Jump to content


 


Register a free account to unlock additional features at BleepingComputer.com
Welcome to BleepingComputer, a free community where people like yourself come together to discuss and learn how to use their computers. Using the site is easy and fun. As a guest, you can browse and view the various discussions in the forums, but can not create a new topic or reply to an existing one unless you are logged in. Other benefits of registering an account are subscribing to topics and forums, creating a blog, and having no ads shown anywhere on the site.


Click here to Register a free account now! or read our Welcome Guide to learn how to use this site.

Photo

Faulty graphics card, or faulty memory?


  • Please log in to reply
34 replies to this topic

#1 stebbinsd

stebbinsd

  • Members
  • 211 posts
  • OFFLINE
  •  
  • Local time:09:58 PM

Posted 20 December 2017 - 05:59 PM

Ok, so the other day, I was playing a game and the computer spontaneously crashed on me. In particular, the screen suddenly became overcome with red and orange pixels. That's the best I can describe here. Otherwise, I'd be looking up an image on Google to show you.

In any event, the pixellation never subsided. Instead, I had to flip the switch on my power supply and boot the system back up.

Except it never came back on. I had spinning fans and internal lights coming on, but no images appeared on-screen.

Now, I've had that problem before, and I know that it's one of the most general symptoms out there. When that happens, nearly anything could be the problem.

Well, I have a motherboard speakers that issues beeping error codes. I don't get any beeps when I try to boot the machine for real. However, when I take my memory out, I get repeated beeps. About 3-4 beeps per second.

This suggests to me that the memory is faulty.

However, I'm also worried that my graphics card might have crapped the bed. For starters, my mouse and keyboard are clearly coming on during the boot, suggesting to me that the machine is indeed booting, but without a feed to the monitor.

Second, I have experienced this pixellation before, but in smaller doses. The pixellation would usually subside. When it did, I got a message in the bottom-right corner of the screen saying that my "kernel mode drivers" had failed. That is unequivocally my graphics card, right?

Third, I had two (2) ram sticks in the system before it crashed. Removing BOTH of them causes the aformentioned beep code. However, inserting just one back into the system - it doesn't matter which one - causes the system to give no beep codes and simply not boot. Bear in mind that my system worked just fine one minute, and then crashed without warning. I find it rather unlikely that both ram sticks would crap the bed at the same time.

Fourth, remember that this crash happened during the middle of a game. While the screen was pixelated, I could still hear the game's music playing. Nothing I did with the keyboard seemed to create any sound effects, though. This suggests to me that the game was at least partially functioning in the background. This would suggest that the RAM was at least partially functional, since shouldn't have been working at all if the RAM completely died. Then again, that doesn't necessarily  mean the RAM can fully post; maybe the only parts of the RAM that were functional were those the computer only uses after a successful post.

But if the graphics card is faulty, that would not explain why removing the ram sticks actually caused a noticeable difference in my posting; namely, that I actually get a beep code by removing the ram and only the ram.

My motherboard doesn't have its own separate VGA port, so I have no means of testing my computer without the graphics card.

Is there any way, short of paying through the nose for a professional diagnosis, for me to tell for certain if the card is faulty?


My current build:

 

Motherboard: Gigabyte GA-970A-UD3P
Processor: AMD FX-8350
RAM: 12GB (2x2GB+8GB)
Video Card: Geforce GTX 750ti

OS: Windows 7 64-bit Ultimate


BC AdBot (Login to Remove)

 


#2 Platypus

Platypus

  • Moderator
  • 14,684 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:Australia
  • Local time:12:58 PM

Posted 20 December 2017 - 06:44 PM

Leave the RAM in but remove the video card, see if you get a different fault code. If you do, the RAM is passing its check and the mainboard is complaining that it didn't then find a video card. The best way to absolutely prove the issue then would be to find a way to borrow a known good video card. (Although when you substitute a component, there's always a very slight possibility that a mainboard fault has damaged the part, and the test part could also be harmed.)

 

If the POST still fails to progress with RAM but no video card fitted, something ahead of the card in the sequence is causing it, harder to diagnose but maybe mainboard fault.


Edited by Platypus, 20 December 2017 - 06:46 PM.

Top 5 things that never get done:

1.


#3 stebbinsd

stebbinsd
  • Topic Starter

  • Members
  • 211 posts
  • OFFLINE
  •  
  • Local time:09:58 PM

Posted 20 December 2017 - 07:01 PM

Leave the RAM in but remove the video card, see if you get a different fault code. If you do, the RAM is passing its check and the mainboard is complaining that it didn't then find a video card. The best way to absolutely prove the issue then would be to find a way to borrow a known good video card. (Although when you substitute a component, there's always a very slight possibility that a mainboard fault has damaged the part, and the test part could also be harmed.)

 

If the POST still fails to progress with RAM but no video card fitted, something ahead of the card in the sequence is causing it, harder to diagnose but maybe mainboard fault.

If I do that, and I get a beep code without the graphics card, how do I know it's because of a card failure, and not the simple fact that my monitor isn't plugged in?

 

Bear in mind: My motherboard doesn't have its own VGA port! I can't hook the monitor directly into the motherboard!

 

So getting a beep code in that scenario could easily mean either one.


My current build:

 

Motherboard: Gigabyte GA-970A-UD3P
Processor: AMD FX-8350
RAM: 12GB (2x2GB+8GB)
Video Card: Geforce GTX 750ti

OS: Windows 7 64-bit Ultimate


#4 Platypus

Platypus

  • Moderator
  • 14,684 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:Australia
  • Local time:12:58 PM

Posted 20 December 2017 - 07:32 PM

You don't get an error code from not having the monitor connected. A computer can boot completely without a monitor.

 

As I mentioned, if you get the different code without the card (and don't have a listing of beep code meanings) to prove definitely that it is the card itself, substituting a known good video card would be the normal procedure.

 

To reduce potential alternative causes, you can disconnect all drives as well, and any peripheral cards if they are fitted.

 

As the simplest first step though, just removing the card and seeing if the error code or boot behavior changes, points you which way to go from there.


Edited by Platypus, 20 December 2017 - 07:38 PM.

Top 5 things that never get done:

1.


#5 stebbinsd

stebbinsd
  • Topic Starter

  • Members
  • 211 posts
  • OFFLINE
  •  
  • Local time:09:58 PM

Posted 20 December 2017 - 09:20 PM

Ok, I put a ram stick back in the machine. I then tried to post both without a graphics card in the PCI slot, as well as having the card in the slot, but without the power supply cable plugged into the top.

 

In both scenarios, I got no beep codes whatsoever.

 

For the record, if I get a single, short beep, that usually means the post test was successful and the computer should boot normally. Right now, I'm not even getting that much.

 

Currently, sans-memory is the only setup I've encountered thus far where I get any beeps whatsoever, in any combination.


My current build:

 

Motherboard: Gigabyte GA-970A-UD3P
Processor: AMD FX-8350
RAM: 12GB (2x2GB+8GB)
Video Card: Geforce GTX 750ti

OS: Windows 7 64-bit Ultimate


#6 Platypus

Platypus

  • Moderator
  • 14,684 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:Australia
  • Local time:12:58 PM

Posted 21 December 2017 - 05:42 AM

So, just to confirm, this is not the system you have in your signature? (An ECS A960M-M4 has an onboard VGA port.)


Top 5 things that never get done:

1.


#7 stebbinsd

stebbinsd
  • Topic Starter

  • Members
  • 211 posts
  • OFFLINE
  •  
  • Local time:09:58 PM

Posted 21 December 2017 - 08:41 AM

So, just to confirm, this is not the system you have in your signature? (An ECS A960M-M4 has an onboard VGA port.)

Ok, this just pisses me off.

 

I very clearly stated in m OP that my motherboard doesn't have its own VGA port. If the ECS A960M-m4 was my current motherboard (the keyword being "current"), then by process of elimination, that isn't my current board, is it?!

 

Use your head!


My current build:

 

Motherboard: Gigabyte GA-970A-UD3P
Processor: AMD FX-8350
RAM: 12GB (2x2GB+8GB)
Video Card: Geforce GTX 750ti

OS: Windows 7 64-bit Ultimate


#8 stebbinsd

stebbinsd
  • Topic Starter

  • Members
  • 211 posts
  • OFFLINE
  •  
  • Local time:09:58 PM

Posted 21 December 2017 - 09:21 AM

So, just to confirm, this is not the system you have in your signature? (An ECS A960M-M4 has an onboard VGA port.)

Look, I'm doing my best to cooperate with your suggestions. But honestly, should I even have had to answer that? How did you not put two and two together?

 

More importantly, how does this even remotely contribute to solving my problem? I very clearly stated that my motherboard didn't have its own VGA port. Does it really matter whether I changed motherboards, or maybe the ECS board had its VGA port break down already! Maybe it's either! Maybe it's both! Does it really matter to the extent my current dilemma is concerned?


My current build:

 

Motherboard: Gigabyte GA-970A-UD3P
Processor: AMD FX-8350
RAM: 12GB (2x2GB+8GB)
Video Card: Geforce GTX 750ti

OS: Windows 7 64-bit Ultimate


#9 Platypus

Platypus

  • Moderator
  • 14,684 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:Australia
  • Local time:12:58 PM

Posted 21 December 2017 - 07:37 PM

At the top of the forum there's a pinned topic titled: Announcement: Please provide system specifications when posting a problem.

In it it mentions: "Having this information available will allow our helpers to provider faster and more accurate advice."

When I saw your topic I was pleased that I didn't have to request this information as your system spec appeared helpfully in your sig.
 

How did you not put two and two together?

 

Even an experienced technician has no reason to know whether a mainboard has onboard video just from its model number. After the initial standard diagnostic procedure that would be applied to any system didn't give a clear result, I downloaded a manual for the ECS board, hoping to find a BIOS error beep code listed. There was none, but I noticed in the listing of I/O ports, it has VGA and HDMI. So I developed the suspicion I was researching the wrong board. However, that was not the only possibility, since ECS/PCChips had a way of producing extra budget releases of their boards, lacking some features. If you'd been prepared to make a civil response to my question, you might have said "Yes, that's the board, but it definitely doesn't have video, there's just a blank in that spot." Or you might have replied, "Correct, it's a Brand X Model Z, and I could have downloaded the manual for that board and gone from there...
 

Does it really matter whether I changed motherboards, or maybe the ECS board had its VGA port break down already!... Does it really matter to the extent my current dilemma is concerned?


Yes, if we were dealing with a system that had a video card fitted because the onboard video failed, knowing that could make all the difference. Also, knowing whether we are dealing with an ECS board would help with potential faults. ECS boards are more prone to developing capacitor and onboard regulator faults than some other brands are.

 

If you'd like me to continue working with you, feel free to post the actual specification of the system we're dealing with, so I can research correctly. If not, I'd simply observe that generically, symptoms like you've experienced can also be caused by a PSU that has been developing bad electrolytic capacitors, and one of the caps has finally given up. (To diagnose, substitute a known good PSU.)
 


Top 5 things that never get done:

1.


#10 stebbinsd

stebbinsd
  • Topic Starter

  • Members
  • 211 posts
  • OFFLINE
  •  
  • Local time:09:58 PM

Posted 21 December 2017 - 09:05 PM

My current rig is thus:

 

Gigabyte GA-970A-UD3P

10GB (2GB+8GB) DDR3 Ram

AMD FX-8350 (that's eighty-three fifty, not forty-three fifty).

Geforce GTX 750ti


My current build:

 

Motherboard: Gigabyte GA-970A-UD3P
Processor: AMD FX-8350
RAM: 12GB (2x2GB+8GB)
Video Card: Geforce GTX 750ti

OS: Windows 7 64-bit Ultimate


#11 Platypus

Platypus

  • Moderator
  • 14,684 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:Australia
  • Local time:12:58 PM

Posted 22 December 2017 - 07:57 AM

Unfortunately, the manual for the Gigabyte board doesn't provide a BIOS beep code. However it does note that the BIOS is AMI, so on the assumption it uses the default AMI code, three short beeps repeated would make sense for when the RAM is not there. AMI doesn't have a specific "RAM not found" code, but 3 beeps is "A failure has occurred within the first 64 KB of memory", which would certainly be the case if there is no memory present. 4 beeps is "System Timer failure", which shouldn't be influenced by presence or absence of RAM.

If the POST was able to proceed with RAM fitted but no video card, there should be an 8 beep code for "The video adapter is missing", so it looks like the fault lies between RAM test and Video Card test. Except for the possibility I mentioned of a PSU fault, it's probably going to come down to CPU or mainboard, as long as all peripherals that could possibly stall the POST if they were faulty, such as drives, are disconnected for the tests.

CPU includes the internals of the CPU itself and the connection to it, so examine for discoloration of the CPU with cooler and thermal paste removed, or any sign of a burnt pin indicating a bad contact in the socket (which may or may not damage the CPU). This would also count as a mainboard fault, as the socket at least would need to be replaced, which may come down to board replacement.

Mainboard includes the componentry and connectors, so examine for any signs of burnt or dislodged components, or discolored or burnt pins in power connectors.

Beyond that, it becomes a matter of proving by substitution, i.e. OK looking CPU needs to be proved in a good mainboard, video card needs to be proved in another system etc.

Probably the worst possible outcome would be if the video card has had a failure that has damaged the mainboard. That would mean if the video card is fitted to another system, it might damage it too. That's one of the risks that is built into the charges you pay when you hand a job over to the workshop.

Top 5 things that never get done:

1.


#12 stebbinsd

stebbinsd
  • Topic Starter

  • Members
  • 211 posts
  • OFFLINE
  •  
  • Local time:09:58 PM

Posted 22 December 2017 - 09:32 AM

Unfortunately, the manual for the Gigabyte board doesn't provide a BIOS beep code. However it does note that the BIOS is AMI, so on the assumption it uses the default AMI code, three short beeps repeated would make sense for when the RAM is not there. AMI doesn't have a specific "RAM not found" code, but 3 beeps is "A failure has occurred within the first 64 KB of memory", which would certainly be the case if there is no memory present. 4 beeps is "System Timer failure", which shouldn't be influenced by presence or absence of RAM.

It's a constant, repeating series of beeps when i take the ram out. About 3-4 beeps per second.

 

 

If the POST was able to proceed with RAM fitted but no video card, there should be an 8 beep code for "The video adapter is missing", so it looks like the fault lies between RAM test and Video Card test. Except for the possibility I mentioned of a PSU fault, it's probably going to come down to CPU or mainboard, as long as all peripherals that could possibly stall the POST if they were faulty, such as drives, are disconnected for the tests.

I told you ... if there's memory but no video card, I don't get a beep code at all!


My current build:

 

Motherboard: Gigabyte GA-970A-UD3P
Processor: AMD FX-8350
RAM: 12GB (2x2GB+8GB)
Video Card: Geforce GTX 750ti

OS: Windows 7 64-bit Ultimate


#13 Platypus

Platypus

  • Moderator
  • 14,684 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:Australia
  • Local time:12:58 PM

Posted 22 December 2017 - 05:20 PM

I told you ... if there's memory but no video card, I don't get a beep code at all!


That's what I said, the fact that there's no response to the absence of video card, when there should supposedly be 8 beeps, but there is a response to no RAM which is an earlier test, suggests that the POST is failing to proceed as far as identifying the absent video. When the built-in test fails or stalls, we're left with a process of elimination with parts substitution to identify where the fault lies.

Another alternative that a workshop or serious enthusiast may have, is a diagnostic board that can be fitted into a bus slot, and takes over from the mainboard POST process. By that means, it may be possible to identify if the problem is something like a CPU or PCIe bus fault (e.g. maybe a fault in the video card has damaged the PCIe 16x bus channels.)

Top 5 things that never get done:

1.


#14 stebbinsd

stebbinsd
  • Topic Starter

  • Members
  • 211 posts
  • OFFLINE
  •  
  • Local time:09:58 PM

Posted 22 December 2017 - 06:02 PM

 

That's what I said, the fact that there's no response to the absence of video card, when there should supposedly be 8 beeps, but there is a response to no RAM which is an earlier test, suggests that the POST is failing to proceed as far as identifying the absent video.

So is it safe to assume, at the very least, that the RAM is faulty, even if it's not the only problem?


My current build:

 

Motherboard: Gigabyte GA-970A-UD3P
Processor: AMD FX-8350
RAM: 12GB (2x2GB+8GB)
Video Card: Geforce GTX 750ti

OS: Windows 7 64-bit Ultimate


#15 Platypus

Platypus

  • Moderator
  • 14,684 posts
  • OFFLINE
  •  
  • Gender:Male
  • Location:Australia
  • Local time:12:58 PM

Posted 22 December 2017 - 07:26 PM

It's not impossible, but it seems pretty unlikely that both DIMMs suddenly failed. It's more likely that when the RAM is fitted, it is passing as OK and the POST is getting stuck further on. There is a continuous beep error code that appears in some (but not all) AMI code tables for "memory not found", so unless both DIMMS are so badly faulty they actually stop the POST, it looks like the mainboard complains when there is no memory, doesn't complain when the memory is there, so it's probably OK.

But the only way to be absolutely certain about the memory is to test it in a known good system, or to try a known good DIMM in your system.

If both DIMMS have failed, there would be a suspicion that the mainboard has done it, e.g. by over-volting the memory.


Edited by Platypus, 22 December 2017 - 07:31 PM.

Top 5 things that never get done:

1.





0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users