Partially, but it also appears that Asus boards may be applying higher SoC voltages than others when voltage is set to auto in the bios, which may lead to earlier failure
-
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
- Thread starter Jamin280672
- Start date
Partially, but it also appears that Asus boards may be applying higher SoC voltages than others when voltage is set to auto in the bios, which may lead to earlier failure
The GN video has confirmed that for nearly all (if not all) the CPU failures out there, that we know about, excessive SoC voltage is the root cause of failure.
As shown, ASUS boards in particular (but others too) have been pumping out very high SoC voltages, as high as 1.45v - even though the actual setting may have been 1.35v they have been pushing beyond that due to, as I previously discovered, variations in their load line calibration resulting in higher voltages under load.
GN also confirmed that time is the primary factor with high SoC voltages, the longer you have been running at the higher voltages the greater the degradation and the greater chance of failure....you can of course dramatically reduce the MTBF by running even higher voltages still, as GN did, to get the CPU to fail.
There are a few subsequent scenarios after the initial failure of the CPU due to excessive Soc:
1 - The CPU fails, internally shorts, the CPU overheats and the CPU tells the motherboard to shut down before catastrophic damage occurs (burning sockets)
2 - The CPU fails, fails to shut down as it's completely burnt out it's thermal monitoring system. Board OCP does not intervene, CPU burns up, de-solders itself, burns the substrate and takes the socket with it.
3 - The CPU fails as in 1, system shuts down. User resets power trying to get system to boot - motherboard OCP does not intervene and CPU proceeds to incinerate and damage the socket.
So excessive SoC is the root cause of failure - but the catastrophic burning up and damaging the board socket is a secondary failure as a result of improper OCP protection of the motherboard.
As I previously stated it seems most of the bios changes at present, outside of the slight lowering of the SoC bios setting, seem to be fixing OCP so the motherboard doesn't get taken out when the CPU fails.
I still think 1.3v is too high an SoC voltage, especially if they have not sorted out the LLC and the actual real voltage applied to the CPU is overshooting still under certain load conditions.
I urge people to manually set SoC to a maximum of 1.25v and confirm with HWiNFO at least that it is staying below 1.3v under loads like prime 95 (large FFT) and other loads.
As shown, ASUS boards in particular (but others too) have been pumping out very high SoC voltages, as high as 1.45v - even though the actual setting may have been 1.35v they have been pushing beyond that due to, as I previously discovered, variations in their load line calibration resulting in higher voltages under load.
GN also confirmed that time is the primary factor with high SoC voltages, the longer you have been running at the higher voltages the greater the degradation and the greater chance of failure....you can of course dramatically reduce the MTBF by running even higher voltages still, as GN did, to get the CPU to fail.
There are a few subsequent scenarios after the initial failure of the CPU due to excessive Soc:
1 - The CPU fails, internally shorts, the CPU overheats and the CPU tells the motherboard to shut down before catastrophic damage occurs (burning sockets)
2 - The CPU fails, fails to shut down as it's completely burnt out it's thermal monitoring system. Board OCP does not intervene, CPU burns up, de-solders itself, burns the substrate and takes the socket with it.
3 - The CPU fails as in 1, system shuts down. User resets power trying to get system to boot - motherboard OCP does not intervene and CPU proceeds to incinerate and damage the socket.
So excessive SoC is the root cause of failure - but the catastrophic burning up and damaging the board socket is a secondary failure as a result of improper OCP protection of the motherboard.
As I previously stated it seems most of the bios changes at present, outside of the slight lowering of the SoC bios setting, seem to be fixing OCP so the motherboard doesn't get taken out when the CPU fails.
I still think 1.3v is too high an SoC voltage, especially if they have not sorted out the LLC and the actual real voltage applied to the CPU is overshooting still under certain load conditions.
I urge people to manually set SoC to a maximum of 1.25v and confirm with HWiNFO at least that it is staying below 1.3v under loads like prime 95 (large FFT) and other loads.
Last edited:
Caporegime
I haven't watched it all yet....
youtu.be
Sort of, yes and no, the Asus motherboard was pushing 450 watts in to the CPU, should it be AMD's responsibility to say no that's too high in the code? Perhaps, then again how much power will an LN2 cooled CPU in extreme overclocking pull? i'm willing to bet 450 watts is not unreasonable.
So should AMD say no you're not doing LN2 overclocking any more, i don't really get what you're trying to say here Steve?
The last time i had an Asus motherboard very high voltages were locked behind and extreme overclocking option, if you unlocked it you were greeted with a great big red warring about killing your CPU, if you didn't unlock that option it didn't allow you to set very high voltages, if its even doing this automatically while just setting XMP that's on Asus, what are AMD to do about that? Tell Asus not to set 1.45v SoC? Do you think AMD don't have guidance for this?
We Exploded the AMD Ryzen 7 7800X3D & Melted the Motherboard
GRAB THE LIMITED GN15 ANNIVERSARY FOIL SHIRT! https://store.gamersnexus.net/products/limited-edition-foil-gn15-tshirt-design | Grab one of the brand new, LIM...
youtu.be
Sort of, yes and no, the Asus motherboard was pushing 450 watts in to the CPU, should it be AMD's responsibility to say no that's too high in the code? Perhaps, then again how much power will an LN2 cooled CPU in extreme overclocking pull? i'm willing to bet 450 watts is not unreasonable.
So should AMD say no you're not doing LN2 overclocking any more, i don't really get what you're trying to say here Steve?
The last time i had an Asus motherboard very high voltages were locked behind and extreme overclocking option, if you unlocked it you were greeted with a great big red warring about killing your CPU, if you didn't unlock that option it didn't allow you to set very high voltages, if its even doing this automatically while just setting XMP that's on Asus, what are AMD to do about that? Tell Asus not to set 1.45v SoC? Do you think AMD don't have guidance for this?
Last edited:
Caporegime
Good Grief.... 1.35v set automatically by Asus, 1.41v actual read.
That will degrade the CPU over time, IE dramatically shorten its life span, yes this is laziness but also recklessness, if AMD say maximum 1.3v and no higher then its 1.3v and no higher, to just ignore that and ram it full of volts.... i don't know what to say about that, what is the quality of the sort of people they are hiring to program this BIOS?
youtu.be
That will degrade the CPU over time, IE dramatically shorten its life span, yes this is laziness but also recklessness, if AMD say maximum 1.3v and no higher then its 1.3v and no higher, to just ignore that and ram it full of volts.... i don't know what to say about that, what is the quality of the sort of people they are hiring to program this BIOS?
We Exploded the AMD Ryzen 7 7800X3D & Melted the Motherboard
GRAB THE LIMITED GN15 ANNIVERSARY FOIL SHIRT! https://store.gamersnexus.net/products/limited-edition-foil-gn15-tshirt-design | Grab one of the brand new, LIM...
youtu.be
Ignore him ffs, wouldn't matter if the president of asus came out and said 'we ****** up', he'd still say its an amd issue.
Last edited:
Associate
There's a second video incoming of all the motherboard bugs they found, not just ASUS, Gigabyte also mentioned plus others.
This is great because reviewers will now know this is something to check for in future motherboard review/follow up and kicks all the vendors in the butt. Because this shouldn't happen. Don't add any extra bell and whistles until the basics are solid.
Check out my race lines on my car!!! The brakes don't work but that's a minor issue. So pleased ASUS, Gigbabyte, MSI, et al don't make cars
This is great because reviewers will now know this is something to check for in future motherboard review/follow up and kicks all the vendors in the butt. Because this shouldn't happen. Don't add any extra bell and whistles until the basics are solid.
Check out my race lines on my car!!! The brakes don't work but that's a minor issue. So pleased ASUS, Gigbabyte, MSI, et al don't make cars
This is after running Cinebench R23, then Prime 95 small fft then large fft and then blend for 5 minutes each
VDDCR_SOC Voltage never went above or below 1.25v
This is on a ASUS X670E Hero motherboard and a BIOS from December 2022
Last edited:
Doesnt just affect asus though. Other mobo vendors have same issue and as gamers nexus mentioned its could be issue with the agesa which amd hands out or its issues with amd not talking to the mobo vendors about cpu voltage specs.
Either way amd fanbois seem to back amd even when they detonate these CPU bombs in the fanboi homes.
If this was an intel issue guaranteed the backlash would be 10 times more severe.
This is why I will be trying to return my system as from memory my motherboard was reporting 1.358v for the soc and I’ve been running it like that for almost a month.Good Grief.... 1.35v set automatically by Asus, 1.41v actual read.
That will degrade the CPU over time, IE dramatically shorten its life span, yes this is laziness but also recklessness, if AMD say maximum 1.3v and no higher then its 1.3v and no higher, to just ignore that and ram it full of volts.... i don't know what to say about that, what is the quality of the sort of people they are hiring to program this BIOS?
We Exploded the AMD Ryzen 7 7800X3D & Melted the Motherboard
GRAB THE LIMITED GN15 ANNIVERSARY FOIL SHIRT! https://store.gamersnexus.net/products/limited-edition-foil-gn15-tshirt-design | Grab one of the brand new, LIM...
Who knows what the actual voltage was, could be well over 1.4v so my cpu might be severely degraded already and its lifespan massively reduced.
How was I supposed to know this.?
Applied XMP just like everyone else and everything was working fine.
This is after running Cinebench R23, then Prime 95 small fft then large fft and then blend for 5 minutes each
VDDCR_SOC Voltage never went above or below 1.25v
This is on a ASUS X670E Hero motherboard and a BIOS from December 2022
The problem is that what motherboard is reporting isn’t what it’s actually delivering to the cpu as proven by GM.
so it was an issue with amd's agesa.
not sure if it was gamers nexus video or another which mentioned that the mobo powers off everything but was still shoving high amps into the cpu, with water coolers the pump will be off which will cause massive temp spike, air cooler may still be ok due to the heatsink mass doing some cooling work.
what happens if aio waterblock is subjected to high temps with no pump running? will it boil the fluid inside? perma damage to cooler?
Last edited:
Associate
There's AGESA that is AMD's responsibility.
There's the implementation of AGESA that is the vendors responsibility.
Both failed from the sounds of it, resulting in this. Number of 7800X3Ds sold (we don't know - a lot) / failed units (we don't know - not many) = failure rate (we don't know - not high).
There's the implementation of AGESA that is the vendors responsibility.
Both failed from the sounds of it, resulting in this. Number of 7800X3Ds sold (we don't know - a lot) / failed units (we don't know - not many) = failure rate (we don't know - not high).
Last edited:
Associate
Turning into a mess this.
If I’d had purchased right off the bat and had been running for months with code that could pump dangerous levels of volts and power. I’d be wanting a proactive solution from AMD to replace my CPU - who knows, suspect they don’t, how much damage has already been done by the code.
I feel kinda lucky, mine was running at 1.3v (1.4 if GM is correct) for a very short time, days before I tuned voltages.
My machine is currently turned off though till the release fixed code is released.
If I’d had purchased right off the bat and had been running for months with code that could pump dangerous levels of volts and power. I’d be wanting a proactive solution from AMD to replace my CPU - who knows, suspect they don’t, how much damage has already been done by the code.
I feel kinda lucky, mine was running at 1.3v (1.4 if GM is correct) for a very short time, days before I tuned voltages.
My machine is currently turned off though till the release fixed code is released.
I'm running g.skill trident z neo 6000mhz CL32 with the soc at 1.15v and and the MC voltage at 1.25v using buildzoids easy DDR5 timings.
EDIT lol corrected the voltage lol
Last edited:
Maybe the BIOS from December didn’t. Who knows.
Exactly who knows. Sorry but that’s not good enough for me.