Hey iamyila, I can't help you but I have the same problem, also on a Mac mini 2018 with a Fireface 802. Did you get it to work yet?

Today's OSX update seems to have improved things, it only happened twice and not every 30 mins.

There are still dropouts, and some not stable behaviour continuing.

Is there any official support present on this page?

happened twice today. Improvement with OS update but no fix

except this seems to be focusing on MacBooks and battery

I have a Mac mini.

It's surprising there is no official support reply on this forum, I used RME for shows and studio setups for stability and now, well what now? Is anyone going to suggest anything?

Ok, thanks for the reply and explaining. I did go through the other thread, but I don't have location services turned on at all, system time is not synced, no wifi, no a battery and I was also on the latest OSX and drivers.

It always seems to come every 30 mins or so and my console shows that it happened again exactly 30mins after logging on, here's the log it went wrong at 11:46

https://www.dropbox.com/s/huwgy0deuin6d … 8.png?dl=0
https://www.dropbox.com/s/dy1fhl3sjbhgb … 0.png?dl=0
https://www.dropbox.com/s/huwgy0deuin6d … 8.png?dl=0

No major change in 10.14.2, from what I hear. So we'll have to wait. This is not something a simple driver update can fix completely.

Regards
Daniel Fuchs
RME

On our Mac Mini 2018 the problem is definitely the timed function, around every 30 minutes. We noticed that our former workaround is not always triggered here (it is triggered in Apple's own Class Compliant driver), and will extend the driver to also handle this error reliably. Which means that instead of lasting bitcrushing you get a short glitch (same as in Apple's Class Compliant driver). This glitch will not go away until Apple fixes it.

The updated driver is available here:

https://archiv.rme-audio.de/download/dr … ac_311.zip

Thanks for continuing the conversation RME.  Are you able to clarify whether you've submitted a developer's bug report to Apple about this timed function?

We've been editing and mixing numerous projects on tight deadlines for the last two weeks and the ADI-2 Pro glitches are so frustrating we may switch back to our UCX. But are you saying the problem will affect devices on the 3.11 driver as badly as class compliant mode?

The glitches did not change at all. Means you had them before but didn't notice them.

We're getting drop-outs once every 20min or so.

The UFX was much more stable when we used it for a recording session last week, which included playback.  Maybe I was lucky with the timing on each take?

Bingo, you are right. The current scheme will cause a reset no matter what error shows up (basically you can not differ between them all). The former driver did not react with reset on an error that causes the effective buffer size to get smaller. With this kind of error there is no audible click when the error happens, but at some point the buffer size gets so small that you end up in continued distortion. If the sample shift is small then you can continue working without noticing the error. If you do recording without software monitoring the biggest buffer size available will prevent you from running into problems caused from this specific error for a long time.

This leaves me with two comments:

- you could continue to use driver 3.10 if it suits your work and setup better (but you will have the bitcrushing sooner or later)

- although inaudible the mentioned error is an OS/hardware problem that could (once again) only be fixed by Apple

According to our research (still ongoing) Thunderbolt and everything behind it (PCI, PCIe adapter enclosures, or docking stations/dongles with USB derived from the TB part of USB-C) do not show these problems.

...thanks for the quick reply!

At least only using simple USB-C dongles does not help with Macbook Pro. USB-C is just the connector type for what's Thunderbolt 3 in new macs, meaning that it can transport data in different protocols, usb 3.1 (backward compatible to usb 2), but also PCI express, Displayport 1.2, network and power. To my understanding if you use USB-C to USB 2 adapter you just will use USB controller in your mac, and core of the problems is system tasks stressing it so that it can't handle audio properly. Anyway i guess docks that connect to mac via USB-C have their own USB controller in PCI bus (via thunderbolt), and according to MC these don't have the same issues as mac's internal usb controller. I may not be perfectly precise here technically...

Anyway great to know that maybe there's a way around it and also that research is ongoing, eve though best would of course be a proper fix from apple.

Here is some info on the current situation with Apple's Mac Mini 2018 and 10.14.2.

- USB 2 is not usable for audio on any port.*

- USB 3 is not usable for audio on the two USB-A sockets.*

- USB 3 works as expected on all USB-C ports using a USB-C to USB-A adapter**

If you use an USB-C to USB-A adapter and connect a USB 2 device it is internally connected to the same bus/hub where all the other USB 2 sockets are connected to - and thus won't work. You can gain basic knowledge about the internal structure and connections by using the Mac's System Report/System Information, USB. Connect your USB device and look where it is shown in the nested structure.

In clear words: there is currently no way to have a Mac Mini 2018 working without audio errors when using USB 2, neither directly nor with the standard USB-C to USB-A adapters.

But there is a workaround. Devices connected to Thunderbolt do not show audio errors as with the internal USB 2. So all you have to do is use 'external' USB. Two examples:

- A Thunderbolt 3  to PCIe enclosure (Sonnet, NetStore) with a simple USB PCIe card added. Of course that is kinda overkill unless you own such a system already and just have to add the USB card.

- A more simple Thunderbolt 3 'dongle' with a USB port. The one we tested also seems to be the currently cheapest one:

https://www.caldigit.com/thunderbolt-3/ … splayPort/

This device uses the PCIe bus of Thunderbolt and derives USB with its own USB chip, so is totally decoupled from the broken USB 2 in the Mac Mini 2018. Which is why it also doesn't show the infamous USB disconnect problem (USB hangs until reboot).

A word of advice: there are a zilllion cheap docks with USB, HDMI etc available, that often have 'Thunderbolt 3' in their name or description. These are NOT Thunderbolt 3 devices! To recognize real TB3 is quite easy:

- Thunderbolt 3 is mentioned on the package
- the USB-C connector has the Thunderbolt flash symbol on it
- the product description clearly mentions TB3 and 40 GB/s
- the display ports support 60 Hz, not only 30 Hz
- the price is above 100$

Here are some examples of NONE TB3 docks/adapters that won't fix USB audio:

https://www.amazon.de/sicotool-Thunderb … B07FCGFQJP

https://www.amazon.de/CableDeconn-Thund … B07B66LNRS

https://www.amazon.de/CableDeconn-Thund … B07BK4TRG8

https://www.amazon.de/UGREEN-Thunderbol … B077FZQS95

https://www.amazon.de/QacQoc-Thunderbol … B07KC5V3GR

This last one from QacQoc is presented as real TB3 solution, but on the second look it's a scam. The flash symbol stands for power (it's not the Intel flash). The first USB-C to USB-C connection is most probably a wired pass-through, violating the TB3 specification (but it's not TB certified anyway). The other ports are connected to the second USB-C and are therefore not derived from TB3.

Alternative to the CalDigit dock (not tested):

https://www.amazon.de/Elgato-Thunderbol … 07GS9XM54/

We are still continuing our tests.

* All USB 2, and USB3 in the two USB-A sockets, is derived from the Intel chipset,  Device ID 0xA36D, Cannon Lake PCH USB 3.1 xHCI Host Controller.

** USB 3 within USB-C is derived from the TB3 chip, Device ID 0x15EC, a Titan-Ridge Thunderbolt 3 Controller.

Thanks, so that answers one question, and yuck having to buy a dock/dongle! in the meantime can you answer these to which working solution is better:

3.11. Will have dropouts, but will these problems affect recording and will audio quality change over time, is something happening to the sample rate? Or is the driver re-setting itself? - Can this problem be ignored and continue to work without something happening deeper in the audio?

3.0 major but crushing fail sound, reset sample rate of system, back to normal until next crash, but are they any effects on playback of audio / recording until the failure?

I have that dock and it works for me. But for me also the double dongles that don't work for you do work. So there might be something different going on with your FireWire problem.

I just connected my UFX II today, in the exact same setup as the UC it replaced. Playback works fine, but recording doesn't (recorded audio gets bitcrush distortion). Any idea why this might affect recording only? I have updated to the latest drivers.

The interface is running on a USB2-hub (same as the UC was)

Thanks MC for such thorough research.  This explains why our MADIface Pro (USB 2) was causing problems and the MADIface XT (USB 3) was fine, even for larger channel counts.

It would help if you amend the official statement to include 2018 MacBook Pros, and I think it's still important we hound Apple with Console logs etc so they identify and fix the shortcomings sooner than later.

Will the Thunderbolt 3 solution fix glitches on the ADI-2 Pro and other class compliant devices too?

Nobdy said that the driver workarounds remove audio errors. They just make them less annoying.

I am using a MacMini 2018 with an older Thunderbolt Dock by Akitio (connected via Apple's Thunderbolt 3 to Thunderbolt 2 Adapter and a 5m Thunderbolt Cable by Corning). I did not experience any distortion or dropouts so far with my Digiface USB (Driver v3.10) plugged into one of the USB 3 ports of the Dock.

Yes, that's the setup I am working with (except that I have a MBP 15", 2018). I still have audio dropouts due to CPU spikes. But at least it is functional. When I connected my FF 802 via USB it was impossible to work or stream any audio with it.

You have to look at the CPU meter directly in Logic. It occurs at low buffer size. But maybe that is a Macbook specific issue.
You can watch this video:
https://www.youtube.com/watch?v=ZXGyq6EsECk
This guy reproduces the bug on a MacBook Pro 2018. I don't know if this is the same on the Mac mini.