November 4th, 2022 is a day I will never forget. On that day the unthinkable happened: my custom built PC, Twin Turbo, the computer I named after an Uma Musume character, after almost a year of running perfectly fine, literally broke.
It was a quiet Friday evening. I was relaxing at home after work, with the weekend ahead of me. Earlier in the evening I had bought myself some Tiramisu cake, as a small celebration for putting out my original character to the public. That evening I was just wasting time on the internet, as usual, and I started watching the WAN show, to hear the latest tech happenings and random stuff related to LTT. And in the middle of watching the show, my computer froze and the audio was glitch-looping. Oh shit.
Now, I have had a couple small issues with this computer before, but they were all software-related and were easily fixed with a system restore, and I had assumed that this was another one of those. After trying the latest restore point that was available, I somehow ended up with Firefox.exe literally somehow gone even though Firefox was running perfectly fine during the time when the restore point was automatically created. So I tried the only other system restore that was available and my system was seemingly running fine. I thought this was it, but after running the computer for a few minutes it froze again. I later got the latest system restore working correctly, but it didn't solve the root issue. During and after all this system restore stuff my computer literally wouldn't boot sometimes and I also got multiple blue screens of death happen. The error codes on them were "Machine Check Exception" and "WHEA uncorrectable error", which indicate a hardware issue. Fuuuuuck.
I messed around a lot trying to see if I could figure out the problem through just software, but there was no chance I could have done that with the problem I was having. Assuming I could get the computer to even boot, it managed to freeze at all points of use: the startup screens, the recovery screen, or usually a few minutes after being in Windows. However, during this messing around I managed to get a single interesting lead: I could intentionally get the computer to freeze or BSOD by unplugging or replugging the ethernet cable (and in rare cases, with USB drives). This likely indicated a problem with the motherboard, but I wasn't 100% certain. There was another clue I had available that I noticed later.
My MSI motherboard has a series of LED lights that indicate a specific hardware problem that the computer may have. These "EZ Debug LEDs" light up in a normal predictable pattern and then turn off when booting up normally, but in my situation, when my computer failed to boot, one of them remained on. When I looked at these LEDs during a boot failure, the CPU light was on. So this was a CPU problem according to the motherboard? I googled this specific problem looking for a fix, and oh my gosh, I found forum threads indicating that literally every single component in the system could somehow light up that specific LED, even if the CPU was working fine. Also, in rare cases, during a boot failure, the RAM LED lit up on my motherboard. The official troubleshooting solution for all these problems is to basically take your computer apart and put it back together. I tried that, and my computer problems didn't go away unfortunately. Thing is, this official troubleshooting solution is more for people who are assembling a new computer with fresh components and can't get it to POST. This solution was unlikely to help me because my computer booted fine on the first try and was running fine for almost a year.
Given all this information I had, I narrowed down the problem to be either the motherboard, the CPU, or (huge stretch) the RAM or the GPU. My intuition was screaming to me that this was 99% likely a motherboard problem, but I had to be 100% certain before I could proceed. I asked on a couple tech support forums that focus on PC-building what my problem could be, but I either got unhelpful or zero answers. Looks like I was on my own with this issue. So I did probably the dumbest thing I did in this whole fiasco and went to a computer store in my city to buy myself a different CPU that was compatible with this motherboard.
The new CPU did not solve my problem. However, it did allow me to eliminate the GPU as a problem because it came with integrated graphics that my original CPU (Intel i5-10400F) did not possess, letting me know that my Nvidia 1050 Ti was not causing any problems. I returned the CPU to the store on the same day and thankfully got all my money back. Now at this point I was basically 100% certain that my motherboard was the problem and I needed to get it replaced. Figuring out the "replacement" took longer than I expected and the solution was unexpectedly simple in the end.
But first, I want to share the wisdom of the ancients or whatever and let my readers know about which PC components are most likely to fail. Don't ask how I know this, my source is that I've been interested in tech & PC-building since 2008, so I've picked up on some useful information. As far as I know the components that are most likely to fail in a PC are either the motherboard, the GPU, or the PSU. The power supply is an easy problem to eliminate if you buy a higher-tier one or go with a high quality reputable brand. Thankfully I had a Seasonic (The Heart of Your System™) power supply. If you have a low quality power supply it will likely fail early in its lifetime, or will fail if you get too close to its rated wattage (when you put your computer under a heavy load). If you have a high quality power supply, it will likely fail years after its warranty ends or if you go above its rated wattage. If the GPU fails it will cause graphical glitches, won't output any video, or BSOD. CPUs and RAM can break/fail, but do so very rarely (they can basically last forever), and will cause freezing and BSODs if they do fail. Motherboards, the component that ties the whole system together don't boot, freeze, and/or BSOD if they break. This was basically the situation in my case. The CPU and RAM lights plus the ethernet port causing freezing in my situation all pointed to some component on my motherboard not functioning correctly. Maybe it was the chipset?? Anyways, I had to find a replacement for my motherboard.
I basically had three options for replacing my motherboard: buying a new motherboard that is compatible with my CPU, buying a motherboard that isn't compatible with my CPU plus a new CPU that is compatible with it, or sending in my broken motherboard for an RMA (return merchandise authorization) which is effectively just me sending in my broken motherboard for an in-warranty repair. The RMA I was kind of opposed to because I wasn't sure if MSI would really fix my motherboard, and also my introverted ass didn't want to call support to start the RMA process. An RMA would be the cheapest option though. The new, compatible motherboard was not a great option either, because basically all the compatible mini-itx motherboards were either very expensive high-end ones, a weird ASRock motherboard that I didn't like and also seemingly couldn't buy new, or literally the same one I had, just new. The best option of these was re-buying the motherboard I own, and I really didn't want to have this failure happen again with the same new motherboard. Buying a new, incompatible motherboard plus a new CPU would have put me more at ease with worrying about getting another motherboard failure, and I had way more choice and variety with mini-itx motherboards, especially with the Intel LGA 1700 socket. This option however was the most expensive, and I really wanted to spend the least amount of money fixing my computer. So, what option did I pick?
I decided to go with the RMA. Because of the whole calling in thing, I put it off for a couple days. BUT! I managed to figure out that I could do the whole RMA process without ever interacting with a human at MSI. Turns out, MSI had an online form I could fill out with my information to start the RMA process. All I had to do after that was put my motherboard in a box, write my RMA number on the box, and send it to an address in Markham, Ontario that MSI provided, and wait until they fix it and send it back. I had to wait weeks to months for the repairs to happen according to MSI. Though later I found out that MSI had a special little webapp that let me see the progress of my RMA, which was nice.
One thing I'm kind of kicking myself for is not getting one of those in-store warranties when buying my motherboard. Basically if you've ever bought some tech thing and got an offer to buy a warranty that let you replace your tech thing if it breaks for free, that's what I'm talking about. Had I bought that for my motherboard, I would have gotten a brand new one probably quicker than the RMA. If the RMA'd motherboard didn't work when it arrived, or broke again later, my plan is to re-buy the same motherboard and get the replacement warranty with that so I could get it replace just in case it also failed.
While I was waiting for my RMA to happen I needed to use some computer to use. Ain't no way I was going to use my phone or someone else's computer while I waited a month for my motherboard to come back. The best option I had was my mini PC that I modified and played around with Linux on. The fact that literally the month before my computer broke I had basically brought an acceptable replacement home was really great timing! I started using the mini PC the same day my computer broke. And through using my mini PC for a month I managed to learn some interesting things that I hadn't really thought about.
At the time of my computer failure, I had OpenSUSE Linux installed on my mini PC, which so far is my most favourite Linux distro I have used so far. Despite it being my favourite Linux distro so far, I grew frustrated with the idiosyncrasies of Linux and didn't want to spend my time figuring out how to get things working the way I wanted to, so I decided to install Windows 10 on the mini PC after a few days of being on Linux. It turns out that creating a bootable USB stick with the Windows installer ISO on it isn't a straightforward process on Linux, and I couldn't get it to work and boot to the USB. The most convenient option was to use the official Windows media creation tool from Microsoft. Problem is, it only runs on Windows, and I had no Windows machines at home. So I had to go to my parents house and use their Windows PC (that I built for them 😉) to create my Windows install USB.
The Windows installation went smoothly and I was finally back to an operating system I've been familiar with basically my whole life. All I had to do get it set up just the way I like it and I finally had a computer to get by with while I waited for my RMA. Unfortunately this mini PC is really slow and it was mildly frustrating waiting for things to load longer than I'm used to. I had to turn transparency effects off because the GPU was so weak. All this slowness sadly is being caused by modern software being made for the latest powerful hardware and not being optimized for the highest performance which would benefit older, slower hardware like my mini PC. It sucks that just simple web browsing was so slow in current year even though this mini PC is basically a supercomputer compared to average computers a decade or more older, which didn't really do much fancier things than I do today other than higher resolution displays and video.
Speaking of video, my mini PC lagged playing high resolution YouTube videos for some reason and I managed to figure out why. What happened was that 1080p60 videos on YouTube would stutter a lot, and only played smoothly at 720p60, although regular 1080p30 played fine. All this was caused by the video codec that YouTube was using. YouTube serves different codecs for different devices, which either use less bandwidth but require newer devices (VP9) or use more bandwidth but can run on old devices (h264). Basically YouTube decided to use the VP9 codec for me on the web, but my mini PC doesn't have a hardware decoder for that because its processor was made a year before VP9 hardware decoders were a thing. So the weak sauce CPU in my mini PC had to decode it through software, and didn't have enough processing power to smoothly decode VP9. I fixed this by installing a browser extension called h264ify that forces YouTube to use the h264 codec, which my mini PC did have a hardware decoder for, and my 1080p60 videos stopped lagging!
My main computer breaking fortunately didn't result in any data loss because it was an issue of it not being able to run, not it running "normally" and corrupting data. I had a months-old backup (yikes), and I managed to quickly get the most important files off of it in the brief moments I got into the OS before it froze on me (whew). But then I found out that I had some files on my drive that technically weren't super important, and I could get them elsewhere or recreate them, but that would be a pain in my behind, and I didn't want to wait until I got my computer fixed to access them again maybe. Because this drive was an M.2 NVMe drive, I needed to get an adapter for it to use it as a regular USB drive to get the files off of it. This was unexpectedly a bumpy road.
I ordered a Sabrent NVMe enclosure off of Amazon assuming it would work perfectly right away, since it's a brand I'm familiar with and haven't really heard any complaints about. It basically did not work. Windows did recognize some kind of drive was plugged in, but Windows explorer either gave some error when opening the drive, or froze/crashed trying to open the drive. WTF? I honestly was super worried that my drive somehow died and all the files on it were gone because of this. Welp, since the Sabrent enclosure didn't work, I returned it and bought a different NVMe enclosure from a company I have never heard of but seemed legit because it was being sold at the computer store I went to. This new enclosure from a company named Vantec looked way more interesting and came with more things in the box than the Sabrent enclosure. Most importantly though, it worked right away! I got access to all the files on it and transferred them off of the drive in case I need to wipe it later.
This Vantec company seems really strange. The bad copywriting and grammar on the box and on their website makes it seem like this is came from some random Chinese AliExpress seller, but no, Vantec is an American company that's been around for decades, and the NVMe enclosure was made in Taiwan. I'm really struggling to come up with some coherent theory for how an "American" hardware company managed to have bad copywriting on their products. Like does the company literally not employ any native English speakers who would likely come across the bad writing and point out how it looks unprofessional and sketchy? It's genuinely mind-boggling.
Anyway, after about 3 weeks without my motherboard, I finally got it back from MSI a few days ago. After putting everything back together my PC booted fine, and so far hasn't had any issues. The motherboard came back in a really cool looking black "MSI motherboard" box that I wish was the normal packaging. In the shipping box I had a paper that says I got a new serial number and in the remarks section it just said "swapped". I have no idea if this means I got a completely new motherboard or if I got my old one back with some part of it replaced. On the one hand I got exactly all the hardware I sent to MSI back, not a brand new motherboard in fresh packaging. But on the other hand, all my BIOS settings were reset and Windows showed a "setting up hardware" screen during the first boot, which could indicate a new motherboard. A replacement and/or reset of my old motherboard theoretically could have also caused that. Well, regardless of what MSI did during the RMA, I hope I never have to worry about my motherboard breaking again and can peacefully use this computer until I get a newer one. Please don't fail me, MSI!
Alright, that's all I have to say about this whole thing. Thanks for reading. What do you think about all this? Have you ever had your computer break and fixed it?