News Users Report Ryzen 7000X3D Chips Burning Out, Killing Motherboards

[Admin]

Redditors warn that some AMD's Ryzen 7000X3D CPUs can burn down on Asus and MSI motherboards.

Users Report Ryzen 7000X3D Chips Burning Out, Killing Motherboards : Read more

 

[InvalidError]

It is extremely unlikely that the substrate would bulge like this from a bad socket contact. If socket contact was the cause, everything would have been visibly discolored by the heat.

Something must have shorted out, caused one or more traces inside the substrate to burn out, vaporizing substrate material and the trapped gasses formed the blister.

 

[drivinfast247]

FYI: Der8auer raised core voltage to an unsafe and damaging level. And the x3d CPU's do NOT boost like the non 3d chips. They have a much lower maximum clock.

 

[Deleted member 2731765]

More likely an overvoltage issue, at least on ASUS mobos.

Particularly in Der8auer's case, it was reported that the motherboard allowed for voltages beyond what AMD 3D V-Cache CPUs could handle and that led to the chip killing itself.

During the overclocking session, the ASUS ROG X670E Crosshair Extreme motherboard allowed voltage tuning but not clock tuning on the AMD Ryzen 9 7950X3D CPU. So voltages were pushed to 1.35V and after a successful boot under LN2 cooling, the voltages were pushed further to 1.55V which seemed to be just too much for the chip to handle.

Since the 3D V-Cache is a sensitive piece of silicon, it is possible that the higher voltages passing through it may have been just enough to kill it entirely as the chip was unable to boot afterward.

Unfortunately, this was where the testing would halt, as the system flatlined with the motherboard reading the "00" error code. It might also be possible that the voltage not being locked by AMD to board partners was unintentional, and could have been something AMD missed upon sending samples ??

 

[TerryLaze]

It is extremely unlikely that the substrate would bulge like this from a bad socket contact. If socket contact was the cause, everything would have been visibly discolored by the heat.

Something must have shorted out, caused one or more traces inside the substrate to burn out, vaporizing substrate material and the trapped gasses formed the blister.

Where do they say that it's socket contact?!
They say the CPU burns out, and that can result in the pins shorting.

 

[helper800]

It is extremely unlikely that the substrate would bulge like this from a bad socket contact. If socket contact was the cause, everything would have been visibly discolored by the heat.

Something must have shorted out, caused one or more traces inside the substrate to burn out, vaporizing substrate material and the trapped gasses formed the blister.

I would assume a socket defect causing a short would lead to what you said, no? For instance if some of the pins were not long enough for contact but long enough for arcing, you have your short. Since sockets are mass produced it would make sense that if this defect did exist it would be in multiple boards from the same manufacturers.

 

[hotaru251]

GamersNexus and the reddit OP are in contact so we can expect a deeper dive into the issue in a future video.


also I doubt Derbaur's is related to this. He purposefully pushed it into dangerous territory knowing they arent made to do so.

 

[InvalidError]

I would assume a socket defect causing a short would lead to what you said, no? For instance if some of the pins were not long enough for contact but long enough for arcing, you have your short.

A short on the pins would have caused the substrate to get burnt to a crisp from the outside, allowing fumes to escape and cause the burn to crater instead, same with arcing. For the substrate to bulge without obvious external discoloration, the heat source has to be internal to the substrate.

Since they also say other instances of this issue have blisters in the same location also with X3D chips, there may be issues with V-cache chiplets burning out or drawing excessive current from whichever rail powers them, taking out the substrate trace that powers it along with the socket it sits on.

 

[InvalidError]

Since the 3D V-Cache is a sensitive piece of silicon, it is possible that the higher voltages passing through it may have been just enough to kill it entirely as the chip was unable to boot afterward.

SRAM leakage current scales a lot like CPU OC when you are at the end of the diminishing return voltage curve: nudge voltage up and current/power flies off the scale. I can imagine how OC experiments could readily fry V-cache chips like that.

 

[Deleted member 14196]

You just got to understand, overclockers just love the burn up stuff

 

[laxman10100]

GamersNexus and the reddit OP are in contact so we can expect a deeper dive into the issue in a future video.

That's good to hear.

I've been with Intel chips for such a long time now and finally decided to grab a 7800X3D and jump ships to AMD for better gaming performance than my 12900K could handle. I am kinda glad that I decided to snag an ASRock motherboard and presumably won't have to deal with this issue. ASRock did, however, release a 4/20 beta BIOS that has "more RAM compatibility" and AGESA 1.0.0.6....and yet it shafted the compatibility for my 6400M/T CL30 XMP memory sticks - even when I set it set to ignore AGESA parameters... I ultimately had to lower it to 6200M/T and raise the CL to 32 in order to boot (but with poor latency). And oddly enough, no other frequencies work (6000 and lower won't post).

But on the flip side, my CPU has a lower average temperature now even though my voltages are the same... Weird...

 

[ottonis]

I think it's pretty obvious that the 7000X3D chips are much more susceptible to the build-up of dangerous heat spots.
The regular Ryzen 7X00s (without the 3D cash) are known to have plenty of temperature-sensors across the CPU, which are crucial in the swift and timely down-and upregulation of dynamic CPU clock (i.e. throttling/overclocking) in order to keep optimal performance depending on thermal conditions.

That being said, the 3D cash might not have as many sensors or they are not ideally implemented, or the temperature build-up on the on the CPU surface beneath the 3D silucon is much faster in the 7X003D chips than the throttling circuits have been originally designed to catch up with.

So, I guess that AMD will probably have to provide a firmware update (AGESA), that will tighten the throttling process.

Even in the cases with burned 3D chips due to manual ocerclocking, AMD would have to consider restricting that even further.

 

[InvalidError]

That being said, the 3D cash might not have as many sensors or they are not ideally implemented, or the temperature build-up on the on the CPU surface beneath the 3D silucon is much faster in the 7X003D chips than the throttling circuits have been originally designed to catch up with.

I doubt heat buildup in or under the X3D cache has anything to do with it: cache has very low power density (~1W/cm^2 vs ~100W/cm^2 for compute logic) and the chip is layered on top of the CCD L2/L3 caches. Nothing substantial happening there under normal operating conditions.

 

[hannibal]

It is interesting to know how many of these chips were overclocked…
Or does the Asus just make bad bios that overclocks automatically?

Well I am sure that we will out sooner than later.

 

[zorgan.roman]

Is it possible that the significant area of the cpu, together on the M/B faces the deformation of the surface, the bowing of the cpu and the am5 socket? Similar to how it was with Intel 13gen?

In that case, it's a problem above and below.

 

[DookieDraws]

I built a new system back in January with a Ryzen 7600X and ASUS B650E-F motherboard, and haven't had any issues so far. Hope it stays this way! I have noticed ASUS has been frequently putting BIOS' out, and I've updated mine less than a month ago. I usually wait awhile before updating, and only do so if there is a good reason to do so. I no longer see those older BIOS' listed. Just 2 recent BIOS' as of today.

Stories like this make me nervous, especially being a Guinee pig early adopter to the AM5 platform.

 

[gendalf]

X3D are more power efficient tho..

 

[Amdlova]

Ryzen = problems ...

 

[outsider2k21]

I built a new system back in January with a Ryzen 7600X and ASUS B650E-F motherboard, and haven't had any issues so far. Hope it stays this way! I have noticed ASUS has been frequently putting BIOS' out, and I've updated mine less than a month ago. I usually wait awhile before updating, and only do so if there is a good reason to do so. I no longer see those older BIOS' listed. Just 2 recent BIOS' as of today.

Stories like this make me nervous, especially being a Guinee pig early adopter to the AM5 platform.

I build my 7950x/asus x670e hero and Dec last year and I haven't have any issue. Thou I only update the bios to 0922 and haven't update again since i don't plan to purchase x3d, I am also running my 7950x at 105w eco mode instead OC.

I wonder if it has anything to do with the motherboard software. I installed asus rog armory crate before due to Asus decide to move fan control from AI suite to armory. After install armory crate my fans will run at 100% after boot into windows even thou cpu tempture were low. I removed armory crate since I can't stand fan running 100% all the time. Later I tried to install armory crate few time and get same result, so i decide to give up armory crate for good.

I do notice at my last try with armory crate, my cpu core voltage was always set at 1.3x even thou I set cpu voltage to auto at bios and without armory crate my cpu core voltage jump between 1.0x to 1.3x base on load.

 

[JTWrenn]

I feel like this headline should say "overclockers" rather than "users". This originally sounded like normal usage in spec was causing burnouts but it sounds like this is just an overclocking issue.

 

[Deleted member 431422]

I feel like this headline should say "overclockers" rather than "users". This originally sounded like normal usage in spec was causing burnouts but it sounds like this is just an overclocking issue.

It's strange those chips are going to some rapair shop and not the manufacturer. People don't send stuff under warranty to some random guy with a soldering iron. Even if that was a certified repair shop, they wouldn't go releasing it to the public just like that.

 

[SunMaster]

Ryzen = problems ...

Oh, our resident cpu guru chips in with tidbits of his infinite wisdom.

Something like your comment does warrant a snide remark.

 

[InvalidError]

It's strange those chips are going to some rapair shop and not the manufacturer. People don't send stuff under warranty to some random guy with a soldering iron. Even if that was a certified repair shop, they wouldn't go releasing it to the public just like that.

If their RMA got denied for whatever reason, they might send it to a repair shop for a 2nd opinion before going back to the manufacturers and telling whichever party is most likely responsible to fix it if they don't want to deal with a small-claims court case that can easily cost them more than the repair or replacement would. If there are enough people with the exact same issue, it could become a recall or class action matter.

Denying problems existed is what AMD initially did with the malfunctioning RX7900 vapor chambers too.

 

[drivannega]

Nope,That's it, 😅 I am buying only all Intel n
Now on

 

[UWguy]

Doubt it’s the motherboard, more likely the CPU.

Now tell me why Intel hasn’t had similar issues?