For those on AM5, keep an eye out for BIOS updates in the coming days, and do them

[Illusion of Progress]

I don't often do disclaimer type stuff, but I felt this one might be worthwhile if it saves a CPU or spreads awareness. I haven't seen it mentioned by anyone here in the last few days, but I presume a few may be aware of it anyway. Some may not be.

There have been some reports on AM5 CPUs, at least the 7800X3D, developing problems with the bottom of the CPU visibly bulging and the motherboard also having damage (discoloration or burn marks).

While it's thus far been a small number of people, apparently the problem has been considered a risk for all Ryzen 7000 series CPUs, namely when using Expo (in short, memory overclocking, like DOCP/XMP). The SoC voltage is getting high and I guess that's part of what's causing the issues, so the fix is to limit this voltage, which the upcoming BIOS updates will do. So the good news is a fix is coming for a problem that has been identified.

If you're on AM5, I would heavily recommend keeping an eye on your board manufacturer's website for your particular motherboard, and wait for the BIOS update, and then apply it. I would do this whether you've been having issues or not.

And I, personally, would probably remove the CPU from the board and inspect both the CPU and motherboard socket for damage too, especially if you have a 7800X3D, 7900X3D, or 7950X3D, or if you've been using Expo (or otherwise overclocking the memory and thus SoC), or PBO or adjusting voltages from default. I'd probably even check it regardless if I wasn't doing any of those things and just had a regular Ryzen 7000 series CPU just to be sure.

https://www.tomshardware.com/news/amd-issues-follow-up-statement-on-ryzen-burnout-issues-limits-soc-voltages

In the meantime, you may also want to keep an eye on your SoC voltages, and if they go above ~1.3V is when problems can supposedly occur. I don't have a Ryzen 7000 series, but this is the voltage to keep an eye on. If you see it at or going above 1.3V commonly (mine mostly stays static but I'm not sure if it moves on Ryzen 7000), then it's probably not good and I'd double recommend physically inspecting for damage.

https://i.imgur.com/eqYbb4G.png

Use Libre Hardware Monitor, Ryzen Master, HwInfo64, or possibly others if you already have them on hand to check the SoC voltage.

[Phénomènes Mystiques]

It was the same for Ryzen Gen 1... it took some BIOS manufacturers up to 3 years to stabilize.

I would also wait until 2-3 months after BIOS updates are released because those can break the system. I seen it happen for Gen 1.

Also, do not forget about the AMD Chipset driver. And of course, save the current working BIOS somewhere before updating, just in case it fails.

And please read all red lettering disclaimers for your motherboards and follow those directions carefully and fully. Do not skip anything! Pay attention...

[skOsH♥]

Lol

"Hey guys we know you recently upgraded but anyone who bought this series cpu, um, your board and cpu may be permanently ruined. But, we fixed it."

[emoticorpse]

d¢ eredeti hozzászólása:

Lol

"Hey guys we know you recently upgraded but anyone who bought this series cpu, um, your board and cpu may be permanently ruined. But, we fixed it."

I don't mind the incompetence in this specific situation. Knowing my board might melt/cave in at any moment? it's the kind of exhilaration that beating a game just can't get me.

[Illusion of Progress]

d¢ eredeti hozzászólása:

Lol

"Hey guys we know you recently upgraded but anyone who bought this series cpu, um, your board and cpu may be permanently ruined. But, we fixed it."

The situation is unfortunate, but then what are you suggesting?

That it was done with malicious intent? I find this unlikely. Incompetence, maybe.

That AMD isn't responding satisfactorily? AMD is offering support to anyone affected. They've also issued a new AGESA/BIOS to stop the boards from giving the SoC too much voltage when Expo is used (almost like when they say memory overclocking is overclocking, they mean it or something, who knew?). If you're not affected, update your BIOS to mitigate the risk going forward.

That things should always be perfect and mistakes should never happen? As much as I wish it with you, that's not the reality of our world. I said in my last post but I've watched this happen across the gamut of companies and the gamut of PC parts (or rather, gamut of markets because it happens everywhere).

And at least it was addressed this time. I've seen plenty of "responses" where the issue gets swept under the rug. Sometimes the part gets a revision and those on the old one are left dealing with it. Been there done that recently with Intel. Samsung recently did it with SSDs. nVidia's done it. Microsoft has done it. Has anyone NOT made a mistake?

[emoticorpse]

Seems like that bios just came out for my motherboard between yesterday when I checked and it wasn't available and today that I checked again. Will update it right now.

[Rod]

Soc and ram related? My rams 1.35v. Does this soc limit of 1.3 effect the ram voltage?

Thats the expo it was sold under, 1.35v 6000mhz expo.

[Illusion of Progress]

RAM voltage is different. Your RAM itself is fine running 1.35V.

The SoC is the "system on a chip". It's part of the CPU itself. The IMC (integrated memory controller) is part of that, and it will need more voltage with more demanding memory configurations. For example, I see mine (on both Zen 2 and 3 CPUs, so not Zen 4) go to ~1.1V (see picture in my first post) instead of a more normal 1.05V for those CPUs, probably because I have four modules of dual rank RAM running at 3,600 MHz. Apparently this runs much higher with Zen 4 though.

So when Expo was enabled, the IMC was being given these higher voltages to be able to run with the RAM at that speed. Apparently there were also spikes ABOVE this 1.35V which is, from the way I'm interpreting it, what was really contributing to the problem.

While the problem has been identified that it could affect any AM5 CPUs, the fact that it infamously started with the 7800X3D, a recent CPU, means the damage was done in short order. So if you have a non-X3D CPU and apply the BIOS update, I really wouldn't worry too much.

[󠀡󠀡]

Thank You!! Will Keep Look Out For This.

[r.linder]

SOC used to be defaulting to around 1.1V and they recommended staying below 1.2V, not sure why it's such an issue now.

[Illusion of Progress]

That was Zen 2/3 on AM4 I presume? My Zen 2 and Zen 3 CPUs both ran at 1.1V on the SoC (with a rather memory heavy configuration).

It seems typical for it to run higher on Zen 4 though. I'm not sure if there's any tolerance difference but there's still going to be a limit of course, and apparently they are flirting with it. Running at 1.35V+ long term and/or spiking even higher (1.4V or 1.5V) is risking burning them up.

[r.linder]

Illusion of Progress eredeti hozzászólása:

That was Zen 2/3 on AM4 I presume? My Zen 2 and Zen 3 CPUs both ran at 1.1V on the SoC (with a rather memory heavy configuration).

It seems typical for it to run higher on Zen 4 though. I'm not sure if there's any tolerance difference but there's still going to be a limit of course, and apparently they are flirting with it. Running at 1.35V+ long term and/or spiking even higher (1.4V or 1.5V) is risking burning them up.

Yeah earlier gens ran really low on SOC

[_I_]

Rod eredeti hozzászólása:

Soc and ram related? My rams 1.35v. Does this soc limit of 1.3 effect the ram voltage?

Thats the expo it was sold under, 1.35v 6000mhz expo.

ram has its own vrm
the imc on the cpu is different

[skOsH♥]

Mine was always hard set to 1.15 so I think I am good

And apparently it's only for the 3d versions

[PopinFRESH]

d¢ eredeti hozzászólása:

Mine was always hard set to 1.15 so I think I am good

And apparently it's only for the 3d versions

There are multiple simultaneous issues which are getting conflated together. The issue on the CPU side is not only on the X3D versions but they are more sensitive to the issue.

The issues on the motherboard side for those that have them is that the voltage set (regardless of being set to Auto or manually set to a specific voltage) isn't respected and runs beyond the safe spec provided by AMD.

[_I_]

d¢ eredeti hozzászólása:

Mine was always hard set to 1.15 so I think I am good

And apparently it's only for the 3d versions

its the boards monitoring thats wrong, and a safe setting it in bios can allow it to be off by enough to kill the cpu or burn the board

the boards mfg are being lazy and not forcing safe voltages or current protections
at 1v, pulling 200w is 0.005ohm, nearly a dead short, which is what the board would need to detect after things go horribly wrong to protect the board and its socket