From: "Jeff" on Tues, Oct 11 2005 2:47 pm


Im confused regarding the choice of iron powder toroid for a 7Mhz high pass
filter. I need to use a small size (T68) toroid. For the freq range, some
reference texts say T68-2 is the one, others say T68-6" grade. I see
differing "best Q range" specs for the 6
grade.....2-30Mhz......10-50Mhz...etc.
Have similar need for 20Mhz LPF, but assume "6" is really the choice there.
Any suggestions as to the relative differences and reasons to choose between
grades 2 and 6 for the HPF in T68 size?

Been there, done that. In the practical case, what you can expect
is a very slight change in insertion loss in the passband, perhaps
a less sharp transition of attenuation at cutoff, depending on the
type of highpass. It can be modeled in any SPICE analysis program
with accuracy if you make a special model that adds a series R
element computed from reactance divided by Q.

For a highpass filter, the usual filter type has the inductor in
shunt. In that configuration the inductor has a reactance
directly proportional to frequency and will have little effect on
the insertion loss in the passband. The equivalent series R due to
Q won't matter much at 1/3 to 1/4 the cutoff frequency. It might
matter on attenuation in the stopband region but the analysis on
that is more difficult and a practical build-and-measure is the
quicker way to go.

For a lowpass filter, the usual configuration has inductors in
series and there the Q of the inductors will affect insertion loss
more. Those would use the lower frequency range for powder mix.

In a highpass filter it is important that the capacitors have a
minimum series inductance so that the 5x to 10x cutoff frequency
isn't disturbed. Since capacitors are usually in series, their
Q will effect the insertion loss. Fortunately, most capacitors
will have a good Q up around 500+ and won't be a factor.

Use what you have and measure the results (you've got a spectrum
analyzer handy so that's taken care of).

I made a fairly good, practical Synthesis-Analysis program for
L-C filters that includes automatic modeling of (separately) all
capacitor Qs and inductor Qs. I can attach that to private mail
if you want it. Freeware. Proven by practical test comparison.