!BEWARE! ASRock B650E Steel Legend CPU Destroying Bios BUG!

[Zach_01]

OCCT uses HWInfo

What do you mean it uses HWiNFO?

 

[Vya Domus]

Power can be roughly estimated as a function of capacitance, frequency and voltage in ICs, P = c * v^2 * f, voltage is squared and has a non linear effect on power (and temperature).

For example let's assume this CPU runs at 1.3 V and uses 70W which is normal for this CPU, so 70 = 1.3^2 * c * f, so the capacitance and frequency are equal to 41.4.

If the voltage increases to 2 that would mean 2^2 * 41.4 = 165.6W, way more than 70 not to mention the temperature would skyrocket way more than 84C or even 100.

This isn't by any means a very accurate estimation but these figures do not look right.

EDITED : redid the calculation for 1.3V which is more reasonable for this CPU

 

[3x0]

What do you mean it uses HWiNFO?

OCCT sensors are based on HWInfo code.

 

[Frick]

i flashed the bios when i bought the bundle a week after new year and it worked without any problems until now. it imediately happened after just activating this.
i already RMAd the CPU today and i should get a replacement until friday or saturday.

Presumably Asrock or AMD would be interested in the CPU to find out what exactly killed it...

 

[Zach_01]

And there is also this
Dont know if there is any relation

 

[erocker]

Seems every board vendor is having issues with AMD cpu's and bios in one way or the other. I can remember a story with the (I think) x5xx chipset where vendors were initially getting no information from AMD. Seems things haven't gotten much better. I'm on a 3 month old beta bios for AM5 with my board because it's the only one that works for me currently.

 

[Dragam1337]

Please read the following thread on HWiNFO forums

HWiNFO64 CPU Die (Average) Spike?

As you can see from the image below, I had HWiNFO64 open and minimized to tray for a while. Wasn't playing games or anything but when I opened it back up it showed that the Maximum CPU Die (Average) was 104.2c. How could this be possible? As you can see from my other temps from the sensors...

www.hwinfo.com

This is an issue of almost all Zen3/4 CPUs probably on some particular SMU.
Its false CPU reporting of some readings where it sends exactly double values

This is going for 6+ months and still no fix

View attachment 334557

View attachment 334558

Different systems have different value errors
Mine for example never reported high SoC voltage but others do

This is happening as far as I know on almost all vendor boards (ASUS, GB, AsRock, MSI)

I seeing this randomly (10% or less) after Gaming
And usually its just the temp (Die avg) for me
The x2 mem clock/ratio only happened a couple of times

Other systems may see different x2 error reporting values

If it was a simple missreport, then it wouldn't have killed his cpu...

 

[3x0]

If it was a simple missreport, then it wouldn't have killed his cpu...

The missreport and the dead cpu are very likely unrelated, as many users with functional cpus had the same bugged readouts

 

[ThrashZone]

Hi,
Yeah some of the newer versions of hwinfo are pretty weird
Noticed one got pulled from the server the other day and some just flat out freeze just resetting the readings
I'm using 766 atm not sure why I'll likely go back to 622 or 626 hehe at least they worked properly.

This comes to mind

 

[Dragam1337]

The missreport and the dead cpu are very likely unrelated, as many users with functional cpus had the same bugged readouts

Yeah, sure... cpu works fine, he activates a profile that makes his voltages go wack, and then his cpu dies... totally unrelated... amazing logics...

 

[sethmatrix7]

Asus

ASRock

 

[ShrimpBrime]

OCCT uses HWInfo

So does Benchmate.

 

[3x0]

Yeah, sure... cpu works fine, he activates a profile that makes his voltages go wack, and then his cpu dies... totally unrelated... amazing logics...

Several users in this thread with the same bugged readout and a functional CPU, read the thread.

 

[Zach_01]

If it was a simple missreport, then it wouldn't have killed his cpu...

I think the x2 bug reports and the CPU damage are not related if you saw the GN video.
CPUs can be killed with 1.4-1.5V volts on SoC when a "high" amount of current goes through in certain situations and board CPU protection has "holes" according to GN findings 9months ago.

Now as for HWiNFO readings, 2+Volts on CPU cores and 2.4V on SoC its simply a CPU report bug and not related to the above. If those values were true he wouldn't be able to see those values after they happened = instant kill no matter what current.

 

[Dragam1337]

I think the x2 bug reports and the CPU damage are not related if you saw the GN video.
CPUs can be killed with 1.4-1.5V volts on SoC when a "high" amount of current goes through in certain situations and board CPU protection has "holes" according to GN findings 9months ago.

Now as for HWiNFO readings, 2+Volts on CPU cores and 2.4V on SoC its simply a CPU report bug and not related to the above. If those values were true he wouldn't be able to see those values after they happened = instant kill no matter what current.

1.4-1.5v wouldn't be insta death though - it took a while with a continuous very high voltage before it died when steve was doing the "testing", in an absolute worst case scenario.

OP said it briefly spiked to very high voltage 3 times before it died, which seems very plausible. Could hwinfo have been reading the voltage wrong, reporting it higher than it really was ? Sure, but it absolutely did get too much voltage, otherwise it wouldn't have died.

 

[chrcoluk]

Sadly ASRock I think are more concerned about appearing to lead benchmarks now than staying within reasonable tolerances.

On 3 boards if I enable XMP SA and IO get pretty high voltages, on my current main board, the first bios I used set TJMAX to 15C over intel's spec alongside stock settings that inflated voltage, and now this thread.

What is going on.

 

[ir_cow]

On 3 boards if I enable XMP SA and IO get pretty high voltages, on my current main board, the first bios I used set TJMAX to 15C over intel's spec alongside stock settings that inflated voltage, and now this thread.

What is going on.

Yep the Z690 PG Phantom had a 115c limit set by default. Not a good look for ASRock

 

[Zach_01]

1.4-1.5v wouldn't be insta death though - it took a while with a continuous very high voltage before it died when steve was doing the "testing", in an absolute worst case scenario.

OP said it briefly spiked to very high voltage 3 times before it died, which seems very plausible. Could hwinfo have been reading the voltage wrong, reporting it higher than it really was ? Sure, but it absolutely did get too much voltage, otherwise it wouldn't have died.

Yeah if the CPU died it was 99.9% from high voltage/current scenario.
I just dont believe it took 2.4V on SoC. That was just an erroneous report from the CPU's telemetry. For some reason it sends out untrue x2 (exactly) values.

2 different things happened to happen around the same time, Im saying.
Otherwise burned CPUs would be all over the place by now.
I think if every AM4/5 user is using HWiNFO 24/7 a lot will see those x2 values, because AMD did not solve this report bug yet.

But I thought that the other issue in GN's video was history by now with newer BIOS versions.

 

[Dr. Dro]

Burning up is normal for 7000x3d's.

While not unheard of (due to AGESA bugs, no less), I wouldn't call it exactly normal. Thankfully AMD owned up to the issue and is issuing refunds for CPUs that burned up. I don't think this is specifically an AsRock issue either, the old mantra of "beta BIOS updates aren't really beta" are going to apply here IMO. Everything is controlled by AGESA, after all.

 

[GerKNG]

Yeah if the CPU died it was 99.9% from high voltage/current scenario.
I just dont believe it took 2.4V on SoC. That was just an erroneous report from the CPU's telemetry. For some reason it sends out untrue x2 (exactly) values.

2 different things happened to happen around the same time, Im saying.
Otherwise burned CPUs would be all over the place by now.
I think if every AM4/5 user is using HWiNFO 24/7 a lot will see those x2 values, because AMD did not solve this report bug yet.

But I thought that the other issue in GN's video was history by now with newer BIOS versions.

The IGPU hit almost 100°C at the "2.4V SoC Voltage" within 2 seconds. (if i run the IGP in Timespy it peaks at 44°C)
i have AM5 since it launched and other boards (incl. X670) and i NEVER saw doubled voltages in HWInfo. And i run HWinfo 24/7 365 Day a Year on a second Monitor.

 

[Redwoodz]

Were you unable to read what he wrote ? It came as spikes during load... how the f*** was he supposed to verify that beforehand...

And half the speed... did you fail math aswell as reading ? Went from aprox 1600 rpm to aprox 1200 rpm...

Not that fanspeed has ANY relevance for the cpu being blasted with 2.4v...

cpu1 fan 872min to 897 max, might be pump. Was 1000+ on the prior screen

The IGPU hit almost 100°C at the "2.4V SoC Voltage" within 2 seconds. (if i run the IGP in Timespy it peaks at 44°C)
i have AM5 since it launched and other boards (incl. X670) and i NEVER saw doubled voltages in HWInfo. And i run HWinfo 24/7 365 Day a Year on a second Monitor.

Short to ground and of course the sensors are affected. You may have damaged the cpu prior running 1.25v 24/7 and when it failed you got the weird readings. All I am saying is you should make sure of what happened before you start blaming. From what I see that is the temp of the 7900XTX on the last line.

 

[A Computer Guy]

ASrock B650E Steel Legend latest Bios (non beta) Version 2.02
Ryzen 7 7800X3D
Corsair Vengeance RGB 2x16GB 6000 CL30

The Issue:
As soon as you go into the RAM Settings and Enable change the Performance Mode from "AMD AGESA Default" to "Aggressive" the CPU VCORE and VSOC gets DOUBLED in short spikes constantly. (even the IGP at the desktop shoots up to 100°C in two seconds.

The 7800X3D has been completely destroyed within 15 Minutes (crashed, locked up and then it never posted again.)


View attachment 334512

Here is a 10 and a half HOUR long HWINFO session i fortunately had left over after making a full screenshot in a game with just EXPO enabled and nothing else.. (All voltages are completely normal, the First CPU VDDCR Voltage at 1.131V "Current" is the VCore, below that is the SOC Voltage...)
View attachment 334514

And here is a screenshot in a game a couple minutes after booting with the aggressive timings preset while playing a game. (The Voltages as well as the IGPU Temperature at the bottom of the list is highlighted in Blue.)
View attachment 334515

Oh no! This reminds me of when I blew up my motherboard however the circumstances here are different. Luckily my CPU survived a partially melted socket although I will always be suspect of it every time something goes wrong but I've managed to stabilize it for now.

currently i have another 7700X and it works perfectly fine in the board.

Oh good your board survived. Reminds me of a song, something like "...country boards will survive..."

 

[Zach_01]

The IGPU hit almost 100°C at the "2.4V SoC Voltage" within 2 seconds. (if i run the IGP in Timespy it peaks at 44°C)
i have AM5 since it launched and other boards (incl. X670) and i NEVER saw doubled voltages in HWInfo. And i run HWinfo 24/7 365 Day a Year on a second Monitor.

Yet your CPU core temp seems very normal (84.5C) with 2.01V, right?

I quess you didnt read the HWiNFO thread I posted. If you did you would have seen this below already
And this CPU didnt fry...
Pay attention also to IOD temp (40C = normal) and SoC Power SVI3 TFN (43.520W = 2 x 21.76 which is the normal value)
I'll say it again, Its an entirely different issue the faulty x2 values (= just a reporting bug), and the X3Ds frying SoC

 

[Dragam1337]

cpu1 fan 872min to 897 max, might be pump. Was 1000+ on the prior screen

Yes, going from 1005 to 872 rpm is totally half... /facepalm

 

[Redwoodz]

Yes, going from 1005 to 872 rpm is totally half... /facepalm

It was not a factual statement, it was a general statement. I posted the correct stats, wtf is your problem?