Hi
For live shows I've put together a system built around 2 Macbook Pros, a master and redundant safety.
Each mac runs about 140 virtual synths and drums within Logic Pro 9 triggered by a master keyboard and Roland Td20 kit.
There's a little bit of backing too.(The Logic enviroment is a thing to behold built by our fab keyboard player with a little help from me.)
The hub of the system lives in a small rack containing a Motu Midi Time Piece Av for midi routing and sync duties.
A Maudio Profire 610 audio interface for the Master Laptop. A Maudio 410 for the Safety Laptop.
And finally a Radial Pro D8 passive DI to connect to the outside world.
The great thing about the radial is it's redundant inputs and although I lose about 7db of level through passive mixing, the sound and signal to noise remain excellent.

The Master Mac is setup to send MTC to the MTP which is set to Jam sync infinitely.
The Safety Mac externally syncs to this newly generated timecode.
An MMC stop command from the master thankfully halts the MTPs Infinite jam sync at the end of each song allowing it to sync a fresh to the next number.
Now to acheive almost seemless take over of the master by the safety in the event of a crash I send an audio control signal from the master into the safety interfaces input.
Over all 7 Outputs (Backtrack,Drums,Synths,Drummers click) on the Safety mac I placed an instantiation of Softube excellent Valley Dynamite plugin working as duckers and keyed from the Masters audio control signal.
If the master goes down, the control signal stops and opens the duckers on the safety which keeps playing in sync from the MTP Jam synced MTC.
Frustratingly (one for Desmond here?) it appears Logic Pro 9 Mutes any input at the moment the safety picks up sync from the MTP. This causes the Duckers to open for a fraction of a second and for both the master and safety to be heard FOH.
To get around this I created an additional audio control track this time residing on the safety.This one is routed internally to Noise Gates over all 7 safety outputs and also delayed by a 16th note such that when the safety picks up, although the inputs are lost and the duckers open, the delayed safety control track doesn't open the gates until the inputs are back online and the Master control signal can shut the duckers.
Is this making any sense ?

I used to get crosstalk of the master control signal on main outputs but removed this with the aid of an isolating 1:1 transformer between the master 610 output and the safety 410 input.

Anyway the whole thing works really well except for one aspect and here I must confess to being electronically challenged.
Both audio interfaces take their power from the firewire ports of their respective macs and I assume their earth too.
The Radial is Passive.
When one of the macs is not powered I regain the 7db loss of gain through the radials passive mixer.
This I think I understand as the radial is a voltage divider.
However what baffles me is that in this situation I get crosstalk across all odd and all even outputs of the Radial independantly (Could this be crosstalk within the audio interface?).This is particularly annoying as the drummers click is then potentially heard at FOH.
Now you might ask, what are the chances of one of the Macs crashing in a way that all power is lost including the Battery power and thus power to its audio interface?
Still it would be excellent if I understood why this is happening and if there is a workaround so the show would hopefully never have to be interupted for technical reasons.
I tried powering both interfaces with their own power supplies incase this is some earth related issue.
Unfortunately regardless of having independant power, they both appear to require power from the mac firewire busses to fully function.

The Profire 610 has balanced jack outputs and the Firewire 410 has unbalanced outputs. The Radial has unbalanced jack inputs, main and redundant. I use unbalanced TS jacks for all these connections.

Any advice and an explanation on this mystery would be marvelously received

Cheers
Eddy

The output stage of the audio interfaces is probably something like an opamp into a low value series resistor, which presents a low impedance as long as the feedback around the opamp is functional (so as long as the opamp is powered), remove this power and the source impedance will rise, which causes your gain change.

Worse this may be asymmetric and may cause weird issues with the common ground on the unbalanced interface (which may actually be floating providing a common 'mix' into all the inputs when the power goes down.

There are two obvious fixes: A simple two input active line mixer, maybe something from RDL? Or shunt the output of the audio interfaces with a comparatively low value resistor (maybe 1K would do), this will prevent the unpowered output rising to a high impedance level.

Me, I would probably go the active mixer route, but solder does not scare me and half a dozen or so 5532 in a box seems like a reasonable fix.

HTH.

73 Dan.

--------------------
Audiophiles use phono leads because they are unbalanced people!

can't offer any help, way beyond me, but I can't help but wonder what kind of music needs 280 virtual syths?

Hi fletcher,
Thankfully 280 don't run at the same time
It's about 5-10 per song on the master, duplicated on the safety, with all others bypass automated so the macs can handle the load.
And generally a set might be 14 songs from an available larger selection so maybe 80-100 depending on set length,song choice.

Blimey, and to think that when I gigged in the 80s I used to have enough trouble remembering which presets to use on my Juno 60 and what settings were needed for the Roland string machine......probably would have been easier if I'd taken the ray-bans off

Cheers Dmills,
Is it possible that replacing the unbalanced 410 with another balanced profire 610 could help even though I'm using TS jacks?
Or are there superior interfaces out there that would not suffer in their design from this?
It does seem frustrating to have to get an active line mixer when I'd hoped the Radial Pro d8 was designed for this job even though it is passive.
I don't mind the gain change if in the extremely unlikey event we lost power to one interface as our keyboard player has an emergency override for that .
It is the click crosstalk that would halt the show whilst I disconnected at least the redundant safety click.

Hi Pink Fluid
Our drummer's in charge of remembering which patch change !
Changing a patch(song title) on the Td20 automates changing song on the macs.
So thankfully not a lot to remember

I think Dan has already answered the level loss question, but to be honest, clever though your solution is, I wouldn't do it that way myself. Passively mixing two inputs with one set held off by software gates is always going to present problems one way or another...

I appreciate the additional cost of changing direction could be an issue, but maybe you could recoup some dosh by selling the Radial box. However, if I were doing this I would be looking for an active DI with relay switching unit incorporating some form of auto-sensing... and I happen to know that such a unit is available from Orchid Electronics:

http://www.orchid-electronics.co.uk/Switching_DI.htm

This unit employs the same excellent active DI circuitry of the other Orchid boxes, but with main/backup channel switching for eight outputs controlled manually from a front panel switch, or from a footswitch, or -- importantly for you -- when a reference tone goes missing from the master system.

So instead of all that faffing about with software gates and worrying about unloaded level changes, all you need to do is place a constant tone on a spare output of the master machine and feed that to the control input of the Orchid box. If the master stops for any reason the tone stops and the relays switch over. Job done.

hugh

--------------------
Technical Editor, Sound On Sound

Thanks Hugh.
Yes indeed a lot of faffing about was had
I agree that your proposal is definetely the way to go if you can and indeed Radial offer a Switcher with active relays too although the audio path is passive and it's quite a lot more than the pro d8.
The system really came about from bits of gear we already owned and trying to think how to utilise them for seemless redundancy. Did lead to a lot of problem solving which I'm sure the Orchid would remove.

Just seen the price of the Orchid... Very reasonable

Yes, the Orchid gear is extremely competitively priced, and it is all very well designed and built. I've tested quite a few Orchid units now, and I've seen (but not used) the box in question -- and I remain very impressed. I believe the switching box was commissioned by a big PA company and has been well proven in real applications.

hugh

--------------------
Technical Editor, Sound On Sound

Good to know, I was starting to feel in-adequate, I'm alergic to midi as it is.

I know of some big name acts using the Radial Switching DI for backing track failover.

The orchard thing looks good too - ta.