The Dead CPU Isn't the Only Thing Dead | RMA Rescue 1

Summary

Highlights

• GNCA's RMA Rescue series aims to help users with faulty hardware by purchasing and investigating the issues 🎯.
• The episode focuses on a Gigabyte motherboard with issues similar to the ASRock and 9800 X3D problems 📉.
• Bent pins were discovered during troubleshooting, halting further investigation of the initial issue ⚠️.
• Viewers are encouraged to participate by reporting their hardware issues for a chance to be featured 💌.
• The video discusses practical steps users can take to address similar hardware problems themselves 🛠️.

Key Takeaways

• RMA Rescue is a new initiative to help users by buying their faulty hardware for testing 🛠️.
• Gigabyte and ASRock motherboards have known issues with booting and stability, often related to BIOS or voltage 🤔.
• Troubleshooting can reveal hidden problems, like bent pins, which prevent further analysis 🔍.
• Users are encouraged to follow specific RAM installation slots for Gigabyte motherboards for optimal performance 🧩.
• RMA Rescue aims to test company response to hardware issues anonymously 👀.

Overview

Chapters

• 00:00 - 00:30: Introduction The introduction of the chapter welcomes viewers back to GNCA and mentions a playful engagement with the audience by referencing a desktop icon on the presenter's t-shirt. The audience is encouraged to comment, although the interaction has no material reward beyond a symbolic acknowledgment. The chapter sets the stage for a deeper exploration of ASRock and 9800 X3D issues, referencing past reports from the Gamers Nexus channel. It introduces a new series known as the 'RMA rescue' where the team is actively purchasing parts from viewers to investigate these issues.
• 00:30 - 02:30: RMA Rescue Series Overview The chapter discusses the Rescue Series program, highlighting its purpose and functioning. The program buys potentially broken items from individuals, allowing them to use the money as they wish. Even if an item cannot be repaired or used after examination, the individuals keep the money. The goal of the program is to help people avoid the burdens of the RMA (Return Merchandise Authorization) process.
• 02:30 - 05:00: User's Troubleshooting Experience The chapter titled 'User's Troubleshooting Experience' discusses issues faced by users with a Gigabyte motherboard, specifically the Gigabyte X870 Eagle Wi-Fi 7. It highlights how the problems encountered are similar to those experienced with ASRock and 9800 XPD, where users face stability and boot issues, often suspecting the components to be faulty.
• 05:00 - 07:00: Initial Findings and Speculation The chapter titled 'Initial Findings and Speculation' discusses issues related to BIOS for ASRock, which is a part of their RMA (Return Merchandise Authorization) rescue series. The speaker mentions that there will be a dedicated web page where users can learn about the RMA rescue process. Users interested in participating can email RMA rescue at the provided email address.
• 07:00 - 11:00: Board Inspection and Damage Discovery The chapter titled 'Board Inspection and Damage Discovery' discusses the challenges associated with troubleshooting certain types of computer hardware, specifically focusing on a board they purchased. The discussion highlights specific issues related to boot and stability, particularly when using a 9800 X3D. The chapter emphasizes the importance of following specific templates and formats when filtering information. It also outlines the team's testing and troubleshooting processes, which are crucial for identifying and solving the issues with the hardware. The narrative suggests that while troubleshooting can be demanding, the team is committed to thoroughly examining and addressing the problems.
• 11:00 - 15:00: Troubleshooting Steps and User Guidance This chapter provides a guide on troubleshooting a recurring technical problem, specifically focusing on issues with a 9800 X3D and a Gigabyte X870 motherboard, as reported by a user. It compares the issue with similar problems previously identified with ASRock motherboards. The chapter aims to offer a step-by-step checklist for users to identify and resolve such issues.
• 15:00 - 21:00: Conclusion and Wrap-Up The chapter titled "Conclusion and Wrap-Up" discusses various technical issues related to motherboards, with focus on Azrock and Gigabyte brands. The narrator explains differences in reports seen online about ASRock, clarifies the misidentification with Gigabyte, and highlights discrepancies in what precedes these issues. Furthermore, the chapter covers how error LEDs on motherboards attempt to indicate problems, noting these are simply LEDs rather than more advanced seven-segment displays.

The Dead CPU Isn't the Only Thing Dead | RMA Rescue 1 Transcription

• 00:00 - 00:30 Welcome back to GNCA. Bonus points to anybody who recognizes the desktop icon on my t-shirt. You can post the comments below. The bonus points mean nothing, but you'll get a heart on your comment. And YouTube now sent you a notification for that. So, I'm sorry. All right. So, in our attempts to better understand what's going on with the ASRock and 9800 X3D situation, which we reported on the Gamers Nexus channel, we have started buying parts from viewers. And this is part of our new RMA rescue series where
• 00:30 - 01:00 we'll talk about this more as we continue to develop it. But basically, we buy the stuff out when people say it's broken. Uh they get to then go spend that money on whatever they want or keep it. I don't really care. It's their money now. We get a part that hopefully has a problem. And in instances like today where we can't do anything with it after we investigated, well, you still get to keep the money and we still get to keep the part. So, uh, the goal is to help people get out from under needings, go through an RMA process, and
• 01:00 - 01:30 then we get to do the process instead. And that means companies like Gigabyte have more to think about now. So, this is what we're looking at today. This Gigabyte motherboard, which is a Gigabyte X870 Eagle Wi-Fi 7, it has a problem. And the problem really seems to mirror a lot of what we were talking about with the ASRock and 9800 XPD issues where people are just having stability issues right now or boot issues or they think that the part has died. In some cases it has like the CPUs
• 01:30 - 02:00 in very few cases have definitely just incinerated but in most of them it appears to be something related to BIOS for ASRock. Now we cover that separately again that's on Gamers Nexus but because this is part of our RMA rescue series or it's supposed to be uh we're covering it here. So, first of all, if you want to be part of this, we're going to have a dedicated web page on the site. I'll link it below. You can learn about the RMA rescue process and what you need to do and how it works and all that stuff. And then, if you want to participate, you can email RMA rescue gamers nexus.net. Uh although it would really
• 02:00 - 02:30 help us if you check the web page that I'm posting just cuz it'll help us filter them faster. Uh so, we may not read yours with high enough volume if it doesn't follow a certain template and format. So, anyway, this is the board we bought. Now, uh, it exhibits a lot of the same behaviors with boot, with stability, and it was specifically with a 9800 X3D initially. Now, unfortunately, troubleshooting issues that happen sometimes can be kind of a challenge, but we're going to do our best with it. So, to go over some of our findings today, part of this video is going to explain our testing process and our uh troubleshooting process. So, for
• 02:30 - 03:00 this type of problem, if you ever run into it, hopefully you learn something, you take home, a list you can kind of use for your own stuff to do step by step and try to figure out what's going on. So, the user's complaints are up first. So, the user recently emailed us detailing a sudden failure with a 9800 X3D and a Gigabyte X870 motherboard after several months of use. His report resembled the ones we've seen with ASRock motherboards that we'd previously covered. So, we wanted to get it into the office for hands-on testing. Now, there are two key differences between what was going on with most of the
• 03:00 - 03:30 Azrock reports we've seen and what happened with this one. The first one is pretty obvious. It's Gigabyte. It's not ASRock. But we have seen gigabyte reports in those same uh types of threads about ASRock online. Now the second difference we saw was what led up to the issue but also the reading on the error LEDs that are on the motherboard. So the board's got a couple of LEDs on it that will read out and attempt to identify what the issue is. That seems to differ. And actually, unfortunately, it is just LEDs, not a seven segment
• 03:30 - 04:00 display, because that would cost more money and then they would make less. Even though they're going to spend a lot more on RMAs and supports from lacking a seven segment display. I don't know why they do this. I do. It's money. That's That's why they do this. Now, the users who experienced issues on ASRock motherboards typically mentioned the problem arising either after a BIOS update or randomly with no obvious changes before the failure. Additionally, those same ASRock board
• 04:00 - 04:30 users usually noted a zero debug code indicating an issue with the CPU. This user, on the other hand, observed that his PC would no longer boot after reinstalling his RAM modules. He also described his motherboard illuminating a red DRAM status LED. Now, all this is already great for us because again, if we want to do an RMA rescue on this, we've got all kinds of difficult to identify problems and we could send it out to Gigabyte and then test their RMA process in addition to trying to solve
• 04:30 - 05:00 this thing. So, that's kind of the that's the hope we had going into it. We'll it's not going to work out that way. That's the that's the spoiler alert. You'll learn why as we go through. So, here's the timeline of how the user dealt with this issue. Before we received the system, the user did some of his own troubleshooting to assist in identifying the issue, and they did a pretty good job here. The user originally believed this to be a failure caused by either the 9800 X2D, the RAM, the motherboard, or a specific combination between those three components. The user separately tested his RAM, which we didn't buy, and the
• 05:00 - 05:30 9800 X2D in another system, which both functioned as expected. This allowed him to rule out those two parts as being faulty and identify his motherboard, which we now have, as presumably being the cause of the issue. So, this is already great troubleshooting steps. If you have multiple boards or systems available, this is exactly what you should do. And uh this is already setting a good example for how to approach it. The user then describes that when he swapped his previous processor into the Gigabyte board, experiencing the failure, it worked. In a later email to us, however, he mentioned that the previous CPU that had
• 05:30 - 06:00 once worked in the problematic Gigabyte board was now no longer booting with that board after installing it a separate time. Now, this instantly stood out to us because that shouldn't happen. There's not really a great reason for that to happen unless something scorches, and that would normally be obvious through burn marks or something like that. Uh, and because it was swapping to a different CPU from the 98X 2D, it also stood out to us. So for those experiencing the failures on Azrock motherboards, they typically
• 06:00 - 06:30 mention that their board performed normally with processors uh even after their 9800 X3D failure had occurred. So they could in other words put in a 7600 or something and it would all boot up just fine. But that wasn't happening here. And that makes this user situation different again from the one we initially set out to investigate. But it makes this kind of fun. So, uh, the key difference is that the board went from some level of function with some CPUs to no functionality at all after
• 06:30 - 07:00 troubleshooting. Now, for our part, some speculation from us on what we think the initial problem was, kind of ignoring that last bit of information. So, we believe the user was facing issues similar to the boot issues that we've seen with the ASRock and 9800 XPD reports. And uh based on our prior research, we wanted to determine if this is something related to the board's BIOS that could be easily rectified. So for example, in some situations with ASRock, the issues with booting were because the voltage was too low. This isn't all of them. Uh and so boosting the voltage
• 07:00 - 07:30 would help get it uh you know into a bootable state and that could be done through a BIOS update from ASRock. some of the other issues were potentially a voltage too high or at least that was some of the speculation. And uh if this either one of these were the case, then we'd first identify exactly how the issue appears and then investigate further with a whole litany of tests to discover what the problem was with this board and the CPU. Now, ultimately, we didn't get to do our usual in-depth
• 07:30 - 08:00 testing and uh that's that's because well, we'll get that it's there was there was a bent pin there. There was a pin that was bent in half, but the original issue the user had was still a problem. We just we think that in their troubleshooting process a pin got crunched and now there's a new problem that unfortunately is not that actionable for us. But we're still going to walk through all the stuff we planned out because this will
• 08:00 - 08:30 help you with troubleshooting steps. It's going to be basically a road map to test for the same thing. So if you're having instability issues, a boot problem, really just any problem at all with a CPU and a motherboard, this basic list and test plan will kind of walk you through what you can try and troubleshoot. Uh and also we have some additional discussion on this particular board and on the pin. We'll show the pins. We'll show the damage. It is impressive in one spot. But anyway, in terms of our own troubleshooting due to the user reporting DRAM status LED uh
• 08:30 - 09:00 being illuminated, we first wanted to cross off anything related to RAM. And so for us, this includes testing the RAM in each socket, one stick, testing different sticks, and basically doing that whole combination of things with all the permutations of it. A DRAM LED lighting up commonly indicates RAM modules being installed in an incorrect slot. Though it could also be a compatibility issue or defective memory channels or something. So first steps were we put together the test plan of swapping different sticks in different slots, but also making sure that the
• 09:00 - 09:30 stick was installed in the correct slot first in the right order because board manufacturers require that well typically to get the best functionality that the memor is in certain slots if you only have say one or two sticks. Additionally, what we did was uh give the system sufficient time to do memory training. So on AMD especially, they can be slow sometimes booting up first time, especially with higher capacities of memory. And what it's doing is it's just kind of cycling through figuring out, probing, and setting all the subtimings and trying to get it bootable. And that can take time. So manufacturers include
• 09:30 - 10:00 a recommended memory configuration in the manual regarding which slots should be used based on the stick count. Some board vendors, including Gigabyte, also note the order you need to install the sticks in. They say you should install an A2 first. Finally, motherboard vendors maintain a QVL or qualified vendor list for tested memory that they know will work in their boards. This should help narrow issues with broaden compatibility of a particular DRAM package or speed. Now, memory training is expected. We give it 15 minutes at sort of the worst end of things. It has taken longer than that. It's pretty rare, typically with a thread ripper
• 10:00 - 10:30 system with tons of memory. Uh, normally this is something that happens in like one to three minutes, sometimes five. Going more than that's rare. We gave it well we we were planning to give it 15 if we could have put a CPU in it. From there we planned on testing different configurations of RAM using single sticks in separate channels. After that we intended to clear CMOS and then update to the most recent version of BIOS which stated quote optimized memory compatibility end quote. Now unfortunately during our troubleshooting we hit a hard stop. And so if you
• 10:30 - 11:00 remember that the user said the board was working and then suddenly it wasn't working at all in any capacity. That's kind of what we ran into. And the first thing we do when we buy a product that is in any capacity used and has a reported failure is we physically inspect it before ever plugging it into anything. And that's so we can look for things like scorch marks for damage caused by the reported failure. So if a user says their 9800 X3D has blown up, the first thing we would do before socketing it is carefully inspect the
• 11:00 - 11:30 whole CPU and the socket and look for what that damage might have been. because as soon as we try to use it, we no longer know the state it was in. And that's not to look for issues with how the user was approaching it, although those can be discovered there. It's to make sure that when we find an issue, we know it was there and that plugging it in didn't cause some new separate issue. So, we went through that process, looked for physical damage, and part of that process uh meant that unfortunately, we discovered this, which is two bent pins
• 11:30 - 12:00 that were almost unnoticeable without an extremely close inspection. Moving the board under the microscope and again this is out of the box uh testing gave a better idea of what we were dealing with. As seen here two adjacent pins near the bottom lefth hand corner of the AM5 socket were almost completely bent backwards. Using this AM5 socket pinout diagram we identified the two pins affected which corresponded to the X48M_X1 signal used for the integrated clock generator and a VSSs signal used
• 12:00 - 12:30 for ground. Bent pins can be resolved with a magnifying instrument. patience and a needle or a similar tool. The pins are extremely delicate and lose some of their limited durability with every adjustment made. You might see where this is going. One of the pins was completely bent over backwards. So, it was 180° the wrong direction and it was effectively creased at that center line where it was folded. And we tried our best to bend it back as carefully and gently as possible. But man, that gold
• 12:30 - 13:00 is really thin and uh it is really fragile. And when it's bent that badly, like when it has just simply been crushed for whatever reason, uh the opposite direction, there's really not a lot of recovering is possible. But in this case, the pin snapped. Now, in some cases, boards are still able to boot with a single broken pin if that pin happens to correlate to a ground or reserved pin pad or something not used. Sadly, our broken pin wasn't anything that wasn't used. So, that wasn't the case for us. And as a quick aside, there was a time when I had a broken pin on an
• 13:00 - 13:30 AMD PGA CPU that would have been the FX Bulldozer lineup, I think. And in that situation, what I actually did was I took the pin, I put it in the socket, and then I clamped it down, and I was like, "Please work." And it did. You can't do that with LGA, though. And really, you shouldn't be doing it anyway. Like, it shouldn't happen, but it's a good trick to know. Anyway, it's not entirely clear, but we believe that this is a case where originally the motherboard issue was similar to what people are seeing online. And
• 13:30 - 14:00 unfortunately, just through the multiple testing configurations and installations and a removal of the various processors uh during the users troubleshooting process, the pins got bent. Uh and how's that for passive voice? Trying to make it nice for everybody. It's an accident. People can make these mistakes. Happens all the time. We get it. It's okay. And the user gets to keep the money. It's fine. I don't I'm not worried about it. um and they shouldn't get any flak for it because there was a real problem here and the real problem is something that's affecting lots of people like at least hundreds of them that are known. So
• 14:00 - 14:30 unfortunately though we could not proceed with this one to troubleshoot that issue because we became a different issue and additionally the CPU and memory status LED um they're the same color and they're adjacent to each other on the affected board. And so in that situation, it uh it would be easy to misread an LED as DRAM after uh CPU failure or something, assuming you even get power to it at that point. So anyway, ASUS has a similar approach to Gigabyte status LEDs with its own QLEDs. They're less specific than seven segment debug displays. Unfortunately, the LEDs
• 14:30 - 15:00 can highlight the component causing failure. Sometimes they're wrong. A main difference is that ASUS' Q LEDs use different colors for each component. And in a situation like this, it makes sense why. Now ultimately again we think this was the problem we wanted to see if you're experiencing a problem like what was originally described where it's some kind of just issue booting or whatever and really any problem related uh especially to computer hardware but we'll take other electronics too. We have a limited amount of time and actual budget to buy this stuff obviously. So, we're not going to answer every email.
• 15:00 - 15:30 But what you can do if you have an issue is again check the link below for the templating, the format. That will help us a lot with filtering and especially for active investigations we're doing versus sort of stuff that might be new. Both are interesting to us. Uh you can email RMA rescue, one word, RMA rescue gamers nexus.net and we'll look through those. We again we will not reply to all of them but the approach is just like with this user if we talk to you we think okay this is something we'd like to look into or it's something we don't
• 15:30 - 16:00 need to look into but it's a failure that we'd like to do an RMA test of. So for example if you're looking at your board you're like yeah that that cap is literally exploded. There's not much to investigate. Does GN still want it? The answer is yes. And the reason the answer is yes for those is because what we could do instead of testing it is send it out to the companies under an alias. Once we buy it from you, we take ownership of it. We'll ask you for the receipt in case we need it to prove to the manufacturer that we own the device and we have the receipt and uh then
• 16:00 - 16:30 we'll do RMA testing under that alias. So you'll be completely out of this process. We'll be anonymized and we get to test how they react and treat a real user just like we did with our ASUS ROG Ally stuff. So that is the RMA rescue series. this is your kind of first look at it. And in those situations where we want it, uh even if we're unable to recreate the problem you're talking about, the transaction's done. So just like this guy, you keep the money once we buy it, we don't care. And the reason is uh one, we're trying to help out.
• 16:30 - 17:00 Two, even if we can't recreate the issue, you know, we might learn something. Hopefully, we get a video out of it, even if it's like this one, and we can recoup some of that cost or all of it. And if we don't, that's fine. uh I'll give the part to like the Craden Institute or something and they can use it in uh either their e-waste recycling, their educational programs or whatever. It's okay. It's not a big deal. So, um that's the program and part of RMA rescue is that it's meant to be a pretty low hassle approach. So, if we choose the components, we take the full responsibility of them. We take
• 17:00 - 17:30 ownership of them and then we can probe for the RMAs and see how the companies behave. Uh and we'll buy stuff that's not computer parts, but we're going to start with computer parts to get the series going because that's what we know the best. Now, if you're observing these boot failures on your own system and you're using a 9800 X3D, maybe even an ASRock motherboard, we'd suggest trying Flashback. So, ASRock in particular, we'd suggest trying to flash back to BIOS 3.1, which has worked for some users. It's going to be different for different boards, of course. And for more information on that specific configuration, you can check out our past report on the matter where we
• 17:30 - 18:00 compiled several user accounts and shared our findings. If you're encountering issues like were originally described on this board, then we've got a short list of things you can try. There's a lot more you could do, but this is based on typical troubleshooting exercises and also things we found online that work that are specific to these issues. So, Gigabyte says you really should install the RAM and DDR5 A2 first. So, easy enough to do. Might as well try it. We'd recommend isolating to a single stick of RAM and swapping those sticks and then swapping the slots, but do it in a way where you can
• 18:00 - 18:30 keep track of what you've tested. We'd also recommend updating the BIOS and the drivers, including the chipset if you're able to boot at all. uh attempting flashbacks. If you can't boot, you can still flash the BIOS and specifically previous versions. So on Gigabyte, it appears F2 or F3A and we already talked about ASRock. You can also disable fast boot and switch PCIe slot link speed from auto to either Gen 3 or four in BIOS. Some users have reported that this helps them. Switching SATA mode in BIOS from RAID to AHCI,
• 18:30 - 19:00 uh some users have also said helps. You should be aware of what that does beyond that though. And then finally, removing the NVME drive from the PCIe 3.0 slot. Some users have said helped with this specific board. Now, unfortunately, a lot of these are temporary or workarounds. And like most things, you really just wait for the manufacturer to actually fix it. But hopefully that helps somebody. RMAs usually don't cover damage related to bent pins. I I'm not aware of a single one that does. If you know of one, please tell us because that would be an awesome test. But we don't know of one unless maybe it comes out of
• 19:00 - 19:30 the factory that way and you can somehow prove it or they just take your side. So, unfortunately, we just will not be able to use this motherboard for the RMA probing process. We still got to do our RMA rescue. We bailed the guy out, but we're not going to be able to send it to Gigabyte and see what they do because I mean, they they should say the pin is bent. We can't help you or it's going to cost money. And that is a completely reasonable response. So, uh we're looking for other stuff to come in at this point. We have plenty more actually uh that I've gotten intake. Now, in
• 19:30 - 20:00 regards to our imminent RMA rescue series, we recently opened that separate inbox for it. It's for filtering and for sorting faulty parts and for conducting RMAs for them and uh and doing stuff like this where we just try to figure it out and do investigations into issues. Really fun. This kind of stuff's awesome. It's lab work. We get to do the testing and maybe it doesn't fit for Gamers Nexus like this one didn't, especially because that channel, we're so busy right now on that one. But it goes on this one too. And these again, like we said in the past, they are both main channels to us. just different types of content, different flow,
• 20:00 - 20:30 different tone, different style. And for me personally, I I love doing this stuff so far. So, we've shot the one that's public, we have another one that's not up yet, and then we have this one. So, I've shot three so far for this. It's so much fun for me cuz even like the backdrop here suggests, and this table, which are the original two set pieces, uh, after I'd kind of gotten off the card table that we deployed back when I was still working out of the house. This is fun for me because it is a it's a throwback to that and the style just feels like uh like it was back then, but
• 20:30 - 21:00 GN is huge amounts of fun for all the ultra technical stuff. So, they're they're really fun for different reasons. This allows me to explore uh a different hosting style, presentation style, and we get to be a little more casual with it and with people watching it. So, thank you for joining and for checking this out. We'll do more of these now. Uh yeah, RMA rescue at gamersex.net. If you want to support us, you can go to store.gamersex.net. It's all in the same store for the two channels. You can grab a shirt, not like this one. This I'm not going to reveal
• 21:00 - 21:30 it cuz I said I would give out YouTube heart icons that like have value. I don't know, maybe in our post-apocalyptic hellscape dystopian future, you can trade those for currency or something. Invest in in YouTube heart icons, I guess. That's it. See you next time. I like the game. [Music]