I have boxes of boards that didn't quite make the grade and I now find since I have little time to build new, this is a perfect time to fix some of those boards. It is all there -- they just don't work.
Here is a case of a candidate board where I tried to skate by with a 4-pole homebrew crystal filter and it had a few hiccups. This is a ripe selection.
A Board that might become a Rig.
So, what really happens with a filter of less than 5 or 6 crystals. Ten Tec sold the early Triton I and II radios with 4 pole filters and early on KVG made a 5 pole and 8 pole 9MHz filters. [I can really tell my Triton II only has a four-pole filter. The Triton IV has an 8-pole filter.]
The 1st problem is that the lesser pole filters likely have a shape factor such that the skirts are not approaching vertical but more like gently rolling hills.
A plot from my Nano VNA (one of two) -- more poles would give you skirts that would have a greater depth of null before flaring out. The pass band ripple isn't too shabby, but the broader skirts mean adjacent channels will be heard.
The filter is not symmetrical and favors LSB (right side) as that skirt is more vertical than the left-hand side (USB). Will it work -- yes. Will you get reports of being wide or distorted -- most definitely! But hey you are making contacts with a marginal 4 pole Crystal Filter.
Placing the LO above the Crystal Filter frequency would cause sideband inversion and now that more vertical right side could be used for USB.
The board itself has a steerable module (2 Relays), an ADE-1 (PD/BM), and SBL-1 (Rx/TX Mixer), the 4-pole 4.9512MHz Filter (SMD Caps backside) and two DGM's fashioned from pairs of J310's with a tuned tank, and an on-board BFO with two SMD 2N3904's. A couple of trim pots sets the stage gain (Gate #2) of the DGM's. There is input matching to the DGM's.
I have no recollection why this didn't work but am not seeing the Crystal Filter matching transformers. (Must have been liberated for another project.) Or the input match to the DGM's takes care of that issue since the cores are FT-37-43s.
If this were to be fixed for use (FFU) I would remove the BFO board and use the Arduino + Si5351.
Today's post demonstrates that sometimes old non-working projects possibly can be resurrected and that I do have a Nano VNA useful in looking at filter pass bands plus knowing how to build crystal BFOs.
While some would look at the plot as having low pass band ripple, I see a filter likely to show its shortcomings on a contest weekend on 20M and the need to have the LO above the filter frequency to take advantage of the LSB slope.
Check your junk box as you just may have a rig awaiting resurrection.
73's
Pete N6QW