The following is extracted from a July/August 1985 NCJ article (or was it Sept/Oct 1981?? The page I have pasted in my notebook has dates from both issues, neither of which I have any more. The original article was written by N3RD and K3ND.
Maybe later I will draw up some pretty .gifs of these, or someone could scan the originals. But for now bear with my cheap text drawing.
C1 C2 C3 C4
IN ]---------||-----------||-----------||-----------||--------[ OUT
| | | | | |
| | | | | |
| | | | | |
=== $ L1 === $ L2 === $ L3
C5 | $ C6 | $ C7 | $
| | | | | |
| | | | | |
---------------------------------------Ground
Band C1 C2 C3 C4 C5 C6 C7 L1,2,3 Core Turns
---- ---- ---- ---- ---- ---- ---- ---- ------- ---- -----
160 600 100 100 400 720 1200 1000 5.3uH T50-2 30
80 300 50 50 200 360 600 500 2.7 T50-2 17
40 150 24 24 100 180 300 240 1.4 T50-6 17
20 75 12 12 51 91 150 120 .7 T50-10 15
15 51 8 10 36 62 110 80 .45 T50-10 10
10 38 6 6 26 46 75 60 .35 T50-10 7
All capacitors are Silver Mica.
Toroid core numbers are Amidon designation.
Wire for coils is #24 enameled
Claimed Characteristics:
Other nice thing I have found. They have a very high input impedance outside of their passband. This makes them useful for splitting Beverage antennas between multiple receivers.
Construction is simple, use small aluminum miniboxes and a perf board prototyping board. Use Phono jacks for In and Out, keep leads short. The recommended layout has the string of C1-C4 going up the middle of the board with L1 and L3 on one side and L2 on the other side of them to maximize the distance between the inductors. The whole thing can be fit on a board about 2 1/2" long and 1 3/4" wide.
Tune by spreading turns on the toroids until the bandpass is centered in the middle of the band. I had to adjust the turns in some of the coils to get them to line up right, may be due to variations in capacitors or toroid cores. A spectrum analyzer would be nice. I had good results using a sweep generator and O-Scope and a reciever as a marker generator for the scope.
The following came from an internet e-mail message and appears to be an updated design:
W3LPL RECEIVING BANDPASS FILTERS
Receive only filters optimized for minimal loss and very high rejection of
frequencies below 75% of the filter center frequency.
These filters use very high "Q" high impedance resonators consisting
of powdered iron core inductors and silver mica capacitors. The resonators
are lightly top coupled to each other with minimum value silver mica
capacitors.
C1/C2 and C6/C7 form capacitive voltage dividers to match the input and
output resonators to 50 ohms.
Be sure to use good quality RF connectors (UHF, BNC or N; not phono!)
Do not substitute cheap inductors or capacitors! The components specified
below are not difficult to find, nor are they expensive.
50 ohms --- C1 ------------ C3 ------------ C5 ------------- C7 --- 50 ohms
| | | | | |
| | | | | |
C2 L C4 L C6 L
| | | | | |
|____| |____| |____|
| | |
| | |
Ground Ground Ground
TABLE OF VALUES
INDUCTOR (all three identical)
Band C1 C2 C3 C4 C5 C6 C7 turns wire size powdered iron core
10 36 47 7 75 6 56 27 8 16 T50-10 BLACK
15 51 62 10 110 8 75 36 10 18 T50-10 BLACK
20 75 91 15 150 12 120 51 13 20 T50-10 BLACK
40 150 180 27 300 24 240 100 18 20 T50-6 YELLOW
80 300 360 51 620 47 470 200 23 22 T50-2 RED
160 600 720 100 1240 100 1000 400 28 24 T50-2 RED
Tune filters by squeezing or spreading turns and/or adding or subtracting
turns
Nominal 3 dB bandwidth: 10% of center freq
Nominal midband loss: 1 dB
Attenuation is greater than 40 dB above 125% of center freq
Attenuation is greater than 60 dB below 75% of center freq