Thanks Lee and Rick for your suggestions.
I solved the problem this morning - I hardly dare admit that it was
mostly "technician error"!
I decided to check the obvious, and measured the final 20m LO injection
frequency. It was in error by 130Hz - which handily ruled out the BFO
frequencies. I checked the LO injection frequency on 2MHz, 14MHz and
30MHz and discovered that the offset was not fixed - it varied in
proportion to the LO frequency. So my earlier posting was rubbish - I
hadn't checked carefully enough!
This pointed to the TCXO being slightly off; so rather than trust my
(now suspect) frequency counter I tuned in a 40m broadcast band AM
station which was playing music. Assuming the station would be on a
"1KHz multiple", I tuned the Omni so that the last two digits were
zeroes and then adjusted the TCXO until the music sounded OK. I find
this a much more accurate way of detecting "zero beat", because the ear
is very sensitive to any frequency error in musical harmonics.
So everything is now fine. But I'm left with the problem that, without a
really accurate frequency reference, I could still be slightly out on
the TCXO and be compensating with a slight shift on the BFO. Which leads
me to ask if anyone knows of a reasonably priced off-air frequency
standard which would give me something against which to calibrate my
test gear for the future?
73,
Steve
|