• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD Ryzen 7 7800X3D CPU Burns Up

I'm supprised the usual youtube suspects are not doing videos about this, but to be fair it's an unfrorseen problem in a lot of respects, so they are probably and correctly waiting untill more information is available.
 


It's not an Asus thing, its an AMD voltage spec issue.

But I supose if you overclock you don't get warranty, that stands for XMP and EXPO, you're running outside the standard JDEC RAM spec.

That said, we all buy RAM on the basis of XMP or EXPO overclocked speeds, so we expect to auto overclock the RAM as a rule, it's assumed to be safe.

Maybe it's a consumer awareness issue, but also a retail issue... stop peddaling super fast RAM if if it's gonna melt your CPU.

Back to the good old days of overclocking when if you push it it too hard your CPU will burn out.
 
Last edited:
I have a suspicion that GN have uncovered some of the wonky LLC issues and the deviations from specified SoC voltage to actual. I know that some of the ASUS ROG series were running SoC as high as 1.45v, and I suspect spiking considerably higher.
 
I have a suspicion that GN have uncovered some of the wonky LLC issues and the deviations from specified SoC voltage to actual. I know that some of the ASUS ROG series were running SoC as high as 1.45v, and I suspect spiking considerably higher.


I think the rabbit hole goes a bit deeper than that. People keep mentioning ASUS mobos, but these seem to be the just the first examples.

Plenty of reports of all MOBO vendors rushing out BIOS patches.

It's an AMD CPU issue, which I'm sure will be patched soon, but with lower RAM voltages/speeds.
 
I've seen this issue now occurs on Asus, asrock,'gigabyte and MSI. Asus just sells the most boards so it always looks like they have a biggest issue

Board makers are tripping over their own feet to try and fix an issue they don't yet understand. And releasing new bios so fast = new bugs don't get picked up and we've already seen today Gigabyte pulled it's new bios that it released just yesterday because it was full of bugs
 
Last edited:
I think the rabbit hole goes a bit deeper than that. People keep mentioning ASUS mobos, but these seem to be the just the first examples.

Plenty of reports of all MOBO vendors rushing out BIOS patches.

It's an AMD CPU issue, which I'm sure will be patched soon, but with lower RAM voltages/speeds.

Oh yes, it affects all motherboards and all 7000 series CPU's

However it affects X3D far more than X series, and also there is a far greater failure rate on Asus motherboards than others - looking at all the failed reports out there.

The issue, as far as AMD and mobo manufacturers are concerned is too high an SoC voltage.
X3D chips are far more sensitive to this, and it looks like some ASUS boards were going just that one step further on the SoC overvolt than other boards.
They all have been overvolting too far when EXPO is enabled, just it appears some more than others (Asus).
 
Last edited:
Oh yes, it affects all motherboards and all 7000 series CPU's

However it affects X3D far more than X series, and also there is a far greater failure rate on Asus motherboards than others - looking at all the failed reports out there.

The issue, as far as AMD and mobo manufacturers are concerned is too high an SoC voltage.
X3D chips are far more sensitive to this, and it looks like some ASUS boards were going just that one step further on the SoC overvolt than other boards.
They all have been overvolting too far when EXPO is enabled, just it appears some more than others (Asus).


As said though we don't know the failure rate. You just see more Asus user complaints because 1) Asus has something like 50% of the entire motherboard market locked up, it's share of sales far exceeds other brands and 2) Asus sells a lot of boards to high end users - their high end Maximus/Crosshair boards sell far more than competing boards from other brands and naturally these are users who will use more high end components and are more likely to do overclocking and tweaking
 
Last edited:
Well, overclocking always carries risks, hence why no warranty.

The strange thing for me is that XMP or EXPO overclocks are taken as a given...

I wonder and think RAM vendors might need to be a little bit more carefull in future.

Especially as AMD CPU's seem to be particulary tetchy when it come to RAM.

By tetchy, I mean setting themselves on fire.
 
Last edited:
Well, overclocking always carries risks, hence why no warranty.

The strange thing for me is that XMP or EXPO overclocks are taken as a given...

I wonder and think RAM vendors might need to be a little bit more carefull in future.

Indeed - you get a warning when you go into the 'AMD overclocking' section of the bios.....maybe you should have a warning when enabling EXPO, as no where does it say EXPO settings are overclocking and may invalidate your CPU warranty. I've not read the small print of the bit of paper with my RAM but i doubt it says something to that effect.
 
It's not an Asus thing, its an AMD voltage spec issue.

Ehhhhh lets see what actually comes out. The failure is so rare that I think more than one issue has to collide to blow up the chip.

What GN is saying is that Asus is blaming AMD for an Asus problem. This doesn't mean there aren't other problems which are AMD problems.
 
Ehhhhh lets see what actually comes out. The failure is so rare that I think more than one issue has to collide to blow up the chip.

What GN is saying is that Asus is blaming AMD for an Asus problem. This doesn't mean there aren't other problems which are AMD problems.

Yeah I'm inclined to agree... I doubt it's an issue for the average user
 
Zen 3 chips were very sensitive to SoC voltages, and anything above 1.2v was considered unsafe and could kill the CPU. Why Zen 4 all of a sudden is any different and was thought to be able to handle 1.4v, i don't know
Zen 3's IO die was manufactured on GlobalFoundries 12nm which was sensitive to high voltage. Zen 4's IO die is manufactured on TSMC 6nm which can take a lot higher voltage, though apparently not as much as they thought.
 
Last edited:
Buildzoid has a very interesting video, he is actually wondering why board manufacturers are all panicking about SOC voltage when the bulging a burning is clearly under the core die not the IO die and is blaming vcore voltage not SOC voltage.

 
Buildzoid has a very interesting video, he is actually wondering why board manufacturers are all panicking about SOC voltage when the bulging a burning is clearly under the core die not the IO die and is blaming vcore voltage not SOC voltage.


Buildzoid is a bit out of date there....that video was done before the statements/articles started coming out explaining how the excessive SoC was burning out the thermal & vcore monitoring systems.

Ryzen 7000 CPU boost algorithm increases power and clock speed until they hit defined limits (thermal, vcore, power etc) - if the limiter/monitor has gone (because it has burnt out in the IO core) then it will just scale up power (vcore VRM requests) until the cores burn out.

As I posted earlier, the current thinking on the sequence of failure is:

SoC overvoltage kills internal CPU monitoring -> MB overvolts CPU core on request of the CPU -> CPU incinerates itself -> CPU burns the MB socket
 
This could just be defective CPUs or even the mobo socket.
I remember when getting a socket 1156 mobo for my i5 750 nehalem cpu there were 2 manufacturers of the socket, foxconn and lotes. The lotes socket was better since if didnt have the issue of poor contact of some pins like the foxconn socket had.

However i wouldn't put defective cpu out of the question. Amd had a compiler bug with first gen ryzen which amd would never admit to and just fixed the issue in cpus manufactured after a certain date and just kept rma-ing the faulty ones. Even to this day they kept silent on the cause of the issue, very shady.
 
The common theme seems to be, windows 11, Asus, RGB, hardcore memory overclocking and redit. Avoiding any two of the five is probably good practice. Someone should make a video.
 
Back
Top Bottom