X870E TOMAHAWK WIFI M.2 issues

gjetryc15a902e2

New member
Joined
Jan 26, 2026
Messages
25
Hello.
I've encountered an issue that I thought was fixed, but it seems like it really wasn't. After finishing my build on a X870E Tomahawk WIFI motherboard I noticed that my Lexar NM990 speeds in slot m.2_1 were not okay. As it turns out, each reboot changes the speed of the slot. Out of about 10 reboots I managed to get PCIe 5.0 x4 speeds once. The rest was either PCIe 3.0 x4 or PCIe 1.0 x4. Changing the settings for this slot in BIOS doesn't do anything. I tested it on BIOS 7E59v2A84 as well as 7E59v2A90, 7E59v2A91 and 7E59v2A95. Each gave the same result.
 

Attachments

  • DiskInfo64_cqXiHbgztL.png
    DiskInfo64_cqXiHbgztL.png
    55.7 KB · Views: 117
Last edited:
I reseated it and encountered very interesting results. I rebooted my PC 10 times after that and 8/10 were PCIe 5.0 with appropriate speeds, BUT 2 of those were 1.0 with 5.0 speeds.

This cannot be a BIOS issue.
This indicates a CPU socket contact issue or bad CPU. ;)
 
We weren't able to replicate the issue on my brother's PC with a 4.0 drive in the first slot. I will now proceed to reseat the CPU. If that fails I'll return both the board and the CPU. It's weird, because I've never had any issues with this CPU before. It works just fine and performs well. No anomalies, artifacts, weird voltages etc. Could it be damaged from earlier BIOS versions when there was this whole burnt CPU fiasco or possibly a bad batch? I've had it since June last year.
 
Nope. It looks like it's an issue with the motherboard or the BIOS. I'm terribly disappointed as I really liked the Tomahawk in every other aspect, but at the same time I'm relieved as the mobo is the only thing with which I'm still in the return period.
 
If we assume that the board/BIOS causes this, then there are two possibilities. 1) It's something with your individual board, then the replacement board should fix it. But 2), if it's a "serial defect" of sorts, then most, if not all such boards would behave similarly, and replacing the board would've been a waste of time and resources. And let's not forget, there's also an outside chance of something being wrong with the CPU, then it might even carry over to a board from a different brand.

I don't think the CPU was damaged from running on earlier BIOS versions. This is not Intel we're talking about, where CPUs could experience deterioration over time from some bugs in the microcode that was bundled in older BIOS versions. With AMD, there were problems before, but nothing like "your M.2 PCIe link speed will drop randomly", or delayed problems like this.

Ideal would be to test with a different AM5 CPU, which is always the most hassle. But that would truly be able to narrow it down to the board.
 
I disassembled my PC and will return this board today. If the issues carry over then we'll be able to eliminate the SSD and motherboard. We'll be left with a CPU.
 
SSD is ruled out, from it happening with two completely different models, the Samsung being a very popular one where this should definitely not occur. With the Lexar I had my doubts due to the controller they like to use on those, but with the Samsung something like this is not known to happen. So strictly speaking it's already between the board or CPU (or the contact between them, but I assume you carefully checked the pins).
 
Also this might sound stupid, but it might lead to something bigger. I recently upgraded my PC and built my brothers PC. Both X870E Tomahawks. His looked perfect without any abnormalities. Mine had a few small, but noticable defects such as missing silver blobs around the screw holes, foggy spots on the back etc. I didn't think of them as meaningful, because why would I care about silver blobs on the motherboard when the socket and everything else looks okay lol. My theory is that it indicates a quality control issue on this specific board that I got, because my brother's Tomahawk never had any issues and booted the same drive properly 20+ times while mine failed to boot with PCI-e 4.0 at 9'th attempt.
 
If I encounter the same issues on the second motherboard that I'll get probably tomorrow then it'll lead me to a faulty CPU which I really wouldn't want to happen. If that happens to be the case then I'll RMA my CPU and buy myself a 9800X3D since they got cheaper due to 9850X3D in my region. In any case I'll update the thread to share my failure or success.
 

Highly unlikely. On there, you can already see from the PCB design that this is not a legit 990 PRO. They just copied the sticker design and put it on some low-end junk with a modded firmware. You would notice this already from the SMART data in CrystalDiskInfo, by which data IDs are available and how it's structured. Most likely they can't fake all that, probably just the name and the firmware version. And of course you would notice it during use, when the speeds are as low as the first ever 3.5" HDDs. But at least it's done much more professionally than this, because that can also happen sometimes.
 
Mine is a legit one. I updated my firmware on it and tested it on my brother's Tomahawk where it reached the advertised speeds 20+ times in a row. My Tomahawk failed at the 9'th attempt with the same BIOS. Seems like I just got a faulty board looking at the things I mentioned above. I'm currently debating which motherboard to order, but I can't find anything better than the Tomahawk. I wish the refresh would be out sooner as I really like the aesthetics of it.
 
Looking at this topic I think it is worthwhile to have a photo of the CPU pins on the motherboard. Spotting bent pins is not an easy task by any standard and having extra pair(s) of eyes inspecting them will help to see whether bent pins are a factor or not.
 
I've been very busy lately, but I finally built my PC on the new board. After 20+ reboots with a new 9850x3d that replaced my 9700x I always got PCI-e 4.0. I hope it stays that way, but for now I've got nothing to complain about with the exception of some driver issues and other windows crap. So far so good.
 
Back
Top