Carbon S/pdif input distorted?

[lilith_envy]

Hi,
Running a Joe Meek Twin Q via SPDIF light pipe. I have the SPDIF selected on ADAT 2. (ADAT 1 = Focusrite Scarlett Octo Pre, which sounds great)
The audio coming in on the Joe Meek is very crunchy. Not aud distorted or clipped waveform but very digital.
Do I need to link the Joe Meek via bnc word clock?
I can't seem to find any volume input control on the ADAT in?
It's strange because the ADAT 1 sounds really good.
Mini Mac 2 - Carbon - Joe Meek

[dominicperry]

With digital clocking, only one device can be master. Everything else must receive clock. Those devices can receive clock using bnc wordclock or via ADAT.

If you have ADAT running on one direction only, the device that is sending MUST be the master. This is obvious if you think about it - ADAT out is not receiving any data, just sending.

Anything which is ADAT out only will need to be the master clock unless it has word clock. This is also the device which will set the sample rate.

In your case, you have two devices (Focusrite and JoeMeek) which are ADAT out only and they are effectively both trying to be the master clock. Luckily, they both have wordclock inputs, so you should instead make the Carbon the Master, and sync both Focusrite and JoeMeek to the Carbon using wordclock.

Theoretically, you can make any one of those devices the Master as long as they are all clocked together, but given that you will have to use wordclock anyway, you might as well make the Carbon the Master. It will have the best clock, and it means you can set the sample rate in PT, the Carbon will then communicate that to the other two devices via wordclock.

So. Carbon Master.
Wordclock out on the Carbon to Wordclock in on the other two.
You'll have to look up whether you need to do a hub and spoke or a chain.

On the last point - no ADAT volume control. It's a full scale digital transfer, so there should be no need. RME provide ADAT volumes, but they're extra-clever. It's quite usual for there to be no volume control - it's not like an analog connection where you can overload the input because of mismatched levels or impedance problems. It's not the issue here - that's a clock problem as discussed.

Dominic

[dominicperry]

OK, someone is going to pop up here and say you shouldn't daisychain OR split bnc wordclock and you need a wordclock distribution box.
Across short distances, splitting bnc wordclock is often fine. Daisy chain and split are effectively the same, and create the same theoretical problem.

So, one possible solution is to leave the Focusrite as Master, leave the ADAT out on the focusrite to clock the Carbon, and use the wordclock out of the Focusrite to the wordclock in on the JoeMeek to clock the JoeMeek.

The advantage of this is that you only need one more cable. The disadvantage is that you now need to set the sample rate on the Focusrite and let everything else follow. (But on the plus side, it won't upset the wordclock purists).

Dominic

[JFreak]

SPDIF carries clock, it is embedded into the audio signal just like AES.

There can only be one Lord of the Rings, only one who can bend them to will, and he does not share power. So be mindful of which direction the signal flows, because that is also the direction where clocking flows.

[noah330]

I had a bunch of clocking issues when attempting to use my Carbon and SSL AlphaLink,

I came from an Omni and had been using BNC. Removing the WC cable solved the problem. Carbon was master on both.

[JFreak]

I never had omni and do not have Carbon either, but this... "removing WC solved the problem" makes me think if there is a problem with WC termination. This may be a RTFM thing

[dominicperry]

Originally Posted by noah330

I had a bunch of clocking issues when attempting to use my Carbon and SSL AlphaLink,

I came from an Omni and had been using BNC. Removing the WC cable solved the problem. Carbon was master on both.

Not really relevant here. The Alphalink can send and receive clock on MADI/AES /ADAT whereas, in this case, the Focusrite and JoeMeek cannot receive clock via ADAT (neither have ADAT IN) and so they must receive clock via BNC, or be master.

In your case you probably had two lots of clocking - one via AES(S/DIF) or ADAT, and the other by bnc word clock. Removing WC meant there was only one clocking mechanism left.