AMD Ryzen 7 7800X3D CPU Burns Up

[Silent_Scone]

I mean it was GN who measured the ASUS ones. I think I trust them enough to get the correct value. Level1tech measured the others, and well I trust him too to be fair. He also knows his stuff.

This wasn't just some random youtubers.

It wasn't an electrical engineer, either. There's no evidence in the video to suggest it was taken from the right place, and I highly doubt he would have asked given the controversial nature of the situation. Given according to their tests there *is* a 50mV difference between what is set, a discerning mind would tell you they may not have been doing it correctly - as this is the kind of behaviour one can expect from measuring elsewhere. That's all I am saying. Unfortunately for ASUS, the auto values were still higher regardless. That doesn't excuse the wrong methodology, though.

 

[Troezar]

Didn't GN say they didn't even test ASRock or MSI boards?
So really we have no idea what they're like.

This is part of my issue, I'd send my Asus back, but what do I get instead?

Personally I think you'll be fine. If you really don't want the Asus look at equivalent Gigabyte or MSI boards. To be fair there's very little in it these days. Check a few reviews on whichever takes your fancy, job done. Build and enjoy

 

[Harlequin]

Didn't GN say they didn't even test ASRock or MSI boards?
So really we have no idea what they're like.

This is part of my issue, I'd send my Asus back, but what do I get instead?

MSI, and part 1 and the breaking news previously said it was 1 from the over vendors.

 

[Rawr]

Didn't GN say they didn't even test ASRock or MSI boards?
So really we have no idea what they're like.

This is part of my issue, I'd send my Asus back, but what do I get instead?

They had level1techs measure the SoC voltage differences between what is reported and what is supplied of MSI, Gigabyte and Asrock boards. So we know that part.



Biostar was also reported to be under reporting too but I don't think we have any confirmed values, just Steve making a offhand comment.

We know Gigabyte didn't manage to screw up the OCP as the board was fine after the CPU got 'melted' on that one. However that CPU seemed to fail due to a bugged bios rather than the way the other CPUs did. Something which apparently is still happening as buildzoid also made a video of how the board was adjusting the SoC in weird ways even without EXPO. Though gigabyte have since taken quite a few bios down off their pages so they might be working on it.

MSI and Asrock are unknowns on the OCP, they could have also messed it up but we don't really know.

The temperature protection is an issue on all boards, with it being set too high at 115C, which is the non X3D value.

 

[Nelly]

My MSI Z77 MPower motherboard had V-Check Connectors and LED debug panel, OC Genie/Power/Reset buttons on it for a costly sum of £150, bought a multimeter just for that. Imagine how much tax the motherboard manufacturers would add on just for having V-Check Connectors added on future boards, they feature tax them enough with the current AMD B650/X670 for having just an LED debug panel.

 

[Rawr]

It wasn't an electrical engineer, either. There's no evidence in the video to suggest it was taken from the right place, and I highly doubt he would have asked given the controversial nature of the situation. Given according to their tests there *is* a 50mV difference between what is set, a discerning mind would tell you they may not have been doing it correctly - as this is the kind of behaviour one can expect from measuring elsewhere. That's all I am saying. Unfortunately for ASUS, the auto values were still higher regardless. That doesn't excuse the wrong methodology, though.

I mean, they have people on the team who do have plenty of experience with electrical engineering though and contacts they can phone to get second opinions. This isn't their first explosion rodeo.

I understand being concerned about most youtubers as they don't have the necessary qualifications, however Gamersnexus isn't one of those. At this point they've proved their experience a number of times.

Edit: To add, while I don't have any time stamps, I'm pretty sure Steve @ GN even commented on what you said about there being different places to measure it when he was on the PCWorld podcast the other day talking about the issue in length. I doubt that would have been commented on if he, or their team, wasn't aware.

 

[Sythe79]

The only thing we really know STILL is the bioses are buggy and badly written as hell, and motherboard vendors don't test their settings. Which is the conclusion a few people here came too before the videos.

Pumping enough power into a CPU with no cooling, so it gets to 200c will kill any CPU that's not the issue. So far, no one seems to have proven there is an endemic hardware problem with either these CPUs or motherboards, again just badly written code.

 

[GoogalyMoogaly]

Personally I think you'll be fine. If you really don't want the Asus look at equivalent Gigabyte or MSI boards. To be fair there's very little in it these days. Check a few reviews on whichever takes your fancy, job done. Build and enjoy

Oh I've been looking, but I don't want to pay Gigabyte prices, ASRock seem to have issues with m.2 speeds and MSI boards seem to have a thing with sharing PCIe lanes with m.2 drives, so with the B650 Carbon it seems 50% of the m.2 drives will reduce the 1st PCIe slot to 8x.

 

[Silent_Scone]

I mean, they have people on the team who do have plenty of experience with electrical engineering though and contacts they can phone to get second opinions. This isn't their first explosion rodeo.

I understand being concerned about most youtubers as they don't have the necessary qualifications, however Gamersnexus isn't one of those. At this point they've proved their experience a number of times.

Edit: To add, while I don't have any time stamps, I'm pretty sure Steve @ GN even commented on what you said about there being different places to measure it when he was on the PCWorld podcast the other day talking about the issue in length. I doubt that would have been commented on if he, or their team, wasn't aware.

I admire the faith here, but faith is based on belief rather than empirical evidence. I've alluded to what happens when measuring from lower on the power plane due to [accounting for] resisted losses, and how it aligns with what GN discovered. In my experience, it doesn't sound like it was measured at the pin pad - at least not for the initial findings.

 

[Rawr]

I admire the faith here, but faith is based on belief rather than empirical evidence. I've alluded to what happens when measuring from lower on the power plane due to [accounting for] resisted losses, and how it aligns with what GN discovered. In my experience, it doesn't sound like it was measured at the pin pad - at least not for the initial findings.

I mean I just went back and checked the video, and they explicitly mention they checked at the pad. So I'm not entirely sure what you're getting at, or even why you're so distrusting of GN. I could understand it if it was a youtuber who doesn't do investigative deep dives into technical/electrical issues, like Linus or Jayz2cents, but GN do have a good history of it.

Oh I've been looking, but I don't want to pay Gigabyte prices, ASRock seem to have issues with m.2 speeds and MSI boards seem to have a thing with sharing PCIe lanes with m.2 drives, so with the B650 Carbon it seems 50% of the m.2 drives will reduce the 1st PCIe slot to 8x.

I noticed that issue with Asus as well when I was looking (before this drama started). Only their most expensive b650 didn't do that and at that point you might as well go for a x670e board because it was near to the price of the cheapest ones of those like the x670e tomahawk.

 

[eddiew]

I would go for MSI if buying now. My PC has been fine so far, I went with 5600 RAM as it was less likely to have issues and I plan to get a 64GB kit at some point. The SOC has always been < 1.28v

Would or wouldn't? Also is that the B650E-E in your signature?

For better or worse, I've already bought 2x32 of 6000c30, which I've since discovered not all boards have on their QVL. I do wonder if QVLs are entirely valid if we bring the SoC down from whatever silly level it already was, but it does seem like there's more safety in picking one that says it does, and MSI's B650 Edge Wifi is in that group. I'm open to other options though

 

[Troezar]

Oh I've been looking, but I don't want to pay Gigabyte prices, ASRock seem to have issues with m.2 speeds and MSI boards seem to have a thing with sharing PCIe lanes with m.2 drives, so with the B650 Carbon it seems 50% of the m.2 drives will reduce the 1st PCIe slot to 8x.

Which models are you looking at? The Asus tax is why I rarely ever buy their stuff.

 

[FredFlint]

Would or wouldn't? Also is that the B650E-E in your signature?

For better or worse, I've already bought 2x32 of 6000c30, which I've since discovered not all boards have on their QVL. I do wonder if QVLs are entirely valid if we bring the SoC down from whatever silly level it already was, but it does seem like there's more safety in picking one that says it does, and MSI's B650 Edge Wifi is in that group. I'm open to other options though

Yes, Asus B650E-E, had it since late November. I never look at the QVL lists, just get whatever RAM is not crazy expensive and hope it works. Sort of think board makers use it to get RAM companies to pay them to get added to the list, any DDR5 should work. Thinking of getting a 6400 48GB or 96GB kit as the new high-density chips are looking good, will probably run it at 6000 though. Should be lots of kits out in the next month or so.

 

[Rawr]

Would or wouldn't? Also is that the B650E-E in your signature?

For better or worse, I've already bought 2x32 of 6000c30, which I've since discovered not all boards have on their QVL. I do wonder if QVLs are entirely valid if we bring the SoC down from whatever silly level it already was, but it does seem like there's more safety in picking one that says it does, and MSI's B650 Edge Wifi is in that group. I'm open to other options though

I think the SoC values boards were applying were far beyond what were actually needed. I've since dropped my SoC to 1.15v from the 1.35v they were originally applying(1.3v with new bios) for 2x16gb 6000c30 expo and I've not noticed any issues at all on a 7950x3d.

Which models are you looking at? The Asus tax is why I rarely ever buy their stuff.

Yeah, that's why I've not really had much Asus stuff too. I never really had anything against them, they were just too expensive every time I checked. I think I had an Asus motherboard with one of my early builds over 10 years ago but that's about it. My last two were gigabyte. This one is my first time with a MSI board and I'm reasonably happy with it.

 

[Silent_Scone]

I mean I just went back and checked the video, and they explicitly mention they checked at the pad. So I'm not entirely sure what you're getting at, or even why you're so distrusting of GN.

Not sure what to tell you. Questioning results doesn't automatically mean distrusting, maybe it does for the Reddit crowd. People make mistakes. I've explained what can be expected from measuring off the power plane and why it would result in what GN discovered. Often, the most obvious reasons are the correct ones (from an electrical perspective, this is it). If you want to weigh in on why else you think there's a 50mV differential you're welcome to do so, there seems to be a lot of speculation going around already surrounding various aspects of this case.

EDIT: Found a moment yesterday to go back over the video, he specifically mentions the motherboard SOC pad which is the probe point on the power plane. There's even an on-screen blurb. It does not reference the correct pins, or even the CPU pad itself.

 

[Harlequin]

A Gamers Nexus investigation forced a manufacturer to recall an entire product line and fix the problem

 

[Jamin280672]

GN said that Asus in the biggest culprit , wait for part 3 and you`ll understand. Those `multiple vendors` are single or low number cases and require stupid steps and disabling protections to replicate; with Asus boards they just go bang out of the box, because asus doing what asus always has done. Asus boards overvolting isnt a new thing.

Ummm no, Buildzoid did a video with a Gigabyte motherboard not sticking to set voltages, and GN's first video which also contained a Gigabyte motherboard show'd the board (no matter what you set the voltage too) spiking 1.45v SOC into the CPU at random, its at 1min 8 seconds of the video, happens right there in the beginning on the GN video, so you dont even have to watch the whole thing, so its far from ASUS alone, no vid's have shown how ASRock and MSI boards behave yet, but one of those destroyed chips came from an ASRock board.

 

[Harlequin]

Ummm no, Buildzoid did a video with a Gigabyte motherboard not sticking to set voltages, and GN's first video which also contained a Gigabyte motherboard show'd the board (no matter what you set the voltage too) spiking 1.45v SOC into the CPU at random, its at 1min 8 seconds of the video, happens right there in the beginning on the GN video, so you dont even have to watch the whole thing, so its far from ASUS alone, no vid's have shown how ASRock and MSI boards behave yet, but one of those destroyed chips came from an ASRock board.

So thank you for confirming exactly what i said then.

 

[GoogalyMoogaly]

Which models are you looking at? The Asus tax is why I rarely ever buy their stuff.

For ASRock I was looking at the X670E Steel Legend.
For Gigabyte anything X670E is just silly prices.
For MSI I was looking at the B650E Carbon WiFi.

Currently have an Asus Strix X670E-F.

 

[01001101]

Didn't GN say they didn't even test ASRock or MSI boards?
So really we have no idea what they're like.

This is part of my issue, I'd send my Asus back, but what do I get instead?

Personally, I would have gone with MSI if I was in your position.