[Moon-net] Scanning For JT65 signals with an FT-847 (And other rigs I presume)

[Bob Baker]

Dear Moonnet,

This email is intended to describe how to use some features of an FT-847 (I
strongly suspect that other rigs have similar capabilities) to facilitate
unassisted JT65 QSOs. It may be useful for CW, but I do not know. Perhaps
many of you are already using a procedure like this one, or even a better
one (if so, please let me know). But I suspect this procedure will be
helpful to some on moon-net, especially given the capabilities of 5.9.0.

This is all very preliminary, I just started monkeying around with it last
night. I am sure many on moon-net are capable of improving this procedure, I
intend to keep working on it. I am not an expert, I am a rookie. And I don't
like reading hundred page users manuals. I assume others on moon-net share
my dislike of long manuals. But sometimes they should be read.

BACKGROUND
Ever since the W3SZ moonnet email regarding use of Linrad to find JT65
signals I have become fixated on the concept (I get obsessed sometimes). My
approach is not as good as Linrad, not even close. It is cheaper and much
easier however. Jeremy's email for manual jt65 searching was also
interesting and helpful, although I think I may be a bit slower, less
coordinated and lazier than Jeremy.  Unassisted digital operation would be
an excellent opportunity for more qsos whether in contest or not in a
contest. Plus it is "purer". Do not get me wrong, I have nothing against
loggers, and enjoy the chance to say hi to my friends when the loggers are
not too busy.

FT-847 FREQUENCY INFO
The FT-847 menu option 3 allows you to set the MEM/VFO CH step size to 1khz
(see ft-847 users guide page 83). For many jt65 ft-847 users this mode of
operations may present an improvement over tuning the main dial since the
MEM/VFO CH knob has stops which facilitate quick qsy in even khz increments.

Unfortunately for me, my ft-847 has a bias of approximately -400hz. (I read
an interesting email on moon net on how to correct this bias, but absent a
frequency counter, this seemed a bit dangerous to me and I never did it).
This bias hasn't really been a problem for me in the past, I just set the
VFO 400hz high, then I was more or less on frequency (note to yaesu
engineers- please give us a menu option to correct bias in frequency
display). However, tuning with the MEM/VFO CH knob sets the HZ values to
000, so if I use this knob for tuning, I am 400hz off frequency.

The way I got around this problem was to use the memory channels. there are
78 of them, 80 if you include L and U. So I set MEM channel 1 to
144.101.400, MEM channel 2 to 144.102.400, .... MEM channel 78 to
144.178.400. (Can set MEM channel L to 144.100.400 and MEM channel U to
144.179.400). The band is more or less covered. See page 72 and 73.

SCANNING
All the above stuff may be useful to some of you who don't like reading
manuals. And hopefully it will help some of you to manually scan for JT65
signals. But it is still very time consuming to search the band, 160 minutes
for a once over 144.100 to 144.180 odd and even periods. (Could we all agree
on a non-contest jt65 CQ range?? Say 144.140 to 144.150??? IMPORTANT NOTE: I
am not trying to tell The big gun pioneers of EME where to CQ, I would never
do that! I am just a rookie, everybody knows the big gun frequencies, that
is good, I am only trying to spur discussion of where us little guys can
find each other).

Last night I conducted an experiment to determine if a CQ could be decoded
with less than 60 seconds of monitoring. I am just a two yagi (4 wavelength
booms) station. I cheated for purposes of experiment and went on the logger
and found YU1CF CQing on 120. (Thanks Goran). YU1CF is big. 16x7, 25dbi,
1500W. In memory scan mode, the FT847 can be set in menu 26 to pause on a
frequency for 0, 3, 5 or 10 seconds, 5 second default (ssb). 10 seconds is
pretty short, I would have preferred longer, but that is what I had. The
results are listed below:

051800  0  -25  1.5  159  3 *      CQ YU1CF KN03             0  6
051900  0  -33 -0.3  140 34                                  0  0
052000  2  -18  1.5  159  3 *      CQ YU1CF KN03             1 10

The 0520 decode was for a full minute. The 0518 decode was for 10 seconds.
You can see the 7 db decrease in signal strength, about what you would
expect given the time ratios. Don't know why I only had a sync 2 on Goran
for the full minute, obviously I need more tests. But even with sync 0,
5.9.0 was able to pull out the call with deep search. (I had set it to
enhanced). Obviously when Goran was sending OOO to another station I could
not decode him. But the pattern on the waterfall was still pretty
suspicious. Shorthands are very noticeable visually.

This is not a set it and let the computer do all the work procedure. When a
minute ends, you need to double click on all suspicious streaks in each of
the 6 periods (For 10 second scans). This could not have been done with
earlier wsjt versions. My computer is pretty fast, 2.0 ghz with a big cache,
but this would have been hopeless with 5.8.6. Based on this one test I feel
pretty confident that if I moved up on the learning curve a bit I could
decode shorthands visually from most 4 yagi stations with a 10 second scan
dwell period, faraday willing. CQ's are more problematic with a 10 second
scan dwell time, but even without a decode, I think I would be able to
detect many of them (flag them for longer decoding). If you are a 4 yagi
station you should be able to do much better than me. I had the waterfall
set to speed 3. This was all done visually. I intend to conduct a future
scan test using the waterfall and headphones. (I may not get much benefit
from the headphones however, a recent moonnet email mentioned the effects of
Cream, Hendrix and Joplin, to them I must add The Who, Stones, and Zepplin,
but I will try).

The above test was conducted with a manual 10 second scan, I did have
similar results with the Scan set on automatic (4 scan button, 6 PMS,
squelch off, page 80).


SOMETHING MUCH BETTER?
Perhaps using the programmable memory scan is not the best way to go
especially for smaller stations such as myself. The 10 second dwell time may
be just too short. Perhaps scanning the memory channels manually with a 15
or 20 second dwell time would be better for me. I read enough of the CAT
(Computer Aided Transceiver) section of the users guide to realize that CAT
is probably the best way to go. Perhaps a standalone program with Scan
start/stop frequencies and user selected dwell time? Coordinated with the
computer clock? Automatically set up to cover both the even and odd periods?
Set up to account for the less than 60 second xmit time of jt65? Other
considerations? I also learned enough from reading this section to know that
I cannot do it. I do not have the programming skills.

CONCLUSION
This may not be very useful, maybe just a sign of my obsession, but I
suspect that it will be helpful for me in next years contest. I had fun in
the assisted category, but maybe next year I will try unassisted. Or maybe
stay assisted but use scanning to find those stations who are not posting
CQs on the loggers. For non-contest periods, if we could finally agree on a
JT65 CQ band segment, it would be fun, and maybe get me some initials. Any
feedback would be appreciated (try to be nice).

73,
Bob
KD3UY