ADuM1402:  Is there any specified or typical jitter performance?

Document created by analog-archivist Employee on Feb 23, 2016
Version 1Show Document
  • View in full screen mode

Is there any specified or typical jitter performance? We need to transfer a
clock signal (MCLK) to the AD7766-2 using the ADuM1402, and this clock should
have low jitter for optical SNR. Would the speed grade of the ADuM ('a', 'b' or
'c') affect the jitter performance?


We do not have a great deal of detailed information on Jitter in our iCoupler
devices. Jitter is rolled up in the PWD (Pulse Width Distorsion) number on the
data sheet.  Some of PWD is related to changes in the prop delay due to
temperature and supply voltages, and some of it is Jitter due to internal noise
effecting timing slightly.  Measurement  on ADuM1100 are 50pS jitter but
ADuM1100 has only one channel and so has no channel to channel noise.  The
ADuM1400 has much worse jitter and it is highly dependant on what the other
channels are doing, there could be as much as 1nS of jitter in a ADuM1400. If
you need a low jitter channel, we would recommend the ADuM1100.

The easiest way to transmit the clock across the isolation barrier is to use a
transformer - it then becomes a problem of selecting a transformer which
- sufficient isolation voltage
- sufficient mechanical clearance (creepage and clearance distances required in
your design)
- sufficient frequency response
The Analog-Digital Conversion guide  shows a suitable solution in section 9
page 9.40 figure 9.28.

Ideally you would generate the clock on the isolated side, if the
jitter-critical devices are located on the isolated side, and then transmit to
the non-isolated side.