MSI MPG X670E CARBON WIFI - PCIe slot no longer works after BIOS update

[ndepoe159002db]

Hi,

At the launch of AM5 I purchased an MSI MPG X670E CARBON WIFI for myself along with a Ryzen 7 7700X and 2x 32 GB DDR5-5600 RAM. For the first few months it has been running pretty smoothly with no major issues.

However last weekend I decided to update the BIOS to the latest version (7D70v12) and while it offers some nice quality-of-life improvements, it did also introduce one major issue for me. In PCIe slots 2 and 3, I have installed two capture cards: an Elgato Game Capture 4K60 Pro MK.2 and an Elgato Game Capture HD60 Pro, respectively. I use these side-by-side for different purposes. After the BIOS update, the HD60 Pro capture card in the PCI_E3 slot from the Chipset is no longer being detected by Windows. There are no errors, no unknown devices in Device Manager, it's like the card isn't there at all.

I've already tried various things: updating drivers for the motherboard including the Chipset drivers, installing the latest drivers and software for the capture cards, updating Windows 10 to the latest version and installing all optional updates. I also reset the BIOS settings to the defaults and checked with that. None of those things solved the problem. At this point I'm very sure it's not a software or settings problem.

To eliminate the possibility of a failure with the capture card itself, I also swapped around the cards around so that the HD60 Pro is in the PCI_E2 slot. And sure enough: now the HD60 Pro is being detected again while the 4K60 Pro card is no longer visible in Windows. So I'm very certain that the problem is with the PCI_E3 slot (the only slot to come from the Chipset and not from the CPU) and it cannot be a coincidence that the problem appeared immediately after updating the BIOS. My guess is that this BIOS version introduced a bug in the Chipset firmware that causes the attached PCIe slot to fail initialization.

I would like to downgrade the BIOS again to the version that came preinstalled on the motherboard (version 110) just to see if that solves the problem, but that version does not appear to be available for download anywhere.

P.S.: I also noticed that after the BIOS update, the Sleep option disappeared in Windows. This bug is already covered and acknowledged in a different thread on this forum and I've confirmed that disabling A-XMP in the BIOS makes the Sleep option reappear, so this seems to be unrelated to my PCIe slot issue described above.

 

[lysy15e102f0]

Added to Report
14 Dec

I replaced the memory and the computer is working fine now.

AMD Ryzen 9 7950X @ 5397.38 MHz - CPU-Z VALIDATOR

[n0wps9] Validated Dump by lysyn (2022-12-13 17:56:24) - MB: MSI MPG X670E CARBON WIFI (MS-7D70) - RAM: 32768 MB

valid.x86.fr

The option to change the memory voltage in the bios does not work. So it has to be?

 

[zerocoolriddle15b702ea]

Hynix 16Gbit M-Die
VDD should be adjustable up to 1.43v.

 

[lysy15e102f0]

Everything works fine with these memories. Hibernation too.

 

[Pixeljunkie]

Everything works fine with these memories. Hibernation too.

I have the exact same RAM as you. What problems are you experiencing with your build besides the known slow boot ?

 

[lysy15e102f0]

I am testing at the moment. I don't see any problems.

 

[lysy15e102f0]

Bios 142

Sleep Mode Disable Fix - Google Drive

drive.google.com

 

[junkreceive15b302e9]

Added To Report
20 Dec
Alan J T
---------------

E7D70AMS.110 - working PCIE_3
E7D70AMS.120 - Broken PCIE_3
E7D70AMS.124 - Broken PCIE_3
E7D70AMS.132 - Broken PCIE_3
E7D70AMS.141 - Unable to boot Windows 11 or the Automatic Repair properly @ 4800 XMP off or 6000 XMP on
E7D70AMS.142 - Unable to boot Windows 11 or the Automatic Repair properly @ 4800 XMP off or 6000 XMP on

One of the 2 sticks of RAM (both same except consecutive serial number).

For reference, on 110, the following stability tests were passed successfully before I even considered the system ready after being build with RAM at XMP speed of 6000:
MemTest86+ : 12+ hours
Prime 95 : Small : 6 hours
Prime 95 : Blend : 6 hours
Prime 95 : Large : 6 hours
Intel Burn Test : Standard : 6 hours
Intel Burn Test : Maximum : 10 hours

EDIT: Note: while I don't believe the issue to be the RAM, I am switching to F5-6000J3040G32GX2-TZ5NR before the new year as I am still in the return window for the current RAM and that will arrive before the end of the year. F5-6000J3040G32GX2-TZ5NR is on the current compatibility list (things always changing). At least by having RAM firmly on the compatibility list at an acknowledged date and time I can future pursue the issue from that regard. Bonus that the exchange will also return $140 to my pocket since the F5-6000J3040G32GX2-TZ5NR has lowered in price compared to what is in the system now.

 

[walkman]

I updated to 142, I am very happy that sleep working but have some issues:
1. When sleep the first time then wake up it, fans of case run max speed (It seems that the power supplied is full 12V for fan, so very loud noise ), but sleep from second then wake up it, fans run normal.
2. if enable EXPO, Appear blue screen when Enable or Disabe "Memory Context Restore". i must set Auto (when set auto: The boot time still slow).

 

[Chris1234]

Added
22 Dec
Alan J T

I'm also having problems with my PCIe_3 slot. It affects my specific PCI card, but also how my pc boots (or rather, fails to boot).

Situation:
I'm using a PCIe 1x card in my third slot. It's the Intel Wigig card for the Vive wireless. When connecting to my VR headset it works, but keeps disconnecting and reconnecting every 5 seconds or so. I did manage to get a stable connection once, but it was lagging heavily and I had a lot of screen tearing every time I moved my head.

What exacerbates the problem is that every time I try to restart my pc it will not boot again. It spends a minute or two doing nothing, then give power to my keyboard, restart by itself again, before settling on the error code A1 and a light in the BOOT debug LED. The only way I could get it to boot into BIOS was by connecting to my CPU's internal graphics and pressing CTRL+ALT+DEL. I could also shutdown the pc completely and wait between 5-10 minutes and it would boot fine, albeit slowly.

Workaround:
Swapping the PCIe card to the second PCI slot has made these problems go away. My VR seems to be connecting and working normally and my pc is able to be restarted, even if it takes a while to boot.

Extra info:
The PCI card seems to be powered at all times. There is a light on the card which is always on and I could always start the software and get it connected, even if it disconnected constantly.
I have not overclocked my system and the problem persists regardless of EXPO profile turned on or off. I have also cleared CMOS once, with the same problems.

Hardware:
CPU: Ryzen 9 7950X [100-100000514WOF]
RAM: G.Skill Trident Z5 Neo [F5-6000J3040G32GX2-TZ5N]
GPU: MSI GTX 1080 Gaming X 8G [Unsure]
PSU: Corsair HX1500i [CP-9020215-EU]
SSD: Kingston Fury Renegade 2TB [SFYRD/2000G] (PCIe 4.0 NVMe, placed in slot M.2_2) - primary boot drive
SSD: Kingston Fury Renegade 2TB [SFYRD/2000G] (PCIe 4.0 NVMe, placed in slot M.2_3)
SSD: Samsung 970 EVO 500GB [MZ-V7E500BW] (PCIe 3.0 NVMe, placed in slot M.2_4)
2xSeagate HDDs
PCIe card: Intel Wigig card for the Vive Wireless
BIOS version 7D70v141

CPU-Z validation:

AMD Ryzen 9 7950X @ 5198.37 MHz - CPU-Z VALIDATOR

[bukrbn] Validated Dump by 7950X22 (2022-12-20 18:57:36) - MB: MSI MPG X670E CARBON WIFI (MS-7D70) - RAM: 65536 MB

valid.x86.fr

 

[Pixeljunkie]

I'm also having problems with my PCIe_3 slot. It affects my specific PCI card, but also how my pc boots (or rather, fails to boot).

Situation:
I'm using a PCIe 1x card in my third slot. It's the Intel Wigig card for the Vive wireless. When connecting to my VR headset it works, but keeps disconnecting and reconnecting every 5 seconds or so. I did manage to get a stable connection once, but it was lagging heavily and I had a lot of screen tearing every time I moved my head.

What exacerbates the problem is that every time I try to restart my pc it will not boot again. It spends a minute or two doing nothing, then give power to my keyboard, restart by itself again, before settling on the error code A1 and a light in the BOOT debug LED. The only way I could get it to boot into BIOS was by connecting to my CPU's internal graphics and pressing CTRL+ALT+DEL. I could also shutdown the pc completely and wait between 5-10 minutes and it would boot fine, albeit slowly.

Workaround:
Swapping the PCIe card to the second PCI slot has made these problems go away. My VR seems to be connecting and working normally and my pc is able to be restarted, even if it takes a while to boot.

Extra info:
The PCI card seems to be powered at all times. There is a light on the card which is always on and I could always start the software and get it connected, even if it disconnected constantly.
I have not overclocked my system and the problem persists regardless of EXPO profile turned on or off. I have also cleared CMOS once, with the same problems.

Hardware:
CPU: Ryzen 9 7950X [100-100000514WOF]
RAM: G.Skill Trident Z5 Neo [F5-6000J3040G32GX2-TZ5N]
GPU: MSI GTX 1080 Gaming X 8G [Unsure]
PSU: Corsair HX1500i [CP-9020215-EU]
SSD: Kingston Fury Renegade 2TB [SFYRD/2000G] (PCIe 4.0 NVMe, placed in slot M.2_2) - primary boot drive
SSD: Kingston Fury Renegade 2TB [SFYRD/2000G] (PCIe 4.0 NVMe, placed in slot M.2_3)
SSD: Samsung 970 EVO 500GB [MZ-V7E500BW] (PCIe 3.0 NVMe, placed in slot M.2_4)
2xSeagate HDDs
PCIe card: Intel Wigig card for the Vive Wireless
BIOS version 7D70v141

View attachment 167015

CPU-Z validation:

AMD Ryzen 9 7950X @ 5198.37 MHz - CPU-Z VALIDATOR

[bukrbn] Validated Dump by 7950X22 (2022-12-20 18:57:36) - MB: MSI MPG X670E CARBON WIFI (MS-7D70) - RAM: 65536 MB

valid.x86.fr

Just a minor question : why aren't you using the M.2_1 slot ?

 

[enearl15b902ef]

I updated to 142, I am very happy that sleep working but have some issues:
1. When sleep the first time, fans of case run max speed (It seems that the power supplied is full 12V for fan, so very loud noise ), but from second, fans run normal.
2. if enable EXPO, Appear blue screen when Enable or Disabe "Memory Context Restore". i must set Auto (when set auto: The boot time still slow).

My memory context restore is also unstable when enabled. I sit on code 15 for around a minute with it set to auto but enabled breaks it altogether.

x670e carbon wifi
ryzen 7950x
4x gskill 6000 cl30 (f5-6000j3038f16gx2-tz5n) some kind of hynix die lul
bios 7D70v141

I am only using the top pcie slot and cpu1 m.2, which are working fine atm.

 

[Chris1234]

Just a minor question : why aren't you using the M.2_1 slot ?

I'm waiting on a PCIe Gen 5 SSD becoming available - then I'll reinstall windows to that and use the Kingstons for games and editing.

 

[Alan J T]

I'm also having problems with my PCIe_3 slot. It affects my specific PCI card, but also how my pc boots (or rather, fails to boot).

Situation:
I'm using a PCIe 1x card in my third slot. It's the Intel Wigig card for the Vive wireless. When connecting to my VR headset it works, but keeps disconnecting and reconnecting every 5 seconds or so. I did manage to get a stable connection once, but it was lagging heavily and I had a lot of screen tearing every time I moved my head.

What exacerbates the problem is that every time I try to restart my pc it will not boot again. It spends a minute or two doing nothing, then give power to my keyboard, restart by itself again, before settling on the error code A1 and a light in the BOOT debug LED. The only way I could get it to boot into BIOS was by connecting to my CPU's internal graphics and pressing CTRL+ALT+DEL. I could also shutdown the pc completely and wait between 5-10 minutes and it would boot fine, albeit slowly.

Workaround:
Swapping the PCIe card to the second PCI slot has made these problems go away. My VR seems to be connecting and working normally and my pc is able to be restarted, even if it takes a while to boot.

Extra info:
The PCI card seems to be powered at all times. There is a light on the card which is always on and I could always start the software and get it connected, even if it disconnected constantly.
I have not overclocked my system and the problem persists regardless of EXPO profile turned on or off. I have also cleared CMOS once, with the same problems.

Hardware:
CPU: Ryzen 9 7950X [100-100000514WOF]
RAM: G.Skill Trident Z5 Neo [F5-6000J3040G32GX2-TZ5N]
GPU: MSI GTX 1080 Gaming X 8G [Unsure]
PSU: Corsair HX1500i [CP-9020215-EU]
SSD: Kingston Fury Renegade 2TB [SFYRD/2000G] (PCIe 4.0 NVMe, placed in slot M.2_2) - primary boot drive
SSD: Kingston Fury Renegade 2TB [SFYRD/2000G] (PCIe 4.0 NVMe, placed in slot M.2_3)
SSD: Samsung 970 EVO 500GB [MZ-V7E500BW] (PCIe 3.0 NVMe, placed in slot M.2_4)
2xSeagate HDDs
PCIe card: Intel Wigig card for the Vive Wireless
BIOS version 7D70v141

View attachment 167015

CPU-Z validation:

AMD Ryzen 9 7950X @ 5198.37 MHz - CPU-Z VALIDATOR

[bukrbn] Validated Dump by 7950X22 (2022-12-20 18:57:36) - MB: MSI MPG X670E CARBON WIFI (MS-7D70) - RAM: 65536 MB

valid.x86.fr

Hi this problem is still under Investigation by MSI
For now, I would leave it in PCI-E-2 slot
The only other option is to roll back to Original Bios
Bios have been Uploaded to the Moderator Bios site
MSI Moderator BETA BIOS Site

 

[Chris1234]

As long as it's being worked on, that's all right. I'll look back in a month to see what's happened. Thank you.

 

[niqingliang20012f202ae]

See post just above yours please

 

[junkreceive15b302e9]

Just reporting in.
I have not yet given the new RAM the full stability tests yet (that will take a few days and determine if I keep the new RAM).

I can, however, report the following with the new RAM:

E7D70AMS.110 - working PCIE_3
E7D70AMS.132 - Broken PCIE_3
E7D70AMS.141 - Unable to boot Windows 11 or the Automatic Repair properly @ 4800 EXPO off or 6000 EXPO on
E7D70AMS.142 - Unable to boot Windows 11 or the Automatic Repair properly @ 4800 EXPO off or 6000 EXPO on

I should also add the following observation. In 132, 141, 142 if I go to BIOS via DEL, then select M-Flash and say yes to reboot, etc the PC does not actually reboot. I have to hit the reset button. 110 correctly reboots the PC from BIOS to M-Flash. This happened both with the old and new RAM. All BIOS versions correctly reboot when saving changes and exiting, it is just the M-FLASH reboot that does not take place when that option is chosen.

 

[zerocoolriddle15b702ea]

Did you disable "Memory Context Restore"? If not, bad things are going to happen...
Looks like they messed it up.
EDIT:
Auto is the only stable setting for me.

PS:
Nice shiny Hynix 16Gbit A-Die. Overclockers favorite IC.

 

[burn172155a02d8]

Added to
Report
07 Jan
Alan J T
-------------


Hi there,

I have the same issue - PCI_E3 does not work when PCI_E2 is populated.
This happens on BIOSes versions above E7D70AMS.110.

Additionally, while on BIOS v110 sometimes PCI_E2 or 3 runs at x1 speed for some reason and while they're at x1 the OS is extremely slow.

I will leave my spec below, for future reference.

Motherboard: MSI MPG X670E CARBON WIFI [911-7D70-001]
CPU: AMD Ryzen 7 7950X [100-100000514WOF]
RAM: Corsair CMK32GX5M2X6600C32 [48-013794 AA] (2x 16GB DDR5-6600; 32-39-39-76; SK-Hynix M-Die 4.D)
GPU: AORUS GeForce RTX™ 4090 XTREME WATERFORCE 24G [GV-N4090AORUSX W-24GD] (PCI_E1)
PSU: Corsair AX850 [CP-9020151-EU]
SSD: Samsung 980 Pro 1TB [MZ-V8P1T0BW] (PCIe 4.0 NVMe/M.2_1)
SSD: Samsung 980 Pro 500GB [MZ-V8P500BW] (PCIe 4.0 NVMe/M.2_2)
SSD: Samsung 970 EVO 1TB [MZ-V7E1T0BW] (PCIe 3.0 NVMe/M.2_3)
SSD: PNY XLR8 CS3030 2TB [M280CS3030-2TB-RB] (PCIe 3.0 NVMe/M.2_4)
Capture card: Elgato Game Capture 4K60 Pro MK.2 [10GAS9901] (PCI_E2)
Sound card: Sound Blaster Z SE [70SB150000004] (PCI_E3)

OS: Windows 10 Pro 22H2 (19045.2364 but it happened previous W10 versions as well)

Highest stable OC was @6200/30 timings but the PCI_E3 issue is not OC-related, I checked.

Memory sticker photo and data:

DDR5 32GB (2X16GB)
6600 Mhz
32-39-39-76
1.40V
ver 5.43.13
CMH32GX5M2X6600C32

 

[Alan J T]

Hi there,

I have the same issue - PCI_E3 does not work when PCI_E2 is populated.
This happens on BIOSes versions above E7D70AMS.110.

Additionally, while on BIOS v110 sometimes PCI_E2 or 3 runs at x1 speed for some reason and while they're at x1 the OS is extremely slow.

I will leave my spec below, for future reference.

Motherboard: MSI MPG X670E CARBON WIFI [911-7D70-001]
CPU: AMD Ryzen 7 7950X [100-100000514WOF]
RAM: Corsair CMK32GX5M2X6600C32 [48-013794 AA] (2x 16GB DDR5-6600; 32-39-39-76; SK-Hynix M-Die 4.D)
GPU: AORUS GeForce RTX™ 4090 XTREME WATERFORCE 24G [GV-N4090AORUSX W-24GD] (PCI_E1)
PSU: Corsair AX850 [CP-9020151-EU]
SSD: Samsung 980 Pro 1TB [MZ-V8P1T0BW] (PCIe 4.0 NVMe/M.2_1)
SSD: Samsung 980 Pro 500GB [MZ-V8P500BW] (PCIe 4.0 NVMe/M.2_2)
SSD: Samsung 970 EVO 1TB [MZ-V7E1T0BW] (PCIe 3.0 NVMe/M.2_3)
SSD: PNY XLR8 CS3030 2TB [M280CS3030-2TB-RB] (PCIe 3.0 NVMe/M.2_4)
Capture card: Elgato Game Capture 4K60 Pro MK.2 [10GAS9901] (PCI_E2)
Sound card: Sound Blaster Z SE [70SB150000004] (PCI_E3)

OS: Windows 10 Pro 22H2 (19045.2364 but it happened previous W10 versions as well)

Highest stable OC was @6200/30 timings but the PCI_E3 issue is not OC-related, I checked.

Thanks for the Info
Could you add a Pick of your Memory Sticker to your post please as it will help in getting this mess sorted.

 

[0gRanid153d02d8]

Guys, hello from Russia. Sorry for the offtopic. I bought this board before I read this topic. I am so disappointed in the poor quality of this product.