I made the attached two plots of the bandpass of the crystal filter. U1 and U2 were removed from my test board. Tracking generator was fed into pins 4&5 of U1 socket and input to spectrum analyzer was taken at pins 1&3 of U2 socket.
The 1.BMP file used my widest ranging crystals 28, 30, 43 and 45 tens of Hz. The 2.BPM file uses crystals 35,35,35 and 36.
Plot 2 using matched crystals should have been better than the first that uses crystal differing by as much about 170 Hz. Maybe I don't know what to look for but the two plots look about the same to me. Next step will be listening to the two filters to see if I hear any difference. I don't have the VFO built on the board but at this point plan to try a DDS VFO. I will compare that to what I hear with a matched crystal filter from a stock NC40A build.
While it has been interesting matching crystal I do wonder how critical it is. I see posts of people wanting to buy and other offering to provide matched sets so it seems there is concern that without matched sets the radio will not work. I am trying to get some practical insight on how true that is.
If the attached files don't make it through with this text I will post them separately.
73/72 - Mike WA8BXN
