I think this is pretty much an open and shut case of motherboard failure, but want
some expert confirmation before I return it to Amazon for replacement.

Motherboard: MSI MPG-Z590 Gaming Plus NEW
CPU: Intel Core i7-11700 NEW
CPU Cooling: Cooler Master 212 EVO V2 NEW
RAM: Corsair Vengeance LPX DDR4 3200MHz (8GB x2) NEW
GPU: MSI Nvidia RTX3050 NEW
PSU: Corsair RM850x NEW
NVMe: Samsung 970 EVO Plus - 250GB (C-System Drive) EXISTING
HDDs: WD Gold - 1TB (E-Editing Drive) EXISTING
WD Black - 1TB (F-Storage Drive) EXISTING
Optical Drives: LG WS16NS40 Blu-ray Writer EXISTING
Lite-On iHAS124 EXISTING
OS: Windows 10 Pro EXISTING INSTALL

Once everything was put together, appropriate drivers installed and setup done, the
system appeared to run mostly well. There were some small glitches such as a few
things not running exactly as expected, BIOS setting changes not taking effect
immediately, and a 2-3 second delay in getting sound via S/PDIF (line out worked, but
sounded a tiny bit fuzzy). RAM initially ran at 2133MHz, but XMP finally took effect
and it was running at 3200MHz. Over the next day, I noticed more unexpected
happenings such as brief cutouts. I found a download at MSI that was supposed to fix
the S/PDIF problem, but it didn't. The board came with the 3rd BIOS installed; the
most recent is the 8th. So I eventually flashed that in hopes it would solve the
issues. The flash completed with no problems, but all it appeared to do was change
some of my BIOS settings back to default. They were mostly simple ones such as
disabling the MSI logo at startup and disabling TPM (I don't want Microsoft trying to
sneak Windows 11 onto my PC). I did turn XMP back on and it showed as being enabled
and supporting 3200MHz, but RAM stayed at 2133MHz. Going back into Windows was
simple as usual. I checked when I got back in and, sure enough, 2133MHz RAM speed
was reported in Task Manager Performance. After several hours, I rebooted and went
back into the BIOS to recheck RAM speed there. It had moved to 3200MHz then. I made
no changes, but when I started to leave there were a number of strange looking
changes indicated. Couldn't make heads or tails of what looked like gibberish. I
cleared that mess by clicking on "No" and proceded to get out of the BIOS and back
into Windows. No go! I got a notice that Windows couldn't start and I needed to go
into Recovery Mode, but keyboard selections weren't responding as they should. I
could get into some pages from the initial screen, but not others. Then things fell
apart very quickly. I couldn't get back into the BIOS. My keyboard stopped being
responded to. The reset button wouldn't work. In the end, only the power button
would work. USB keyboard lights didn't come on (it's backlit) nor did any of its
LEDs. The monitor screen stayed off. The power button started at least some part of
the system (indicator LED on, but drive activity LED stayed off), and a hold of
several seconds turned it off. Nothing else worked.

I removed the problem board and put my old one back in (ASUS Prime Z390-A) with the
new PSU and GPU and old CPU, CPU cooler, RAM, NVMe, HDDs and optical drives.
Everything works as it should. The installation of Windows that couldn't start from
the new board started with no problem other than a normal pause while Windows
installed devices for the old one. I haven't pulled the new CPU, RAM and CPU cooler
off the MSI board yet.

My thinking is that the new board was flawed from the start, couldn't tolerate
changes and just got progressively worse until it's now more or less bricked.

I've been building PCs for many years and have never experienced anything like this
before. In the past they either ran or they didn't. There was never a slow death.

Any thoughts about anything other than the motherboard being at fault? Except for an
outright PSU failure, I can't think of anything other than the motherboard that could
make USB inoperable, especially with it happening over a short period rather than
suddenly. Not RAM and certainly not the CPU.

Larc

On 6/14/2023 9:07 PM, Larc wrote:
> I think this is pretty much an open and shut case of motherboard failure, but want
> some expert confirmation before I return it to Amazon for replacement.
>
> Motherboard: MSI MPG-Z590 Gaming Plus NEW
> CPU: Intel Core i7-11700 NEW
> CPU Cooling: Cooler Master 212 EVO V2 NEW
> RAM: Corsair Vengeance LPX DDR4 3200MHz (8GB x2) NEW
> GPU: MSI Nvidia RTX3050 NEW
> PSU: Corsair RM850x NEW
> NVMe: Samsung 970 EVO Plus - 250GB (C-System Drive) EXISTING
> HDDs: WD Gold - 1TB (E-Editing Drive) EXISTING
> WD Black - 1TB (F-Storage Drive) EXISTING
> Optical Drives: LG WS16NS40 Blu-ray Writer EXISTING
> Lite-On iHAS124 EXISTING
> OS: Windows 10 Pro EXISTING INSTALL
>
> Once everything was put together, appropriate drivers installed and setup done, the
> system appeared to run mostly well. There were some small glitches such as a few
> things not running exactly as expected, BIOS setting changes not taking effect
> immediately, and a 2-3 second delay in getting sound via S/PDIF (line out worked, but
> sounded a tiny bit fuzzy). RAM initially ran at 2133MHz, but XMP finally took effect
> and it was running at 3200MHz. Over the next day, I noticed more unexpected
> happenings such as brief cutouts. I found a download at MSI that was supposed to fix
> the S/PDIF problem, but it didn't. The board came with the 3rd BIOS installed; the
> most recent is the 8th. So I eventually flashed that in hopes it would solve the
> issues. The flash completed with no problems, but all it appeared to do was change
> some of my BIOS settings back to default. They were mostly simple ones such as
> disabling the MSI logo at startup and disabling TPM (I don't want Microsoft trying to
> sneak Windows 11 onto my PC). I did turn XMP back on and it showed as being enabled
> and supporting 3200MHz, but RAM stayed at 2133MHz. Going back into Windows was
> simple as usual. I checked when I got back in and, sure enough, 2133MHz RAM speed
> was reported in Task Manager Performance. After several hours, I rebooted and went
> back into the BIOS to recheck RAM speed there. It had moved to 3200MHz then. I made
> no changes, but when I started to leave there were a number of strange looking
> changes indicated. Couldn't make heads or tails of what looked like gibberish. I
> cleared that mess by clicking on "No" and proceded to get out of the BIOS and back
> into Windows. No go! I got a notice that Windows couldn't start and I needed to go
> into Recovery Mode, but keyboard selections weren't responding as they should. I
> could get into some pages from the initial screen, but not others. Then things fell
> apart very quickly. I couldn't get back into the BIOS. My keyboard stopped being
> responded to. The reset button wouldn't work. In the end, only the power button
> would work. USB keyboard lights didn't come on (it's backlit) nor did any of its
> LEDs. The monitor screen stayed off. The power button started at least some part of
> the system (indicator LED on, but drive activity LED stayed off), and a hold of
> several seconds turned it off. Nothing else worked.
>
> I removed the problem board and put my old one back in (ASUS Prime Z390-A) with the
> new PSU and GPU and old CPU, CPU cooler, RAM, NVMe, HDDs and optical drives.
> Everything works as it should. The installation of Windows that couldn't start from
> the new board started with no problem other than a normal pause while Windows
> installed devices for the old one. I haven't pulled the new CPU, RAM and CPU cooler
> off the MSI board yet.
>
> My thinking is that the new board was flawed from the start, couldn't tolerate
> changes and just got progressively worse until it's now more or less bricked.
>
> I've been building PCs for many years and have never experienced anything like this
> before. In the past they either ran or they didn't. There was never a slow death.
>
> Any thoughts about anything other than the motherboard being at fault? Except for an
> outright PSU failure, I can't think of anything other than the motherboard that could
> make USB inoperable, especially with it happening over a short period rather than
> suddenly. Not RAM and certainly not the CPU.
>
> Larc
>

What you know is, the hardware works. It would not have survived all of that,
without crashing, if there was a problem at the hardware level.

There was some issue in the UEFI. By observation, the UEFI is its own OS,
running out of firmware. It's not a BIOS. It is not a Basic Input Output
Subroutine, by any stretch of the imagination. It is a base of operations
for somebody.

And you know from the recent Gigabyte issue, there is PLENTY of evil
these UEFIs can get up to . They can carry out exploits on the running OS.

One thing I noticed, even on the seven year old Test Machine, is the
RESET signal is no longer a hardware reset. One day, the Test Machine took
30 SECONDs to respond to reset. That is not hardware control. That is a joke.

Four days ago, I changed my CPU here. On an MSI Gaming series for AMD.
The system normally boots in seconds, maybe ten seconds or so. When
the new processor was installed, there was a delay of a minute or two at
UEFI level, before I could even get into settings. Responses were sluggish
(as if maybe CPU cache was turned off), but I carried on, removed the
Full Screen Logo as you did. As the settings got reset by the CPU change.
I expect the BIOS flashes the NVRAM and clears it out, after a significant
hardware change, and that could be what it was doing during that delay.

It took maybe three power cycles, before, magically, the proper ten
second boot sequence suddenly came back. It was not completely normal
on the second attempt at starting. I seemed to be fumbling about
at finding something to boot from, but the fumbling took way way too long.

*******

Summary: I would return that motherboard to the seller as quickly as possible,
so you could get a refund. Change motherboard models when you
acquire a replacement. Try not to get saddled with the same model.

You may want to Google that model of motherboard, and see if anyone
else reports symptoms like this.

Gigabyte's issue, is they installed a software with behavior similar
to Computrace. They injected something into Windows. This would have
been, um, "fine", if the method was not insecure. The mechanism
could be exploited by Black Hats.

In the case of MSI, they "lost" their signing keys or something.
They had a breach where some materials used to secure things like
BIOS, have gone missing. And, I don't think there is any mechanism
for them to recover from this. They should have had an invalidation
procedure, so the breached materials could be disabled, but I don't
think they have that.

I would not say these companies are sloppy necessarily, but
the tech is getting too complicated for its own good. Imagine if the
plan to remove CSM and go only Secure Boot had been carried out.
We would be in a hell of a lot of trouble.

As for RESET, if you have any doubt about hardware sanity,
today that means turning off all power. Switch off, pull the plug,
push the front power button to drain. That's the best I can offer.
At best now, that so-called RESET signal, is just an interrupt of some
sort. Maybe it's an SMI that the UEFI picks up, and the response to
the SMI is "takes its own sweet time" in the UEFI. I will not know
what they're doing, until I can get my hands on a reference schematic.

While the OS is running, UEFI continues to run. I do not know the
exact mechanism, but on my board, it has closed loop control of
temp, CPU fan, voltage, frequency, power limiters, and so on. Mine
seems to have set a target for itself, of 70C at the silicon die level.
That's what its closed loop control is doing. My processor might be
able to run at 5GHz, but with my cooling system, I get maybe 4.0
to 4.7GHz or so. If I'd bought "the big Noctua", like a D15, maybe
it would run at 5GHz, but I really don't care about that, as the
monstrous cooler takes up all available space (it also torques the
board a bit because of the weight), and may even prevent
VCore from receiving cooling. I'm using a C14S instead, a blow-down
cooler with only one fan. There may not be clearance to one of the
heatsinks, to allow adding a second fan. I purchased the blow down,
because I figured VCore needs it, but VCore does not need it because
of the closed loop control. (If VCore converter gets too hot, I think
it just turns down the CPU to compensate. That's why it can be seen
running at only 4GHz.)

If I reach into the BIOS and turn off the current and power limiters,
then, whatever happens, is on my head. And with the labyrinth of
UEFI screen settings, it's almost impossible to find those settings
anyway.

Paul

On Thu, 15 Jun 2023 05:10:13 -0400, Paul <nospam@needed.invalid> wrote:

| Summary: I would return that motherboard to the seller as quickly as possible,
| so you could get a refund. Change motherboard models when you
| acquire a replacement. Try not to get saddled with the same model.
|
| You may want to Google that model of motherboard, and see if anyone
| else reports symptoms like this.
|
| Gigabyte's issue, is they installed a software with behavior similar
| to Computrace. They injected something into Windows. This would have
| been, um, "fine", if the method was not insecure. The mechanism
| could be exploited by Black Hats.
|
| In the case of MSI, they "lost" their signing keys or something.
| They had a breach where some materials used to secure things like
| BIOS, have gone missing. And, I don't think there is any mechanism
| for them to recover from this. They should have had an invalidation
| procedure, so the breached materials could be disabled, but I don't
| think they have that.
|
| I would not say these companies are sloppy necessarily, but
| the tech is getting too complicated for its own good. Imagine if the
| plan to remove CSM and go only Secure Boot had been carried out.
| We would be in a hell of a lot of trouble.
|
| As for RESET, if you have any doubt about hardware sanity,
| today that means turning off all power. Switch off, pull the plug,
| push the front power button to drain. That's the best I can offer.
| At best now, that so-called RESET signal, is just an interrupt of some
| sort. Maybe it's an SMI that the UEFI picks up, and the response to
| the SMI is "takes its own sweet time" in the UEFI. I will not know
| what they're doing, until I can get my hands on a reference schematic.
|
| While the OS is running, UEFI continues to run. I do not know the
| exact mechanism, but on my board, it has closed loop control of
| temp, CPU fan, voltage, frequency, power limiters, and so on. Mine
| seems to have set a target for itself, of 70C at the silicon die level.
| That's what its closed loop control is doing. My processor might be
| able to run at 5GHz, but with my cooling system, I get maybe 4.0
| to 4.7GHz or so. If I'd bought "the big Noctua", like a D15, maybe
| it would run at 5GHz, but I really don't care about that, as the
| monstrous cooler takes up all available space (it also torques the
| board a bit because of the weight), and may even prevent
| VCore from receiving cooling. I'm using a C14S instead, a blow-down
| cooler with only one fan. There may not be clearance to one of the
| heatsinks, to allow adding a second fan. I purchased the blow down,
| because I figured VCore needs it, but VCore does not need it because
| of the closed loop control. (If VCore converter gets too hot, I think
| it just turns down the CPU to compensate. That's why it can be seen
| running at only 4GHz.)
|
| If I reach into the BIOS and turn off the current and power limiters,
| then, whatever happens, is on my head. And with the labyrinth of
| UEFI screen settings, it's almost impossible to find those settings
| anyway.
|
| Paul

Thanks very much for the getback, Paul. You confirm my thinking. The MSI board is a
thing of beauty, very solidly built and heavy. At least in the appearance
department, it was a bargain at only $150. But with events this crazy, I think a
different brand might be best. I've been an ASUS fan for years and have generally
been happy with their products. Even they have had some serious problems recently,
but mainly with Z690 boards. I haven't heard about any particular problems with
Z590. The ROG Strix Z590-E is $50 more than the MSI, but I think I may head in that
direction or maybe toward the TUF Z-590 Plus. Like the MSI, the ROG Strix does have
BIOS Flashback, which I like.

I did check some Amazon reviews for the MSI MPG-Z590 Gaming Plus. They are generally
quite good with most problems being either DOA or failing after months of use. I
didn't see any with problems similar to mine, although I didn't read them all. 82%
gave the board 5 stars, only 1 less than #1 rated Z590 board ROG Strix at 83%.

Larc