MP 1,1-5,1 Is there any hardware issues that could cause a black screen after sleep?

[pywm]

Hi hive mind,

I have a Mac Pro 4,1 -> 5,1, Dual X5690, Gop Enabled on motherboard, RX580 unflashed, Martin Lo's OpenCore 0.9.2, Monterey 12.6.8, 256GB RAM. The machine was working fine before, including normal sleep in both MacOS and Windows 10, until a few days ago.

Since two days ago, the Mac Pro is showing a black screen after sleep. It went into sleep normal, as I can see the lights were out and fan stopped. It seems to resume normal as I can see the light back on and the fan started spinning. But there was nothing on the screen. It is black but not lack of signal, otherwise the monitor would indicate no signal.

The two changes that might be related were below. But these two may not be related to the problem.
1. I upgraded Monterey from 12.6.7 to 12.6.8
2. I started playing games in Windows 10.


I have done the following tests but none have improved the situation:
1. Change RAM to 4 sticks of 2GB each so 8GB total;
2. Try to sleep in Windows 10 but same issues;
3. I have another Mac Pro with 96 RAM and very similar specs. It was also upgraded to Monterey 12.6.8 recently, it is sleeping fine, but at resume, the monitor did have a few seconds of artefacts before resuming normal;
4. I moved the RX580 from Slot 1 to Slot 2;
5. I took out all other PCIe cards;
6. I used High Sierra without OpenCore;
7. I checked the BootROM, looks about the same as when I flashed Enabled Gop 1.3 three months ago;
8. I used a GT 710 but it refused to go to sleep, maybe due to a lack of driver under MacOS. In Windows 10 it had the same sleep problem.

It seems like a hardware problem, any ideas what else to check? Is there any log on MacOS and/or Windows that I can check for errors?

 

[tsialex]

Check the BR2032 RTC battery.

 

[pywm]

Check the BR2032 RTC battery.

Thanks. BR2032 3.17V, I must have replaced it not long ago. This didn't fix the problem.

I found the command "pmset -g log" from the internet, and saw this "Failure during sleep: 0x0000001F : EFI/Bootrom Failure after last point of entry to sleep." The first of these errors started yesterday, so I guess my next step is to reflash the BootROM.

Not sure why it suddenly happened. The only new thing that happened was I put in a new ASM3142 USB3.2 PCI card in slot 4. Could that cause the bootROM to fail?

 

[tsialex]

Could that cause the bootROM to fail?

When the garbage collection already failed and the VSS store is full anything, even a reboot, can cause a NVRAM volume failure, happens a lot with cross-flashed early-2009s.

So, it’s more probable that was just an unfortunate coincidence that it failed after you installed the USB card.

 

[pywm]

I have flashed the BootROM now. I think it looks healthy.

Unfortunately, it is still showing the same error message in "pmset -g log".

"Failure during sleep: 0x0000001F : EFI/Bootrom Failure after last point of entry to sleep"

Any other ideas?

 

[tsialex]

I have flashed the BootROM now. I think it looks healthy. View attachment 2242630
Unfortunately, it is still showing the same error message in "pmset -g log".

"Failure during sleep: 0x0000001F : EFI/Bootrom Failure after last point of entry to sleep"

Any other ideas?

I'd do three four things, testing one by one:

• Remove any PCIe cards besides the GPU and anything connected via FW/USB, to eliminate any buggy OROM blocking the sleep,
• Replace the 32GB DIMMs and test with High Sierra booting fully native, to eliminate any OC related issues,
• High Sierra should still support the GT 710, it's a Fermi GPU that should still have support with 10.13.6, so you can also eliminate GPU issues. If not, install an AppleOEM GPU.
• Temporarily flash the generic firmware upgrade image MP51.fd from 10.14.6 to eliminate any BootROM related issues, since this BootROM image of yours is ancient and the hardware descriptor base_17 is the buggiest version ever released. Also, being a cross-flashed early-2009 do not help.

You can also try AHT/ASD - download the MP5,1 versions since you cross-flashed - and see if anything wrong is shown. You will need to install one of the AppleOEM GPUs to run AHT/ASD.

 

[pywm]

I'd do three four things, testing one by one:

• Remove any PCIe cards besides the GPU and anything connected via FW/USB, to eliminate any buggy OROM blocking the sleep,
• Replace the 32GB DIMMs and test with High Sierra booting fully native, to eliminate any OC related issues,
• High Sierra should still support the GT 710, it's a Fermi GPU that should still have support with 10.13.6, so you can also eliminate GPU issues. If not, install an AppleOEM GPU.
• Temporarily flash the generic firmware upgrade image MP51.fd from 10.14.6 to eliminate any BootROM related issues, since this BootROM image of yours is ancient and the hardware descriptor base_17 is the buggiest version ever released. Also, being a cross-flashed early-2009 do not help.

You can also try AHT/ASD - download the MP5,1 versions since you cross-flashed - and see if anything wrong is shown. You will need to install one of the AppleOEM GPUs to run AHT/ASD.

So I did the following, all together:
* Remove all PCIe cards except the GPU, no USB except the keyboard and ethernet cable
* Swap GPU to GT120
* Swap RAM to 2GB x4 ECC (total 8GB)
* Remove all hard drives/NVME, except one HDD with High Sierra

Still the same problem.
The "pmset -g log" error message under High Sierra is different:
"Sleep Failure [code:0xFFFFFFFF0400001F]"

When I googled this error message, there was one comment in reddit saying a faulty USB port caused it. I checked all USB ports, and they all seem to work normally. There were quite a few comments about this error after upgrading to a different version of Mac OS, like Catalina. If we see more people having sleep problems after upgrading to 12.6.8, then it could be the newer version related.

I also booted into Windows 10, but the sleep button was somehow missing, and using the command line to suspend to sleep didn't have any effect.

I forgot to reflash to the original BootROM. Maybe I still try it in my next scheduled reflash.

 

[pywm]

Thank you tsialex for your help. No need to do further work on it now as the mac pro just died.

It started with no responding to power button after shutdown. If leave disconnected for a while it booted up ok. Last night it just did a sudden power off, which in my experience is a faulty backplane. In hindsight the sleep problem is the early warning of a dying backplane.

I have two other faulty backplanes with similar sudden poweroff, any chance to fix them electrically by swaping components?

 

[tsialex]

Thank you tsialex for your help. No need to do further work on it now as the mac pro just died.

It started with no responding to power button after shutdown. If leave disconnected for a while it booted up ok. Last night it just did a sudden power off, which in my experience is a faulty backplane.

In my experience, this is more like a PSU dying than a defective backplane, but yes, can also be a corrupted BootROM image, dead SPI, a burned southbridge or even a burned SMD fuse.

In hindsight the sleep problem is the early warning of a dying backplane.

I have my doubts, if this Mac Pro was here I'll check everything else before looking at the backplane itself. The first thing to do after being sure that everything else is working fine besides the backplane, is to test the backplane with a MATT card or replace the SPI.

I have two other faulty backplanes with similar sudden poweroff, any chance to fix them electrically by swaping components?

Sure, but it's economically reasonable to do so? If you don't have the training, equipment and resources, it's not viable to acquire that for something that can be bought as little as $75 with shipping from eBay. Just an IR nozzle adequate for the south bridge will cost the same as a used backplane.

 

[pywm]

In my experience, this is more like a PSU dying than a defective backplane, but yes, can also be a corrupted BootROM image, dead SPI, a burned southbridge or even a burned SMD fuse.



I have my doubts, if this Mac Pro was here I'll check everything else before looking at the backplane itself. The first thing to do after being sure that everything else is working fine besides the backplane, is to test the backplane with a MATT card or replace the SPI.



Sure, but it's economically reasonable to do so? If you don't have the training, equipment and resources, it's not viable to acquire that for something that can be bought as little as $75 with shipping from eBay. Just an IR nozzle adequate for the south bridge will cost the same as a used backplane.

Isolation tests indeed confirmed the problem with the backplane. So time to part out the components.

Thank you again for your help!