Daisy Chaining MADI devices

[jaminem]

Hi there, QQ on MADI.

I currently have an RME UFX+ and a Ferrofish A32, if I wanted to add another MADI converter Is it as simple as the following:

Current: RME OUT > A32 IN, A32 OUT > RME IN
Additional converter: RME OUT > A32 IN, A32 OUT > New converter IN, New Converter OUT > RME IN

Guess I'll need single MADI cables rather than the paired one I use now?
Anything else to consider?

Elf you have a similar setup I think?

[Hugh Robjohns]

You need to make sure your new converter can be configured to fill the vacant channels in the MADI stream, so 33-64 if the Ferro is using 1-32.

Correct clocking between the converters is also essential, of course.

[The Elf]

Yep, you've got it.

I don't recall how I set the second converter up to be channels 33-64, though. I set it up years ago and forgot it was there. That's the joy of the A32s and an RME interface, I suppose!

I'm sure it's in the manual somewhere.

EDIT:
Hugh, I think the clocking takes care of itself, but again the memory is misted. I can see I have no additional clock cables between my interface and A32s, so it's probably just a parameter thing at most.

[Hugh Robjohns]

MADI carries embedded clocks, so depending on the system there may be no need for external word-clock cables. However, setting the clocking parameters for each converter is critical for the system to work.

[jaminem]

Nice one chaps thanks, I have a master clock is it'll be word clock off that, so not to worried about that bit.
One other Q, if I went for an A16 I assume I can daisy chain another one later if I wanted to add the final 16 channels?

[The Elf]

Sorry, can't help on the A16 - it's many years since I used one.

[Hugh Robjohns]

One other Q, if I went for an A16 I assume I can daisy chain another one later if I wanted to add the final 16 channels?

The A16 has been replaced with the Pulse 16. Up to four Pulse 16s can be chained to access all 64 MADI channels. If I recall correctly, you need to configure the unit with its position in the chain (1 to 4) as the MADI conversion imposes a 2 sample delay, so the unit needs to know how many units are in front of it so it can introduce the appropriate synchronisation delay.

[jaminem]

Lovely, yes I meant Pulse 16, so in my case the chained Pulse 16 would be device 2 and another would be device 3, thereby giving me access to all 64 channels.

Thank you gentlemen

[jaminem]

Ah actually just checked, this is not right. The A32 only allows delay compensation for 2 devices not 4, so I guess this means I can have 1 Pulse 16 as device 2, or 1 A32 as device 2.

I'll keep saving then....

[Matt Houghton]

I believe the Pulse 16 (like the A16 before it) also acts as an ADAT<>MADI converter, and an ADAT<>Analogue converter. So if you were to need further expansion, then I think you should be able to hook up an ADAT converter to your Pulse to give you your 64 channels of MADI<>Analogue conversion in total.

[Hugh Robjohns]

Ah actually just checked, this is not right. The A32 only allows delay compensation for 2 devices not 4...

Yes, because it handles 32 MADI channels, so there would only be two units in the chain.

I believe the A32 MADI latency is 3 samples... just to be awkward

I guess this means I can have 1 Pulse 16 as device 2, or 1 A32 as device 2.

I don't see why you couldn't configure the Pulse units as 3/4 in the chain. You'd just need to be aware of a one sample delay/offset between MADI channels 1-32 and 33-64 -- so don't put a stereo array across channels 32/33! Other than that, it should work fine.

But it might be worth dropping Ferrofish a note to ask their advice.

[jaminem]

Just to circle back round incase anyone does a search, I have spoken to Ferrofish - they say it can be done!