Proper Interface Clocking

[EBikales]

Hello all,
I have a little mystery going on in my studio as to the proper method for clocking all my interfaces---and a Presonus Digimax (48k). I have:
(1) 24o8 mk.III which has a word-in jack, and a word-out jack. I have (2) 24 i/o's. Each has a word clock dual in/out jack. And a Digimax that has a word-in and word-out jack.
I want to be able to use the ADAT out of the Presonus Digimax. Right now it's word clock is not hooked up, and I'm getting lots of clicks. I also noticed that only one of my 24 i/o's is clocked to the 2408, which serves as my master clock. The other one, which most of my synth modules go into, isn't even connected for word clock. (don't ask me why?)
So, does all of it need to be connected together, and if so, can I just daisy-chain the 2 24 i/o's with the Presonus?
Thanks!

[XYZ]

Your best bet is to buy a clock and then hook them all up in a star configuration to that clock, so everything is slaved to that clock. But that is not necessary.

What you can do if you don't want to spend big bucks on a pristine clock, is determine which of your units has the best clock. Then hook that word clock output to all the other's word clock input. All will then be synchronized, and all will work well.

So to be clear, buy 3 BNC "T" connectors, 2 BNC terminators [75 ohms], and 2 BNC cables. Now hook one T up to one of your 2408 clock output, the second "T" to your second 2408, and the third "T" to your presonus. On the master 2408, hook up one terminator and one BNC cable, the other end of this cable hook to the T on the second slave 2408. Then connect the other side of the slave 2408 to the BNC cable, which will then go to the T on the presonus. Then put the second terminator on the one last open connector T on the slave presonus.

You probably don't need the terminators, but if you do, this might not work, and they are cheap, so why not just avoid impedance problems in the first place. If the presonus is internally terminated, you don't need the T and terminator for it, just hook the BNC right to it.

[EBikales]

Hi XYZ,
Thanks for your advice. I always wonder if an expensive clock might improve my audio, but I'm not really in a position to buy one now anyway. I'll collect all the parts and hook it up the way you suggest.
Thanks!

[XYZ]

The specs on the black lion are pretty good, and it is ~ 500 bucks, if you wanted to upgrade. My guess is, if you are going to need pristine sound, a big ben might be the way to go so you won't ever buy another one. But first let's get rid of pops and clicks, and synchronizing all your boxes should do that. Make sure the driver is set to "word clock in."

[ajdelange]

So to be clear, buy 3 BNC "T" connectors, 2 BNC terminators [75 ohms], and 2 BNC cables. Now hook one T up to one of your 2408 clock output, the second "T" to your second 2408, and the third "T" to your presonus. On the master 2408, hook up one terminator and one BNC cable, the other end of this cable hook to the T on the second slave 2408. Then connect the other side of the slave 2408 to the BNC cable, which will then go to the T on the presonus. Then put the second terminator on the one last open connector T on the slave presonus.

This may or may not work but it is not good practice. Assuming that the outputs and inputs are all 75 Ohms then the combination of the unit and terminator on the unit at the end of the chain would present 37.5 ohms to the 75 Ohm cable. The impedance at the other end of that cable will depend on the frequency (of each of the harmonics of the clock) but will not be 75 Ohms resistive but more like 37.5 ohms at the fundamental at least. Thus you will have distortion of the clocking waveform due to reflections with unpredictable results. The situation would be the similar looking back from the second unit in the chain to the master. The second unit will present 75 Ohms in parallel with whatever the transformed impedance at the end of the cable to the third unit happens to be which will be about 37.5 ohms at the clock fundamental at least and so the impedance looking into the cable at the master end will be something like 25 Ohms (at the fundamental) which is the load the generator will see. If a 75 Ohm terminator is added here the load on the generator is about 19 Ohms. Assuming again at this point a 75 Ohm source impedance only 20% of the source voltage will be applied to the cable.

OTOH if the driver (clock out) is a 75 ohm or some other low impedance source and the clock inputs on the other units are high impedance then do not use a T or terminator at the generator (master). Run cable to a T on the second unit and another cable to a T with terminator at the last unit. This will present 75 ohms to the generator (master).

Of course I have never seen a spec on input or output impedance on any MOTU piece of equipment so I have no idea as to what they actually do in this regard. As the stuff is used professionally I would expect 75 ohm input and output impedances because preservation of clock shape requires impedance matching between source, load and cable. The proper way to configure such a system is with a clock distribution amplifier. Another method which is not quite as good but which should give adequate results with only two units is to daisy chain master to first slave's clock in, first slaves clock out to second slave's clock in. If the PLL's are decent in both slaves units you shouldn't pick up much additional jitter in this way. MOTU even suggests (while recommending a distro amp) that this scheme is probably OK.

[XYZ]

This would only be the case if the MOTU clock impedance was 75 ohms. I don't remember where I determined this, but the MOTU does not terminate their clock interface with a load--that's why the terminator is a good idea.

Some interfaces have a switch where you can add the load internally , and some can auto sense proper attenuation and tell you [Apogee does this]. But not MOTU.

Also, it will work.

[EBikales]

Thanks ajdelange,
I appreciate your explanation, although I have no idea what the real impedances are, and I suppose I could just ask MOTU what they recommend. I found 2 T connectors here in my studio, so I'll just go buy a couple more coax cables and hook it all together without the terminators, for now. We'll see if that works! Buying a nice quality clock, or a dist. amp. is not an option at the moment. Anyway, thanks for the advice and explanation.
Cheers

[XYZ]

Use only 1 T from your source clock if you don't have terminators. Don't leave an open, unconnected connector on your plugs.

[EBikales]

Thanks XYZ,
My 2408 mkIII has an input and output connector. The 2 24 i/o's only have 1 jack each (an in/out) But the Presonus Digimax has an in and an out, so I guess I could put the T on the output of the 2408, which is to be the master clock, and connect to 1 24 i/o, and the Presonus. Then another cable from the out of the Presonus to the 2nd 24 i/o. That would get them all connected without the need for terminators.
I still need to get a couple more cables tomorrow.

[XYZ]

I'm guessing that the driver software allows you to switch the 24I/O from a slave to a master. To do what you want, you want it to be a slave. If the presonus has an option for internal termination, then hook to that last.

[therealbigd]

the MOTU manual reckons you can run 2 devices using the internal clock. it reckons once you get onto 3 you should run off an external clock.

as somebody said above the BLA clock is pretty good. as is the Aadvark Aadclock. and then up the top end you get things like the Apogee big ben.

though Prism sound think things work better if you clock from the weakest device...

[EBikales]

Thanks therealbigd,
That saves me a call to MOTU. (a considerable investment in time!) I think I'll go ahead and get the necessary cables to hook it all up, just to see if I get lucky with all 4 things connected. Since buying an external clock will have to wait, my other option would be to keep the 3 interfaces hooked up, and simply use the direct outs of the Presonus, like I've been doing up till now. Thanks for the information though. I'll post my results soon.
Thanks.

[XYZ]

Also, maybe you can hook some of them up with ADAT lightpipe [and some with word clock], [if you have that option]. You can synch using adat, although jitter can be increased [adat clock recovery is not as precise as a dedicated word clock].

[EBikales]

Thanks XYZ,
The 2408 has ADAT, but the 24 i/os don't, so I could run ADAT only to the Presonus. But that's the original problem: the Presonus apparently needs to see word clock along with the ADAT, so it doesn't click and pop. I'm going to report back just as soon as I hook this all up a different way.
Thanks again!

[XYZ]

OK, it may be that the presonus can't be re-configured to respond to ADAT clocking--maybe it REQUIRES word clock. If the MOTU doesn't have enough push to drive the other three, you could also go with a distribution amplifier, I think Lucid has one for a couple hundred bucks. You might be able to find it cheaper.