Problems with new gtx 960.

[Vivarso]

Hello!

I have a problem that I have not been able to solve by searching and browsing various forums, and I would like to know if any of you know what might be going on.

Recently I bought a gygabyte gtx 960 4gb, and after installing it and testing it everything was ok (by testing I mean putting a game like project cars and turning up the specifications to the maximum). The problem comes when one day I decide to really play. After 5-10 minutes of playing, my monitor turns off, following the cpu functioning (and the game is still running because the controller continues to vibrate).

The first thing was to reinstall drivers and the problem persisted, the second was to check temperatures and both the graphics and the cpu did not exceed 45º just before failing. The next thing was to try another monitor and another port on the graphics... the same thing.

Could someone enlighten me? Thank you very much in advance.

Intel core i5 3450
Mainboard ASUSp8b75-m LE
8gb RAM DDR3
Nvidia gtx 960 4gb
Monitor Philips 191EL
Source 550w

[Espinetenbolas]

Everything points to a graphics problem, try with OCCT in GPU or PSU mode and if not I would try with the FurMark

I copy and paste from another thread:

With OCCT you will squeeze everything out of the equipment and dissipate the maximum possible heat using a test based on DirectX and an Intel Linkpack test that stresses the CPU:

1.- Install OCCT

2.- Run OCCT

3.- Tab POWER SUPPLY

4.- Select:

*Type of Test-> INFINITE
*DirectX Version-> 11
*Video Card-> Your graphics
*Resolution-> The native of your monitor or the one that gave you that error
*FULLSCREEN
*64BITS LINKPACK
*USE ALL CORES

5.- Launch the Stress Test by clicking ON (GREEN BUTTON)

6.- Check if after at least 15 minutes or half an hour your error has been replicated, if not, leave it on for a couple of hours.

For Furmark which runs OpenGL it is simpler and only stresses the GPU:

1.- Install Furmark

2.- Run Furmark

3.- Select:

*FULLSCREEN
*Resolution-> The native of your monitor or the one that gave you that error
*Anti-aliasing-> Configure it according to how you use it in games or disable it with OFF

4.- Check if after at least 15 minutes or half an hour your error has been replicated, if not, leave it on for a couple of hours.

5.- Launch the Stress Test by clicking on GPU STRESS TEST

To exit in both cases press ESCAPE and the test will stop.

In both cases if the graphics have any problem or the test will fail and it will indicate it or your error will be replicated… otherwise nothing will happen.

It is important to keep the monitor on and that you are controlling how the temperatures evolve or if the screen does anything strange since the test is only limited to forcing the hardware.

It is important to be able to replicate the error to process the warranty since if you say "this game crashes" they will say that their tests do not give errors and the error may be in other components... that's why you have to put it to the test with appropriate tests... it could be the ram or even the cpu... since the graphics heats up little but could be heating up the temperature of your case which can give instability to the memos or the cpu.

Try and comment ;D

[Fassou]

Tell us the brand and model of the font, it would also be helpful :sisi:

Salu2!

[Espinetenbolas]

The PSU I think when it is not specified is that it is generic but even so the equipment is of low-mid range and consumes very little so in principle there should not be much problem… and if there is we will find out with the OCCT :troll:

[Vivarso]

Thanks for your help.

After a brief test with OCCT for 15min as my colleague mentioned, no errors or problems have arisen. Now I'll leave it on for a couple of hours to see how it goes.

The power supply I have is a Hantol ATX Silent 120mm 550w.

Exactly what should I expect with that test? It's curious, I've tested it and it does it with all games regardless of the year and the quality required...

Thanks again.

Edit: the kind of whirlpool of red color has started to appear.

[Fassou]

Let's remember this post from 2012 :facepalm:

HANTOL power supplies - You look at them but you don't buy them - HardLimit

Cheers!

[Vivarso]

:wall:I'm sorry in 2012 let's just say that I and the gaming PCs were not having a good run... you mean that's where we have the problem? if so, what do you recommend?
thank you very much for everything.;D

[Fassou]

It doesn't have to be responsible for your failure, but I wouldn't sleep easy having a source of that brand or quality.

Salu2!

[Vivarso]

After reading the post about that power supply, I'm no longer at ease, so I'll change it... what do you recommend? And going back to the issue, with the diagnostic programs I haven't found anything, I mean I don't really understand what I was looking for.
I could only see how the FPS fluctuated up and down (I don't know if it was normal or not) and that the temperatures didn't go above 70º... but nothing, the monitor was active at all times...
What else can I look at?

[Fassou]

Personally, in addition to changing that font, I would check the BIOS version, see if a later one solves problems with graphics or directly puts the newest one, and if the problem has been having Win10, I would take a look at this recommendation from the Asus website's FAQ.

Salu2!

[Espinetenbolas]

@Vivarso:

Que mas puedo mirar?

Puede parecer una gilipollez pero… ¿has pensado que sea por las opciones de energía del OS o el salva pantallas el que apaga automáticamente?...:wall: lo digo porque si el fallo es a los pocos minutos y no se ha replicado con OCCT dudo que sea cosa del hardware ni de la fuente.

Por cierto la ralentización al cabo de unos minutos en la animación del toroide (donut) es por el sobrecalentamiento de algún componente que te hace thermal throttlingy, cae el rendimiento y por tanto los FPS... bien la CPU, bien la GPU... eso lo tienes que mirar tu para saber que tienes mal refrigerado.

[Vivarso]

Good morning, after investigating and trying one thing after another, I decided to try OCCT without the side panel of the tower. The parameters were similar although slightly lower the CPU temperatures. Having seen this, I tried it with several games and my surprise is that I could play uninterruptedly for more than 1h and I left it because it was late.
Today I will continue with more tests, but I think my achilles heel is the cooling. I have also noticed that no matter how much I push the GPU, the fans rotate timidly… is this normal or could I have a problem.
Thank you for your help, goodbye!

[wanautentico]

It has the same symptoms as the failure of the AMD R9 290. And the problem was that the frequency of the vram was too high. But seeing that you have a very poor quality power supply, in principle I would change that power supply for a corsair or seasonic and if you are a bit tight on money for an OCZ. If with the new power supply it continues to do the same thing what I would do is with msi afterburner I would relax a little the frequency of the vram and try with a game or with furmark. Tell me how it went.

[Espinetenbolas]

wanautentico estás totalmente equivocado con el diagnóstico, Vivarso tiene una mala fuente pero sin OCCT no peta el sistema esa fuente es suficiente y no es el origen de su problema… ni es necesario bajar la frecuencia de la memoria :facepalm:.

Como ha dicho los problemas desaparecen quitando el lateral de la torre... eso es un problema de refrigeración... la cuestión es saber ahora cual es el componente o componentes que se sobrecalientan.

Para ello solo tienes que monitorizar mediante OCCT y controlar cuando se produce una caída de FPS... eso debería de corresponderse con un pico de temperatura y/o con una bajada de Mhz o en la CPU o en la GPU.

Si no aparece nada claro postea capturas de pantalla con los resultados o sube directamente los gráficos en PNG que genera OCCT al terminar el test (recomendado)... nosotros te ayudaremos a interpretarlos ;D

Puedes usar PICSEE o cualquier otro hosting que de el formato de imagen para incrustar directamente en este hilo. *(Create thumbnail (width) [px] 900)
**
No hace falta que hagas nuevamente los test si OCCT ya ha guardado los registros de los test que les has echo … busca en C:\\Users*NOMBRE USUARIO

EJEMPLO DE MI EQUIPO CON UN TEST PSU DE 1,5MIN:

PD: Seria de ayuda que colgaras los resultados de un test OCCT con tapa y sin tapa…y lo indicaras para par ver mejor donde está el fallo... contra más información mejor.

[wanautentico]

@Espinetenbolas:

wanautentico you are totally wrong with the diagnosis, Vivarso has a bad source but without OCCT it doesn't crash the system that source is enough and it's not the origin of his problem… nor is it necessary to lower the memory frequency :facepalm:

As he said the problems disappear by removing the side of the tower... that's a cooling problem... the question is to know now which component or components are overheating.

To do this you just have to monitor through OCCT and control when an FPS drop occurs... that should correspond to a temperature peak and/or a drop in Mhz in the CPU or GPU.

If nothing clear appears post screenshots with the results or upload directly the PNG graphics generated by OCCT at the end of the test (recommended)... we will help you interpret them ;D

You can use PICSEE or any other hosting that gives the image format to embed directly in this thread. *(Create thumbnail (width) [px] 900)
**
It is not necessary to do the tests again if OCCT has already saved the test records that you have done … look in C:\\Users*NOMBRE USUARIO

PD: It would be helpful if you posted the results of an OCCT test with and without the cover… and indicated it to see better where the fault is... the more information the better.

The graphics card says it's new, that is, unless it comes defective it's difficult to overheat.
There is a possibility that it is also poorly plugged in or that in the assembly of the graphics card it doesn't make good air flow inside the case.
If the temperature problem were in the processor it wouldn't do the symptoms he says. The PC would turn off directly to avoid burning the micro or at most it would restart.

The user says the PC is still on but with a black screen…
Although it has a power supply of poor quality but it doesn't crash with OCCT it doesn't mean it's feasible with that graphics card, low amperage in the 12v rail can cause a malfunction of the graphics card.

I still think the same thing. Change the power supply to a certified one, if the problem continues it may be what I say about the VRAM of the graphics card with very high frequencies (it will be defective)
And to test it as easy as putting that graphics card in another PC and see how it behaves with a better power supply and no temperature problems.

A good help is that you install Aida 64 and monitor all temperatures, check if any of them goes out of its limit. Even the chipset of the motherboard or the hard drive.
In these cases to detect the problem what I do is trial and error, to rule out things and isolate the problem.

But I still lean towards a power supply failure and if with a certified power supply it continues to do the same with good temperatures in general, I would try the graphics card in another PC.

If in the other PC it does the same RMA the graphics card.

PD: No one mentioned updating the motherboard BIOS, but I rule this out. In this case it wouldn't detect the graphics card if the BIOS were old.

Best regards.

[whoololon]

I understand that we all do our best to help, the fact is that it is preferable to go step by step in a problem that:

1. Is punctual.
2. Does not affect system stability. (It is not clear if the image returns after a while or not, or if the computer crashes).
3. Is not deliberately reproducible.
4. There is a lot of missing information (mainly brand and model of components and test results).

Let's leave who pees the farthest until Vivarso offers us a more detailed explanation of the problem. ;D

[Espinetenbolas]

whoololon the problem with reason is that it is the most evenly distributed thing in this world… everyone believes they have enough :troll:

As for the problem, it should be said that if the case is poorly ventilated and its internal temperature exceeds 40º, the components with the stock coolers will have problems because that is what the manufacturers' technical documentation says.

Read "Boundary Condition Definition" page 13 -> http://www.intel.com/content/dam/www/public/us/en/documents/design-guides/4th-gen-core-family-desktop-tmsdg.pdf

That is to say… 35ºC of ambient temperature and 5ºC difference with the internal temperature of the case and the environment... hardware working conditions 40ºC... if you exceed them inside the case you are screwed.

40ºC may seem like little but it is supposed that a person at 35ºC of ambient temperature finds it difficult to stay clear-headed/efficient in front of a PC and not stay glued to the chair as if it were the paper of a biscuit... so it is understood that the machine will not work either... or they will put air conditioning

If here the companion opens the tower and does not give the error, several things can happen but everything points to a refrigeration problem that needs to be identified.

By the way Vivarso… what tower do you have?... how many fans have you mounted in it?... what cooler does the CPU have?

Anyway, post results.

[Vivarso]

Thanks for everything really!

Well as soon as I have a little time I'll upload what you ask for.

In response to the fans I can say that it only has one small one attached to the cpu. Previously I had a gtx550ti and I had not had any problems.

And thanks again.

[Espinetenbolas]

@Vivarso:

To the fans' response I can say that you only have one small one attached to the cpu. Previously I had a gtx550ti and I had not had any problems.

That is, your case is not cooled… understand that there is no air flow that enters (COLD FLOW) and another that exits (HOT FLOW).

There are various reasons why that 550ti did not give you those problems... the first is because it had much less graphic power and therefore the CPU was more underused.

Also the 550ti dissipates less heat... so we can talk about 20 - 30 or 50W of difference adding it to what the CPU did not dissipate and what it dissipates now.

That perfectly adds up to the fact that the components are now more balanced which increases the failure rate when operating near or above those 40ºC that you surely have inside the case... this is proven by the fact that when you open the case the failure disappears.

As soon as you upload the screenshots we will know better what is failing but for now it could even be the RAM or the hard drives.

Certainly what it looks like is that the case is poorly ventilated and some component is overheating.

PD: a photo of the tower open would not be a bad idea to see how you have it mounted ;D

[Sylver]

i5 with stock cooler, tragoncilla graphics, no fan and power supply made in Taipei... Inside it fries sardines like atoms in the Chernobyl reactor, but obviously you can't see it

Components to be changed/incorporated in the very near future:

-Processor cooler
-Additional case fans
-Power supply

And depending on the case, the case itself could also be nominated for a change when we see it

Best regards and happy holidays