Re: NEC-LIST: Passive component values

At 11:30 AM 29/06/00 +0200, David de Haaij <david_at_eng.up.ac.za> wrote:
>
>I'm busy building a matching circuit for an antennas, and I came up
>with an answer that consisted of a parallel resonant LC circuit. I
>computed the values, but right now I'm having trouble implementing the
>values in microstrip lines. Can anyone tell me how to implement very
>small L and C values, or perhaps give me some reference.

The method I've used many times, with great success, is to model the
LC circuit on a circuit simulator using ideal discrete components. I
then replace one of the discrete elements with a microstrip, and
adjust its length (and impedance if necessary) to obtain the same
resonant frequency, filter shape or whatever, as with the discrete
component. I then repeat this for each other element in turn.

I've used this technique to design microstrip filters up to 7th order
elliptic, at frequencies up to 1.5 GHz and beyond. When compared to
the ideal discrete filters from which they are derived, their
frequency response is virtually identical, at least up to where the
lines start to become become re-entrant.

Remember that if discrete components are used in the final design,
they MUST be fully modelled (taking into account their physical
dimensions), otherwise the resonant frequencies will a long way
out. For example, an ideal 4.7pF capacitor will often have to be
reduced to 4.3 or 3.9 pF, because of the effect of its case size
(e.g. 0805). The physical interconnects must also be fully modelled as
transmission lines.

Peter Nesbit VK3APN
RF Design Engineer
Imark Communications Pty Ltd
Received on Mon Jul 03 2000 - 05:50:24 EDT