Contribute
Register

[SUCCESS] Gigabyte Designare Z390 (Thunderbolt 3) + i7-9700K + AMD RX 580

Yea, I'm an X299 lurker in this thread mostly watching the TB stuff develop! Still, maybe the problem is the onboard TB vs PCI card?

Here is my SSDT. Hope it can help in any way!
 

Attachments

  • SSDT-X299-TB3HP.aml
    6.3 KB · Views: 93
Clover won't provide any response to F11; it will silently clear NVRAM. I have two Designare Z390 systems, and the update experience on them was different:
  • Test Bench.
    • This was a fresh installation of Catalina. It has very few apps.
    • All kexts are in /CLOVER/kexts/Other.
    • Native NVRAM enabled.
    • Latest Clover, Lilu, WEG, AppleALC, VirtualSMC.
    • Update from 10.15.3 to 10.15.4 was very smooth.
    • No "bless" errors in which macOS is unable to select a startup disk.
    • No leftover "Install macOS" volumes in Clover.
    • AMD RX 580.
  • Main System.
    • This started with Mojave and was upgraded in-place to Catalina.
    • Kexts are in both /L/E and CLOVER/kexts/Other.
    • Clover 5104.
    • Older versions of Lilu, WEG, AppleALC, and VirtualSMC. Two revisions old.
    • Uses EmuVariable and RC scripts. Native NVRAM not enabled.
    • AMD RX 5700 XT.
    • Update from 10.15.3 to 10.15.4 went through, but received "bless" error at the end.
    • "Install macOS" volumes were left over in Clover.
    • Being unable to remove the leftover "Install macOS" volumes, I erased the hard drive and restored it back to 10.15.3 from a full backup made immediately before performing the update.
    • This time I switched to full native NVRAM.
    • Left all kexts and Clover the same.
    • Received "no-entry" sign on restart.
    • Cleared NVRAM in Clover Boot Menu (F11). Tried again.
    • Now the install went through to the end, but received "bless" error again.
    • Leftover "Install macOS" volumes still present.
    • Will try different techniques to remove them.
    • I also re-ran the update with latest Clover and kexts, but again ran into "bless" error at the end.
    • Not sure why this system responded differently from the Test Bench.
I manually updated clover to 5107 and retried the update. This time it took off and went through the process ending with the same bless error as your main system. Repeated Restarts on the preboot volume after the bless error now result in a new error and the system wants to set a boot volume. I was hope that may wrap the install up and close things out.

Then restarting from the macOS Catalina volume brings up what shows as a 10.15.4 environment that is operating. Not sure if this is good or bad, the leftover preboot volumes will bother me.
 
You are both spot on - I have this board too and it was NVM14.
I got ASRock to provide a NVM20 here - https://www.tonymacx86.com/threads/...olt-3-i7-9700k-amd-rx-580.267551/post-2089293
I have safely upgraded to NVM 20 from windows and have dumped the firmware again. Both are in the thread above.
@dgsga


Some success. All credit to @Elias64Fr for the patched firmware and @CaseySJ .

@CaseySJ - Have I got this right ? IOReg (before SSDT and after SSDT) and bootlog attached. I know nothing about an SSDT, so I modified one that I found here with my Thunderbolt DROM. If you see any mistakes - please help me to correct them.
I got the CRC8 and the CRC32 errors - Have corrected them in the SSDT. (thank you, Elias and others for the amazing work and explanation)

Hotplugging, Cold reboot and Thunderbolt bridge works. Will test more tomorrow.
Under the PCI section I see some information missing.

My original DROM is here (NVM20.bin.zip) - https://www.tonymacx86.com/threads/...olt-3-i7-9700k-amd-rx-580.267551/post-2086474

And info about the DROM is here - https://www.tonymacx86.com/threads/...olt-3-i7-9700k-amd-rx-580.267551/post-2086474

Details about how to patch are here - https://www.tonymacx86.com/threads/...tb3-igpu-mojave-sff-build.277418/post-2092659

@rj510
 

Attachments

  • iMac- NVM20.ioreg
    6.9 MB · Views: 89
  • iMac -NVM20-SSDT applied.ioreg
    7.6 MB · Views: 110
  • Screenshot 2020-03-27 at 9.05.37 PM.png
    Screenshot 2020-03-27 at 9.05.37 PM.png
    747.4 KB · Views: 115
  • Screenshot 2020-03-28 at 12.15.26 AM.png
    Screenshot 2020-03-28 at 12.15.26 AM.png
    172.9 KB · Views: 118
  • boot.txt.zip
    2 MB · Views: 84
  • SSDT modified with DROM.aml
    2 KB · Views: 102
  • ASRock_z390_ITX_Orignal_NVM20.bin.bin.zip
    213.7 KB · Views: 75
  • ASRock_z390_ITX_Patched_NVM20_try1.bin.zip
    213.6 KB · Views: 62
Last edited:
Clover won't provide any response to F11; it will silently clear NVRAM. I have two Designare Z390 systems, and the update experience on them was different:
  • Test Bench.
    • This was a fresh installation of Catalina. It has very few apps.
    • All kexts are in /CLOVER/kexts/Other.
    • Native NVRAM enabled.
    • Latest Clover, Lilu, WEG, AppleALC, VirtualSMC.
    • Update from 10.15.3 to 10.15.4 was very smooth.
    • No "bless" errors in which macOS is unable to select a startup disk.
    • No leftover "Install macOS" volumes in Clover.
    • AMD RX 580.
  • Main System.
    • This started with Mojave and was upgraded in-place to Catalina.
    • Kexts are in both /L/E and CLOVER/kexts/Other.
    • Clover 5104.
    • Older versions of Lilu, WEG, AppleALC, and VirtualSMC. Two revisions old.
    • Uses EmuVariable and RC scripts. Native NVRAM not enabled.
    • AMD RX 5700 XT.
    • Update from 10.15.3 to 10.15.4 went through, but received "bless" error at the end.
    • "Install macOS" volumes were left over in Clover.
    • Being unable to remove the leftover "Install macOS" volumes, I erased the hard drive and restored it back to 10.15.3 from a full backup made immediately before performing the update.
    • This time I switched to full native NVRAM.
    • Left all kexts and Clover the same.
    • Received "no-entry" sign on restart.
    • Cleared NVRAM in Clover Boot Menu (F11). Tried again.
    • Now the install went through to the end, but received "bless" error again.
    • Leftover "Install macOS" volumes still present.
    • Will try different techniques to remove them.
    • I also re-ran the update with latest Clover and kexts, but again ran into "bless" error at the end.
    • Not sure why this system responded differently from the Test Bench.

I was reading somewhere else about this issue.

Code:
The specific of what happened:

[LIST]
[*]Apple added Ice Lake and Comet Lake support to the kernel
[*]Problem is it wouldn't fit as-is, so the mask of the processor changed
[*]This then meant that it turned from 16 bits to 20 bits long
[*]This then broke Clover's KernelPM and AppleIntelCPUPM patch
[/LIST]
So because of this change, we have 4 paths to go down:

[LIST]
[*]Disable CFG-Lock in the BIOS
[LIST]
[*]See Disabling CFG-Lock Guide
[/LIST]
[*]Use a CPU ID that still works with the old KernelPM patch
[LIST]
[*]Broadwell and older are fine
[*]Note doing this can partially or completely break XCPM
[/LIST]
[*]Update to Clover with commit version af8a84f or later
[LIST]
[*]r5107 is the minimum version, plus recent Cover updates moved to C++ breaking a lot of things so keep a backup
[/LIST]
[*]Buy a hackintosh that starts with the letter M
[/LIST]
For the CPU ID route, you can spoof to a Broadwell-H CPU with the following:

config.plist -> KernelAndKextPatches -> FakeCPUID -> 0x040674

[LIST]
[*]Remember doing this can partially or even completely break XCPM
[/LIST]
 
Last edited by a moderator:
Some success. All credit to @Elias64Fr for the patched firmware and @CaseySJ .

@CaseySJ - Have I got this right ? IOReg (before SSDT and after SSDT) and bootlog attached. I know nothing about an SSDT, so I modified one that I found here with my Thunderbolt DROM. If you see any mistakes - please help me to correct them.
I got the CRC8 and the CRC32 errors - Have corrected them in the SSDT. (thank you, Elias and others for the amazing work and explanation)

Hotplugging, Cold reboot and Thunderbolt bridge works. Will test more tomorrow.
Under the PCI section I see some information missing.

My original DROM is here (NVM20.bin.zip) - https://www.tonymacx86.com/threads/...olt-3-i7-9700k-amd-rx-580.267551/post-2086474

And info about the DROM is here - https://www.tonymacx86.com/threads/...olt-3-i7-9700k-amd-rx-580.267551/post-2086474

@rj510
Good progress!

Is "NVM20.bin.zip" the modified firmware for ASRock Z390 ITX, or is that the original?
 
Hello Everyone!

Last night I decided to do a fresh install of 10.15.4 using the mini-guide that was very successful for 10.15.3 for me a few months ago. But I am getting the dreaded:

"An error occcurred while preparing the installation. Try running this application again."

I even tried preparing another (new) USB 32GB drive and the same thing. For kicks I went back to my 10.15.3 prepared USB drive (which now has been updated to Clover 5107 as well) and...no problem! Works perfectly.

Also, in both cases (the .3 and the .4 USB drives) Lilu, WG, VSMC and AppleALC are all at the latest.

I don't want to experiment around too much but if I have to I'll just go back to 10.15.3 and wait until we see the first update to .4.

Any thoughts? Thanks!

Lam


I got exactly the same problem when I tried to a fresh install of 10.15.4 using the mini-guide. I tried several times, all ending up with "An error occcurred while preparing the installation. Try running this application again."
Looking forward solutions.
 
Hello Everyone!

Last night I decided to do a fresh install of 10.15.4 using the mini-guide that was very successful for 10.15.3 for me a few months ago. But I am getting the dreaded:

"An error occcurred while preparing the installation. Try running this application again."

I even tried preparing another (new) USB 32GB drive and the same thing. For kicks I went back to my 10.15.3 prepared USB drive (which now has been updated to Clover 5107 as well) and...no problem! Works perfectly.

Also, in both cases (the .3 and the .4 USB drives) Lilu, WG, VSMC and AppleALC are all at the latest.

I don't want to experiment around too much but if I have to I'll just go back to 10.15.3 and wait until we see the first update to .4.

Any thoughts? Thanks!

Lam
I got exactly the same problem when I tried to a fresh install of 10.15.4 using the mini-guide. I tried several times, all ending up with "An error occcurred while preparing the installation. Try running this application again."
Looking forward solutions.
It may not be possible yet to perform a fresh installation of 10.15.4. For the time being I've made the following change to Catalina Mini-Guide:

Screen Shot 2020-03-27 at 4.35.42 PM.png
 
Update on Damaged Asus ThunderboltEX 3

Last week I damaged the Winbond chip on my Asus ThunderboltEX 3 add-in-card while attempting to force-flash a modified Thunderbolt firmware. This was before creating the circuit with a 47uF capacitor and 2K Ohm resistor. Had I used that circuit at the time, I would not have damaged the chip.

Nevertheless, to turn the loss into a gain I decided to desolder and replace the damaged chip. First step: to order a set of replacement chips and desoldering materials, all of which arrived over the past couple of days.

Status update:

The Winbond W25Q80DV comes in two package sizes:
  • W25Q80DVNIG
  • W25Q80DVSIG
The following boards use the larger W25Q80DVSIG:
  • Gigabyte Designare Z390
  • Gigabyte Z390 AORUS Xtreme
  • Gigabyte GC-Titan Ridge
  • Asus ThunderboltEX 3
IMG_1109.jpg

It was easy to flash the standalone chip, but it could only be gripped properly by the Pomona 5250. The cheap clip that accompanies most CH341a Programmers placed too much pressure, bending some of the pins.

Next the existing flash chip on the Asus add-in-card had to be desoldered:
  • A clean solder tip is essential. I had to replace a worn tip that was merely evaporating the solder instead of melting it into tiny beads.
  • Applied liquid flux and used ChipQuik at a temperature of about 250 C. After about 4 or 5 strokes on each side, the chip glided off easily. No hot air this time.
IMG_1113-resized.jpg

Soldering the replacement was fairly easy because of the relatively wide spacing between pins.

IMG_1114-resized.jpg

So after all that, did it work? No! :(

The soldered-on chip does work and is readable and writable by the Raspberry Pi, but it's possible that other components were damaged. Will make one more attempt to repair the board.

Meanwhile, the replacement Asus ThunderboltEX 3 arrived today and it's easily readable and programmable with the 47uF / 2K Ohm circuit.
 
Last edited:
Hi @Casey, wondering if you might know a fix for this problem. When I share a video from the photos app I see a pie shaped progress indicator in the toolbar of the photos app. The VTEncoderXPCService process runs during this time and CPU activity is ~600%. To share one 60 second video takes longer than 60 seconds. I know it's converting it or downscaling it, but that seems extremely slow for such a well specced machine. Is this supposed to be offloaded to my 5700 XT or can you think of anything that might not be optimized for this process?
 
Back
Top