System powers off with overclock (EVGA GTX 1080 FTW)

S

SimianRob

Weaksauce
Joined
Mar 28, 2013
Messages
66
Hey guys,

I just picked up an EVGA 1080 FTW about a week ago to replace my R9 290. The card works perfectly 99% of the time, but I've had 3 times now when the system has just completey powered off, and started itself back up a few seconds later while gaming for about an hour+ while the card was overclocked. The settings I used were +100mhz core, +100mhz memory and 120% power limit. Now, the obvious guess that a lot of people would go to is the power supply, but I just find it hard to believe because I never had this issue with my heavily overclocked R9 290 (though that was an 8+6pin vs this card which is 8+8) that it replaced. The R9 290 tended to be far more power hungry though, but maybe stressed the power supply differently. So is the issue the card or the PSU? I was able to replicate the issue by running furmark with the overclock settings. I guess one of the obvious solutions is to back the overclock settings off, but if this overclock has exposed a weakness/failing PSU, I would rather deal with it now. What do you guys think? RMA the psu? Is there an issue with the card? The other odd thing I've found with the card is that when overclocking, I rarely ever see artifacting, the program will be running absolutely fine and then the program will either crash or the video driver will stop responding, which isn't quite the same thing that I used to see in older GPUs.

Also wanted to add that the card has the new bios with the increased fan curve as well as the thermal pads for VRM cooling, at least according to the EVGA website (serial # check).

My specs:

i5 4670K @ 4.2ghz
Gigabyte GA-Z87X-D3H mobo
8GB DDR3
EVGA GTX 1080 FTW
HX 850 Corsair power supply
 
Last edited:
Just had these same symptoms. Installed new EVGA FTW DT and out of the blue, random reboots.
Thought it was video card.
Then motherboard.
Turned out to be power supply.
Think back. Think really hard. I bet it has happened before but you blamed it on the game you were playing or some other random thing.


I know it seems like a coincidence. You install video card and suddenly things go wonky.

Also, random reboots are almost always either psu or memory. imho...
 
Last edited:
In my case (no pun intended), I went from 970 to an 1080, ran without a hitch for about half a year before the random reboots started occuring. It used to run 970 SLI without a hitch, and the system was specced with 970 SLI in mind, so a single 1080 should not even be a problem. Or so I thought.

Eventually the reboots ended up from once a day to several minutes after starting up, and eventually it stopped rebooting, I had to switch the PSU off and back on again before I could switch it on.

I did everything I can to find out what caused it, but replacing the PSU did the trick. I didn't think it was the PSU because it took longer for 970 to trigger the reboot than the 1080 did, especially the fact that I could load the 970 before triggering the shutdown, but I could not even load the 1080 properly before triggering it. The OS definitely wasn't what caused it because I reinstalled my OS several times, even reverting to Win 7, no dice.

The windows event viewer didn't say anything beyond the fact that it experienced a sudden shutdown, so it wasn't software related.

I ended up relenting and bought another PSU to check if it was the issue, with PSU being the part I could test it without wasting as much money, and it ran without a hitch from here onwards.

My PSU was SeaSonic X-850, replaced it with SuperFlower Leadex 650W.
 
How long did you have the R290 running heavily OC'ed on that power supply? Sure it might have been a greater load than the current 1080 setup but look at it this way instead - that heavy 290 load wore out your PS and when you put in the lighter loading 1080, the PS finally gave up the ghost. The 1080 likely didn't do anything to your system but rather was put in just as the PS was about to go out.
 
The power supply is from 2015. So it's been running about 2 years like that.

I set my CPU back to stock (just to rule it out) and left the GPU overclocked, and after playing a game for about 45 mins the system shut off. I'm going to RMA the power supply and see if that resolves it.
 
DrLobotomy said:
RMA that Seasonic and then you will have a new spare.
Click to expand...
I have been meaning to do that actually, still haven't gotten around to doing it. I don't have a particular need to do so yet.

If I ever go for 1080 SLI though, but rewiring the thing is a PITA (Leadex and SeaSonic uses different cables at the PSU, so I have to detach and re-attach every cable...)
 
For what it's worth I was having the same power issues in my PC when I went from SLI GTX 970 to a single GTX Titan X. Turned out to be a loose heatsink over the VRM on my motherboard.
 
At first I thought PSU but you have a plenty powerful, high quality PSU and previously loaded it harder with a 290.. so, maybe not.

+100 is actually a moderately high OC for a FTW.. but, the power consumption is only maybe 5-10w more.

Run OCCT w / Error Checking enabled and see if it's clear. I bet not.
I also have a 1080 FTW and never, ever see artifacts. If its unstable, it just crashes.. either the whole PC or drivers. Most I can get is 2088 w/out errors at 1.093v.

EDIT: Just saw your reply, "I set my CPU back to stock (just to rule it out) and left the GPU overclocked, and after playing a game for about 45 mins the system shut off. I'm going to RMA the power supply and see if that resolves it."

This pretty much confirms its NOT the PSU at fault. My bet is its just an unstable overclock.
 
45454asasa said:
At first I thought PSU but you have a plenty powerful, high quality PSU and previously loaded it harder with a 290.. so, maybe not.

+100 is actually a moderately high OC for a FTW.. but, the power consumption is only maybe 5-10w more.

Run OCCT w / Error Checking enabled and see if it's clear. I bet not.
I also have a 1080 FTW and never, ever see artifacts. If its unstable, it just crashes.. either the whole PC or drivers. Most I can get is 2088 w/out errors at 1.093v.

EDIT: Just saw your reply, "I set my CPU back to stock (just to rule it out) and left the GPU overclocked, and after playing a game for about 45 mins the system shut off. I'm going to RMA the power supply and see if that resolves it."

This pretty much confirms its NOT the PSU at fault. My bet is its just an unstable overclock.
Click to expand...

That doesn't confirm anything. He's still seeing random reboots, regardless of the clocks. Even with the OC on both CPU and GPU he was still pretty damn far from the PSU limits.

I had a similar issue, same symptoms, with the PSU (Corsair HX850) that turned out to be a broken wire in the 24-pin ATX connector.
 
45454asasa said:
At first I thought PSU but you have a plenty powerful, high quality PSU and previously loaded it harder with a 290.. so, maybe not.

+100 is actually a moderately high OC for a FTW.. but, the power consumption is only maybe 5-10w more.

Run OCCT w / Error Checking enabled and see if it's clear. I bet not.
I also have a 1080 FTW and never, ever see artifacts. If its unstable, it just crashes.. either the whole PC or drivers. Most I can get is 2088 w/out errors at 1.093v.

EDIT: Just saw your reply, "I set my CPU back to stock (just to rule it out) and left the GPU overclocked, and after playing a game for about 45 mins the system shut off. I'm going to RMA the power supply and see if that resolves it."

This pretty much confirms its NOT the PSU at fault. My bet is its just an unstable overclock.
Click to expand...

But does the whole system typically just literally turn off with an unstable GPU overclock? I can imagine the whole system crashing (I've seen blue screens under windows 10 with unstable CPU overclocks) but I've never had the system completely turn off with a GPU overclock. Keep in mind I'm not applying any additional voltage, just the raised power limit. I tried lowering the overclock to +85 and I'm still getting the same behavior (card is boosting to around 2025 in games).
 
Last edited:
Bandalo said:
That doesn't confirm anything. He's still seeing random reboots, regardless of the clocks. Even with the OC on both CPU and GPU he was still pretty damn far from the PSU limits.

I had a similar issue, same symptoms, with the PSU (Corsair HX850) that turned out to be a broken wire in the 24-pin ATX connector.
Click to expand...
Only with the GPU oc'd is it randomly rebooting unless I'm misreading something.

SimianRob said:
But does the whole system typically just literally turn off with an unstable GPU overclock? I can imagine the whole system crashing (I've seen blue screens under windows 10 with unstable CPU overclocks) but I've never had the system completely turn off with a GPU overclock. Keep in mind I'm not applying any additional voltage, just the raised power limit. I tried lowering the overclock to +85 and I'm still getting the same behavior (card is boosting to around 2025 in games).
Click to expand...

I'm not sure. If you arent bothered by RMAing the PSU go ahead; that just sounds like a nuisance to me. The only test I can imagine doing is flashing the Asus Strix T4 bios to your FTW (here: http://forum.hwbot.org/showthread.php?t=159025) and setting a custom voltage/frequency curve, say 1.2V and 1900MHz. If it doesn't crash there, but does at say 1.2V and 2050MHz, then it's probably not the PSU. I run that bios fine, btw. It unlocks voltage. I think one of the displayport ports stop working with it but thats not a real problem.

The point of that test being to put more power into the GPU without raising frequency.
 
maybe reseat all power connections? It's a pain but another box to check off the list.
 
45454asasa said:
Only with the GPU oc'd is it randomly rebooting unless I'm misreading something.



I'm not sure. If you arent bothered by RMAing the PSU go ahead; that just sounds like a nuisance to me. The only test I can imagine doing is flashing the Asus Strix T4 bios to your FTW (here: http://forum.hwbot.org/showthread.php?t=159025) and setting a custom voltage/frequency curve, say 1.2V and 1900MHz. If it doesn't crash there, but does at say 1.2V and 2050MHz, then it's probably not the PSU. I run that bios fine, btw. It unlocks voltage. I think one of the displayport ports stop working with it but thats not a real problem.

The point of that test being to put more power into the GPU without raising frequency.
Click to expand...

Corsair offers express RMA, so I wouldn't have any downtime. If I set the voltage without raising the clock, it's not necessarily going to draw the full 1.2 unless it needs to will it? I'm just wondering if that will stress the PSU in the right way. But I get what you're saying, maybe this is an unstable GPU clock, just never seen this behavior from overclocking before.
 
So, I got the issue to happen again but this time without any overclock, just an increased power limit. I set the limit to 130% (with the alternate bios switch on the EVGA FTW) and ran OCCT with error checking. System turned off after 2 minutes. Then I lowered the power limit to 120%, but still stock clocks (+0/0) and system turned itself off after a few minutes of running OCCT again. Haven't been able to replicate this with 100% power limit but haven't run the test for more than 4-5 minutes. Does this seem normal? Any chance that this is actually the GPU? I guess if I RMA the PSU I can at least rule it out.
 
If you have a second set of PCIe 6pin and 8pin cables (and you should since it's a 850W unit), try swapping them in. Make sure the plugs are fully seated in the video and power supply sides.
 
Azrak said:
If you have a second set of PCIe 6pin and 8pin cables (and you should since it's a 850W unit), try swapping them in. Make sure the plugs are fully seated in the video and power supply sides.
Click to expand...

My PSU is only semi-modular. So I was using the PCIE 6+2 pin cables already coming out of the unit to power the card. I just tried using 2 seperate PCIe cables plugged into the additional PCI-e ouputs on the modular portion of the PSU but the system still powered off during the tests :(
 
SimianRob said:
So, I got the issue to happen again but this time without any overclock, just an increased power limit. I set the limit to 130% (with the alternate bios switch on the EVGA FTW) and ran OCCT with error checking. System turned off after 2 minutes. Then I lowered the power limit to 120%, but still stock clocks (+0/0) and system turned itself off after a few minutes of running OCCT again. Haven't been able to replicate this with 100% power limit but haven't run the test for more than 4-5 minutes. Does this seem normal? Any chance that this is actually the GPU? I guess if I RMA the PSU I can at least rule it out.
Click to expand...
That's not normal lol. Yeah, go ahead with the PSU RMA
 
Azrak said:
If you have a second set of PCIe 6pin and 8pin cables (and you should since it's a 850W unit), try swapping them in. Make sure the plugs are fully seated in the video and power supply sides.
Click to expand...
Doesn't the FTW use 2 sets of 8 pin?
 
Sparky said:
Doesn't the FTW use 2 sets of 8 pin?
Click to expand...
Yes it does, but I did not look up the specs of that particular card until after I posted. OP swapped cables and it did not help with the shutdowns.
 
You must log in or register to reply here.
Back
Top