Some particular observations and improvements in reference to them. #142
Description
I leave here a document made by me specifically in my case with NUC10i7FNH updated to Sequoia 15.3.2 (24D81) concerning some observations and improvements that have made my system (almost) completely perfect based on the efi provided by this repository.
*1 Symptoms:
Bluetooth is sometimes not recognized.
Fix:
Disable the xhciportlimit quirk and use a usb mapping that respects the 15 ports including bluetooth (as internal). You can try to use the one I provide here for NUC10i7FNH.
Possible Complications:
This map does ### NOT ### include internal headers. So if you use them for some particular reason this map is not good for you.
UTBMap.kext.zip
Just replace the kext with the one already present in EFI / OC / KEXT
I also recommend using the NVMeFix.kext if you are using a non-natively supported NVMe drive.
*2 Symptoms:
The nuc wakes up as soon as it enters sleep in reference to this and this .
Fix:
To solve this problem once and for all I chose a very traditional solution that prevents the device from waking up due to the USB ports; I had this problem because of my Apple bluetooth keyboard that caused continuous waking up problems, probably due to incompatibility between the bluetooth kext and the native Sleep/Wake functions.
You can use the provided patch and integrate it by opening the config.plist with OpenCore Configurator in the ACPI -> Patch sections right click and click "Import set of patches" select GPRW_UPRW-Renames.plist then in the patch list disable everything and activate only Change GPRW to XPRW
GPRW_UPRW-Renames.plist.zip
Possible Complications:
NUC WILL NOT WAKE UP FROM USB DEVICES. Example: mouse clicks, keyboard etc.
*3 Symptoms:
The NUC fails to shut down completely with the fan on and the power LED light but no sign of activity forcing a forced shutdown.
Observations:
I have not had this problem with the NUC10i5FNH (until "ventura" then abandoned) but I started having it with the NUC10i7FNH from Sequoia.
It seems to be caused by the thunderbolt port used as USB4 by an IOCREST adapter with ACQ113 chip for a 10gbps fiber connection (unconfirmed cause), so to solve this problem I referred to this .
I DO NOT USE AN EGPU, I use the thunderbolt port with an IOCREST USB4 adapter compatible with Thunderbolt 3/4 with which I can use all the 10gbps bandwidth currently provided by my ISP.
Fix:
So you can integrate in ACPI -> Patch the patch FixShutdown-Patch.plist and then enable it and then add the attached files in EFI/OC/ACPI and you have to remember to add them also to the config.plist file using OpenCore Configurator and then of course enable SW-Thunderbolt.aml and SW-USB.aml
REMEMBER TO ALSO ENABLE _PTS to ZPTS Patch as shown a couple of screenshots above.
*4 Undervolt:
In reference to this .
General observations:
The bios is read-only from version 0055 (excluded) onwards. So to make the related changes you can downgrade using the bios present in this same repository.
Note that downgrading is a procedure that should not happen so complications may not be excluded, so if you do not know what you are doing stop immediately. This is only a guide for informational purposes where you assume personal responsibility for what you are doing.
Guide:
Flash the bios 0055 and in the same USB formatted in FAT32 and put the RU.efi file attached here.
RU.efi.zip
(An alternative version of the tool provided by James Wang)
The root of the USB must be EFI/boot/bootx64.efi
Reboot the NUC and boot into the USB , In reference to this
we will focus on this point:
Once it's booted you will be presented with an ancient looking interface. Hit Esc to exit the info window then click on "Alt" + "=". You should see a menu of options. Select "CpuSetup". You should be presented with a grid of hex values with 0x00 to 0x0F along the top for the columns and 0x00 to 0xF0 along the left for the rows. Navigate to the location of the bit. Mine was 0xEB so i went to the 0xE0 row then the 0x0B column. there should be a 01 there. Change that to 00 and click "ctrl + w" to save then "alt+q" to exit.
In our case the "interesting" values are 0xDA for the Over/Underclock and 0x3E for the CFG unlock So knowing this and using the guide mentioned above we will modify these values accordingly
Cfg variable present in this bios
Undervolt lock variable present in this bios (it's the same as before).
Reboot your system and put the VoltageShift.kext in EFI/OC/Kext and move the voltageshift file to your Home directory ( chmod +x voltageshift ), reboot your system once again and run the voltageshift script from terminal via ./voltageshift info you should be presented with a screen similar to this
zearp@nuc ~ % ./voltageshift info
VoltageShift Info Tool
CPU voltage offset: 0mv
GPU voltage offset: 0mv
CPU Cache voltage offset: 0mv
System Agency offset: 0mv
Analogy I/O: 0mv
OC mailbox cmd failed
Digital I/O: 0mv
CPU BaseFreq: 2300, CPU MaxFreq(1/2/4): 3800/3800/3600 (mhz)
PL1: 30W PL2: 64W
CPU Freq: 0.8ghz, Voltage: 0.6144v, Power:pkg 3.53w /core 0.80w,Temp: 94c
if here
PL1: 30W PL2: 64W
is written
OC_LOCKED PL1: 30W PL2: 64W
the unlocking was not performed correctly.
Once unlocked you can have fun looking for the best combinations to get the best performance and the least thermal/energy waste.
I recommend carefully following the rest of the guide .
Note that even native tools like Volta work correctly (In Sequoia), however in my case the undervolt via this tool does not exceed -120 so I fully used voltageshift also to control this.
Managing to obtain in my specific case
Note that the PL1 and PL2 values can be changed from the BIOS but it is not recommended to do so if you are using the stock cooling system.
Of course I didn't write it every time but every time you change the config.plist file you need to reset the Nvram. (Patches added or removed, kexts added or removed etc)
For the rest, thanks for your splendid work , Goodbye 🌋
P.s. I'm not very familiar with GitHub and this might not be an "issue" so forgive me.