Video hardware error, or is it?

[banger]

Hi

This is the second video card I have had this problem with, First was an MSI GT610 1GDDR3 passive cooled card on a MSI P35 Neo motherboard. Unfortunately hot plugging this card into another monitor while trying to diagnose the problem blew the DVI port so it was replaced with an Asus GT610 1GDDR3 also passively cooled but the problem has re-occured.

I was running an old version of prime95 to stress my E8400 CPU to check load temperatures. But when I closed prime95 the screen went black then came back on with lots of the graphics on programs I was running missing.

Also a bubble popped up saying the nvidia driver had stopped responding and recovered after the black screen. I cant believe two video cards, different makes are faulty. In the reliability section of action centre it says video hardware error. In event viewer it says kernal error at the time of this black screen.

Is it a driver problem or hardware? The motherboard is from 2008 and used 24/7 til now and I have heard the capacitors are not of the highest quality. Below is a screenshot of the reliabilty error details. Any ideas? I am now using the latest prime95 and cant reproduce the problem.

 

[Itaregid]

Unfortunately hot plugging this card into another monitor

FTR, I know of no graphics card or monitor that supports hot swapping. While HDMI cables are supposed to allow hot swapping, in practice, not all equipment does. The problem is the pins are too many and too close and if the user does not perfectly align the plug before connecting, or does not have a rock-steady hand, a short or arcing may occur. I have seen where this practice took out monitors and TVs too. Not just the card (or motherboard, in the case of integrated graphics).

I don't recommend the use of Prime95 for troubleshooting. Prime95 is designed to stress hardware to its breaking point, subjecting your components to abuse, and potentially damaging high heat.

Below is a screenshot of the reliabilty error details. Any ideas? I am now using the latest prime95 and cant reproduce the problem.

If you cannot reproduce the problem, then maybe it was a one-off problem. I note that was from the 28th so I assume it has been working fine since.

The motherboard is from 2008 and used 24/7 til now and I have heard the capacitors are not of the highest quality.

:huh: 24/7 for 7 years running surely indicates there is nothing wrong with the quality of the components used on that board.

There is nothing wrong with passively cooled graphics cards either. Just remember, it is the computer case's responsibility to provide a sufficient supply of cool air flowing through the case. This may be more important for passively cooled cards. And note there are many more heat sensitive devices inside the computer case in need of adequate cooling besides the CPU and GPU. So you might consider adding a case fan, or replacing your existing fans with fans that move more air.

By the way, the only passively cooled (no fan) ASUS and MSI GT610 cards I can find are "low profile" cards. If your case is a slim (low profile) case, then case cooling may very well be part of your stability problem.

Finally, I ALWAYS want to ensure I am providing good, clean, stable power when troubleshooting hardware issues - especially problems that cannot be or are difficult to explain. So I would swap in a known good PSU before replacing any other hardware.

 

[banger]

By the way, the only passively cooled (no fan) ASUS and MSI GT610 cards I can find are "low profile" cards. If your case is a slim (low profile) case, then case cooling may very well be part of your stability problem.

Both are low profile cards but fitted with a standard bracket each with 3 connectors, VGA, HDMI and DVI. The MSI came with 3 brackets, 2 low profile and a standard one. The Asus with a standard bracket.

I have two 12mm case fans, one push one pull and I regularly remove the dust inside. My E8400 stepping C0 CPU idles at 45 c and has a load temperature of 70 c. I have a program called real temp that shuts the PC down if either GPU or CPU reach 90 c.

Yes this all happened on 28 March, but I am not convinced it is a hardware problem, I suspect Nvidia drivers of which I have the latest although I have rolled back to the previous ones after reading of problems with the latest on the Nvidia forum.

PSU is a 500 watt Corsair and is only 2 years old. Thanks for the detailed reply.

 

[Itaregid]

It is not the brackets I am concerned about, it is the size of the case.

I am familiar with Real Temp and it is good - though I personally prefer CoreTemp.

I would not be happy with 70°C for the CPU. I don't like my CPU temps to sit above 60°C for more than a few seconds. If you are overclocking, I would stop. While no damage is likely to occur, stability issues may result. And while GPUs typically run hotter, 90°C is pretty warm. That is still within the 610's maximum of 102°C in terms of damage, but stability may still be an issue.

If your case supports another or larger fans, you might consider adding one or buying larger fans, or buying new 120mm (12cm or 120mm, not 12mm) fans that move more air. When you say one push and one pull, I am assuming that is 1 in front pulling cool air in, and 1 in backing pushing heated air out to support the desired front-to-back flow through the case. If your case cannot support additional cooling, you may need to consider another case, or a double-wide graphics card that exhausts the GPU's and graphics RAM's heat directly out the back.

 

[banger]

It is not the brackets I am concerned about, it is the size of the case.

It is a full height case and yes I meant 12 cm fans one at front pulling and rear pushing.

I am familiar with Real Temp and it is good - though I personally prefer CoreTemp.

I have both and just noticed Core Temp supports over heat protection for CPU but not GPU.

TJ max is 100 c for my CPU. 70 c is rarely reached except under extreme load when stress testing with prime95 and real temp, a good 20 c headroom even so my PC will shut down if either processor reaches 90 c.

There isnt room for more fans and the two 12 cm fans I have are quiet and are OEM replacement fans after the case supplied fans became noisy. If you can recommend any 12 cm fans that are quiet and reliable I will certainly look at them. I am not overclocking although CMOS settings are Optimal instead of default.

 

[Itaregid]

There are lots of fans out there. Sleeve bearing fans tend to be quieter, but unless top quality, don't last as long so I usually go for fluid or precision ball bearings. You want the highest CFM (cubic feet per minute) with the lowest dB (decibels). That said, you don't get both, except with larger fans.

Note TJ max is the point the CPU will shut down to keep from burning up. That does not mean it will be stable up to that point.

It is not so much the height of the case (though taller do tend to support more fans) but the width as wider cases support larger fans.

You might also look at Speccy
(from the makers of CCleaner). While it does not have real time system tray component it will tell you about your System (chipset) temp too (if your motherboard has a sensor there).

 

[banger]

Thanks for the tips looking at Speccy now.

 

[banger]

You might also look at Speccy
(from the makers of CCleaner). While it does not have real time system tray component it will tell you about your System (chipset) temp too (if your motherboard has a sensor there).

It has a tray option and I am now monitoring my motherboard chipset temp. in real time from the tray.

 

[Itaregid]

I don't know when they added that, but thanks for pointing it out!

Looks like it was with v1.22.537 (18 Jun 2013) when they added Temperature sensor output to System Tray.

 

[banger]

Tray sensor gets stuck sometimes. Opening and re-minimizing fixes;

 

[Itaregid]

What do you mean by stuck? And what sensor are you monitoring? I have a quad core CPU and I am monitoring the average. Seems to keep changing regularly as it should. Using W8.1.

 

[banger]

I am monitoring the MB sensor and for about 3 minutes I noticed it said 57. I am monitoring in 1 sec steps. I opened the main window and MB was 46 then it started working again.

 

[Itaregid]

Well, that makes sense to me. The motherboard's work load does not vary like the CPU load does. The CPU can go from idle to 100% and back to idle in just a second or two. I would not worry about the motherboard sensor.

 

[banger]

MB is currently idling at 50 c but when I do any disk intensive operations its up to 60 c

 

[Itaregid]

Okay. That is not hot.