Status
Not open for further replies.
Oct 11, 2020
9
0
20
Hi, I've been having some trouble with my computer giving out a VGA error light whenever I try to boot. I've tried switching out my GPU, motherboard, and have performed troubleshooting to the farthest extent that I possibly can. I've tried looking online for a knowledgable source or database which lists possible or known culprits of this problem. The most I could find was "There is a problem with your graphics card" which is obviously not the case for me.

I'm thinking the only two things that could possibly be causing this issue for me that I haven't tested replacing yet would be the power supply or the CPU. Here's a link to a post I made over on LTT forums that goes into more depth about my problem. Since then, I've replaced the motherboard and found out that it is a problem across multiple motherboards.

Try PSU or CPU next?

Thank you for your replies, have a good night everyone.
 
First of all, whoever was answering you over on LTT is an idiot, which isn't surprising really over there. That GQ unit couldn't POSSIBLY be more than four or five years old, at most, because the earliest review of any GQ unit is 2015, making it only possible to be five years old or less, not "up to 7 years".

Secondly, aside from a few minor cons, that's a pretty decent unit overall. Of course, that doesn't mean it can't be faulty because ANY power supply can be faulty even right out of the box, but it does make it a LITTLE bit less likely in general, or a LOT less likely than if it was a poor or mediocre quality unit.

There are still quite a few other options for what's going on though. I'm not going to spend all my time over there trying to rehash old information, so please, do us both the favor of explaining EXACTLY what is going on, from start to finish, and what you've already tried that didn't work, in as much detail as possible, within reason of course. I'm sure we don't need a "then I put the next screw in", you understand.

As for the CPU, unless you did something physically damaging to the CPU, like dropping it, or bending some pins while installing it, or overclocking for a period of time with excessive voltage, then it's HIGHLY unlikely to be a CPU fault. CPUs don't generally "go bad" unless something causes them to like a direct short of some kind, a faulty power supply or motherboard or memory taking them out, something that would likely be a pretty obvious contributor to the problem. For a CPU to just fail, would be uncommon unless you caused it to fail or it was VERY old, or has been an overclocked daily driver for quite some time.

Also, please list your FULL hardware specifications.
 
  • Like
Reactions: RainBowCake2010
Oct 11, 2020
9
0
20
First of all, whoever was answering you over on LTT is an idiot, which isn't surprising really over there. That GQ unit couldn't POSSIBLY be more than four or five years old, at most, because the earliest review of any GQ unit is 2015, making it only possible to be five years old or less, not "up to 7 years".

Secondly, aside from a few minor cons, that's a pretty decent unit overall. Of course, that doesn't mean it can't be faulty because ANY power supply can be faulty even right out of the box, but it does make it a LITTLE bit less likely in general, or a LOT less likely than if it was a poor or mediocre quality unit.

There are still quite a few other options for what's going on though. I'm not going to spend all my time over there trying to rehash old information, so please, do us both the favor of explaining EXACTLY what is going on, from start to finish, and what you've already tried that didn't work, in as much detail as possible, within reason of course. I'm sure we don't need a "then I put the next screw in", you understand.

As for the CPU, unless you did something physically damaging to the CPU, like dropping it, or bending some pins while installing it, or overclocking for a period of time with excessive voltage, then it's HIGHLY unlikely to be a CPU fault. CPUs don't generally "go bad" unless something causes them to like a direct short of some kind, a faulty power supply or motherboard or memory taking them out, something that would likely be a pretty obvious contributor to the problem. For a CPU to just fail, would be uncommon unless you caused it to fail or it was VERY old, or has been an overclocked daily driver for quite some time.

Also, please list your FULL hardware specifications.

Hey there, thanks for taking the time to reply. Here are my system specs:

Motherboard: Gigabyte X470 Aorus Ultra Gaming* / ASUS X570-Plus Tuf Gaming (Wifi)
CPU: AMD Ryzen 7 2700x
GPU: Gigabyte G1 Gaming GTX 1070* / ASUS ROG STRIX GTX 1070
RAM: Corsair Vengeance RGB PRO 2x8gb (16gb) DDR4 3000 Mhz C15
PSU: EVGA 750 GQ (750 Watt, 80 Gold, Semi-Modular)
Boot Drive: SP 256gb M.2 NVMe
Extra Drive: Crucial MX500 1TB Sata SSD
Case: Cooler Master H500
CPU Heatsink: Corsair H115i RGB Platinum

"*" indicates which part I was using before my repair attempts.

The first time I started noticing this problem was around a week ago, after I had put my new desk in place. I connected my computer, booted it up, and everything worked as it should have. The next day I woke up and when I tried to turn my computer on, the screen recognized that there was a device connecting to it and it lit up. The screen was still blank though, and it was not displaying an output. So, I waited for a while and decided to look inside of my computer to see if there was anything astray.


The first thing I noticed was that most of the lighting was on, including my AIO and RAM. They usually turned to their default rainbow when the system was booting, then turned off because of the settings that I had them on. The lights were not on for the GPU or the motherboard, but both were turned off by default using their respective software. All fans were spinning in the case, and the GPU’s fans were also spinning. For the error lights, the one that corresponded to VGA was lit up. I tried restarting several times, but the problem persisted.


I then did some research on the cause of this issue, and I found that it was likely due to a faulty GPU. At the time, there was no way of testing this to see if this was the case, so I did some additional troubleshooting. I tried using a different cable to connect to my monitor, reseating the GPU, replacing the power cable, connecting the power cable to VGA2, VGA3, and VGA4 on the power supply, seating the GPU in the other PCIEx8 slot, taking out one stick of ram, then try the other one by itself, reconnecting all power connectors, disconnecting everything that the computer did not need for it to boot, using the wall outlet directly instead of using it through the power strip, and finally letting it sit for a day disconnected from everything then coming back to it. After letting it run for a few minutes, you could feel that my graphics card was getting warm. I was still getting the same VGA error after all of this.


After all of these failed attempts, I called gigabyte to see if they had any suggestions. The lady on the phone told me that she was 99% sure that it was the graphics card that was faulty. So I bought another gtx 1070 to try it out and when it came I was still getting the same error. This time, the fans didn’t spin up but that was actually a feature of the graphics card. The graphics card displayed a white light where the power was connected showing that it was receiving power. I tried all of the steps previously mentioned, plus I reset CMOS. Then, I tried running the two cards in SLI and instead of the constant VGA error light, it started going from CPU to DRAM to VGA in a never-ending loop. Even after I removed one of the cards, this error was still happening. After this, I went to best buy and bought a new motherboard (x570) to see if the problem would persist. It did. This time, I was receiving the constant VGA error light from before. I tried using both cards again, but neither of them worked and both resulted in the same error. I find it highly unlikely that it would be the graphics card since both cards, from different manufacturers, were having the same problem. I gave it a few minutes since it was its first time booting up but nothing ever changed. The lights on the motherboard were on and the fan for the chipset was spinning. Everything else in the system was also on. Still no display.


This is where I am at right now. Either both of the RAM sticks are faulty, the CPU has “gone bad” or the PSU is faulty. I took out my CPU when switching boards and there were no bent pins. Everything looked like it was receiving power, hence the lights and every fan turning on. Both of the sticks of RAM would have to be faulty since neither one of them made the system work when put in by itself.

I should note as well that the lights on the second GPU turned on for an instant when booting but did not turn back on. Now that I think about it, this shouldn't have happened because i never used ASUS software on my system and never set it to turn off.
 
Last edited:
replacing the power cable

Replacing WHAT power cable? The one from the PSU to the wall or the one from the PSU to the graphics card? If the one from the graphics card, did you replace it with one that came WITH the power supply, or some other cable you had laying around?

Did you, at ANY point, between when it WAS and WAS NOT working, use ANY cable between the power supply and any internal component that did not come with the power supply specifically?

How long has that GQ been in service?

Have you tried running the system OUT of the case, in a bench test format?

 

wi5pa

Distinguished
May 20, 2012
189
23
18,665
I had this issue,
and it was faulty ram causing the white light to come on randomly.
I have to say i didnt suspect the ram but now i know that it was passing corrupted data to the graphics card, hence why the white light comes on it confuses the graphic card then the cpu cant deal with it.

Its so random and i go days with no problems, then white light, freezes, blue screens,
it was causing all sorts of mischief with my system.

Please test your ram,
To start with try the more friendly OCCT memory test (set memory to 70%) . if you get a clean run, then
Have a go at running memtest64,

Lets see if that comes up with no errors
 
Last edited:
The obfuscating factor is that the system has worked, fine, with the memory that is installed. Otherwise I'd say that it may simply not be compatible because Ryzen really doesn't like Vengeance memory kits, especially the LPX kits but also a lot of the Vengeance Pro kits as well, and generally won't run at 3000mhz, only at 2933mhz or at 3200mhz or higher.

All of that pretty much goes out the window though once the system has run fine with that kit installed.
 
  • Like
Reactions: wi5pa

wi5pa

Distinguished
May 20, 2012
189
23
18,665
The obfuscating factor is that the system has worked, fine, with the memory that is installed. Otherwise I'd say that it may simply not be compatible because Ryzen really doesn't like Vengeance memory kits, especially the LPX kits but also a lot of the Vengeance Pro kits as well, and generally won't run at 3000mhz, only at 2933mhz or at 3200mhz or higher.

All of that pretty much goes out the window though once the system has run fine with that kit installed.

Funny thing is that my ram failed after i hadnt touched mine for 6 months,
i worked abroad, pc was just sat on desk, no one had entry to my flat,
came back and booted it up and two days later it was having issues.

Id still change the ram for an elimination, just to be sure especially if you say the ryzen can have issues with it
 
Oh, I'm not saying it isn't possible, because it is, and we've seen it happen before, but it's pretty unlikely to have two modules both go bad at the same time from nothing more than moving your system long enough to put a new desk in place. If he can buy, beg , borrow or steal another DIMM to test the system with, that would certainly put that question to rest for certain, but as it stands I believe it is a pretty unlikely eventuality anyhow despite the fact that it is technically a possibility.
 
  • Like
Reactions: wi5pa
Oct 11, 2020
9
0
20
Replacing WHAT power cable? The one from the PSU to the wall or the one from the PSU to the graphics card? If the one from the graphics card, did you replace it with one that came WITH the power supply, or some other cable you had laying around?

Did you, at ANY point, between when it WAS and WAS NOT working, use ANY cable between the power supply and any internal component that did not come with the power supply specifically?

How long has that GQ been in service?

Have you tried running the system OUT of the case, in a bench test format?


I replaced the 8 pin VGA power connector that connects to the graphics card and VGA1, VGA2, VGA3, and VGA4. I replaced it with another cable that was included with my EVGA 750 GQ power supply that had been sitting around in its box, unused/untouched. I also tried another VGA 8 pin power cable, same problem.

I have never used a power cable inside of my system that was not included with my EVGA 750 GQ power supply.

It's been in service for a little bit longer than a year. From late June, 2019 until now.

I have not tried running the system out of the case. I will try doing that.
 
Oct 11, 2020
9
0
20

I used my old, x470 motherboard for the testing out of the case. After i turned it on, I received a CPU error. So i shut it down and looked at the CPU. It had a little bit of thermal past along the base of a few of the outer pins. I use mx-4 which is non-conductive, so it wasn't really a big deal. I cleaned it off using a cotton swap and some 70% isopropyl alcohol. I put it back in once it was dry and i was greeted with the same error i had before. It seems like with the new x570 board i was getting an error where it was a constant VGA light, and with my x470 board i was getting an error where it was cycling CPU -> DRAM -> VGA. Before i tried doing SLI with my x470, I was also getting the constant VGA light though. I'm not sure what caused it to start cycling like that.

I tried my other graphics card and i got the same result. With the ASUS GPU the fans revved up to full speed for a moment but quickly reduced to no fan spin after about 3 seconds.
 
As uncommon as it is, I think you're going to have to consider that possibly it's a memory CPU failure. It's a highly unlikely scenario but it sounds like it's a distinct possibility if you've tried 2 motherboards, 2 graphics cards and the only things you haven't really tried swapping out are the memory and CPU.

I assume you HAVE disconnected ALL of the storage devices while testing this out right? Because if not, you need to do that. A faulty drive could cause a number of issues especially if the problem is that the unit is shorted inside somewhere, or one of the cable is improperly connected or shorted.

You followed my guide for bench testing, exactly as outlined?

If so, then I think your next step is maybe to first TEST the PSU. You'll need to come up with a volt meter though to do that, so either borrow one, or buy a cheap one from Walmart, Harbor freight or one of the home centers etc., if you don't have one already.

 
Oct 11, 2020
9
0
20
As uncommon as it is, I think you're going to have to consider that possibly it's a memory CPU failure. It's a highly unlikely scenario but it sounds like it's a distinct possibility if you've tried 2 motherboards, 2 graphics cards and the only things you haven't really tried swapping out are the memory and CPU.

I assume you HAVE disconnected ALL of the storage devices while testing this out right? Because if not, you need to do that. A faulty drive could cause a number of issues especially if the problem is that the unit is shorted inside somewhere, or one of the cable is improperly connected or shorted.

You followed my guide for bench testing, exactly as outlined?

If so, then I think your next step is maybe to first TEST the PSU. You'll need to come up with a volt meter though to do that, so either borrow one, or buy a cheap one from Walmart, Harbor freight or one of the home centers etc., if you don't have one already.


Here's an update on my situation.

I borrowed some Vengeance 2400mhz c16 2x8 Gb RAM from a friend to swap out with my own. I tried every combination of the two sticks in the slots, but I was receiving the exact same problem as before on my x470 and gigabyte gtx 1070, with it switching between CPU -> DRAM -> VGA.

I ordered a multimeter off of Amazon and it should arrive tomorrow. I will test my power supply once it comes.

Edit: formatting
 
Oct 11, 2020
9
0
20
Good news! I am currently writing this from my computer that was giving me issues. Let me tell you what happened.

TLDR/Others with same problem answer:

I swapped out my graphics card and motherboard and was still getting a series of error lights in a cycle of CPU -> DRAM -> VGA. There was no display on screen. Before this, and on a different x570 motherboard, I was getting a constant VGA error light. I eventually got into UEFI by removing all storage devices. Inspect your drives, especially your boot drive for damage. Try reseating your SSD in the other slot(s). If this doesn't work after waiting a few minutes and your error lights don't go off, try turning your computer off, take all of your drives out, then holding down power + reset button for ten seconds, then release. This will put you into a backup version of your UEFI/BIOS, in my case version F2 for the x470 Aorus Ultra Gaming. From there I was able to adjust settings to their default. I set my RAM frequency to manual, but to the same frequency as auto. I/E I set it to 2133mhz, which was the same frequency as it was on auto. I did not touch my CPU voltage/frequency. I then was able to save the profile, save and exit UEFI, turn my PC off after you get back into the UEFI screen, install the m.2 drive, boot again, and hopefully you'll be into windows.

Full Answer:

When I got home today, I took my PSU out of my system for manual testing. I used the multimeter and followed the video you linked. I tested everything in the video and everything else. Everything was up to spec and all the voltages were spot on.

After this, I tried running my system once again without any storage devices. When I removed my M.2 drive, I saw that there was a warp in it. I knew there was a warp in it before, but I must have disregarded it or forgotten it. Funnily enough, the system ran fine using this bent SSD for months, if not more than a year. I'm not sure if this was causing my problem or if it was the slot. The slots looked like it had a small dent on the armor.

Curved SSD Picture

To my surprise, after 5-10 cycles of the CPU -> DRAM -> VGA error light mystery, the error light section of the motherboard went dark for the first time in weeks. I turned off the system and connected it to a monitor, plugged it in, turned it on, and I was like a child on Christmas when I was able to get into my UEFI. It told me that the settings had been reset (I had cleared CMOS sometime last week) so I went through and set them all up and saved the profile. After this, I saved and exited and turned off the computer.

I put my m.2 drive only back in, but this time using the top slot and it booted into windows. Everything was well and the system seemed normal. I turned it off and went to go eat dinner.

When I came back, my PC was still only half put together, with used thermal paste on the heatsink and stuff inside of the computer being unplugged. I plugged everything in, put my second graphics card in (since I now plan on using it in SLI) plugged in the SLI bridge, and my SATA SSD. When I turned the PC on, it was giving me the same series of error lights, this time not displaying an output and not ending. I took out one of the graphics cards and unplugged the second sdd. Same lights. I take the M.2 SSD out. Same lights. Stumped.

Next, I held down power + reset to get into the backup bios. I was able to get in with no problems. I adjusted settings, but this time I set the RAM speed to 2133 instead of "Auto (2133)" Then I restarted to see if I could get back there. I could. I turned the computer off and plugged the M.2 drive in, and the system booted back up into windows. This time, I tried to gather as much info as I could while I had access to the OS. I ran a benchmark when I got into Windows 10 the second time. It says I am running 3200mhz RAM, but I assure you that I am not.

I went back into UEFI to set my XMP profile and up the frequency on my CPU to 4.0Ghz. When I got back into Windows, I ran another benchmark. This time it was a little.. <Mod Edit>. I'm not sure what happened but I ran the benchmark twice and I got the same weird results.

No matter what the problem is, I plan on purchasing a new boot drive. Probably a 970 EVO. If I'm receiving weird benchmarks with my SSD and the PCB is physically curved, It's probably not the best idea to continue using it. I guess that's what I get for cheaping out on my computer components O.O

I'm going to try setting up SLI again and see what happens. Thank you so much for all of your help, moral support, and patience with me! I know this is probably a weird problem to have. If I receive any other problems, I'll reply to your latest comment again. Thanks again!:D

Edit: TLDR/Best answer
 
Last edited by a moderator:
Status
Not open for further replies.