Question Asus mobos burning x3D CpuS?

[VirtualLarry]

Point being, its really my "home production box" and I want stable and efficient.

That's kind of why I'm thinking to keep my 7600 rig dGPU-less, and just stay on the Ryzen 7000-series 2CU IGPU.

 

[Markfw]

Gotcha, my system is not as fancy as yours but I use some unrelated software to give me notifications. my stuff is on auto pay but I get text alerts whenever a payment goes out. half the time I don't know what time of the month it is and I worry about being late. 3% is to cover what the processing fees?

Using any type of credit card, debit or otherwise. When you are talking about over $6,000 thats a lot of money.

 

[jarablue]

I'll stick with Intel. Been using it since 1990 with 0 issues. 1990, I was 15 and first got introduced to pc gaming. Thanks Chris Natello!

Big Blue baby!

 

[Schmide]

I'll stick with Intel. Been using it since 1990 with 0 issues. 1990, I was 15 and first got introduced to pc gaming. Thanks Chris Natello!

Big Blue baby!

Duh Big Blue is IBM. intel is chipzilla

 

[Markfw]

I'll stick with Intel. Been using it since 1990 with 0 issues. 1990, I was 15 and first got introduced to pc gaming. Thanks Chris Natello!

Big Blue baby!

Not only are you wrong about big blue, but this is an AMD thread. Take your BS elsewhere.

 

[aigomorla]

home production box" and I want stable and efficient.

Thats supermicro for you.
I have not owned a board which could out endurance a supermicro board in crunching or being on full boat load 100% 24/7.
Not even Tyan which comes close second was that stable, which is why i absolutely love supermicro when it comes to things like that.

But i want RGB, i want a rave going on inside my case, and i most definitely want to overclock my HEDT.

 

[Markfw]

What about this ? This is my 9654 96C/192T

 

[DAPUNISHER]

This thread has made things as clear as mud for me. I would appreciate someone explaining a couple of things to me.

1. If GN is wrong, what is the correct answer to why the CPUs, and a few boards, are burning up?

2. Whatever the alternative answers to the above question are, would you call it a consensus opinion? Excluding GN of course.

 

[Markfw]

This thread has made things as clear as mud for me. I would appreciate someone explaining a couple of things to me.

1. If GN is wrong, what is the correct answer to why the CPUs, and a few boards, are burning up?

2. Whatever the alternative answers to the above question are, would you call it a consensus opinion? Excluding GN of course.

My personal opinion. 1) ASUS mobos (on older bios) auto-set the vsoc voltage too high. Fixed in newer bios.
2) update your bios, and no overclocking (except some people call EXPO OC'ing, that should be allowed)

Consensus ? lets see what people say.

 

[AdamK47]

My personal opinion. 1) ASUS mobos (on older bios) auto-set the vsoc voltage too high. Fixed in newer bios.
2) update your bios, and no overclocking (except some people call EXPO OC'ing, that should be allowed)

Consensus ? lets see what people say.

I should also add that Asus was applying the high VSOC when EXPO was enabled. It's not in the profile. It's just what Asus did when selected.

 

[IEC]

This thread has made things as clear as mud for me. I would appreciate someone explaining a couple of things to me.

1. If GN is wrong, what is the correct answer to why the CPUs, and a few boards, are burning up?

2. Whatever the alternative answers to the above question are, would you call it a consensus opinion? Excluding GN of course.

GN is only arguably wrong about the board applying +0.05V over what is set, as the measurement may have been taken in a location where it was reading falsely high compared to what is actually delivered to the CPU. It's not wrong about the premise that too much voltage was applied one way or another, as we've already established that ASUS were defaulting up to 1.4V vSOC in some presets. Which we now know is outside "safe" limits. ASUS obviously tried to punt the blame to AMD. Whether that's deserved or not remains to be seen (my guess is no, given ASUS tries to blame the customer all the time in cases of RMA).

Regardless of the details, the mode of failure makes it apparent the ASUS implementation had neither 1) a working OCP nor 2) a working OTP. Either of those should have stopped the runaway input of current over a low resistance path that leads to the spectacularly cooked chip. There are cases where only the CPU died and motherboard was spared, but the majority of cases where the board also got fried appears to be an ASUS board.

I'll be interested to see what the final details are after investigation, but it's apparent to me that excess voltage was fed on one or more rails causing damage first to the CPU. I would speculate this damage resulted in a short circuit. In the cases of burned CPU+board it is suggestive of a low resistance short where without proper OCP the board dumps enough power into the CPU to cause the catastrophic failures seen.

 

[DAPUNISHER]

Thanks for the replies. I wasn't watching that youtube vid from some rando I never heard of, and therefore have zero trust in. So I had no idea what the contention with GN was all about. As everything I read previously indicated they were on point.

Also, with freelance social media managers being a dime a dozen and working for chicken feed, it can be tough to cut through all the attempts at astroturfing, disinformation campaigns, and of course, the haterade guzzlers.

The internet lynch mob forcing Asus' hand, was the anticipated outcome. I was shocked Asus resisted as long as they did. Reading around, they turned off a lot of potential customers. Not just for boards, but all things Asus. Seems there are still a lot of newer DIYers that did not know about their infamously bad CS. If nothing else good comes from this fiasco, that is indeed the silver lining.

 

[H433x0n]

GN is only arguably wrong about the board applying +0.05V over what is set, as the measurement may have been taken in a location where it was reading falsely high compared to what is actually delivered to the CPU. It's not wrong about the premise that too much voltage was applied one way or another, as we've already established that ASUS were defaulting up to 1.4V vSOC in some presets. Which we now know is outside "safe" limits. ASUS obviously tried to punt the blame to AMD. Whether that's deserved or not remains to be seen (my guess is no, given ASUS tries to blame the customer all the time in cases of RMA).

Regardless of the details, the mode of failure makes it apparent the ASUS implementation had neither 1) a working OCP nor 2) a working OTP. Either of those should have stopped the runaway input of current over a low resistance path that leads to the spectacularly cooked chip. There are cases where only the CPU died and motherboard was spared, but the majority of cases where the board also got fried appears to be an ASUS board.

I'll be interested to see what the final details are after investigation, but it's apparent to me that excess voltage was fed on one or more rails causing damage first to the CPU. I would speculate this damage resulted in a short circuit. In the cases of burned CPU+board it is suggestive of a low resistance short where without proper OCP the board dumps enough power into the CPU to cause the catastrophic failures seen.

I think I can make a compelling case it’s not the fault of motherboard vendors. Let alone a single motherboard vendor since I can find identical failures across all 4 vendors. While some vendors (Asus) may have more reported cases, we don’t know if it’s disproportionate to market share.

First, the setting of 1.4V was not out of spec, it was the maximum allowable value from AGESA. This would lead one to assume having a 1.4V setpoint was assumed as a safe value through the stated warranty window. Some motherboard vendors were more aggressive than others, but again it was a value within allowable ranges.

What I feel is a valid analogy is this: Michelin issues a speed safety rating of 200km/h for the first 80,000km of tire wear. BMW decides to set a vmax of 200km/h on their sports cars. A year later, tires begin blowing out on the autobahn, would this be BMW’s fault? To me, that’s a grey area at best, with most of the blame being on Michelin.

We’ll likely never know the root cause other than vSOC above 1.3V is a contributing factor. It doesn’t appear to be something the enthusiast community has the tools or IP to figure out themselves. As of now only AMD and motherboard vendors know all of the mechanics behind the failures and the motherboard vendors can’t tell us because they signed an NDA from AMD to receive that information.

Considering we’ve been waiting a month for AMD to provide updated AGESA suggests to me it’s not the motherboard vendors. If it was just as simple as vSOC is too high, this could’ve been fixed using the existing AGESA.

 

[dlerious]

This thread has made things as clear as mud for me. I would appreciate someone explaining a couple of things to me.

1. If GN is wrong, what is the correct answer to why the CPUs, and a few boards, are burning up?

2. Whatever the alternative answers to the above question are, would you call it a consensus opinion? Excluding GN of course.

Another youtuber weighing in on ASUS in general

 

[DAPUNISHER]

I think I can make a compelling case it’s not the fault of motherboard vendors. Let alone a single motherboard vendor since I can find identical failures across all 4 vendors. While some vendors (Asus) may have more reported cases, we don’t know if it’s disproportionate to market share.

First, the setting of 1.4V was not out of spec, it was the maximum allowable value from AGESA. This would lead one to assume having a 1.4V setpoint was assumed as a safe value through the stated warranty window. Some motherboard vendors were more aggressive than others, but again it was a value within allowable ranges.

What I feel is a valid analogy is this: Michelin issues a speed safety rating of 200km/h for the first 80,000km of tire wear. BMW decides to set a vmax of 200km/h on their sports cars. A year later, tires begin blowing out on the autobahn, would this be BMW’s fault? To me, that’s a grey area at best, with most of the blame being on Michelin.

We’ll likely never know the root cause other than vSOC above 1.3V is a contributing factor. It doesn’t appear to be something the enthusiast community has the tools or IP to figure out themselves. As of now only AMD and motherboard vendors know all of the mechanics behind the failures and the motherboard vendors can’t tell us because they signed an NDA from AMD to receive that information.

Considering we’ve been waiting a month for AMD to provide updated AGESA suggests to me it’s not the motherboard vendors. If it was just as simple as vSOC is too high, this could’ve been fixed using the existing AGESA.

By my standards, your contention lacks sufficient facts and evidence. It relies almost entirely on opinion, conjecture, and argument from ignorance. Consequently I find it neither compelling nor convincing.

Any attempt to absolve mobo makers of blame falls flat for me. They have been "overclocking" their boards for benchmark advantages since way back when. This issue is no different. This does not absolve AMD of any culpability. But no way am I buying the board makers are blameless.

 

[aigomorla]

Here's how it feels to me.
I think ASUS is trying to do a massive nuclear waste cleanup at the cost of all of its influencers to try to discredit them in anyway with manipulating both hardware and results.

Soon they are going to ask JPL to launch a motherboard into space and see if the VSOC has any impact in 0G, and somehow corelate that to all the influencers being just straight up wrong, because they worked with NASA's JPL to get results.

I really don't think there is any cleanup in this regards other then them saying they have let go of there PR team.
Also they need to publicly state they are restructuring the RMA department, and QC department.
They will then need to reinvite all the said influencers to there warehouse to so the changes made.

This will probably be the only cleanup most of us will accept.
Otherwise ASUS will always mean "ROG" - Royally Overpriced Garbage.

 

[A///]

Sigh you really have no idea about ASUS RMA.
They will look at burnt socket.
RMA Request: Burn Socket.

Technician replys.
RMA Rejection: Bent CPU Pins... (ignoring the burnt socket)

Welcome to ASUS RMA.

Even tho they say they will accept it.
FAT chance your not going to have any problems actually getting it done.

Brought to you by LLM AI.

In shadows cast by tales untold,
A tragic saga, now to unfold.
Where Asus dwells, a woeful plight,
Explosive motherboards, a chilling sight.

Their fiery core, a menacing blaze,
As Asus' boards ignite in fiery haze.
Consuming dreams, they wreak havoc and pain,
Devouring AMD Ryzen, a cruel disdain.

Oh, Asus, once adorned in grace,
Now stained with fire's merciless embrace.
Your promises, a smokescreen of lies,
As motherboards feast, consuming skies.

They detonate with an ominous roar,
Reducing dreams to ash on the floor.
AMD Ryzen processors, helpless prey,
Swallowed whole, devoured in dismay.

Consumers tremble, their spirits torn,
By Asus' inferno, mercilessly borne.
The fragility of trust, so swiftly burned,
Leaving naught but sorrow, lessons unlearned.

Oh, victims caught in this fiery maw,
May justice reign, and hold Asus in awe.
For safety's sake, this plea shall ring,
End the devastation these boards bring.

Let not the flames engulf us all,
Let not destruction be our downfall.
For in the ashes, a chance to rise,
To find resilience, to claim the skies.

In somber verses, this tale must unfold,
Of Asus' motherboards, explosive and cold.
May truth prevail, exposing their might,
To protect our dreams, in radiant light.

 

[A///]

Asus as I said the other day didn't think like most companies now a days that this story would amount to being trouble for them. BIG MISTAKE. of course it did, which should not surprised anyone here. now that they've been dropped by Jay and I think Linus tech tips and possibly others in the near future they're beginning to scramble. most of the aic companies made big headway when influeners began marketing and reviewing their products on youtube. they pay a set sponsor cost and the views can range from almost nothing to millions of views. their money goes a lot further than a spendy ad campaign. now that these youtubers are saying no to them asus is realising it might be back to online adverts or print adverts to get the word out. who subscribes to their aic youtube pages if they have any or visit their disorganised and terribly designed websites? few people if any can be bored enough to.

 

[moinmoin]

I think I can make a compelling case it’s not the fault of motherboard vendors.
(...)
First, the setting of 1.4V was not out of spec, it was the maximum allowable value from AGESA. This would lead one to assume having a 1.4V setpoint was assumed as a safe value through the stated warranty window.

What do you think warranty voiding OC'ing entails then?

 

[H433x0n]

What do you think warranty voiding OC'ing entails then?

Technically in most cases if you're running an X670 or Z790 chipset motherboard you're 'voiding' the warranty by just installing the CPU and modifying absolutely nothing in the BIOS. Most motherboard vendors have their own form of "Multicore Enhancement' enabled by default that pushes clock multipliers and voltages beyond the VF curve located in the CPU's registers. So warranty coverage on overclockable CPUs relies on the good faith of AMD/Intel.

 

[moinmoin]

Technically in most cases if you're running an X670 or Z790 chipset motherboard you're 'voiding' the warranty by just installing the CPU and modifying absolutely nothing in the BIOS. Most motherboard vendors have their own form of "Multicore Enhancement' enabled by default that pushes clock multipliers and voltages beyond the VF curve located in the CPU's registers. So warranty coverage on overclockable CPUs relies on the good faith of AMD/Intel.

You didn't answer my question. You first claim "the setting of 1.4V was not out of spec, it was the maximum allowable value from AGESA" thus "1.4V setpoint was assumed as a safe value through the stated warranty window" and now you claim motherboard vendors void the warranty with the defaults they set in the BIOS. This flies in the face with the first part of your first claim. You can't have both, pick your poison.

 

[aigomorla]

So warranty coverage on overclockable CPUs relies on the good faith of AMD/Intel.

You do know there is almost 0 times i can think of when intel said sorry you voided your warrenty because of overclocking.

Instead it was more like... "oh you had fun with that chip... well, please be kinder to this one, here is the tracking number."

And if you got the overclocking warrenty they used to offer, it was more

Intel Support

Intel® Customer Support home page

www.intel.com

"like, thats it? the IHS came intact? Weak sauce, here ya go, round 2."

The only way you could void a Intel warrenty, well even if you did, i heard a lot of people still got replacements, was if you delided the cpu, and had liquid metal damage.

But again, ive heard in like 70% of the time, they still replaced the cpu if you had the performance warrenty.

If AMD can't step up to that level of service, well, unless were talking a massive performance AMD has over intel. Like how Core2Duo was wasting Phenoms.
Its not significant enough to counter customer service you get with the other brand.

So AMD better be replacing these chips with no questions asked, even if it came as a pile of molten slag.

 

[H433x0n]

You didn't answer my question. You first claim "the setting of 1.4V was not out of spec, it was the maximum allowable value from AGESA" thus "1.4V setpoint was assumed as a safe value through the stated warranty window" and now you claim motherboard vendors void the warranty with the defaults they set in the BIOS. This flies in the face with the first part of your first claim. You can't have both, pick your poison.

I’m not sure what you mean. I believe anything above 1.1V voids the warranty from a technical standpoint.

Clearly 1.4V was within the threshold of “safe”, otherwise they wouldn’t have lowered it from 1.4V -> 1.3V and rewritten the AGESA. It’s not as if AMD was oblivious to motherboards running >1.3V vSOC. At least I would hope that didn’t catch them by surprise that their largest motherboard partner was shipping BIOS with these values on their flagship motherboards. Somebody on AMD"s QA team at one point had a Zen 4 CPU and an Asus X670E Crosshair for validation testing. The press kit that was sent out to tech reviewers was a Aorus Master X670E with the GSkill DDR5-6000 32-38-38-96 kit (F5-6000J3038F16GX2-TZ5N). This motherboard and ram kit defaults to a vSOC setpoint of 1.35V when enabling Expo. What I'm trying to say is that nobody (including AMD) knew that a vSOC of >1.30V was an issue until 6-8 weeks ago. This somehow resulted in Asus in particular getting singled out, which I guess it sucks to be them.

Also, as far as I know Asus never actually had a default vSOC target of 1.4V, it only appeared that way depending on where the voltage was measured and what sensor was referenced in HwInfo. I could be wrong about this, but I couldn't find anything other than 1.35V for DDR5-6000 Expo.

 

[biostud]

I’m not sure what you mean. I believe anything above 1.1V voids the warranty from a technical standpoint.

Clearly 1.4V was within the threshold of “safe”, otherwise they wouldn’t have lowered it from 1.4V -> 1.3V and rewritten the AGESA. It’s not as if AMD was oblivious to motherboards running >1.3V vSOC. At least I would hope that didn’t catch them by surprise that their largest motherboard partner was shipping BIOS with these values on their flagship motherboards. Somebody on AMD"s QA team at one point had a Zen 4 CPU and an Asus X670E Crosshair for validation testing. The press kit that was sent out to tech reviewers was a Aorus Master X670E with the GSkill DDR5-6000 32-38-38-39 kit (F5-6000J3038F16GX2-TZ5N). This motherboard and ram kit defaults to a vSOC setpoint of 1.35V when enabling Expo. What I'm trying to say is that nobody (including AMD) knew that a vSOC of >1.30V was an issue until 6-8 weeks ago. This somehow resulted in Asus in particular getting singled out, which I guess it sucks to be them.

Also, as far as I know Asus never actually had a default vSOC target of 1.4V, it only appeared that way depending on where the voltage was measured and what sensor was referenced in HwInfo. I could be wrong about this, but I couldn't find anything other than 1.35V for DDR5-6000 Expo.

You do know that some settings are present for those who wants to do extreme overclocking, and are not considered to be used for normal use.
I don't think you should try and maximize all your voltage settings in the BIOS and then expect them all to be safe...

 

[dccc]

Not a fan of influencers that plays a toxic twat character to get views. Looking at Stevie nexus here.
If he is wrong, just admit it and mend the bridges. Do not rile up your internet army to incite hatred and negativity