Dead stick of ram

[WoloWizardHQ]

Hi everyone, sorry for the long post and for writing in English; I just don't speak German
A little backstory, my ram is Fury 32GB (4x8GB) that's certified for 3200MHz CL16, and I was running DOCP with that profile.
Today I decided to learn more and overclock it, and I did so successfully (even though it's Samsung C-die) the process was messed up but I made it to a stable 3600MHz CL18, with not so bad timings, Vdimm was about 1.3 I think, so it's below the die threshold. (for reference, the mobo is the ROG Crosshair VIII Dark Hero.)
Everything was running fine, did a lot of testing, benchmarks and so on. Afterwards, I didn't play but I used the pc for some general stuff for quite a while.
I found out then that I had a mobo update pending, so I though I'd inatall it. Installed the BIOS update, it finished and was supposed to restart, but it didn't POST afterwards.
The Qled had 0d so I checked it was reserved for future events so that was a dead end.
The process being a BIOS update, I didn't think of checking the issue according to the post leds (it was stuck on dram). The mobo has a flashback feature so I reverted the BIOS to the version I had, but the PC was stuck on the same error with no POST.
I took off all the ram and tried posting with one, it worked. I then kept adding one by one until I found the faulty one. After that I checked different combinations to make sure of the ram, and that the issue is not the dimms, and that was it, a dead stick that the PC won't post if it's installed, no matter how many others are installed, or in what dim.
Anyone had a similar issue?
Could it be corrupted EEPROM?
I would really appreciate any help, and thanks so much for reading as well.
TLDR; a stick of overclocked Samsung C-the ram died after bios update.

 

[emissary42]

Just removed the [Übersicht] tag, since your thread does not qualify for it.

Regarding your question: It is possible that something went wrong during the update process, that damaged the module. Maybe it was just bad luck. It is neigh impossible to verify with certainty, if the system refuses to POST with the module installed. I have heard of similar cases with other motherboards (Gigabyte, maybe also MSI) a couple of times, but to me this is the first time an ASUS motherboard was involved. In general it is a good idea to revert to defaults before updating the BIOS, just in case.

You could try to contact Kingston's tech support about a possible RMA of the kit. As long as there is no visible damage to the modules they might be willing to replace it.

P.S. Welcome to Hardwareluxx

 

[WoloWizardHQ]

Thanks so much! It's too late for RMA and I don't have boxes/receipts anymore.
Tbh, I don't feel like it's dead, I'm trying a dangerous stunt with it right now, going to try to rewrite the eeprom in a dumb way, but it's already dead, so nothing to lose.
I'll report back with what happens tho!!

Beitrag automatisch zusammengeführt: Heute um 02:42

So... Update: My dumb shit worked! That was definitely a win
First, I installed only the other piece of the kit, then, not wanting to pay for Taiphoon or wait, I started checking spd repos, until I stumbled into this [https://github.com/integralfx/DDR4XMPEditor]. Downloaded it and dumped the RAM's SPD (HEX data). Then, the dangerous move lol, I alligned the dead RAM with the DIMM with the PC on, and pushed it straight so that I wouldn't short anything if pushing it slowly or on the sides. To my surprise, the PC didn't restart!! So from the same repo, there's an SPD writer, I checked it and it read the dead stick and showed that the SPD is corrupt, so I wrote the other stick's SPD dump to the dead stick lol.
I restarted the PC and it worked lmaoooo.
Now, both sticks had the same serial number which is not so good for my OCD, so I checked Taiphoon's logs, since I used it before overclocking. I was lucky enough to find the dead stick's hex dump but in text. I opened the working dump and the text in VS Code, but editing was tedious so I sent the text to chatgpt to convert it to a .bin file, and it did. To verify I opened it in VS code and found a line of zeroes missing so I corrected it, then tried to write it. Before writing it showed me a checksum error, so I opened the bin file with the editor that's in the repo (it doesn't have a serial field, so that's why I took the long way), double checked the data and timing and all was correct. So I saved the file from the editor and that seemed to fix the checksum bug. Wrote the new SPD to the stick and Voila!!

 

[WoloWizardHQ]

I just noticed how the XMP profile seems messed up in the previous photo, that's just Taiphoon's gooneyness I guess
Here's a fresh screenshot, and a screenshot from ZenTimings, maybe someone can help improve as today was the first time ever I overclocked ram
Cheers!