V.2 No 1 |
41 |
|
Mismatched
ladder filters |
||
|
||
The
same as in case of unloaded filter, at the transparency band we see n resonance
peaks (taking into account the nth shorted node). The conditions of
their appearance are determined now by the vanishing denominator of (16). The
filter input impedance written as sc for the filter with the shorted output
also essentially differs from the results obtained by the two-port method.
Its value will be determined now by the following system of expressions: at the pass band of the
filter, el
<1 |
||
|
(19) |
|
at the stop band, el >1 |
||
|
(20) |
|
and at the cutoff
frequency, el
=1 |
||
|
(21) |
|
We
see from (19) – (21) that at the pass band the input impedance will have (n
–1) resonance peaks. At the stop band the impedance falls with frequency
proportionally to 1/el+, and with the boundary frequency it is
equal in amplitude to the impedance only at large n. As
the results have been refined, the relation between the ladder filter natural
impedance and the filter input impedance (being important for calculations)
will essentially change. We can determine it by an analogy with electrical
transmission lines as follows (see, e.g., [4, p. 572]): |
||
. |
(22) |
|
Substituting
sequentially (13) – (15) and (19) – (21) into (22), we can see that in
general case the conventional equality (22) is not true. It speaks that for
ladder filters this method to find the natural impedance is unacceptable.
Although for the single filters the conventional computations retain true,
the same as the two-port method itself. The
reason of discrepancy of the results obtained with the help of DEMA and by
the two-port method is the same. The same as (16) – (18), solutions (9) –
(11) describe the standing wave – just the process which has to take place
when the load mismatched and in this connection the waves reflected from the
ends of the filter are produced. The two-port method does not take it into
account. And in the finite filters we also cannot obtain the solution on the
basis of a simple superposition of direct and reverse waves, since in these
filters the multiple reflections from both ends take place. So, in general
case, we cannot use for sure the idle run and shorting regimes to find the
filter natural impedance. We have to determine it on the basis of complete
analytical solutions – just so as we did it with the help of DEMA. 4. An arbitrarily loaded finite ladder filter In
ladder filters practical calculations the most important is the case when the
filter is loaded arbitrarily. As we saw in the introduction, generally the
two-port method cannot solve this problem, because its application requires,
the filter output impedance to be matched with the load. The assemblage of
the method DEMA with the method to obtain exact analytical solutions for
mechanical analogues enables us to find the solutions so necessary for the
practice. |
||
Contents / 35 /36 / 37 / 38 / 39 / 40 / 41 / 42 / 43 / 44 / 45 / 46 / 47 /