Potentially broken Rampage IV Formula (BF3 edition)

[Galeton]

Hi,

To just start with it: I've got a Rampage IV Formula (BF3 edition) that's been on the fritz since I got a new GPU three years ago. Specifically, when using my desktop there is a big chance that after a while my screen turns itself because my GPU doesn't give any output. Usually one would think this is a GPU problem, however, the GPU works perfectly fine in other builds (tested it at two other rigs). But to go into more detail, as I've tried a lot, here it goes.

My current setup:
Rampage IV Formula (BF3 edition), with a fully updated BIOS
i7 3820 boxed CPU with Scyte Mugen 3 cooler
Gigabyte GV-N780OC-3GD GPU
16 GB Corsair Vengeance memory
Corsair HX1050 modular psu

The issue:
At first I got the BSOD's while putting (minimal) stress on the GPU. So during gaming, while using GPU heavy programs, benchmarking etc. What would happen is that my screen goes black, the GPU fans start blowing at full speed and the only solution was to reset the entire computer. This has very recently started to increase in frequency, for example, now it also happens while just using VLC player or the internet. In addition, when resetting the computer, it gives the BIOS beep error code that there is no GPU output (1 long, 3 short beeps). The only way to circumvent that issue is to reset the mobo and hope for the best.
In the windows 10 log it always gives the same critical error: error 41 - Kernel-Power. So it is definitely a power issue.

This is a video of the GPU going mental (mainly about the sound as you don't really see much).

What I've tried:
Initially, the solution was to simply switch PCI-e lanes. I switched from lane 1 to lane 2 and the problem was gone for three years. Now it has returned, and it doesn't matter which of the four lanes I use, the same problem happens in all lanes.

• I've switched to all PCI-e lanes.
  
• I've tried to put the GPU in other setups, there is no issue there.
  
• I tried to put in an ASUS GeForce G210 GPU that does not require extra juice from the psu, also without problems.
  
• I have tested RAM, no problem there either (jesus that takes long to test btw).
  
• I reïnstalled all drivers, and updated the BIOS.
  
• I've switched out the cables of the modular psu to make sure it isn't a bad cable.
  

My conclusion:
It's most likely a power issue. However, since there is no constant problem I don't think it's the psu. Therefore, it must be the mobo, and specifically the powersupply to the PCI-e lanes. I think this for multiple reasons:

• first of all, I had this problem a few years back as stated before, and the solution was to switch PCI-e lanes.
  
• second, sometimes there is video output even when the GPU is stressed, so it's not consistent. If it were the psu, I think it would simply shut down as soon as I ask anything demanding of the GPU. It's almost as if the mobo somehow shortcircuits from time to time.
  
• third, some of the USB outlets are also acting up from time to time, not powering my mouse (razer naga), headset (razer megalodon), and keyboard (filco majestouch II).
  

Ideally I would like to test both my psu and mobo separately, but I don't have a spare available of either, nor do I know of any other way to test either component. Therefore I want to simply switch out the broken component. However, before switching out the mobo, I want to verify if you guys think I'm right. I'm too broke to switch out a mobo only to find out it's the psu (or the other way around).

Any ideas or solutions? Is there something I forgot or that I should test before buying new things?

[Arne_Saknussemm]

In the end it can be PSU MB or CPU or RAM...of those I would say PSU and MB as prime suspects and there is no way around it but to swap out components. RAM can test OK but still be the problem....sometimes the CPU IMC is failing but rarer.

I presume this is testing at BIOS defaults...F5 defaults nothing Overclocked etc. no RAM XMP....

Have you tried removing CPU and checking pins in socket....carefully reseating CPU and cooler?

You have ALL ATX (24pin 8pin and 4pin) and ezplug power connectors connected? if applicable...

No friends with a PSU they can lend you....friendly PC hardware store?

[Galeton]

Arne Saknussemm wrote:
In the end it can be PSU MB or CPU or RAM...of those I would say PSU and MB as prime suspects and there is no way around it but to swap out components. RAM can test OK but still be the problem....sometimes the CPU IMC is failing but rarer.
I presume this is testing at BIOS defaults...F5 defaults nothing Overclocked etc. no RAM XMP....

Yes, I reset the CMOS, updated the BIOS and left them at default settings. Nothing is overclocked or tweaked at the moment (although the GPU is standard OC edition).

Have you tried removing CPU and checking pins in socket....carefully reseating CPU and cooler?

I have, nothing seems wrong there. As you said, I don't think it's the CPU.

You have ALL ATX (24pin 8pin and 4pin) and ezplug power connectors connected? if applicable...

I have everything connected as should for as far as I know. The only thing that I can think of that I can still test is to swap the power connector of the mobo itself.

No friends with a PSU they can lend you....friendly PC hardware store?

Hardware stores are scarce these days, let alone friendly ones. I can ask friends, but most have gaming laptops or do not live in the vicinity. Thats why I was hoping you guys could help me out based on this story.

[Arne_Saknussemm]

When you tried GPU in other setups....I presume with different PSU in those cases?

Have you tried a fresh OS install and fresh drivers on fresh install?

Tried performance power plan and turning off any energy saving eup or whatever they're called?

Just out of interest...have you tried removing factory OC on GPU? Downclocking it to stock values with PrecisionX for example...in the interests of being thorough...

[Galeton]

Arne Saknussemm wrote:
When you tried GPU in other setups....I presume with different PSU in those cases?

Yes, simply took my GPU with me and put it in their systems. Less easy to do with mobo / psu.

Have you tried a fresh OS install and fresh drivers on fresh install?

Several times, even tried downgrading driver initially because the very first instance of a BSOD three years ago coïncided with a driver update. At the time I thought it was the new driver, but it wasn't. Also tried multiple OS (Linux Kali, Windows 7, Windows 8, and Winows 10).

Tried performance power plan and turning off any energy saving eup or whatever they're called?

No, but I'm also not sure how I could do that? For as far as I know there is an energy saving on the GPU itself, but not in general. So is there some kind of pogram to do that wit (GPU-z or PrecisionX for example)?

Just out of interest...have you tried removing factory OC on GPU? Downclocking it to stock values with PrecisionX for example...in the interests of being thorough...

Also haven't tried that, more so because I'm not very familliar with overclocking etc. (hence the standard OC version of the GPU).

Do you know of any useful guide that can help me with your last two potential solution factors?

[Arne_Saknussemm]

Ok understood....so may still be PSU issue...

The power plan stuff is usually in BIOS...erp eup or whatever it's called...you would have to enable it so if you haven't all good....the other power related stuff is Windows power plan...in system settings...control panel stuff

I don't know of a guide for downclocking soo....we'll make one 😉

Your GPU is rated 954base / 1006 boost. The standard 780 is 863base / 900 boost
Nvidia always does this in steps of 13 so you would have to downclock 7x13...91

All you have to do is download something like Precision X and select GPU clock offset and press right arrow on keyboard or slide with mouse -91and hit apply...then use the GPU


You can then either click default after a while to return to normal card settings or simply restart and PrecisionX will disappear and GPU will be back to factory OC..

Sometimes instability with factory OC can be shown like this since some cards running standard frequency do not fail while at the factory OC they do fail....just a question of exchanging cards then since the factory OC should be stable. Worth a shot?

[Galeton]

Thanks for the guide 🙂 Much easier than I thought 😛 😉

The problem persisted, but it did provide some additional information. Since I kept the changelog, beneath are the stats at the moment of the BSOD. While nothing special is going on, perhaps it can help exclude some of the possibilities! I'm not sure if asus AI suite also logs its data somewhere? I can't find it, but if it exists I can post the voltage and temp of the mobo as well at the moment of the BSOD.

Power: 61
GPU clock: 1005
Memory clock: 3004
GPU Temp: 60
GPU usage: 10
FB usage: 5
VID usage: 0
BUS usage: 3
Memory usage: 1455
GPU voltage: 1162
Fan: speed: 51
Fan tachometer: 2229
Power limit: 0
Voltage limit: 1
OV max limit: 0
CPU temperature: 48
CPU1 temperature: 44
CPU2 temperature: 45
CPU3 temperature: 45
CPU4 temperature: 48
Framerate: 13
Frametime: 76

[Arne_Saknussemm]

When you say bsod....did the screen go black as before or this time a bsod....any code for the bsod?

[Galeton]

Arne Saknussemm wrote:
When you say bsod....did the screen go black as before or this time a bsod....any code for the bsod?

It's actually no real BSOD. It just turns off (so the video output is interupted). So the same as before happened: at stress the screen turns off, the fans at the GPU blow at full speed, and the only way to counter this is to simply hit the reset button.

[Arne_Saknussemm]

Yeah right you are....so really we're pretty much back where you started...PSU or MOBO as main suspects...

I must say I reckon PSU is main suspect for me since you ran another GPU in the board OK

But at the end of the day it's going to come down to swapping something out... to tell for sure

The PCIe controler on the CPU is a suspect the MB the PSU

😞