Computer Help Please! My computer keeps crashing after installing new GPU.

(1/2)

Hi all,

I upgraded my graphics card in January, and I am unable to play FFXIV because I will randomly lose my display and nothing I have tried to do has resolved the issue permanently. I will have to split this post to accommodate the forum restrictions. The rest of this post basically goes over what I HAVE done to try to resolve the issue.

Original computer specs were that of an HP Pavilion 500-054, purchased in 2013. Link to full specs here.

Current computer specs:
OS: Windows 8.1, 64 bit
Motherboard: MS-7778 JasmineR
PSU: Rosewell Photon 850 80 Plus Gold
CPU: AMD A8-5500
Ram: 8 GB DDR3-1600
GPU: XFX Radeon 560 RX (4GB DDR5 virtual memory) (driver 17.12)
HDD: 1.5 TB SATA
Monitor: Samsung S22C300H 21.5” LED TN
We purchased the GPU in 2017. We tried to install it when it was purchased, but my computer's original PSU wasn't strong enough to run it, and the case was too small to fit a larger PSU.

Fast forward to 2019. I bought a CoolerMaster Masterbox K500 (Link here) so that I could fit a larger power supply in with all the other computer innards. I bought an EVGA 600 w 80 Plus PSU to help run the new GPU, took my old computer apart, put everything in the new case, hooked everything back up, and turned it on.

At first, I couldn't get my display to run. My husband had the idea of turning off the RGB lights on my case and then restarting the computer. Lo and behold, I had a display again. However, I really didn't want to have to live without the lights. So I bought a stronger PSU. Went to a Rosewell Photon 850 80 Plus Gold PSU. I was finally able to get a display AND the lights on my computer. After that, I tried to switch from the integrated graphics card to the one I just installed, and I encountered issues with that too. One of the things I tried was changing the secure boot configuration in my BiOS to “Legacy”. That seemed to work because I was able to see the GPU and install the driver software.

But now I am encountering a different problem, one which is still occurring and is driving me up a wall. This is the issue I am trying to solve.

I can play FFXIV and have it on the max settings without any FPS lag.

BUT.

I will randomly loose my display. Losing the display is preceded by the screen turning one single, solid color (it varies every time), the sound glitches out, and the display goes dark and starts flipping between HDMI and analog, like it's looking for a signal. Then it goes dormant after a few seconds of trying, and failing, to find a signal.

Sounds like your GPU is dying.

(2/2)

Here is a list of things I have tried in an attempt to fix the problem:

• switching out the HDMI cable for a different one.
• Unplugging the HDMI and plugging it back in.
• Doing a hard boot on the computer, waiting a few seconds, and then restarting it.
• adding additional RAM to my computer so that I had a total of 16GB of DDR3 RAM (the other two sticks were Team Vulcan 2x4GB DDR3-1600.)
• When I kept having the same issue I ran a memory check and my computer found nothing wrong with my RAM.
• I removed the additional RAM b/c FFXIV is fine with 8GB.
• I checked the windows event log to see if there was something that was happening consistently right before the crash. The only thing I found was that the last thing that windows logged before the crashes was Bonjour Service having a critical error. But that might have been because of the thing that was making me lose my display, not the cause of it.
• contacting AMD support about which driver I should use since they no longer make drivers for Windows 8.1, and their response was “upgrade to windows 10”. -_-
• Trying different AMD driver versions (17.1, 17.7, 17.11, and 17.12)
• Installing MSI Afterburner, and changing the settings for the display to “force constant voltage”.
• Using MSI Afterburner to tell my computer not to allow the CPU to get over 120 degrees F. (This one actually seemed to have the most impact. I was able to log into the game and play it for a few hours without losing my display. However, not allowing the CPU to get very hot made it so the graphics card couldn’t function as well as it normally would and I got fairly significant lag in some spots.)
• Adjusting the temp settings to allow for a PSU temp of 153 degrees F so that I could play without lag (I tried this on Friday night, and the last reading I saw on the MSI Afterburner HUD was 136 degrees F. Shortly after that, I tried to run the benchmark and I lost my display again and couldn’t get it back, even after letting the computer sit for a few minutes.)

What I am wondering, is, if temp seems to be the primary cause of losing my display, and if I have to let the computer cool ALL THE WAY back to room temp before I can restart it and have my display come back on, does that mean that all of these problems are most likely related to the fact that my CPU is from 2013 and it is old? I don’t want to spend money on a new CPU if it isn’t going to fix the problem, but I don’t know how to diagnose the issue to determine if it really is the CPU.

Plus, I don’t know if it is possible for me to get a new CPU and have it work with my old motherboard. I also don’t know if it is possible to find a CPU/motherboard combo for under $300.

I’m hoping someone on here might be able to help me fix this so that I can play the damn game without worrying that my computer is going to crash on me. Ask questions if you need more information; I’ll try to answer them as best as I can.

Check the CPU temp first by downloading a utility like coretemp or hwinfo, turn the game on and monitor the temperature. For hwinfo use the sensors table. If it rises really rapidly the heatsink might be loose.

If the entire computer is shutting down the cpu is probably getting too hot. According to the spec sheet, the cpu will automatically turn itself off at 71C/160F. The cpu is kinda weak so it might be fully loaded when running FFXIV.

It's also a possibility that you have a dud power supply but less likely since using the integrated graphics does the same thing. (That processor and a 560 let alone just the processor by itself is barely using an actual working 600-850W 80 Plus)

Check cpu temps first. If they are ok since you have an extra PSU I would try to switch to the other one. If all that doesn't work it might be the GPU. You can also remove the card and run it off the integrated for a bit and let it sit there to see if it is stable (it will lag but just check for stability/no crashing for a couple minutes) to check if it's the gpu.

I had a similar problem years ago when i blew my 780ti screen would go black but i had sound in 14 still (could move around etc just not see) signaling a gpu failure, but if the whole thing just shuts off its most likely way to taxing on your cpu.

Ok complete nm

Solution... AMD is rife with issues and it's the bargain for a reason, switch to Nvidia.

Did you do anything in the BIOS? Like did you overclock the CPU using BIOS settings? Also did you build the computer? If you did did you connect computer fan to the "CPU FAN" pins? Check that the CPU has a CPU fan connected properly.

That sounds like you're losing a graphics card. The description here sounds like a hardware fault of some kind, not a driver problem. Also you were right to scoff at the "switch to Windows 10" suggestion from AMD. That isn't going to help you. I mean you should do it if you can because 8.1 is a pile and 10 is not, but that's beside the point.

You can check temps on your CPU and GPU with a utility like SpeedFan. If your CPU were overheating the entire computer would shut down. If your GPU is overheating it could manifest itself like this. It's worth checking to see what kind of temps you're getting when the game is running. But honestly, it sounds like you've just got a dying card.

NB: Mmm also, please report computer temperatures in Celsius. It's a far more standard scale even in the US when measuring something like this. Easier to tell at a glance if something's up without having to convert first. As a rule a computer will shut itself down over 100C and anything over 90C should be concerning.

Cousin says might be heat issue on your graphics card.

He said if you have adrenalin 18.4.1 get rid of it OC is bad.
He said usually says sometimes the thermal paste is crap on cards that changing it helps also, if nothing else helps time to change it out for nvidia as AMD graphic cards are junk.

Make sure you have proper cooling as HP computers tend to overheat as well windows 8 also outdated platform.

10 gets more support which means hardware support and devices, but if not its probably a issue with overheating.

Noctua NT-H1 Thermal Compound

I think one thing that is going on here is a failure of your motherboard.
If you couldn't have an RGB light with 600W PSU then there is something cheesy with it, maybe draining too much power for some reason, maybe it has too old capacitors which don't hold the voltage too well. If it would be the CPU overheating, then the system would just shut down itself, and if it would be the graphic card, your game would be still running and you would be able to hear the sounds from the game. It may be that motherboard is rusty somewhere or cracked, drains more power and in result overheats in one part or another. Probably also your GPU contributes to your problem, it may drain too much power out of PCIE slot and this could be really heavy on your motherboard power section.

Its a time to change motherboard, CPU and Memory to DDR4, my guess is your setup was never designed to have high power GPU in it and play games, its an office computer made into gaming which never ends well.
I would recommend for you to get i5-9400f with B365 motherboard or Ryzen 5 2600 with B450 motherboard.

Quote:

---

Originally Posted by Nedkel

I think one thing that is going on here is a failure of your motherboard.
If you couldn't have an RGB light with 600W PSU then there is something cheesy with it, maybe draining too much power for some reason, maybe it has too old capacitors which don't hold the voltage too well. If it would be the CPU overheating, then the system would just shut down itself, and if it would be the graphic card, your game would be still running and you would be able to hear the sounds from the game. It may be that motherboard is rusty somewhere or cracked, drains more power and in result overheats in one part or another. Probably also your GPU contributes to your problem, it may drain too much power out of PCIE slot and this could be really heavy on your motherboard power section.

Its a time to change motherboard, CPU and Memory to DDR4, my guess is your setup was never designed to have high power GPU in it and play games, its an office computer made into gaming which never ends well.
I would recommend for you to get i5-9400f with B365 motherboard or Ryzen 5 2600 with B450 motherboard.

---

If that were true they'd get bsod instantly instead of screen shutting down.

though they could use a debugger but not sure if they know how to work it.

Quote:

---

Originally Posted by Zeonx

If that were true they'd get bsod instantly instead of screen shutting down.

though they could use a debugger but not sure if they know how to work it.

---

The best way to test it out, is to put different graphic card, load the game and see if it does anything.

Bsod may occur if there is a failure of the CPU or CPU power section. It may be just a motherboard power section rippling all around and not giving the constant voltages and power.
XFX Radeon 560 RX drains all its power from the PCIE slot, and it could drain up to 120 W peak and 100W sustained.

Quote:

---

Originally Posted by Nedkel

The best way to test it out, is to put different graphic card, load the game and see if it does anything.

Bsod may occur if there is a failure of the CPU or CPU power section. It may be just a motherboard power section rippling all around and not giving the constant voltages and power.
XFX Radeon 560 RX drains all its power from the PCIE slot, and it could drain up to 120 W peak and 100W sustained.

---

^this is exactly what I was gonna ask about.

Is the card getting sufficient power? GPUs are pretty power-hungry and the pcie slot they're put in matters, although it'd be impossible to have put it into one of the pciex1 slots, as your motherboard appears to only have a single pciex16 slot for a gpu.

Just on hardware side, I'll let you know that another reason for temp go high is the power supply not adequate, dropping voltage, and currents go rampage. Maybe not your case but who knows.

Quote:

---

Originally Posted by Vaer

Check the CPU temp first by downloading a utility like coretemp or hwinfo, turn the game on and monitor the temperature. For hwinfo use the sensors table. If it rises really rapidly the heatsink might be loose.

---

This is a great idea because in addition to helping you determine whether or not there are cooling problems in the first place, it will also pinpoint *where* they are. Is it your CPU that's overheating or your GPU? Some things to check when you hardware is overheating are:
* Fans that stopped running or aren't running fast enough
* The thermal paste between your heatsink and CPU may have degraded, in which case you should buy a higher quality paste and remount the cooler.

All that having been said, you might first want to update your GPU drivers. Sometimes the drivers that ship with it are defective, and updating them is much easier than tinkering with the hardware.

Sounds awfully similar to a problem I had with EVE a couple years back. Thought my Gcard was busted, but it turned out to be the Power Supply.

And when I say similar, I mean virtually identical, right down to the screen going random color blocks.
Initially replaced the 980Ti was using (it had a fan malfunction the previous year, repaired under warranty) thinking it was the Gcard failing again, but continued having the problem.
The PSU I was using was pretty old, but it was a 1250w, so I had more than enough wattage. The problem was simply that due to age and general breakdown, the PSU was fluctuating its supply to the PC.

So my recommendation is to try a new PSU.
Also, its possible to still do the free Windows10 upgrade if you can find the relevant links. Which I would recommend, since previous versions of Windows have had their support dumped by a lot of places.

Quote:

---

Originally Posted by Sylve

Sounds awfully similar to a problem I had with EVE a couple years back. Thought my Gcard was busted, but it turned out to be the Power Supply.

And when I say similar, I mean virtually identical, right down to the screen going random color blocks.
Initially replaced the 980Ti was using (it had a fan malfunction the previous year, repaired under warranty) thinking it was the Gcard failing again, but continued having the problem.
The PSU I was using was pretty old, but it was a 1250w, so I had more than enough wattage. The problem was simply that due to age and general breakdown, the PSU was fluctuating its supply to the PC.

So my recommendation is to try a new PSU.
Also, its possible to still do the free Windows10 upgrade if you can find the relevant links. Which I would recommend, since previous versions of Windows have had their support dumped by a lot of places.

---

This GPU of her drains power from the PCie slot and it could be motherboard that is not able to sustain her GPU power drain. PCie power section with FFXIV launched could get hot and mess up the currents and voltages that is going to the graphic card, it may eventually break the graphic card by doing so.
So i again recommend checking the game with a second graphic card, and consider changing Mobo + CPU all together because its pretty old.

Try an old graphics card and see if you're still having issues. When troubleshooting you start by checking the most obvious problem. Never buy random things to try and fix a solution you're not completely sure about.

I can't tell whether from your post whether or not you plugged in the additional power cable from the power supply. Most video cards nowadays need more power then can be provided from the slot its plugged into. http://cdn.agilitycms.com/xfx/MediaG...SC0132%201.jpg see the port at the top and rear of the card? That should have a cable from the power supply attached to it. Since you have some extra memory lying around might as well use it, it might not help specifically for FFXIV but it should improve the overall performance for other tasks.

Quote:

---

Originally Posted by NekoGato

I can't tell whether from your post whether or not you plugged in the additional power cable from the power supply. Most video cards nowadays need more power then can be provided from the slot its plugged into. http://cdn.agilitycms.com/xfx/MediaG...SC0132%201.jpg see the port at the top and rear of the card? That should have a cable from the power supply attached to it. Since you have some extra memory lying around might as well use it, it might not help specifically for FFXIV but it should improve the overall performance for other tasks.

---

I know the nvidia cards do it, but if you don't have the power plugs properly connected, it gives you a warning when you boot the PC and wont even load to bios. At least in my experience thus far with more modern gcards.

Quote:

---

Originally Posted by Amnmaat

Did you do anything in the BIOS? Like did you overclock the CPU using BIOS settings? Also did you build the computer? If you did did you connect computer fan to the "CPU FAN" pins? Check that the CPU has a CPU fan connected properly.

---

Hi there. The only thing I did to the BIOS was changing the secure boot configuration to “Legacy”. I did that because the computer has having issues "seeing" the new graphics card. I couldn't install the driver software until the computer recognized that the GPU was actually there.

I did build the computer. Sort of. I am reusing the motherboard, CPU, Ram, and HD from my old computer, but I've added a new GPU, PS and case. The CPU fan is plugged in and is functional. The fans on the front and back of my case are also functional. There is a large vent on the front, back, and top of the case that allows for (hopefully) a lot of airflow. I believe there is also a grate right below the power source in the case. The power source had to go in so that the fan faced towards the ground, but the case has about 1 inch (2.54 cm) of clearance between it and the desk thanks to it's "legs".

Hope that helps. Sorry for the delay in responding. I've been avoiding trying to fix this the last few weeks.

It kinda sounds like you might have a dead/defective GPU. I would definitely try to RMA (call the shop you bought it from (Newegg/Amazon, they'll email you a Return Shipping label which you then save on an USB drive, go to an UPS or Fedex office; print the label there and present the box to them. The return label pays for the shipping). But yeah if your monitor is turning different colors that's definitely a defective GPU.

This is my advice, someone else might have other troubleshooting options though.

I'm a computer tech and did my IT degree here in Oxford, UK. I've worked as a computer tech since finishing Uni.

I recently swapped a load of hardware and was having the same kinda issue. In the end, I realised that the CPU temps were way too high and that the PSU I had in the "new: case was far too weak for my system.

From what I am getting from your setup is that you fixed one issue, but it does sound like you've not go good enough airflow around your system and FF14 is causing it to overheat and shutdown. Most people have already pointed out the key issues and their advice is pretty sound. Could parts be on their way out? It's possible, as FF14 stresses out systems like very few games I know (It killed my original PS4!)

I will be honest and say I don't like AMD stuff that much, as they just seem to fail to get the job done as easily as an Intel/Nvidia setup. Most games tend to work better with this combo, as it's been an industry standard until the last few years and now the high end AMD CPU's are beastly. I got myself some quality secondhand parts and built myself a rig that allowed me to use the better parts of my previous system.


Personally, I would ditch Windows 8 ASAP, as it's bloody horrid. Windows 10 makes life much easier and the difference between 8 to 10 is just like Vista to 7. The more modern OS and drivers would help, as they may allow your system to run more effectively.

Hi,

This might sound pretty simple, but I usually see this get a lot of people when I am brought computers where the screen randomly turns off. When you inserted your GPU, verify that the card was fully clicked in. Like I said, I know it sounds basic but sometimes these cards can look like they are plugged in. Also, when you bolt it down, elevate it a bit when your tightening the screws. Sagging GPUs can sometimes cause the pins to not be seated very well.

Other things you can try and that have probably been mentioned before in this thread...
-Check your temps.
-Make sure you didn't accidentally plug the CPU pin into the GPU and the GPU pin in the CPU.
-Try another GPU
-I also seen you mention you got a new PSU. Make sure it is inserted in the correct orientation. The PSU can generate a lot of heat and you want to make sure that does not get dumped back into the system. Though, if the case doesn't allow for that orientation, having positive pressure (slightly, not overkill. EX 3 fans in, 2 out) in the case would probably help. Out fans being on top and rear of the case.
-And finally, delete the old driver from any old hardware.

Good luck!

Sounds like what happened to me with a 280X I had, although it was just the GPU Driver that kept crashing.

RMA and get another, try Sapphire.

Yes, I would definitely install MSI Afterburner and turn on On-Screen display where you can select to display GPU and CPU temeperature in game.
Recently I discoered FFXIV was overheating my CPU because I had my CPU overclocked in the BIOS to 4.7 Ghz. I then went and set the CPU clock to Stock settings and that fixed the CPU overheating. Everything runs smooth now.

Checking temps is very important.