Clocks, AD Converters, Jitter as a consideration for Interfaces

ibleedburgundy

ibleedburgundy

The Anti-Lambo
I was reading an article in SoS about clocks and AD Converters. The jist of the article was that they ran a bunch of tests and under no circumstances did using an external clock improve performance (contrary to what is typically advertised). Higher quality gear was merely better at being the slave unit. This resulted in lower jitter than the cheaper gear, but in in every case it was still a degradation from using the unit's internal clock.

So if you're a guy like me who records 16 tracks simultaneously on a regular basis, and it wouldn't take much adding to have 16 decent preamps on hand, wouldn't you be better off buying an interface with 16+ AD converters such as the apogee symphony, Antelope Orion, or Lynx Aurora 16 rather than getting a UA Apollo/Digi 003/RME Fireface/Prism Orpheus etc and then matching it up via ADAT with some other clock such as an Audient ASP 008 or Alesis ADAT XT or whatever?
 
In theory, yes. The internal clock of a ADC can't really be improved upon with an external clock, and jitter that gets in at that point can't be undone.
 
If your are connecting/synchronizing multiple and/or different digital devices, then an external clock to "rule them all" helps.....just driving one piece with an external clock is pointless.

I run a Lucid SSG 192....but it's the main clock/house sync for 3 converters, a Timeline Microlynx sync/controller that in turn clocks my 2" tape deck...which initially got it's SMPTE stripe from the Micro while locked to the SSG via video blackburst.

So it's several pieces and the SSG is the master clock/house sync.
 
I now use a lynx aurora 16 (and think it's great). Previously, I combined other A/D units and found that I had to use an external clock - slaving one to another didn't work well - lots of clicks and pops. Maybe the signal strength on the wire was too low? I'm speculating. (I have not yet tried slaving other units to the Lynx, btw).

Individually, I didn't notice a difference between using the internal and external clock on any of the units.
 
Depends on how they did their test. using bnc t's in a trunk fashion and termenating the end will be as good as how much drive the clock source has and how well the cables are made. on external clocks they should have an output for each device so that it is driven correctly and be isolated from the other converters (via descrete buffering and opto and data transformers).

now if the computer has a stable clock, you can use the clock out on pci-pcie host cards. also using other connections such as madi or red-net for high density recording is much more of a practice. and syncing to the card without a clock depends on how many devices are hooked to the chain. if the connection is 1 to 1 there won't be a change (in the frame before buffering exiting the last device) so word clocking sometimes has no benefits in those setups.

usb and firwire are natorious for making clock errors happen (because of lack of electrical isolation, the noise from the computer is imposed on the power and grounds) latency buffers have to be increased so that these interfaces can run without errors and dropouts.

sync over ADAT is only as good as the fiber connection and if the master is the converter, will only be reliable if there is no clipping on the ADAT i/o (wich causes too many high states (instances of 1's too close to the end of the digital frame) that throw the clock off on the bit stream).
 
You must log in or register to reply here.
Back
Top