[amsat-bb] Retuning surplus 400 MHz Tubular Bandpass Filters in to 435 MHz
domenico.i8cvs at tin.it
Thu Oct 12 17:08:56 PDT 2006
I got a surplus K&L MICROWAVE ,INC Tubular Bandpass Filters
model 3B340-400/16-N75-SMA50 D.C.-8040 of the type described
in the following URL
The above filter was originally integrated in to a vertical dipole with a
low noise preamplifier at 400 MHz and used for the Maritime Satellite
MARISAT system between shore and shipboard terminals and the satellite.
Many of this filters are still available in the surplus market
3 numbar of sections
B bandpass filter
340 outside diameter 0.75" - 19.05 mm
400 center frequency (MHz)
16 3 dB bandwidth (16 MHz)
N75 input connector (see note)
SMA50 output connector
D.C.8117 D.C. block
Note:The input of the filter is fitted with a 75 ohm N female connector
to be screwed directly at the dipole ( 75 ohm impedance)
Just for curiosity I decided to unscrew both caps with connectors and pull
off the filter from the tube in the hope to understand how a Tubular Filter
works and I realized that this 3 sections Chebychev response filter is made
by a teflon flexible rod in wich are slipped on in sequence
The end coils are 2 turns of enameled nr 18 copper wire while the center
coil is 3 turns
Each concentric capacitor between coils is made with two brass disc as
amature with inside a numbar of disc of low loss teflon foils for
The above filter is wrapped in to a single layer of low loss teflon foil and
than it is pushed inside the aluminum pipe.
The function of the teflon foil is not trivial because it represent the
dielectric to form many small capacitors between each coil and each
series capacitor with respect to ground as required by a bandpass
The aluminum pipe has 3 longitudinal holes each 5.5 mm in diameter and
the holes are drilled just in correspondence to the middle of each coil so
that the turns are visible from the outside allowing the spacing adjustement
possible during tuning at the factory.
Since the difference between the original 400 MHz center frequency and
our 435 MHz frequency band is very small i.e. between 1/12 and 2/12 of
an octave higher i decided to see if spacing the turns a little bit it would
be possible to shift the center frequency 35 MHz higher in frequency
keeping as well the flatness of the bandpass as in the original filter.
The N75 female connector was then removed and replaced with an SMA female
connector and the filter was connected via 6 dB 50 ohm attenuators between
a HP8444A Tracking Generator and a HP8555A Spectrum Analyzer
A small tool made of insulating epoxy G10 for PCBs was shaped in the
form and size of a hard and sharp-pointed toothpick with the copper removed
from it and it was found to be OK to fine adjust the spacing of the coils.
During retuning step by step it was observed that changing the spacing of
the central coil it mainly shift up and down the whole band while adjusting
the end coils mainly changes the flatness and the shape of the bandpass.
In addition it was observed that introducing or removing the insulated
tool in to the holes of the filter while spacing the coils it do not change
at all the previously made adjustement.
If the adjustement of the spacing in the coils is carefully made it will
converge with easy to translate the original 400 MHz bandpass center in
to the new 435 MHz center frequency with the same flatness in to the same
16 MHz bandwidth as in the original filter.
Here are the results after retuning.
Center frequency 435 MHz
3 dB bandwidth 16 MHz from 427 to 443 MHz
Response 0.05 dB flat from 430 to 441 MHz
Insertion loss center frequency 1.5 dB
VSWR center frequency 1.5:1
435 -1.5 (0.05 dB flat from 427 to 443 MHz)
This low loss bandpass filter is very small can handle 40 watt CW
continuously and can find many applications where low loss at 435 MHz
and high rejection outside the 70 cm band is required.
If you find one of this filter in your junk box do not throw this piece of
cake out of the window.
More information about the AMSAT-BB