AMD Ryzen 7 7800X3D CPU Burns Up

[Jamin280672]

so this is what ive set, going to run memtest pro through it for a few hours now and see if it holds out without any issues.

 

[Jamin280672]

Yikes, terrible news for AMD. This shouldn't happen - rushing products to market and not testing for long enough is responsible.

I've swapped my 4090 over to my 13900k system and shall use that for now. I'll wait until AMD/ASUS release fully updated BIOS's for my 7950X3D/X690E Hero and 6000Mhz C30 EXPO kit before powering them on again. Aware I could manually lower voltages, but don't have time to fiddle/stability test with undervolts with work demands currently.

7950X3D is a fantastic CPU overall though, outclasses Intel with extremely good efficiency and performance, but this will hurt AMD's mindshare/public view.

Not long to wait, new bios for ASUS motherboards with soc voltage protection, should be on the site in the next day or 2 https://docs.google.com/spreadsheets/d/19gqxeeAdUmkwgE0EvH4MlJ7meeDbEHYP-FHYW-PMyOk/edit#gid=5220165...

 

[Jamin280672]

Nice Once it's tested for a few weeks in the wild I'll swap my 4090 back over to my 7950x3D system.

Just manually set your voltages, easier and better safe than sorry.

 

[Jamin280672]

Kinda reserved about lowering voltages. Wouldn't lowering them cause stability issues?

only way to find out is test.

 

[Jamin280672]

I've got two rigs setup, 13900k/z690 hero and 7950x3D/X690e Hero, so I've just swapped my 4090 strix over to the 13900k for now, and unplugged the AMD system.

Manually setting voltages is 99.9% likely to be safe, though I'd rather wait for the official fix and spare myself the extra time tinkering.

Will take a thorough look at my 7950X3D pin pads this weekend, to see if any kind of heat/scorch damage to the pads in question, as in the der8auer video etc.

Yeah, I plan on doing the same thing, I got some new fittings coming for my CPU water block, so need to drain the system and get the block off, might as well check the pads and pins at the same time.

 

[Jamin280672]

And best of all, if you use expo = warranty voided. So for the next while anyone trying to RMA a dead x3d is gonna have extra scrutiny applied to see if expo or overvolting is suspected. That guy on Reddit who's 7800x3d got lucky that gamers nexus bought it from him for because Asus/amd declined warranty due to using expo

LOL, im not using EXPO, ive manually set mine to 6000mhz, 2000mhz FLCK and all related voltages, and manually set all my timings including subtimings or rather Buildzoids easy DDR5 timings, are they going to argue that if something goes wrong ?, because if they ask me if I used expo, the answer will be no and id advise anyone faced with the same question to say no too, tell them everything was manually set especially after the news of these issues arose.

 

[Jamin280672]

Bios 1302 (beta) is avail for the x670e Hero. I will be installing it so will report any issues.

SoC voltage for Ryzen 7000X3D series limited to a maximum of 1.30V to protect the CPU and motherboard.

Before running the USB BIOS Flashback tool, please rename the BIOS file (CX670EH.CAP) using BIOSRenamer.

Please note that this is a beta BIOS version of the motherboard which is still undergoing final testing before its official release. The UEFI, its firmware and all content found on it are provided on an “as is” and “as available” basis. ASUS does not give any warranties, whether express or limited, as to the suitability, compatibility, or usability of the UEFI, its firmware or any of its content. Except as provided in the Product warranty and to the maximum extent permitted by law, ASUS is not responsible for direct, special, incidental or consequential damages resulting from using this beta BIOS.

1.3v still sounds a little high to me, especially when my system is running faultless at 1.15v SOC, MC voltage I've dropped to 1.25v, I tested stability last night although I didn't have much time, I left memtest pro running for 5 hours without any errors and then looped CBr23 for 30mins without any issues.

The beta is there for my x670e gene too, I may flash it, might just wait for the final, I think a lot of us have been guinea pigs enough already.

 

[Jamin280672]

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.

Highly speculative rambling about why Ryzen 7000 CPUs are dying.

My Patreon: https://www.patreon.com/buildzoidTeespring: https://teespring.com/stores/actually-hardcore-overclockingBandcamp: https://machineforscreams.bandca...

www.youtube.com

 

[Jamin280672]

f6b
8.84 MB
2023/04/26

1. Checksum : B2CC
2. Optimized for Ryzen 7000 X3D series CPU

Lol gb added it back on again.

B650 GAMING X AX (rev. 1.0/1.1/1.2) Support | Motherboard - GIGABYTE Global

Lasting Quality from GIGABYTE.GIGABYTE Ultra Durable™ motherboards bring together a unique blend of features and technologies that offer users the absolute ...

www.gigabyte.com

The trouble with these bios fixes is they are all of the 3D cache chips, what about the X and non X chips, where's the fixes for those ? Because it's not just the 3D chips having the issues, it's the whole lineup.

I find it quite funny too how AMD are now saying max safe soc voltage was 1.4v, but now they've dropped it to 1.3v, well thanks AMD, but it's a bit late to tell the ones with dead chips that now then deny them a warranty replacement.

 

[Jamin280672]

I don’t think AMD will deny anyone warranty replacement but motherboard manufacturers might not be as willing.

But even if warranty replacement is readily available then I would think it is an inconvenience at the least but really it is a massive balls up to push out products that can kill itself through no fault of the users.

Makes the 12VHPWR issue looks like child’s play. Also doesn’t help AMD was rubbing salt on that Nvidia issue and now they are reaping the rightful flak

They are denying warranty replacements, because overclocking of anything in your system voids your warranty, this includes using expo on your ram which in turn bumps up the soc voltage and eventually kills the CPU, this is what's in the news everywhere, you've got it the wrong way around, Asus are replacing boards last I read, but amd are denying replacements, stating the warranty is void because expo ram overclocking was used.

 

[Jamin280672]

@pc-guy From Reddit

Reddit - Dive into anything

www.reddit.com

Footnote in small print:

Overclocking and/or undervolting AMD processors and memory, including without limitation, altering clock frequencies / multipliers or memory timing / voltage, to operate outside of AMD’s published specifications will void any applicable AMD product warranty, even when enabled via AMD hardware and/or software

 

[Jamin280672]

What this shows is how close to the edge we are with power limits on silicon. I prefer to back it off 10%, get a non-X and 65/88W. Any reports of a non-X CPU having this problem?

Just changed some fittings on my CPU water block, so as it had to come off, I thought id check my chip over, all is good here and I was running SOC at 1.35v (default with expo enabled) for a few weeks, I was suspicious about that high a voltage as my previous Ryzen max soc was 1.2v, so I did ask the question on another forum, the only response I got was, its safe on Ryzen 7000 upto 1.4v which of course turned out to not be true, I then disabled expo and set buildzoids timings manually, at the same time I reduced the SOC voltage to 1.25v, then this issue arose and I further reduced it to 1.15v, anyway, checking the chip and socket tonight, all is good..........its a Ryzen 7900 non X (65watt part) on an ASUS X670E Gene mobo, I was worried that maybe it would be showing signs of bulging though, wish those parts had of arrived sooner so I could have taken the block off sooner, RM are rubbish though, 4 days for 1st class from OCuk.

 

[Jamin280672]

I've installed the 1413 firmware for the Asus x670e Tuf, which is now out of beta.

No change in soc voltage (1.35) for my 7600x with EXPO enabled. I swapped the 7600x for a 7800x3d (which I had lying around waiting for new Asus firmware). Sure enough, with EXPO enabled, auto SOC limits to ~1.26 for this chip. I've manually set it to 1.2 myseld, and seems to be stable for my use.

The SOC limits seem to be applied to x3d chips only, or higher thresholds are allowed for non-x3d chips. The new bios is welcome, but for the time being I'm still going with manual SOC settings regardless of which chip varient I keep.

AMD have released a new agesa to motherboard partners so you can expect another bios update in the next few days, this update that's come out for Asus boards now that's just come out of beta is still based on agesa 1.0.0.6

 

[Jamin280672]

Pretty sure on AM4 vsoc 1.35 is considered well safe.

nope, 1.2v max 3000 and 5000 series chips.

 

[Jamin280672]

Odd that there’s a later BIOS revision that’s a BETA that came out today. The notes aren’t helpful in deciphering the difference. Will stick with 1303.

its probably AMD's fix, via the new AGESA ComboAM5Pi 1.0.0.7

EDIT: Yup, it is, 1410 : https://docs.google.com/spreadsheet...H4MlJ7meeDbEHYP-FHYW-PMyOk/edit#gid=522016589

 

[Jamin280672]

Ah nice one, will probably wait for it to come out of beta as I’ve only just updated the bios this morning can’t be arsed doing it again today haha.

agreed, I wont flash a beta anyway, not with all the issues, it voids warranty, things are bad enough, it will be a pretty fast release anyway I expect, its in testing now.

 

[Jamin280672]

FYI 1410 is in testing, not even beta. Personally i wouldn't try it.

This is on the ASUS X670E Gene website as a beta right now for my board: https://rog.asus.com/motherboards/rog-crosshair/rog-crosshair-x670e-gene-model/helpdesk_bios/

 

[Jamin280672]

They might not all be doing it. In that case the user set 6400Mhz RAM and the motherboard automatically set the SoC to 1.45v, which IS too high, that's LN2 level volts.

Everyone on any motherboard should be checking their SoC voltages now, don't wait for the outcome of this, absolute maximum should be no more than 1.37v, ideally 1.35v or less.

CPU core voltages are difficult to pin down, it depends on what the CPU is doing, in light loads no more than about 1.5v, in heavy loads no more than about 1.3v, but that's not exact.

Where have you had this info from ???

Previously the max was 1.4v, AMD have revised this and the max is now 1.3v, any higher and it will be damage at your own risk, I have never seen any AM5 board auto the soc to 1.45v, they were autoing it to 1.35v which was too high, the max has been adjusted to 1.3v, it's clearly stated on the agesa 1.0.0.7 update notes, but you can probably run a bit lower than this, between 1.2 and 1.25v

 

[Jamin280672]

wccftech have linked the new beta Asus x670e Agesa 1.0.0.7 firmware here for various motherboard models.

I've flashed the 1601 beta firmware on my x670e Tuf (7800x3d). Just enabled EXPOII, left all other CPU voltages set to auto. Ran a few 3dmark benchmarks - all stable. HW Monitor reports CPU_VDDR_SOC and CPU_VDD_MISC were locked at 1.240 and 1.080 respectively. CPU package temp never broke 80c.

What does it say these voltages are at in the bios ? ive seen HWInfo show some inconsistencies, also what did it default the CPU VDDIO / MC VOLATGE to ?

 

[Jamin280672]

Bios says 1.264v/1.376v respectively on auto. I've lowered the SOC voltage to 1.2 manually - it runs cooler and slightly faster. Would have to run more rigourous stress tests for stability, but everything I run seems to be fine so far.

You can lower that MC voltage too, down to around 1.25v runs perfectly fine.