What's going on with this?

[Puma44]

Logged on about 30 minutes ago and played for about 10 minutes with every thing normal.  Then a very loud constant howl, the fans running at high speed, and a screen freeze.  Nothing would respond.  No cursor, escape, ctl-alt-del.  I had to power down by pressing and holding the power button.  I waited about about 5 minutes and turned the computer on.  The monitor displayed the normal "Checking for signal" for a few seconds and then displayed "No signal, entering power save mode" and turned off.  I shut down with the power button again, waited about 5 minutes and got the same result.

I've never experienced this before.  Any ideas?  Thanks!

[1stpar3]

  No idea...sort of scared too. Doesnt sound very good. Hope its nothing serious. Closest thing I could think of was, this sort of sounds like my old system trying to run AH3. It would load up and be ok till you got to runway, then would shut off due to over heat? Hope that isnt your problem, as you have been playing fine with VR(if memory serves me and we already know about the old memory thing ), that would really suck. So I aint even gonna think about it

[Puma44]

I upgraded to a 1070 card about three months ago and it been running the game just fine.  I'm not into VR, yet.  The last couple of weeks, there has been a slight reduction in frame rate.   But, game play has been normal.

I wonder if it's a power supply issue.  The machine still has its original 750W onboard.  I've never experienced a psi failure.  So, I don't know what I don't know.

[Bizman]

To me the symptoms sound like an overheating processor. Back in the day when overheat protection was made by Intel only, there was a YouTube video about taking the cooler off. The Intel CPU just froze, no response. The AMD died smoking in two seconds. After a while they put the cooler back on the Intel which then continued normally.

What I'm trying to say is that you should check that your CPU cooler sits firmly in its place. Coolers using plastic prongs are somewhat tricky to install and one corner can get loose during transportation or working inside the case - like during a video card change.

[Skuzzy]

Could be overheating, could be power supply.

Heat is easy enough.  Just open the case and point a small fan at the internals and see if the problem occurs.  If not, then heat may be the culprit.

[Puma44]

Thanks Skuzzy and Bizman!  I'll check both and let you know.

[Puma44]

I just checked both of your suggestions.  The cooling fan unit (liquid) is firmly seated. I placed a small fan aimed at the internals, turned the computer on, and the cooling fan went immediately to high speed, like it has been.  The the power supply and cooling fans appear to be spinning freely.  No indication of debris in either fan housing.

I also noticed that the two fans on the 1070 card were intermittently and independently starting to spin and then shut down.  Is this normal or an indication of of a problem?

Thanks for your help gents!

[Bizman]

I suppose that the 1070 fans spin only when needed. IIRC there's been some discussion about them not spinning when idle.

You said that the cooling fan unit is firmly seated. Do you mean the initial block on the processor or the radiator which cools the liquid? The seating of the latter is of less importance as long as there's some airflow through the radiator. Now that you mentioned you have liquid cooling another potential issue raises its head: A minor leak or manufacturing failure may have left a bubble into the system. If such a bubble gets stuck in the CPU block it may cause temporary overheating, making the computer to freeze for protection.

Check the seating of the CPU block. Checking the block and tubes for minor cracks might also be worth the effort. Someone once reported having a minor crack in the block so it's plausible. Further, since the CPU runs quite hot, a tiny leak may vaporise the leaking liquid before it causes any electrical failures.

[Arlo]

Just developed power supply surge issues tonight. One year old case and power supply. AHIII triggers it. Progressively worse (from 15 mins in flight - offline .... to triggering on start-up splash screen). But consistant with attempts to run AHIII (6 attempts). System is identifying shut down triggered by power supply surge safety measures. Have yet to see if all graphic intensive games do this. Will get back later this week to test (after eclipse camp out in Casper).

[Skuzzy]

If you suspect power supply, first thing to do is check the voltages in the BIOS.  It is not always indicated, but of the voltages are way off it is a clear indicator the supply has gone bad.

Careful with bad power supplies, they can take everything down with them.

[Arlo]

It's not AHIII. It crashed the system with Battlefield I (a game I bought but never play because it requires too much umpf for online play). After that the system wouldn't even handle the online browser. I pulled the STRIX video card and I'm just running the motherboard video (ASUS). It can't handle AHIII (not even the splash screen - just white) but it doesn't crash the system (ctrl alt del out of AHIII, though). I'll buy the rest of my new build when I get back from Casper Wednesdayish and make sure the new box/ps can handle the STRIX card (or - rubs hands - new video card ..... the wife has okayed - ).

[Puma44]

Bizman - I've thoroughly checked the cooling unit and see no obvious (to me) issues or disconnects.

Skuzzy - I'm not sure how to check bios voltages.  I assume that the computer has to boot up in order to check bios voltages.  At this point, I can't get there.  When I turn the computer on, the fans go immediately to high speed, the monitor shows "checking for signal" for a few seconds, then shows "going to power saving mode" and shuts down.  At the same time this is happening, the fans are constantly running at high speed and continue to do so until I shut it down with the power button. 

On the back of the power supply I discovered a green led light that stays illuminated steady while the computer is powered ON.  Additionally, there is a small button that I assume is to power test the PSU when the computer is OFF.  With the computer powered OFF, the fans run when this button is depressed.  Fans shut down when the button is released.

I currently have a 750W power supply.  The 1070 card requires at 500W.  If the power supply is the isssue, should I stay with 750W or go with a higher wattage supply.  If so, how much?

Thanks very much for your help and expertise gents.  Once again, I am in the land of "I don't know what I don't know". 

[Bizman]

This is just a wild guess and worth the $1 it costs: Replace the BIOS battery. As I said, it's a wild guess, but as of yesterday I fixed a computer that wouldn't start at all by replacing it.

However, the howling and the following freeze don't fit. It sounds too much like a heat issue: The system notices overheating and tries to fix it by revving the fans. Since every component seems to be properly seated, are you sure that there's any liquid left in your cooling system? A very easy way to test if inadequate cooling is the issue is to replace the water block with a regular air cooler. For just booting it'll suffice if it fits, you won't be trying to get to the extremities.

As for a new power supply, power doesn't tell everything. As you said, the 1070 requires a 500W one. Add some overhead for future aging, let's say 600 to 650W. The more important aspect is the quality of the PSU. Seasonic has a very good reputation, they build PSU's both for themselves and to others. 80+ Gold or Platinum is also somewhat of a quality proof. Plain wattage is not a proof of quality, there's too many ways to measure it. If you like reading: http://www.tomshardware.com/reviews/psu-buying-guide,2916.html

[Puma44]

Thanks Bizman!  I've never replaced a BIOS battery.  In fact, until now didn't realize there is one  . Where is it typically located?

As far as the heating issue is concerned, does it make sense for the fans to immediately go to high speed as soon as it's turned on first thing in the morning after nearly 24 hrs being turned off?  I'm not discounting your advice and expertise, just trying to advance my Neanderthal level of computer knowledge.  Is the possibility of a heating issue logically connect with nothing getting transmitted to the monitor?  Or could it be the PSU, or a combination of both?

Understand about the cooling block.  Since I'll have to buy one, should it be another liquid cooled or straight air cooled?  Any recommendations on brand, type, etc?

Of the two possible issues, which seems to be the most likely culprit in my legacy computer?

I'm trying hard to keep this dinosaur running effectively.  Insalling the 1070 card was a huge improvement with absolutely no adverse issues for the last couple of months.  When I start talking about a new rig, the wife ack is deployed all over the house. 

Thanks for the article!  It answered a lot of questions.

[Bizman]

Alright... The BIOS battery is a 2032 button cell on the motherboard. Most of the times it's sitting flat, on some motherboards it's sideways. In any case it should pop off without violence. There's a spring in the slot, pushing towards it usually does the trick. Note that your settings will be reset. Usually setting to optimal defaults at the BIOS "save" screen are good. Also, at least when you get back to Windows, your time and date will be way off and need to be set correctly for many functions to work properly.

If it's a heat issue caused by a failed cooler, it's instant. It takes a second or two for the CPU to reach its maximum temperature no matter how long you have had your system unpowered.

The best air cooler is made by Noctua but there's a lot of decent ones from other makers, too. Look for airflow versus dB. However, if you aren't going to overclock, basically anything will work. Even the factory coolers are efficient enough and if you're lucky you can get one for next to free. As I understand your rig isn't of the latest fashion, so it's useless to pay big money for the cooler in any case. My Intel e8500 from 2008 overclocks nicely with a $40 cooler. Air cooling is much more fail safe than liquid. If the fan stops whirling, that's visible and easy to fix. With liquid it's either a loss of liquid, blocked circulation, failing fan...

After all this being said, it still can be a PSU issue or a combination of one or several other factors - some of which may be caused by the failing PSU. Because of the fast spinning fans I still vote for insufficient cooling, which also is the easiest and cheapest to replace for testing.