I have a 20m CW transceiver with an option for a wide or a narrow filter. When the narrow filter is selected, a crystal ladder filter is used in the intermediate frequency filter, between two mixers. When the wide filter is selected, this crystal filter is bypassed and there exists a direct link (that is, only a capacitor) between the two mixers.
I use the wide filter to search for signals by rapidly increasing frequency from 14.000MHz to 14.075MHz. When I found a good signal, I tune it to have the right pitch, and switch to the narrow filter to have less noise and other signals.
There are two frequencies I can tune into to find this pitch that is convenient to listen on, with a difference of about twice the pitch frequency.
I noticed that when I use the lower frequency, when I switch to the narrow filter, the signal is weaker than when using the higher frequency, to the point that only very clear signals can be copied. Can this be explained theoretically, and will this happen with any transceiver, or is it more likely that this is specific to this transceiver?
My explanation would be that the higher frequency, yielding a stronger signal on the narrow filter, is the carrier frequency. The other frequency would be the carrier minus twice the pitch frequency, which is then much weakened by the crystal filter. If the filter is bypassed though, the next mixer outputs both the sum and the complement of its inputs, so the two signals are equally strong. Is this correct?
The RX module is from KD1JV's ATS 3B. I'm not sure if there's copyright on it, but the idea is quite straightforward: RF enters the first mixer, an NE602, together with the output of a DDS. Then the output of the mixer passes through the (4.9152MHz) crystal ladder filter (or not) and goes to the next mixer, a second NE602, together with the output of a controllable oscillator (with a trimmer cap), to adjust the pitch.