To check that the induction was right, (I have no induction meter),
I made a parallel ressonant circuit of the coil to be adjusted, with a know capacitor.
say: 82pf and 0.72uH has res at 20.7Mhz,
and: 82pf and 0.85uH has res at 19.1Mhz, easy to measure with a signal generator and a scope.
The coil value was then adjusted with a small ferite, that was glued in the coil.
The capacitors was made of several small values in parallel,
result: lower esr, lower los, higher power capability.
159pF is 47 + 47 + 47 + 18
361pF is 82 + 82 + 82 + 82 + 33
This filter has about 1dB (0.7dB simulated) loss at 14.4Mhz and at mirror frek 38.7MHz -60dB,
this will change to -80dB after adding the 47pF and the 18pF capasitor, look at the bottom schematic.
It's important to construct the filter, so the 3 coils can't get any inducted signals from the other coils.
This is done by making 3 small rooms for the coils, here the filter is shown before the side walls was soldered on to it.
I have tested my filter with 50 W CW at 14.200 for 5 mins it was a bit hot, but not too hot.
With 100 W the glue starts to smell bad !! after 10 sec. So this filter need to be redisigned if the power need is over 50 W CW
This is a simulation of the above construction
using RFSim99 freeware from www.hydesign.co.uk this is the best freeware program I have ever seen !
To make it alot better, just add two capasitors,
47pf for the 2nd harmonic and 18pf for the 3rd harmonic.
Now a 30dB better rejection is achieved.