On 24/02/2019 13:43, Shane Stroud wrote:
This is interesting why would it convert the frequency to be 10 Hz
up? I don't have a VFO correction or offset established that I know of.
Hi Shane,
having got around to looking at the CAT protocol for the FT-747 and the
Hamlib implementation, I can see what is happening. This rig has a very
oddball CAT frequency arrangement which is hard to deal with. The rig
has 25 Hz tuning steps but does not display them, it only displays 100
Hz steps. The CAT command to set the frequency accepts 10 Hz steps (as a
BCD encoded number) but he rig internally converts the 10 of Hertz
digit (least significant BCD digit) to 25 Hz steps (0, 1, 2 -> 0 Hz, 3,
4 -> 25 Hz, 5, 6, 7 -> 50 Hz, and 8, 9 -> 75 Hz). The get frequency CAT
command returns a BCD number for the 100 hz steps but a 25 Hz step using
the least significant two BDC digits (00, 25, 50, 75).
The Hamlib driver attempts to round the frequency to be set to 25 Hz
steps which are then encoded as 10 Hz steps. This makes it very
difficult to round-trip a frequency value since the input and output
steps are different due to the implicit conversion within the rig.
The odd 10 Hz displayed in WSJT-X are not inconsistent with the actual
frequency set on the rig since the rig will round that back down to 0
Hz, so despite the rather strange effect I would ignore it and not worry
about it. OTOH if it does result in a walking frequency continually
stepping up or down then that should be addressed although I am not sure
how. WSJT-X does attempt to determine the rig's frequency resolution but
in doing so it assume that it is possible to round-trip frequency set
and query commands at the determined resolution, which is not the case
with this rig.
The only sure solution I can think of is adjust the frequency resolution
determination in WSJT-X for this edge case and drop back to 50 Hz
resolution since at that round-tip frequency control is possible, even
that would need adjustment to Hamlib to ensure that 0 Hz, 50 Hz, 100 Hz,
... frequency steps were correctly round-tripped via the rig.
73
Bill
G4WJS.
_______________________________________________
wsjt-devel mailing list
wsjt-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/wsjt-devel
