Cascaded UAD Apollo system recording delay offset notes

I posted this on the UAD forum but I think it's worth adding here as it could help out this community also.

I just finished a round of testing the recording delay offset (i.e. the amount that the Apollo misreports to LPX, resulting in recorded audio being placed inaccurately in the Tracks area) in my cascaded UAD Apollo system (16 mk II, 8, 8p) with LPX, and in most cases, delay was identical on all 3 converters (this addressed another scenario in which a user observed a different recording delay between different models in a cascaded system.)

I say "most cases", because on 1 occasion with a 1024 buffer, after I made a recording delay offset adjustment in prefs, the 8 and 8p were 53 TICKS (not samples) later than the 16. After adjusting the buffer and toggling core audio on and off, the interfaces started being exactly in sync again.

I used both the i/o plugin ping (as detailed in the video below) and a manual loopback test. The results agreed exactly.

On my system, at all buffer sizes, the recording delay was 31 samples at 44.1K and 48K, and 3 samples at 88.2K and 96K. So, after determining this number, I entered a recording delay of -31 (at 44.1K), repeated the loopback, and the single sample tick lined up sample-accurately.

Of note: Sometimes after switching the sample rate or buffer size, the i/o plugin would get squirrelly and report HUGE delays. Usually toggling core audio would fix the problem.