Question Asus mobos burning x3D CpuS?

[VirtualLarry]

Is this newsworthy?

 

[Markfw]

Well, I can say only that the 7950x that I am typing on, is on an ASUS motherboard, and I get getting reboots and its a bug check that indicates memory, but this is the same memory that I have on 4 other ASRock motherboards as EXPO 6000 cl30 gskill memory, and just noticed today that the bios I am using has been removed. (1410 is nowhere to be found)

Bottom line ? I suspect ASUS is having BIOS issues with AM5. Not going to get any more ASUS motherboards. PERIOD. Until this situation is resolved.

 

[Tech Junky]

Asus has an issue with a cap being reversed polarity recently causing system fires as well.

 

[OlyAR15]

Yeah, I've been seeing so many posts about burnt AMD chips in Reddit. Scary stuff. At first it seemed to be an issue with Asus boards, but now other manufacturers have been implicated.

 

[VirtualLarry]

Asus has an issue with a cap being reversed polarity recently causing system fires as well.

I thought that was with their Z690 boards, not AM5.

 

[aigomorla]

So im hearing its only on the X3D cpu's.
And its not just ASUS but all the vendors.

I am willing to guess it has to do with that boosting profile the Ryzen 7000 series has.
That mixed with the larger cache on the X3D could trip a massive power draw causing those burnout's as it does not have sufficient enough cooling.

 

[CP5670]

All of the big motherboard brands have had issues in the past of putting too much voltage into the CPU by default, especially with any kind of auto-OC or "gaming mode" feature. I wonder if something like that is causing it.

 

[IEC]

Wouldn't be surprised if they are sending too much voltage with specific MB/BIOS combinations. The limits are different between X3D chips and non-X3D. If they're not careful...

The timing of this story coming to light after 7800X3D launch suggests to me it's specific to some change in either AGESA 1005 or AGESA 1006 based BIOS.

No issues here yet on 3 separate CPUs and mobos.

7800X3D + ASUS X670E Prime Pro
7700X + ASRock X670E PG Lightning
7700X + ASRock B650M-HDV/M2

 

[aigomorla]

Yeah... definitely has to do with the larger cache, and higher boosting draw.
The FPGA i think was not intended or designed for that much amps.
This is why Intel went though god knows how many different pin layout / revisions / updates to perfect that slot.

I remember how many of us wanted to really kill intel when they said no new boards ~
This was on LGA1155, because we all got mad they had to change the entire board for 1 friggin pin.
Then they did it to us again, on LGA1150, but at least it was more then 1 pin... however... again on LGA1151, and yup... guessed it again... LGA1200.

I am guessing Intel saw this coming from LGA1156 -> LGA1155.
Then saw it again on LGA1150, and had to put in a extra for LGA1151.

 

[Markfw]

Yeah... definitely has to do with the larger cache, and higher boosting draw.
The FPGA i think was not intended or designed for that much amps.
This is why Intel went though god knows how many different pin layout / revisions / updates to perfect that slot.

I remember how many of us wanted to really kill intel when they said no new boards ~
This was on LGA1155, because we all got mad they had to change the entire board for 1 friggin pin.
Then they did it to us again, on LGA1150, but at least it was more then 1 pin... however... again on LGA1151, and yup... guessed it again... LGA1200.

I am guessing Intel saw this coming from LGA1156 -> LGA1155.
Then saw it again on LGA1150, and had to put in a extra for LGA1151.

I agree on the larger cache having an effect, but based on personal experience and a combination of 7950x and 7950x3d, I believe its BAD BIOS FROM ASUS. And possible others. But ASRock has been solid so far for me, and ASUS (for memory compatibility) has been bad.

 

[Tech Junky]

I thought that was with their Z690 boards, not AM5.

You're spot on but, who's to say that it didn't happen again. Other boards had UEFI issues. Always a good idea to give things a couple of weeks post release to see the reviews from buyers and not reviewers. Could just be a bad batch from a qa standpoint. To me though it seems like AMD likes guinea pigs on new releases. I'd rather be under promised and over delivered than the other way around. Start stable and add features with UEFI updates as they're tested a bit more thoroughly.

 

[pj-]

it seems like the type of damage happening would have to be due to something other than slightly too aggressive voltages

 

[VirtualLarry]

it seems like the type of damage happening would have to be due to something other than slightly too aggressive voltages

you mean, like a re-hash of Lotes versus foxxconn sockets?

 

[H433x0n]

If I had to guess, it’s pre-X3D bios creating these issues where they would frequently boost SoC voltage amongst other things for Expo compatibility and general stability.

The guy on Reddit that created that thread was running BIOS 0805 when his CPU blew up (early pre-X3D BIOS) since he said the newer releases couldn’t run EXPO.

 

[A///]

it's a pattern with asus fing over amd hardware over the years. ive not heard or read any bad news about asus with intel chipsets for z790. as soresu or ln2cool pointed out to me last week the asus or should I say ass board has been a problem with some big time youtubers like the two cents guy. this generation it also seems like gigabyte is too volt happy on intel sending more than is necessary to maintain high clocks. all can be fixed but this constant over supply from the bioses on these mobos is growing to be a joke now. you don't see problems like this on am5/x670 from other aib's. asus being the biggest would make you think they can attack it with more developers or engineers but no, asus would sooner blame their customrs rather than admitting fault they shipped out janked code.

 

[Ajay]

you mean, like a re-hash of Lotes versus foxxconn sockets?

Lol! Good memory!

 

[TheELF]

If I had to guess, it’s pre-X3D bios creating these issues where they would frequently boost SoC voltage amongst other things for Expo compatibility and general stability.

The guy on Reddit that created that thread was running BIOS 0805 when his CPU blew up (early pre-X3D BIOS) since he said the newer releases couldn’t run EXPO.

The, or any, bios just reveals the issue, if the CPU has no protections against even a little bit more Voltage then that is a serious design flaw. Or cash grab depending on how cynical you are.
If the CPU was fine it would just be a shutdown and not a blow up issue.

 

[tamz_msc]

https://www.reddit.com/r/Amd/comments/12xbgsm

Read the top comment on the reddit thread. The user is implying that Asus and Gigabyte BIOS automatically sets SOC and VDDIO voltages to 1.36-1.4 V, sometimes boosting to even 1.5 V in Windows, when EXPO is enabled. The regular chips can handle it. The X3D cannot.

 

[Makaveli]

Should be fixed with bios updates and its clearly allowing too much voltage on the asus boards.

This kinda stuff is why I try to stay away from Gen 1 items. I'll look at AM5 at Zen 5 at this time motherboard pricing will me much better and most of these issues will be sorted. Should also see the platform using faster than DDR-6000 so will be a win win for those that can wait.

 

[IEC]

https://www.reddit.com/r/Amd/comments/12xbgsm

Read the top comment on the reddit thread. The user is implying that Asus and Gigabyte BIOS automatically sets SOC and VDDIO voltages to 1.36-1.4 V, sometimes boosting to even 1.5 V in Windows, when EXPO is enabled. The regular chips can handle it. The X3D cannot.

I'd be skeptical of that considering there are plenty of people (myself included) who have exceeded those voltages for SOC and IOD/MC for memory overclocking on X3D chips. Though I'm unlikely to be affected now since my recent focus has been minimizing idle power draw so my vSOC is currently 1.1V and IOD/MC is at 1.33V...

VDDCR_VDDM (supplies 3D vcache) and VDDCR_CPU seem like more likely culprits, IMO. Though I don't know what those pins correspond to since they're just labeled "VDDCR" in that pin diagram.

Either way, speculation seems to be a combination of bad BIOS QC (copy pasting non X3D values to be applied to X3D chips 1:1) and using EXPO to enable those bad values. Time will tell if that's the case.

 

[aigomorla]

New Socket for 8000 series Ryzen... Im calling it.

Days of boards lasting more then 1 builds are OVER~

 

[Makaveli]

New Socket for 8000 series Ryzen... Im calling it.

Days of boards lasting more then 1 builds are OVER~

We may see a new revision of the current boards when Zen 5 comes out but i'm not expecting a new socket maybe for Zen 6 and that is still a maybe as they said AM5 will go to 2025+

 

[YBS1]

So glad all this drama is unfolding just days after I completed my build. I have the perfect storm as well Asus X670E/7950X3D/EXPO memory, yay for me I guess! I did update to the latest bios before ever booting though (assuming the latest is an emergency "we f-ed up" fix). After reading about this issue I went and looked and my SOC voltage with EXPO enabled was still ~1.37v. Dialed that back manually to 1.15, guess I'll just hope for the best. Probably will end up being a fluke (bad) luck of the draw situation anyway to have that happen.

 

[aigomorla]

We may see a new revision of the current boards when Zen 5 comes out but i'm not expecting a new socket maybe for Zen 6 and that is still a maybe as they said AM5 will go to 2025+

they are going to go

"tehe pero"


Like how intel did with us.

I remember talking to an intel engineer about the whole pin fiasco.
He told me, they had no choice.
They could not use the same power pads on the DIE, and had to move them for optimization.
This is why we got those annoying LGA sockets with 1 extra PIN, or 1 PIN removed.

I think AMD will need to do the same, and give us that Tehe Pero face with a oops... trace needed to be here.
But the price will most likely be better thermals from optimization, and also most definitely less leaky voltages, and a higher overclock ceiling.

 

[reqq]

My 7950x3D working fine with Asrock board with EXPO. Soc at 1.25 and DRAM 1.35, set by Asrock not me.