Estimation of integral (
) simplifies greatly if
we note that for we have two different characteristic
decay times for waves with small and large k:
The smallest value of k is 
, where N is the total number
of junctions and if it were possible to make the inductances l very
large: 
we would have had N independent
current sources each having characteristic time of l/r and without
any mutual interference. It is technologically infeasible, however, to
make the inductances of such a large value and we can safely assume
that there are plenty of states with 
. We then have
the following crude estimate for the integral
() ( 
):
For times 
first integral in () is exponentially
small and from second integral in () we have:
Comparing () and () we see that in case when a junction ``feels'' the same non-exponential influence as in
() from only 
neighboring junctions.
For a device operating at frequency f we can estimate the
``worst case'' change in the bias current induced by neighboring
junctions through biasing line as:
Here 
is the number of influencing junctions and
ft is the number of flops they made.
Apart from the accumulative changes in bias current happening over
long periods of time there is also an instantaneous change
happening immediately after the flop, at 
. Moment of time
t=0 is another case when integration in () can be done
analytically for all l,L. When we have:
From () it follows that around 
neighboring
junctions experience an increase in bias currents (note the sign in
()). Typically, some of them will have to process the same SFQ
pulse shortly after it passed through junction n=0, so it is natural to
require that this change 
is small:
This condition can be satisfied if we choose
and 
(so that ). Result () (for n>0)
is illustrated in Fig. .
A very unrealistic ``worst case'' scenario in this case of a short-term
positive bursts of bias current would be a simultaneous ( )
phase jump in neighboring junctions on both sides
(or even all junctions) so that the total variation would be
. If 
even
this impossible ``worst case'' would result in an acceptable variation
of the bias current.
Figure: Currents (in units of 
) through
junctions n=1..10
