Contribute
Register

[Success] ASRock Z390 Phantom Gaming-ITX + TB3 + iGPU + Mojave + SFF Build

I can only specify a custom profile, instead of an automatic one.

I didn't realize that in this version of BIOS it's not possible to disable the XMP
Try deactivating it from Clover and see what happens, but I think it will fail anyway.

Clover section: Boot -> XMPDetection = false
 
Last edited:
I've not seen this on my build. You might try disabling the graphics section in Clover/Arbitrary. If that doesn't work, my suspicion is that in "re-using" the SSD, you've got some corrupted files. You may need to do a fresh install.
Hi,
I tried to make a bootable drive with your EFI tree on it and have tried a lot of other ways and most of them shows the same error "...attempting to load /sbin/launchd.
I bought today the same M2 (WD BlacK) and build a fresh install Mojave on it from a MacBook Pro, and put you EFI on it. (I know that it is not the best way to do that.) But, anyway, when the computer restarts, the error was the same.
I don't know what to do anymore. In the other builds I don't have these problems.
 

Attachments

  • WhatsApp Image 2019-07-16 at 21.42.11.jpeg
    WhatsApp Image 2019-07-16 at 21.42.11.jpeg
    389.3 KB · Views: 75
Hey, thanks for this great guide. I'm using your latest EFI and successfully installed Mojave. I can load into the OS via the dGPU. However, it's Nvidia so doesn't have accelerated GTX. When I opt to use the iGPU in the BIOS, the display becomes one big multi-colored mess. Have you seen this happen? TIA
 
@rj510 @jsil

The NGFF wifi card you showed in Spoiler: Replacing the BT PCB section looks a little bit different to the Amazon one.

416928


After noticing the slight difference, I've checked on the Asrock site and found the description about WiFi/BT below:

- 1 x Vertical M.2 Socket (Key E) with the bundled WiFi-802.11ac module and Intel® CNVi (Integrated WiFi/BT) (on the rear I/O)

So not only does the product number has to be the same (BCM94532z m.2 ngff Dw1560), the socket matters as well, correct?

The local retailer tells me that Broadcom BCM94356z NGFF/M.2 is what they've got per se and send me these front and back pics. The product name and the socket are not quite the same, wondering if this one would work on the z390 mb?

416931
416932
 
I didn't realize that in this version of BIOS it's not possible to disable the XMP
Try deactivating it from Clover and see what happens, but I think it will fail anyway.

Clover section: Boot -> XMPDetection = false
Unfortunately, this did not solve my problem. I also tried to include profile number 1 in XMP, but it did not help.
 
Backing up to a simpler question: is anyone having issues with sleep/shutdown? It used to work perfectly for me, but somewhere along the way issues were introduced, and I'm not sure where to look for the mistake:
  1. Hitting 'sleep' makes the screen go blank, fans continue whirring, then after 5-10 seconds, there's a "click", fans stop, the GPU LED goes off, but then 1-2 seconds later the computer gets woken up again.
  2. Shutdown often just restarts instead...
I think this is fairly common and that it's probably an issue with an SSDT patch, but I'm not sure. Has anyone encountered this?
 
Yep, saw it :)

Appreciate the quick reply, of course. Unfortunately, I cannot get back to my i9 during the day to upload my config.plist (for a prior poster) or write up a more thorough response to your post until I'm back at my i9 and have spent more time testing/reading!

I'll also be working on fixes I might come up with / find—ultimately I think the best bet overall for the most Apple-like TB3 functionality will be an SSDT fix similar to the NUC link I provided a few pages back. He specifically accounts for the interrupts that the Alpine Ridge controller sends the system on different events, which he specifically wrote was for the purpose of enabling hot-plugging. If that type of fix works, I suspect it'll take care of 3.0 (and 3.1 gen 2?) speeds by virtue of it hooking into Apple's TB XHC configs. (While I'm not ruling out modifications to the XHC-5 section of the codeless kext, and we may be able to get USB working manually that way.) I didn't have time to test the SSDT thing over the weekend, though, but I will post when I get to it. Also I of course don't expect you to do to it for me (I'm working on this also), but I thought it was better sharing promising notes before I totally checked it out in case you were interested.

One thing to note: I highlighted the electrical current control parameters in the plist only in connection with the HighCurrent USB setting you have set for Clover—I thought you may have enabled it to overcome an "insufficient power" warning, and I wanted to point out that Hackintool provides these values and that with them, I've not seen such issues. I doubt USB 3.0/3.1 speeds are related just to power settings, but I may be wrong. I'm also slightly wary about manipulating power values for a TB3 controller, since it's pretty different than a standard USB bus.

I got the speed of SSP1 to 10Gbps. I'll attach the USBPorts.kext and the SSDT-TB3 once I've refined them. There is interaction between the two. I booted the computer with the TB3 dock attached and the Apollo x6 connected and all on. Turning off the Apollo and back on did not re-attach, so no hot-plugging.

Next, I un-plugged the dock, plugged in the USB-C drive and it automatically mounted and the speed was set to 10Gbps (see attached Hackintools). To get this speed, there IS interaction between the SSP1 being injected and the SSDT-TB3. It is complicated. I eliminated SSP2 and only injected SSP1.

I'm still working on TB3 hot-plugging. It is possible as was recently demonstrated on a German website (but no details were revealed; I requested an SSDT but was respectfully declined).
 

Attachments

  • 10Gbps SSP1.jpg
    10Gbps SSP1.jpg
    210.1 KB · Views: 63
Last edited:
The "guilty" of that is to have active a profile XMP, if you deactivate it disappears the error.

I have XMP active and have no similar errors.
 
Back
Top