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

[CaseySJ]

Okey, so lets wait @CaseyCZ to check it

thanks @qthegaijin

Thunderbolt is on PC01.BR1A.SL01 but we'll need the DSDT to determine if there's a sub-device that needs to be disabled. Please run MaciASL --> File --> New from ACPI --> DSDT, save the file and post.

 

[CaseySJ]

@CaseySJ @Elias64Fr

So my 4MD4KLB-B monitor has arrived and I have started testing with my gc-titan ridge-running patched Designare NVM 33.

So far, the early results are not good. Here's the hardware, and what I have tried:
2 Monitors- 4MD4KLB-B (TB3)/LG UltraSharp UHD (DP)
CalDigit TS3 Plus Dock (TB3)

Boot Attempts (Hardware Config/Result)

- One monitor- 4MD4KLB-B (TB3) connected directly to GC-Titan Ridge - Computer does not post
- One monitor- LG UltraSharp UHD (DP) Connected to TS3 Plus dock - Computer does not post
- One monitor- LG UltraSharp UHD (HDMI to GPU) --> Computer boots to desktop
- Two monitor- LG UltraSharp UHD (HDMI to GPU) --> Computer boots to desktop --> Hot plug 4MD4KLB-B (TB3) --> inconsistent success (2 of 5 attempts) But it was glorious when it worked!

What has been consistent for me since using the patched firmware- I have had to maintain a monitor connected to the HDMI of the GPU. Occasionally, I would see the computer post to the DP monitor (connected to the dock) but maybe 3 out of 10 times. I suspect that if I try the UHD (DP) monitor directly to GPU that both monitors will work, but I'm taking a break at the moment.

Any advice? Is it possible there is a configuration issue?

Couple of suggestions:

• Try enabling CSM Support to see if it makes any difference.
• Thunderbolt monitors might work more reliably if the GC-Titan Ridge is not flashed (i.e. use original firmware). This should be attempted -- we should still bridge pins 3 and 5 to force power to the card if a THB_C header is not available.
• Also try one of the other firmware files (e.g. NVM 23), but I suspect they might not perform differently from NVM 33.

 

[svandive]

Greetings,

If you want to leave the MSR unmodified (no native NVRAM) you'll need the https://www.tonymacx86.com/resources/virtualsmc.477/ kext in your EFI folder. Follow all the latest instructions on running kexts, etc from the EFI and not from L/E. I've also enclosed the @CaseySJ's Define R6 USB setup. I'm using Clover 5116 which runs from the EFI. If you want to use the Fractal Define R6 patch I've attached, you'll need to go get @RehabMan's https://www.tonymacx86.com/resources/usbinjectall.458/ kext updated by @Sniki.

I've had nothing but problems running a Windows SSD attached during either the Z370 or the Z390 builds. I never have the Windows 10 SSD connected at the same time that the MacOS SSD is connected. Just my preference for stability. FileVault and SIP are disabled for now.

.AML files go into EFI/Clover/ACPI/patched, and of course kexts go into EFI/Clover/kexts/Other
Good luck, and make sure the right EFI partition (not the Windows version) is being selected at boot.

I can suggest that you remove the Kext you have that are for Laptops:
HibernationFixup
NoTouchID
SMCBatteryManagement
SMCLightSensor
--
I don't think you need the following:
AirportBrcmFixup
AppleAHCIPortHotPlug
BT4LEContinuityFixup
HWInfo
--

I have a modified version of IntelMausiEthernet and SmallTreeIntel82576, the ones you have will be just fine.

I'm using Native NVRAM (SSDT-NVRAM) which you will absolutely want to setup for (follow the Guide)
I'm also using the USB 15 Port Patch to maximize my USB ports, you can continue to use the the SSDT Map

And last you can clean up you EFI Driver folder, you really don't need all of that in that folder. It's better to run just what you need. I probably have more than what I need.

Figure I should also share my config.plist. Hopefully it helps. My system runs very stable, but please notice I run iMac Pro 1,1 And I have my iGPU disabled in BIOS.
(I removed UUID, Board SN, and system SN)

 

[gandem]

Couple of suggestions:

• Try enabling CSM Support to see if it makes any difference.
• Thunderbolt monitors might work more reliably if the GC-Titan Ridge is not flashed (i.e. use original firmware). This should be attempted -- we should still bridge pins 3 and 5 to force power to the card if a THB_C header is not available.
• Also try one of the other firmware files (e.g. NVM 23), but I suspect they might not perform differently from NVM 33.

I'll try as suggested. I do have a THB_C header and keep it attached. I agree with your assessment that non patched will probably work- that has been my experience with monitors on this system. But I want an oompa loompa now!
Would you be willing to take a look at my efi and ioreg to make sure I'm not missing something?

 

[CaseySJ]

I'll try as suggested. I do have a THB_C header and keep it attached. I agree with your assessment that non patched will probably work- that has been my experience with monitors on this system. But I want an oompa loompa now!
Would you be willing to take a look at my efi and ioreg to make sure I'm not missing something?

I can certainly look at EFI and IOReg. However it's still good to try various alternatives to see if a clue emerges.

 

[svandive]

My PowerColor Red Dragon 5700 XT has not had any issues since the day it was installed (i.e. long time ago).

I think I'm going to pick one up, the power draw and heat from the dual Vega 64s is just excessive in my opinion...

 

[losinka]

For Alpine Ridge it's not strictly necessary to flash the firmware in order to activate Thunderbolt Bus. A properly configured SSDT-TbtOnPCH.aml can do the job for on-board controllers.

At first I need to look the DSDT for this SSDT. My DSDT has some other methods than by dgsga...

@dgsga: Can you please share your DSDT? Thank you!

 

[yfzmaniac]

At this point you might as well try connecting power to PEG as indicated in green. Normally we use that with two or more GPUs, but it doesn't hurt to try. Also ensure that the 12V_2x4 port in red is connected.

View attachment 468323
Also try disabling X.M.P. Profile 1 for DRAM.

Thanks so much for the incredibly fast reply. I plugged power into the "PEG" input and it seems to be booting on the first try more consistently now. It's not 100% but it's better. I'm going to keep tabs on it for a few days and then get back. A couple other thoughts if you have the time:

1. Strangely enough, if I have all 4 antennas installed, it has the most problems powering on. With only two it is fairly consistently powering on. This sounds crazy to me, but it is repeatable. I can't explain that but maybe it makes sense to someone else?
2. I also have windows installed on a separate 2.5" SSD. This couldn't be messing with it could it? All mac drives were removed during installation of windows and again it was booting flawlessly without T919.
3. I have an old wifi card (a TP-LINK I believe) from an old build that I tested out and even though it doesn't work right off the bat in macOS, it does let the machine start up without any issues. Again pointing me back to the T919.
4. Tough question to answer, but do you think it's safe to continue having the card in the computer?

Again, thank you so much for your time. This machine is an absolute dream otherwise!

 

[gandem]

I can certainly look at EFI and IOReg. However it's still good to try various alternatives to see if a clue emerges.

@CaseySJ I have just concluded a second round of testing. At the moment, I have both screens working, but it took some cable swapping to make it happen. Here's the latest:

- The only success I had booting and getting screens active was to boot with HDMI from GPU to UHD monitor. Then TB3 monitor activated. Then swap UHD monitor cable to DP, and it stayed active. So the 4MD4KLB-B is directly connected to GC-Titan Ridge.

I had some screen failures early when looping the TB3 monitor through the dock. Both screens remain dark, but I can remote into the computer with ARD and see the screens are active.

I'm attaching my EFI and IOREG from this boot with both monitors working ATM.

 

[CaseySJ]

Thanks so much for the incredibly fast reply. I plugged power into the "PEG" input and it seems to be booting on the first try more consistently now. It's not 100% but it's better. I'm going to keep tabs on it for a few days and then get back. A couple other thoughts if you have the time:

1. Strangely enough, if I have all 4 antennas installed, it has the most problems powering on. With only two it is fairly consistently powering on. This sounds crazy to me, but it is repeatable. I can't explain that but maybe it makes sense to someone else?
2. I also have windows installed on a separate 2.5" SSD. This couldn't be messing with it could it? All mac drives were removed during installation of windows and again it was booting flawlessly without T919.
3. I have an old wifi card (a TP-LINK I believe) from an old build that I tested out and even though it doesn't work right off the bat in macOS, it does let the machine start up without any issues. Again pointing me back to the T919.
4. Tough question to answer, but do you think it's safe to continue having the card in the computer?

Again, thank you so much for your time. This machine is an absolute dream otherwise!

We have seen a few rare cases of bad Fenvi T919s, but to get two bad cards in a row is unusual, but not impossible. Suggestions:

• Does the system boot reliably when no antennas are connected?
• Your Windows SATA SSD will not affect this. Which SATA port is it connected to? Are you aware of the SATA port limitations when M.2 SSDs are installed in M2M and/or M2P slots?
• Do you have another motherboard on which to install and test the Fenvi?
• I would also consider returning the card and buying a replacement from a different retailer. Perhaps the retailer got a bad batch of cards, hence switch to a different shop.