[RESOLVED] new TP-Link WiFi 6E AX5400 PCIe WiFi Card (Archer TXE75E) V1 knocks out GTX 970

[Latte5570]

System:
MSI Z97-GAMING 5 + i5-4690K 3.5
2x Gigabyte GeForce GTX 970 SLI
EVGA PS 1000 W 80+ Platinum
Samsung 970 Evo 1 TB
Kingston HyperX Fury 32 GB (4x8GB)
Windows 10 Pro 21H2 Build 19044.1706

 

Behavior:
Upon initial physical and driver installation the TXE75E WiFi showed code 20, and both existing GPUs gave out Code 10 in device manager forcing to use the integrated graphics.
However, despite the exclamation mark, the Wifi component was still recognized, and trying to Update Driver through windows indicated that latest one was indeed installed.
Bluetooth portion of the card is recognized with no errors and WORKS.

Troubleshooting performed so far:

1) Disabled both GPUs and rebooted - this removed the Code 20 from the Wifi Card and it started to work, however GPUs kept the error.
2) Enabling both GPUs and rebooting returned the error codes to them, but WiFi continued to work.
3) Disabling just the TXE75E WiFi and reenabling the GPUs did not get rid of the codes on the GPUs
4) tried disconnecting Bluetooth from onboard f_usb (it's powered separately)
5) tried using 2 other available PCIe ports
6) ran windows update (1 recent security package was installed)
7) updated NVIDA drivers to latest available 512.95 (original was 461.92) clean install
8) disabled SLI and removed 1 GPU
9) disconnecting additional internal drives and all USB devices other than m+kb
10) system was clean but blew out PCIe slots with air compressor anyway
 

Despite the above steps 970 remain w exclamation mark Code 10, with NVIDA control panel not loading as long as Archer TXE75E is plugged into any PCIe slot.
But Archer itself works fine, and so does the rest of the system, except for being stuck with IGP. 
Removing it returns everything back to normal.
 

In filing out mfg support form I noticed that V2 was selectable. Mine is V1.
Could this be related?
 

While I'm waiting for mfg to email me back with the usual scripted response that will ignore half of my steps and suggest I try them over the next few days -  

Anything else you recommend I try, or should I just return it?

Thanks for your time.

[AndreiArgeanu]

Perhaps check the bios for PCIE bifurcation options and if you have them try changing them to see if it helps.

[GeorgeMKane]

22 minutes ago, Latte5570 said:

System:
MSI Z97-GAMING 5 + i5-4690K 3.5
2x Gigabyte GeForce GTX 970 SLI
EVGA PS 1000 W 80+ Platinum
Samsung 970 Evo 1 TB
Kingston HyperX Fury 32 GB (4x8GB)
Windows 10 Pro 21H2 Build 19044.1706

 

Behavior:
Upon initial physical and driver installation the TXE75E WiFi showed code 20, and both existing GPUs gave out Code 10 in device manager forcing to use the integrated graphics.
However, despite the exclamation mark, the Wifi component was still recognized, and trying to Update Driver through windows indicated that latest one was indeed installed.
Bluetooth portion of the card is recognized with no errors and WORKS.

Troubleshooting performed so far:

1) Disabled both GPUs and rebooted - this removed the Code 20 from the Wifi Card and it started to work, however GPUs kept the error.
2) Enabling both GPUs and rebooting returned the error codes to them, but WiFi continued to work.
3) Disabling just the TXE75E WiFi and reenabling the GPUs did not get rid of the codes on the GPUs
4) tried disconnecting Bluetooth from onboard f_usb (it's powered separately)
5) tried using 2 other available PCIe ports
6) ran windows update (1 recent security package was installed)
7) updated NVIDA drivers to latest available 512.95 (original was 461.92) clean install
8) disabled SLI and removed 1 GPU
9) disconnecting additional internal drives and all USB devices other than m+kb
10) system was clean but blew out PCIe slots with air compressor anyway
 

Despite the above steps 970 remain w exclamation mark Code 10, with NVIDA control panel not loading as long as Archer TXE75E is plugged into any PCIe slot.
But Archer itself works fine, and so does the rest of the system, except for being stuck with IGP. 
Removing it returns everything back to normal.
 

In filing out mfg support form I noticed that V2 was selectable. Mine is V1.
Could this be related?
 

While I'm waiting for mfg to email me back with the usual scripted response that will ignore half of my steps and suggest I try them over the next few days -  

Anything else you recommend I try, or should I just return it?

Thanks for your time.

It's possible that the 2nd version of the wifi card fixed some connectivity or other hardware issues.

 

Where exactly on the motherboard did you try installing the wi-fi card? It should be in any of the smaller PCIe x1 (NOT the PCIe x16) ports (the smaller ones, not the longer ones). There should be 4 available PCIe x1 ports on the motherboard, in total. So when you said you tried using the other 2, that's when I got confused.

[Latte5570]

8 minutes ago, AndreiArgeanu said:

Perhaps check the bios for PCIE bifurcation options and if you have them try changing them to see if it helps.

i'll look into that on the next reboot.

thank you. 

[Latte5570]

21 minutes ago, GeorgeMKane said:

It's possible that the 2nd version of the wifi card fixed some connectivity or other hardware issues.

 

Where exactly on the motherboard did you try installing the wi-fi card? It should be in any of the smaller PCIe x1 (NOT the PCIe x16) ports (the smaller ones, not the longer ones). There should be 4 available PCIe x1 ports on the motherboard, in total. So when you said you tried using the other 2, that's when I got confused.

As far as slots, I started with the bottom-most which on this board is x16.
The idea was to keep it away from heat generators.
In the course of above mentioned troubleshooting I then tried the x1 slot between the GPUs (surprisingly a clean fit. nothing touched)
and then the x1 between top most GPU and the CPU cooling fans.
neither made a difference. 

I cant try the other remaining x1's because they're obstructed by the GPU, removing which would counter the objective. 

and yeah, I'm hoping that V2 might do the trick, but it doesn't seem that vendors mark them up as V1 or V2.
Even mfg own website doesn't distinguish between them and the only support material mentions V1. 
Also, there is no downloadable firmware. at least not yet.

thanks guys. 

[GeorgeMKane]

5 minutes ago, Latte5570 said:

As far as slots, I started with the bottom-most which on this board is x16.
The idea was to keep it away from heat generators.
In the course of above mentioned troubleshooting I then tried the x1 slot between the GPUs (surprisingly a clean fit. nothing touched)
and then the x1 between top most GPU and the CPU cooling fans.
neither made a difference. 

I cant try the other remaining x1's because they're obstructed by the GPU, removing which would counter the objective. 

and yeah, I'm hoping that V2 might do the trick, but it doesn't seem that vendors mark them up as V1 or V2.
Even mfg own website doesn't distinguish between them and the only support material mentions V1. 
Also, there is no downloadable firmware. at least not yet.

thanks guys. 

They won't work with any of the x16 slots, as it's a PCIe x1 card only.

 

You got the drivers from tp-link's website? https://www.tp-link.com/us/support/download/archer-txe75e/

 

I do see a note: "Currently 6 GHz Band is only supported on Windows 11." You may have to update to Windows 11, if you don't already have that installed. I would also recommend re-watching the setup videos they have there too, to help you out further.

[Guest]

2 hours ago, GeorgeMKane said:

They won't work with any of the x16 slots, as it's a PCIe x1 card only.

 

You got the drivers from tp-link's website? https://www.tp-link.com/us/support/download/archer-txe75e/

 

I do see a note: "Currently 6 GHz Band is only supported on Windows 11." You may have to update to Windows 11, if you don't already have that installed. I would also recommend re-watching the setup videos they have there too, to help you out further.

Pcie devices are designed to be forwards and backwards compatible and they should work no matter which slot you put it in according to spec.

[Latte5570]

3 hours ago, GeorgeMKane said:

They won't work with any of the x16 slots, as it's a PCIe x1 card only.

 

You got the drivers from tp-link's website? https://www.tp-link.com/us/support/download/archer-txe75e/

 

I do see a note: "Currently 6 GHz Band is only supported on Windows 11." You may have to update to Windows 11, if you don't already have that installed. I would also recommend re-watching the setup videos they have there too, to help you out further.

lane count is automatically negotiated during device initialization. while we cant put a wider lane count / length card into a narrower lane count / length slot, the opposite is supposed to be supported by PCIe base spec.
its possible that this card is derping at that some how, but if that's the case then TP-Link violated the standard which would be surprising from this brand. i've used their products since early oughties. 

the drivers that came on the cd were the same version as those available on the website.
as mentioned earlier when poking at the installed but "un-startable" code 20 card in the device manager, the system still was able to interact with it and in a "just in case" driver Update though windows also came back saying that the driver is up to date. i believe that to be reliable because the chip is intel and the driver provided is even by intel. 

The requirement of Windows 11 / Windows 10 Dev Chanel is only for 6 GHz, which I do not intend to use.
The rest of the cards features are supposed to be fully 64Win10 compatible as per mfg.

over all i do wonder if this is some sort of a hiccup in PCIe lane negotiation.
since the same error occurred when i tried using the card in a x1, maybe the root cause lies with the motherboard when its trying to juggle this x1 card and 2x16 GPUs.
if that's the case i might be SOL because i already have the last mobo BIOS released which was from 2014

[Latte5570]

1 hour ago, WickedThunder86 said:

Pcie devices are designed to be forwards and backwards compatible and they should work no matter which slot you put it in according to spec.

well. somewhat "no matter which"
definitely can't put a longer PCIe card into a shorter PCIe slot

but shorter into longer certainly should not have caused errors

[Guest]

5 minutes ago, Latte5570 said:

well. somewhat "no matter which"
definitely can't put a longer PCIe card into a shorter PCIe slot

but shorter into longer certainly should not have caused errors

You actually can. Many server motherboard actually have the X1 and x4 slots open on the back so that you can actually use a x8 or x16 slot. The only real caveat is the reduced performance.

[Latte5570]

47 minutes ago, WickedThunder86 said:

You actually can. Many server motherboard actually have the X1 and x4 slots open on the back so that you can actually use a x8 or x16 slot. The only real caveat is the reduced performance.

i heard about user mods hacking through the far end 'border', which I never considered to be within the intended spec even if the same auto negotiation can scale down when need.
until now having learned that some slots are actually molded that way. 

good to know. o7

[Latte5570]

so just to follow up and mark this as completed - 

all along I've been trying to troubleshoot this by adding and removing components in various combinations and ending up with roughly the same error state result.

as a last resort i decided to simply pull all PCIe cards, including the GPUs, unplug all peripherals, and reboot into blank mobo with IGP.
i kept the drivers as they were from previous state, and just started plugging the cards in one at a time (starting with Archer) and rebooting per each in hopes of identifying exactly where the conflicts starts to occur. 

with this clean staggered reconstruction no errors came up at all and now I'm all set.

¯\_(ツ)_/¯

thanks everyone for chiming in