[SOLVED] Faulty MB PCIE slots or PSU?

[TheKingKaiser]

Was wondering if anyone is aware of a way to check if PCIE slots are likely/certainly faulty?

Both MB PCIEx16 slots not detecting any GPUs (in bios "not present" and nothing in device manager).

PCIEx16 slots can detect my PCIEx1 network card only in bios, shows up device manager but in error and doesnt run properly.

I have also inspected the slots doesn't seem to be damaged at least from what can be seen visually.

I'm not sure what the above test can conclude, does it mean the MB slots are faulty?

Hardware Specs/Age:
MB: Asus Maximus VII Ranger (~6 yrs)
GPUs: RTX 2070 Super (~2 yrs)/GTX 980-spare (~6 yrs)
PSU: Corsair RM 650 (~7 yrs)

Just wanted to ascertain if MB or PSU is the more probable cause. I have suspicious that this may also be because of a faulty PSU. (More details below).

Extra Background:
Had my PC shutdown on me abruptly and when I tried to turn it back on it was no longer having any video signal, MB post code indicates all else good.

To troubleshoot, I changed the HDMI to intergrated GPU slot, it runs fine.
So I swapped in a spare GPU, also no signal. Noticed GPU fans are not spining at all despite power led on GPU lit up.

Previous week, I had a similar experience where PC shutdown abruptly during a bad storm, upon restart had a message abt power surge protection. After that everything was fine, no issues with GPU video signal.

These 2 events may be linked, Is this a good reason for suspecting the PSU being the issue (power output too low)?

Thank you in advance.

 

[Ralston18]

PSU being 7 years old would, in my mind, be the primary suspect.

Especially if supporting to GPU's.

Look in Reliability History and Event Viewer. Either one or both may be capturing some related error codes, warnings, or even informational events.

One indication of a failing PSU is that, overtime, there are an increasing number of errors and likely more varying errors.

Reliability History uses a timeline format that can be very revealing.

= = = =

And if you have a multi-meter and know how to use it you can do some limited testing on the PSU.

https://www.lifewire.com/how-to-manually-test-a-power-supply-with-a-multimeter-2626158

Not a full test because the PSU is not under load. However, any voltages (3.3, 5, 12) out of tolerance are an indication that the PSU is at EOL (End of Life).

FYI:

https://www.tomshardware.com/reviews/best-psus,4229.html

Not that you need to immediately go out and purchase a new PSU. Link is just to provide some additional information and the suggestion to use the calculators to properly size the wattage needed for your build.

 

[geofelt]

Modern graphics cards will require a UEFI bios to be detected.
Your motherboard is old enough to possibly have a legacy bios.
If the motherboard has been reset to legacy, that would explain the gpu detection issue.

Looking at the newegg reviews for your motherboard, it looks to me like it is not one of the more reliable units that we expect from ASUS.

If the motherboard is faulty, I would not replace it, but consider instead an upgrade of the cpu , motherboard and ram.
$100 could buy you a I5-10100 processor that is fully the equivalent of the strongest cpu your motherboard can support(i7-4790K)
I am not recommending the 10100, but only to indicate that current gen processors are a huge upgrade.

 

[TheKingKaiser]

PSU being 7 years old would, in my mind, be the primary suspect.

Especially if supporting to GPU's.

Look in Reliability History and Event Viewer. Either one or both may be capturing some related error codes, warnings, or even informational events.

One indication of a failing PSU is that, overtime, there are an increasing number of errors and likely more varying errors.

Reliability History uses a timeline format that can be very revealing.

= = = =

And if you have a multi-meter and know how to use it you can do some limited testing on the PSU.

https://www.lifewire.com/how-to-manually-test-a-power-supply-with-a-multimeter-2626158

Not a full test because the PSU is not under load. However, any voltages (3.3, 5, 12) out of tolerance are an indication that the PSU is at EOL (End of Life).

FYI:

https://www.tomshardware.com/reviews/best-psus,4229.html

Not that you need to immediately go out and purchase a new PSU. Link is just to provide some additional information and the suggestion to use the calculators to properly size the wattage needed for your build.

Thanks for the advice.

Event Viewer & Reliablity monitor show multiple errors for unexpected shut downs. Even during periods that I didn't observe any abnormalities, maybe this is a symptom of PSU issues?

I don't have access to a multimeter but, within bios the voltage is shown. 3.3, 5, 12 readings all good, within +/-5%. Not sure how reliable this is vs. using multimeter.

 

[TheKingKaiser]

Modern graphics cards will require a UEFI bios to be detected.
Your motherboard is old enough to possibly have a legacy bios.
If the motherboard has been reset to legacy, that would explain the gpu detection issue.

Looking at the newegg reviews for your motherboard, it looks to me like it is not one of the more reliable units that we expect from ASUS.

If the motherboard is faulty, I would not replace it, but consider instead an upgrade of the cpu , motherboard and ram.
$100 could buy you a I5-10100 processor that is fully the equivalent of the strongest cpu your motherboard can support(i7-4790K)
I am not recommending the 10100, but only to indicate that current gen processors are a huge upgrade.

Thanks for the advice.

Have changed the legacy settings for the PCIE slots and set main GPU to run on PCIE, all no luck in detecting external Graphics Cards on both PCIE x16 slots.

As you mentioned, this model of Asus motherboard is not very reliable had to RMA once due to random hard freezes.
Right now maybe there's some issue that caused both PCIEx16 slots to die on me together.

My PC MB/CPU/RAM update is long overdue anyway, same for PSU.
I guess I'll be ordering them once I research abit more on which ones to get.

Probably gonna swap out the parts one at a time to satisfy my curiosity of which component was actually the culprit.

 

[Ralston18]

Before swapping out parts, do a bit of "preparation":

Power down, unplug, open the case.

Clean out dust and debris.

Ensure by sight and feel that all cards, connectors, RAM, and jumpers are fully and firmly in place.

Use a bright flashlight to inspect, inside and out, for signs of damage.

Bare conductor showing, melted insulation, kinked or pinched wires, signs of blackening or browning, swollen, leaking components, any thing at all.

Look for loose screws, damaged fan blades, dangling wires that can move with air flow changes.

Likely a good time for a new CMOS battery....

While doing so consider what all needs to be done when you start swapping components. Plan it all out and be methodical. Change only one thing at a time and keep notes - especially regarding configuration settings. Record before and after changes are made.

And as always: back up all important data at least 2 x to locations off of the computer in question. Verify that the backups are recoverable and readable.

 

[Mr.Tibbs]

You may have two PCI supply connectors since the PSU has 2 rails. You may have one
only without power.

DEVICE MANAGER, does it say, "simple PCI device driver and a different color. You may need a motherboard specific driver only obtainable from the mother board manufacturer.


https://rog.asus.com/motherboards/rog-maximus/rog-maximus-vii-ranger-model/helpdesk_download

 

[geofelt]

Your PSU is of top quality, so I would normally not expect that unit.
But, in time anything can fail.The only way to determine a PSU failure is to test with a known good quality unit of sufficient power.
Testing outside of lab equipment can only confirm a dead psu, but not confirm proper operation.
See if you can't borrow one to test with.
Failing that, buy a replacement from a shop with a good return policy.
Expect to pay a 15% return fee if you want to return it.

Have you ever replaced the cpu in the socket?
A bent socket pin can cause strange symptoms.
It is all too easy to do if the chip is not dropped in cleanly.
Commonly a ram issue like a non functioning slot or no dual channel operation.
Bent pins can sometimes be pushed back into place.

 

[TheKingKaiser]

Before swapping out parts, do a bit of "preparation":

Power down, unplug, open the case.

Clean out dust and debris.

Ensure by sight and feel that all cards, connectors, RAM, and jumpers are fully and firmly in place.

Use a bright flashlight to inspect, inside and out, for signs of damage.

Bare conductor showing, melted insulation, kinked or pinched wires, signs of blackening or browning, swollen, leaking components, any thing at all.

Look for loose screws, damaged fan blades, dangling wires that can move with air flow changes.

Likely a good time for a new CMOS battery....

While doing so consider what all needs to be done when you start swapping components. Plan it all out and be methodical. Change only one thing at a time and keep notes - especially regarding configuration settings. Record before and after changes are made.

And as always: back up all important data at least 2 x to locations off of the computer in question. Verify that the backups are recoverable and readable.

Thank god, I read your post and a thought came to my mind, maybe the power cable to the GPU is failing.

Took the PC apart to inspect each connection and found this:

Melted vs normal pin.

Plastic melted into PSU slot.

Not entirely sure what led to this, a sign the PSU is about to be done or something else? Could a surge cause such melting?

I switched cable and PSU slot everything works now, seems like the MB PCIE was fine after all.

So I'm a bit perplexed why the network card (Wavlink AX3000) wasn't able to run when plugged into a PCIEx16 slot. Shows up in bios and device manager detects it, but message abt hardware error.

 

[geofelt]

Good job in finding the bad cable

The bright flash was your clue.

On the wavlink, have you installed the driver for it?

 

[TheKingKaiser]

Good job in finding the bad cable

The bright flash was your clue.

On the wavlink, have you installed the driver for it?

Now that I think back I have heard crackling sounds at times when I switch off the power, should have checked it out then. Was lucky damage was quite minimal in this case.

Yea the wavlink driver was definitely installed, works perfect when in the PCIEx4 slots.
When moved to the x16 slots, it doesnt work. Tried updating drivers auto troubleshoot etc. no avail.

Are some network card limited to a certain max no. of lanes? To have them running on a x16, it will have to be lowered in bios to x4?