New B550M Pro4 will UEFI boot but doesn't display POST

[than3]

I just ordered one of these on a black friday sale off Newegg for use with an 5950X Ryzen 9 system.
Interesting thing is this mainboard doesn't POST aside from a brief American Megatrends copyright notice in the upper left corner with a BIOS date and what appear to be a string of hex opcodes in the bottom right that appear to be completely undocumented by Asrock. It does load to BIOS, and the boot menu boots into live environments, so its clearly not your average NOPOST issue.
No RAM self check, storage enumeration, etc. Ended up sending the first one back because it failed a BIOS update (P2.80 to P3.20), the replacement has the exact same issue but was able to update to the latest firmware.
First time seeing something like this. Is this common with the newer boards?
I haven't seen anything like this (where the POST screen is just completely blank/gone) in the 9 years I've worked in IT, so as a sanity check, I'm wondering if anyone here has seen anything like this?

 

[than3]

As a follow-up,

Asrock is no longer shipping their mainboards with POST. Only an EFI-based splash screen which contains no information pre-handoff to boot, and most of what you would normally see at POST must now be accessed from the EFI Manager. Some aspects show less granular detail than before such as only the Mfg, Frequency and Capacity for RAM (without any real check), and some parts of the POST related to device enumeration seem to have had information dropped and/or hidden as well.

I just got off the phone with one of their Tech Support representatives to confirm this, John was able to assist with the confirmation.

Many permutations of hardware were tested. All the component hardware pieces were tested working on an independent mainboard with no issues.
So the issues seem to be solely related to this mainboard hardware. Normally I would chock this up to just a defective board and bad QA, and there were other low level issues with the first board I received, but as this is the second replacement (same model), I'll just have to go with a different mfg.

TL;DR

No POST screen. When splash is disabled, aside from a brief AMI copyright logo at power-on (upper left), the BIOS date and internal BIOS revision (upper left), and some undocumented opcode sequences (bottom right), no information is displayed for the power-on-self-test.

This is by design, though it is not documented in any of their written documentation.

---Notes & Observations
Of note, the physical POST LED indicators on-board may not accurately reflect when there are failures/issues with the subsystems being loaded after RAM/CPU lights dim during early boot.
For example, The VGA/Boot LEDs light and dimmed in both cases regardless of if there was a graphics card present (without APU)/functioning. This should have at least produced a clear fault since video, ram, and CPU are needed for standard bringup.

Hopefully this helps someone out, moving forward.

This was a fairly interesting diversion/surprise after committing to an initial outlay, the tech support by email was non-existent. I had to call them and their phone systems were not functioning well.
The tech had to call me back because their PBX was halving the inbound volume signals intermittently.

In my opinion, Dropping POST is a poor business decision on their part, largely because any professional is going to methodically confirm their hardware works when they first purchase it.
Being unable to isolate a video card fault from a mainboard/ram/cpu fault is a complete non-starter.

Device lifecycle management being what it is, there is no doubt you will need to troubleshoot between those types of faults at some point in the devices lifetime.
POST was used for decades, and some professional certifications even test for this knowledge. Dropping it was a bad idea, so for those considering this board. Buyer beware.

 

[tabascosauz]

As a follow-up,

Asrock is no longer shipping their mainboards with POST. Only an EFI-based splash screen which contains no information pre-handoff to boot, and most of what you would normally see at POST must now be accessed from the EFI Manager. Some aspects show less granular detail than before such as only the Mfg, Frequency and Capacity for RAM (without any real check), and some parts of the POST related to device enumeration seem to have had information dropped and/or hidden as well.

I just got off the phone with one of their Tech Support representatives to confirm this, John was able to assist with the confirmation.

Many permutations of hardware were tested. All the component hardware pieces were tested working on an independent mainboard with no issues.
So the issues seem to be solely related to this mainboard hardware. Normally I would chock this up to just a defective board and bad QA, and there were other low level issues with the first board I received, but as this is the second replacement (same model), I'll just have to go with a different mfg.

TL;DR

No POST screen. When splash is disabled, aside from a brief AMI copyright logo at power-on (upper left), the BIOS date and internal BIOS revision (upper left), and some undocumented opcode sequences (bottom right), no information is displayed for the power-on-self-test.

This is by design, though it is not documented in any of their written documentation.

---Notes & Observations
Of note, the physical POST LED indicators on-board may not accurately reflect when there are failures/issues with the subsystems being loaded after RAM/CPU lights dim during early boot.
For example, The VGA/Boot LEDs light and dimmed in both cases regardless of if there was a graphics card present (without APU)/functioning. This should have at least produced a clear fault since video, ram, and CPU are needed for standard bringup.

Hopefully this helps someone out, moving forward.

This was a a fairly interesting diversion/surprise after committing to an initial outlay, the tech support by email was non-existent. I had to call them and their phone systems were not functioning well.
The tech had to call me back because their PBX was halving the inbound volume signals intermittently.

In my opinion, Dropping POST is a poor business decision on their part, largely because any professional is going to methodically confirm their hardware works when they first purchase it.
Being unable to isolate a video card fault from a mainboard/ram/cpu fault is a complete non-starter.

Device lifecycle management being what it is, there is no doubt you will need to troubleshoot between those types of faults at some point in the devices lifetime.
POST was used for decades, and some professional certifications even test for this knowledge. Dropping it was a bad idea, so for those considering this board. Buyer beware.

If you disable splash screen, you get a very brief report with the AMI logo. That's the case for all boards of all brands and has been for years and years, long before B550.

Can't you use an option to extend the POST delay? That might give you more time to read the AMI screen. Otherwise, no, this isn't 20 years ago, it's not going to go intentionally slowly, step by step through a long POST process and show you output every step of the way.

 

[than3]

If you disable splash screen, you get a very brief report with the AMI logo. That's the case for all boards of all brands and has been for years and years, long before B550.

Can't you use an option to extend the POST delay? That might give you more time to read the AMI screen. Otherwise, no, this isn't 20 years ago, it's not going to go step by step through a long POST process.

The POST delay won't actually display the POST process, its effectively a blank screen. This page also ignores pause/break and scroll lock in early boot.

You can extend the delay but there isn't much point since no information is actually displayed for bup, and the inability to quickly determine between if its a mainboard failure or video card failure (both decently expensive components) is an unacceptable compromise.

I've a B450 Plus, AM4 capable board which has the regular POST process (when you disable the splash). It is ASUS, and has other low level issues, but it is recent and follows the standard.

 

[tabascosauz]

The POST delay won't actually display the POST process, its effectively a blank screen. This page also ignores pause/break and scroll lock in early boot.

You can extend the delay but there isn't much point since no information is actually displayed for bup, and the inability to quickly determine between if its a mainboard failure or video card failure (both decently expensive components) is an unacceptable compromise.

I've a B450 Plus, AM4 capable board which has the regular POST process (when you disable the splash). It is ASUS, and has other low level issues, but it is recent and follows the standard.

Ah, I missed your point that the RAM/drive info is missing on ASRock.

I have not had an AM4 board that displays useful info on the GPU. I've been through a B450I Aorus Wifi, B550M Steel Legend, B550M-itx/ac, B550I Edge MAX, B550M TUF, 2 x B550I Aorus AX, B550 Strix-I, C8 Impact and B550 Unify-X. The Asus boards display some storage info but the Gigabyte boards are comparably minimalistic to ASRock with logo disabled, just BIOS version, board name and maybe RAM size. In all my CPU tinkering and mem OC I don't think the POST screen has ever provided useful info.

I hear you, but the POST report is honestly not particularly high up or authoritative in the troubleshooting process for modern hardware. To be honest, I'm not sure what displaying info on the GPU would achieve; these days, either the GPU is dying (in which case you can probably get into Windows to test properly) or it's already dead (in which case either blackscreen (if no iGPU), or iGPU only (if iGPU)).

 

[ir_cow]

I hear you, but the POST report is honestly not particularly high up or authoritative in the troubleshooting process for modern hardware. To be honest, I'm not sure what displaying info on the GPU would achieve; these days....

Yep this isn't the 90s anymore, the post screen is fairly useless now.

 

[damric]

eh off topic but this reminds me that yesterday my father called me to complain about the nice ryzen 5600g build I have him a while back.

"I couldn't find where to insert the CD. I kept looking, and it doesn't look like it has a CD drive or disk drives and I can't install my favorite game."

He also insists on keeping his land line phone even though there is so much static you can barely hear anything. I bought him a cell phone but he refuses to answer it or keep it charged.

 

[than3]

To be honest, I'm not sure what displaying info on the GPU would achieve; these days, either the GPU is dying (in which case you can probably get into Windows to test properly) or it's already dead (in which case either blackscreen (if no iGPU), or iGPU only (if iGPU)).

A bit OT, but to answer;
It was historically used to infer if the failure was the mainboard socket side, where the GPU plugged in, or if it was the GPU itself, without needing to sacrifice a potentially expensive donor card to find out.
One could easily test the card in a separate PC that was known working, but if the mainboard failed, and destroyed the GPU socket in the process, you could easily have both components dead, in addition to any donors you might try to test on the mainboard itself, not getting you closer to a solution without that inference. It may have been specific to an earlier technology. I seem to remember VGA coming up first fairly standalone, followed by the rest of the card, probably not generalizable at this point since there's so little low-level documentation of modern hardware.

Also, the old-style memory self-test on POST could reliably determine if there were bad RAM sockets (as it may only read socket A1, and not A2 during the test, etc and there was a pause or significant delays when one of those sockets were going bad but not yet a full failure).

You also used to be able to attach to a serial console, and later a PCI card for extended debug information early boot.

With the properly implemented POST, serial connections for quite a lot of troubleshooting were not necessary, since the minimal test equipment needed was built-in.
You could quickly and easily determine the origin of most issues and buy the replacement component no fuss or muss instead of having to effectively replace everything component-wise at significant additional cost, and more importantly time (as you try and order each component piecemeal with significant risk of loss for catrastrophic hardware failure).

Most modern cards today don't have more than one of each important socket such as an alternate x16 socket to swap and test, outside very high end mainboard.
Insofar as I've seen, any serial ports today are either not attached to the mainboard, or usually only become available after BUP (usually provided by a SuperIO type chip).

I've seen many socket failures over the years, both the destructive (anything plugged into it shorts), and the non-destructive (bup fails, but no damage to peripherals).
I'm dissapointed at how one used to be able to quickly determine the what and where the issue with your hardware was, and that is being taken away.

If you can't verify it was working correctly, and that it wouldn't fail right outside the return date, why would you pay what the going prices are (they've only increased in cost).
Even RMA's can take 6-9 months, and the replacement which eventually is sent out; its warranty doesn't toll. Its not a good way to do business.

 

[RJARRRPCGP]

Your symptom may be a monitor firmware fault. (Where the monitor skips everything except Windows) It may also be a video cable issue as well.

More likely to be a bad cable, if the monitor keeps going to the dreaded no-signal message, when in the OS, such as when starting a game or other tasks.

 

[DoubleRock485]

Hmm, I just received one of these boards. The sticker on the BIOS shows P2.80, but that's not one of the versions listed on the website. What is going on?

 

[tabascosauz]

Hmm, I just received one of these boards. The sticker on the BIOS shows P2.80, but that's not one of the versions listed on the website. What is going on?

ASRock and Gigabyte like to pull old and in-between BIOS versions when newer ones come out. Not really for any reason other than just for the sake of it, and not done in any logical way.

Especially for Gigabyte, if you count all the beta versions that hit the forum, there are hundreds of BIOSes out there, way more than you'll ever see on the website.