Monday, December 7, 2009

The band is dead? There's an App for that!

VHF Men who have had a chance to read a few issue of DUBUS magazine may have
noted that Meteor Scatter operators in Europe put a lot more emphasis on
manipulating the WSJT software to achieve successful contacts than we do in
North America. That's a very general statement, as there are certainly some
very skilled WSJT operators in the US and Canada as well...but for most it
seems to be a casual experience. They're quite content to sit back and let the
software do most of the work and rarely attempt to manipulate it to their

Others use the software infrequently, perhaps only during major showers, and
don't take the time to unlock it's full potential.

Recently, Harry Johnson, WB3BEL wrote a very compelling response to a
discussion about this topic on the WSJT reflector. For anyone interested in
the mode, it should be required reading. Following some of the techniques that
he outlines will make you a better WSJT operator, and you'll put more contacts
in the logs. With his permission, I'll share an edited version of those
remarks now:

I agree that while the physics and geometry tend to debunk the notion of
"one-way" rocks. The term will endure to describe the frustration due to one
party decoding more pings the other. I will say I am by no means an expert,
but I have a few observations that may benefit others.

It would be true that if the mode were full duplex and both stations had
similar effective ERP and G/T you would both receive reciprocal reflections
off of the same physical meteor trail. But, that really isn't how we run on
MS. We run simplex mode. So it is probable that if the meteor pings are short
and/or few, one station receives more than the other.

However, if this does not appear random over a long time period you may want
to investigate why.. If you more frequently receive fewer pings than your
partner or you often take longer to move to the next message than they do, and
you have similarly equipped station it's worth investigating.

I'd say that most of the major reasons have a common basis; poorer signal to
noise ratio. It may be you have high local noise, it may be that you have less
skill using WSJT controls or decoding using your eyes and ears. Or it may be a
defect in signal quality.

If its local noise you will have a struggle. An antenna with lower sidelobes
may help. Finding a quieter freq may help. Finding the source and neutralizing
it is your best option but may be difficult or an ongoing battle. Sometimes
pointing off direct bearing by both parties can help especially on short

If you have a big frequency error either TX or RX the decoder will be less
sensitive. The solution may be to start a sked with Tol = 400 and progress to
lower Tol (200, 100...etc) after receiving the first bursts and adjusting VFO
if early in sked or RIT if later. What you are trying to do here is get the DF
low and then tighten the Tol. You don't want to change your freq if the other
station has also made a correction. This is a bit of mind game. For example,
if I have only run a few sequences and RX a fragment of other stations TX1,
and see a big DF say +250. I can fairly safely assume that they have not seen
me with a big -DF and counter corrected. I will move my VFO +250. and then
lower Tol to 200. But if its later in the QSO and say I RX the other station
TX2. I might just move the RIT assuming they may also have made a change on
their side. So being a better operator may improve success here.

WSJT is NOT as sensitive in decoding SH tones as your brain using eye and ear.
If you set WSJT to be super sensitive to SH tones (choosing low clip value)
you WILL get false decodes and result in some busted contacts. Tiny noise
spikes will sometimes decode as SH tone. But listening for specific tone using
headphones and looking for a trace line at the right vertical position on the
SpecJT in combination with the WSJT decoder for SH is many more dB sensitive
and less prone to false decode. If in doubt, you can click on the region on
the specjt trace to encourage multiple decode using different start time for
the captured data. Or you can wait for a second confirming ping. I find that
WAY FEWER THAN HALF of decodable SH messages automatically decode in WSJT.
Know how to adjust the RX level close to 0 dB, and understand how some noise
blankers/AGC reduce sensitivity to short fast rise time pings.

I don't fully understand it, but if I am listening using headphones and
discern a definite meteor ping especially TX1 or TX2 with data, I will
relentlessly click on it or slightly before and after it in time until I
squeeze some useful characters out of it. It's like...I know you are hiding in
there...and I am going to beat on you until you give up a few letters. These
reluctant to decode pings come from signals with lots of impairment. They may
be for example (not exhaustive):
1. Very weak (1 or 2 dB),
2. Very strong /without a sharp rise time
3. Have a big DF.. or a lot of time varying doppler..
4. TX/RX station has a lot of hum, distortion, sound card jitter on signal...

So also it depends on how hard the operator on the other end is trying. If you
walk away from the screen to get a coke, watch TV you will get fewer decodes
than a skilled op trying to choke every character string out of the
ether...Some guys want to see the full message in a decode. Other guys will
mentally glue call fragments together ala old SSB days. I think that is all
fine. If you get carried away and look for single isolated characters you may
be over the line. You can click a blank screen enough times and force any
letter to decode standalone.

So while the best remedy is higher power, bigger antenna, better QTH and rarer
DX call, you might still be able to better your odds of logging a station by
perseverance and skill with your tools.

Good luck with the meteors.

-Harry WB3BEL

[Despite his modesty, Harry is an excellent operator, and very successful at
digging out stations with significantly less ERP than his own. He's done a
great service to the VHF Community by presenting some techniques that can help
everyone make more contacts using the WSJT software]

Marshall Williams, K5QE, another top notch VHF Man, also offered his observations on how operators can improve their decodes, and make more QSO's using the software:

Have you noticed what I observed just by luck??  I often will click 10-20 times across a ping, looking for every scrap of information that might be in there.  It is amazing how often you will find one or sometimes both calls in the ping, even though a single click on the middle of the ping nets garbage.

If you have not done this, I suggest that you try it often and see what your results are.....Of course, it can be a schedule that you are not part of...just something to receive.....

No comments: