don't recall any reporting about a PC switching off at Alt-D/-U. Especially the Alt-U could be sabotage by your wife ...

Does this also happen when you use the On-screen keyboard (with the Alt-U etc. on it) ?

But it looks like something operating at the limit. I mean, any of those short cuts imply the start-up of XXHighEnd, which requires some (relatively much) additional power.

Peter

[manisandher]

I don't think I've ever used the on-screen keyboard. But certainly, before my current issues, Attended has worked perfectly. I've been able to change volume or go to next/previous track no problem.

Peter, how much would XTweaks have a part in all of this? My current settings are:

- balanced load = 43
- nervousness rate = 1
- cool when idle = 1 (this always sounded better to me)
- provide stable power = 1
- utilize cores always = 1

I ask because I've had the PC switched on in Normal OS pretty much all morning with no issues. (I'm going to play music in Normal OS and Attended and see what happens.)

Mani.

[manisandher]

So, music is now playing in Normal OS and Attended... Let's see how long it lasts.

Another thing: I've noticed that in MinOS and Unattended, I always loose Win8 Stability (in settings) after playing music. In MinOS Attended, Win8 Stability remains 'stable' once set.

(If the PC remains switched on in Normal OS (I have a feeling it will), I will go ahead and download and install XX2.02. Who knows, maybe things will fix themselves?)

Mani.

[AlainGr]

Hi Mani,

About the Win8 stability, each time you start a session, the default does not necessarily reflect the status. To be sure, you need to press "Refresh" in the settings, so it will check and show the real status. Mine is "stable", but when in doubt, I press "Refresh" and then I can see if it is working or not.

Hope this helps,

Alain

[manisandher]

Thanks Alain. Just for my understanding, when you say "start a session", do you mean whenever you boot up the PC and start XX? If so, then I totally understand.

But what happens if you press 'refresh' and then play music in Unattended? When you press Alt-X and come out of Unattended, does the settings area still show Win8 as 'stable'? For me, it will always say 'use refresh', even though I did this before playing music in Unattended.

Mani.

[PeterSt]

Mani, the Refresh tells that you just need to do that. So actually it is that simple (to begin with).
And it will do that per XXHighEnd session (startup).

Peter

[PeterSt]

Peter, how much would XTweaks have a part in all of this? My current settings are:

- balanced load = 43
- nervousness rate = 1
- cool when idle = 1 (this always sounded better to me)
- provide stable power = 1
- utilize cores always = 1

I ask because I've had the PC switched on in Normal OS pretty much all morning with no issues. (I'm going to play music in Normal OS and Attended and see what happens.)

Quite crucial ?
But anyway you have the settings right for a low power (consumption) profile.
What remains is the Q3,4,5 = 1/1/1 setting, which will theoretically be the most demanding, but which you won't notice by the given XTweaks settings. Still you could try a 0/0/0 to see whether it matters.

Regards,
Peter

[CoenP]

Hi Mani,

I experienced a similar issue with my pc somewhat out of the blue, but not completely. My sudden and unsuccesfull reboots started a while after I rearranged the harddisks on the sata ports.

The root of the problem turned out to be a poor sata connection on the OS harddisk. No more problems since I 'fixed' it by placing the cable in such a way that it has some tension to keep it connected. A 'spit and tape' solution for now.

I would check the sata connection(s) just to be shure there are not involved in the problem.

Regards, Coen

[manisandher]

Thanks for the idea Coen. But it seems the problem is now fixed. (I'll elaborate in the next post.)

Mani.

[manisandher]

I had my PC on all morning in Normal OS mode. There seemed to be no issues whatsoever. And no issues playing music through XX Attended.

I then took the leap of installing XX2.02, whilst the PC was behaving itself. I've had music playing through XX2.02 MinOS/Unattended for a couple of hours now, and again, no issues whatsoever. (My thoughts on the SQ to come in a separate thread/post.)

So what could have caused the issue in the first place? A few thoughts:

- A slightly loose connection somewhere, as Coen suggested? Maybe, but I haven't touched the PC physically (it sits in the basement), either to cause the issue or to fix it.

- A cooling issue? I'm using a Zalman Reserator 1 v2 water-cooler. It's a totally closed system, so I'd be surprised if there were any clogging. And I've been careful to use the correct quantity of anti-corrosion coolant.

- Something in the air? Well, we had a very mild thunder storm over the weekend. I had to reset my ADSL router, though I didn't associate this at the time with the mild storm, and am really not sure they are connected.

- Just one of those things? Yeah, probably.

Mani.

[manisandher]

Just a final point to wrap things up for now. The PC worked fine all day yesterday. At the end of the day I went down into the basement to just check on things and I noticed something that I hadn't noticed before (because I'd simply never checked) - the Teradak's heatsinks were very hot. I suspect this has always been the case and that with the PC sitting in the basement, I'd simply never noticed before. It might well be that some components in the Teradak are 'aging' and losing their tolerance. I don't know.

But I just wanted to mention this in case anyone else is thinking of going the linear power supply route. Depending on how you use XX in MinOS mode, your PC might be quite 'highly strung', placing a higher demand on your PSU (speculation on my part). In which case a SMPS might be a better option.

At some point I will try configuring XX to place less stress on the PC. The SQ will no doubt take a hit. I will then have to make a decision as to whether to keep the LPS and have a sub-optimal XX config, or use a SMPS and use an optimal XX config.

There's a lot of speculation in this post on my part, so please bear this in mind.

Mani.

[AlainGr]

Hi Mani,

I would tend to make the same assumptions as you. As you mentioned before, you left your PC on for quite some time and it was still on when you started listening to music. Then it started to act like you described under usage, so I would have thought the same.

Regards,

Alain

[PeterSt]

Thank you for sharing that, Mani.

Peter

[PeterSt]

Mani,

I have a Teradak ATX 350 running for 3 continuous days now, and the heatsinks are 30 Celcius. I did not measure, but noticed that when the sun is shining on the unit, it gets some warmer (estimate : 40C). Ambient is just over 20C.
During playback nothing changes but if anything it gets cooler (this is no surprise to me).

Keep in mind that my processor runs at 430MHz and that this sure can matter. When idle, this 430MHz is to be compared with 1200MHz in normal circumstances (when you never changed the processor speed in the BIOS). Do notice that when Playback is going on, this is to be regarded as "Idle" (nothing will trottle-up because of it).


What can have happened in your situation is that something got in a loop during your not being present and that one of the processor cores went to 100% - even in turbo mode. Now that core runs at 3800MHz continuously, which your cooling system may be capable of (to keep all sufficiently cool) but which draws the (thermal) 135W of the processor continuously. And *that* will make the PSU hot for (quite) sure.

If you, next time, see the PSU so hot, try to observe TaskManager - Performance Tab and see whether one of the cores is at 100% (this shows as 8.33% with 12 cores ( 100 / 8 )).

Regards,
Peter

[manisandher]

Hi Peter, thanks for this.

I did take a quick look at Performance in Task Manager when the PC first started switching off, but couldn't see anything untoward. But there's obviously something not right at my end, with such high heat sink temperatures. I'm going to reduce the speed in the BIOS and also take a look at temperature readings in general (CPU, etc) whilst I'm in the BIOS. (Will need to reinstall my GPU card to do this - PITA.)

Thanks again.

Mani.

[PeterSt]

Ok Mani - Don't hesitate to ask me to try something if you feel suspicious about "it". I have the unit here up to Sunday.

Regards,
Peter