NetNews Usenet Archive 1993 #3/NN_1993_3.iso/spool/rec/radio/info/53

Internet Message Format | 1993-01-24 | 5.2 KB

Path: sparky!uunet!usc!elroy.jpl.nasa.gov!decwrl!access.usask.ca!kakwa.ucs.ualberta.ca!ersys!ve6mgs!rec-radio-info
From: [e-mail] (Robert Carpenter)
Newsgroups: rec.radio.info
Subject: Meteors de W3OTC
Message-ID: <[e-mail]>
Date: 23 Jan 93 01:16:54 GMT
Followup-To: rec.radio.amateur.misc
Organization: National Institute of Standards and Technology
Lines: 79
Approved: [e-mail]

I'm happy to see a renewed interst in (at least discussion of) meteor-burst
communication. As someone may have noted, I was the other end of the contacts
with Ralph Wallio, W0RPK, a few years ago.  We used 6 meters, 4 to 6 element
beams and 150-250 W output... but also the worst possible modulation scheme,
packetizing, etc,  AX.25.  AFSK-FM gives you a huge performance hit to start
with, and the AX.25 scheme doesn't play well with short bursts.  Some of the
PSK or QPSK modulation schemes would be a much better start.  I haven't
followed AMTOR, but some approach that doesn't require whole, long packets to
be perfect would be a big improvement over AX.25 . Some FEC might be a help.

One of the problems is that at least one of the stations in the QSO has to
transmit ALL THE TIME, to probe for trails. The correlation distance is such
that you wouldn't do well using a TV transmitter some thousands of meters away
to do the probing for you, not at least on the underdense bursts. On the other
hand, in the meteor burst (MB) system I did at NBS (now NIST) back in the late
1950s, most of the data was transfered on the overdense (long) bursts. They
are quite rare in the afternoon and evening. One or two big ones an hour at
40 MHz, in the doldrums.

Another problem with old-type protocols, is that the transmitter probably
doesn't send any info until it has been notified by the other end that the
path exists. This means a round-trip prop time (6 ms for 1000 km path), plus
the startup/decision time of the remote station, etc.  This eats a big hole
out of your garden-variety 100 ms burst. For the DX-hunters' mode, you'd
continuously send the info blind during the probing for a trail - as in EME,
since it is a short message, and wait for a return msg including an ACK and
the other station's info.  I assume the snow-depth and truck-locator people
use some scheme like this. They don't have to transfer much info.
Even for "serious" message traffic this might be a good way, using short
message units and selective ACKs and repeats.  Number each short message
block and send 100 ms of them as the probe. As ACKs for individual blocks
come in, replace them with not-yet-ACKed blocks. On an overdense burst the
ACKs should come back soon enough to allow continuous transmission. Just a
thought.

As for bands. Both my ham and NBS work has been in the 40-50 MHz range. Why?
Meteor sigs and durations decline as the many'th power of the frequency.
It's HARD on two meters, but people do it because of the dearth of other modes.
Why not use ten meters? You could, except that other modes of prop may
obscure the MB sigs and bring you lots of cochannel QRM. Now that we are
going toward a sunspot min, 10 is becoming like 6 was in the best DX years,
and would be good choice for easy MB.  You have to be able to handle
continuous strong prop like Es, of course.
Since the meteor trails are at about the same height as Es, Es can give
really strong continuous sigs over the same paths.

Since computers hardly existed at the time of my MB work at NBS, we had to
use fast-start, fast-stop mag tape machines for transmitting and receiving
buffers. Oh for even a Z-80, back then!!

As a final item, MB didn't make it for communication to polar regions, etc
where the need existed way-back-then. A similar transmitter power, but with
HUGE antennas, would get you well into ionospheric scatter capability, where
sigs exist all the time, but weakly. And, for the period before widely-
available computers for logic and storage, much simpler and cheaper and more
reliable terminal eqpt (remember it was all tubes back then).  Hams on 6
still use ionoscatter - listen around 50.125 any weekend morning for weak
fluttery SSB phone sigs from 1000 km + - away. The terminology ionospheric
scatter vs. meteor scatter to describe this continuous weak propagation shows
political orientation. The NBS line at the time was that it arose from
turbulence, etc., in the D/E layer.  The Stanford line was that it was caused
by lots of very small meteors.  But maybe the meteors caused the turbulence??
ALASCOM, the Alaskan phone system, uses "MB" equipment, though I would call it
ionoscatter.

Of the main NBS people on the first IEEE IONOscatter paper, Ross Bateman, W4AO
is dead, Dick Kirby, W0LCT (one of my bosses on the MB work) is president of
ITU in Geneva, and I've lost track of Dana Bailey, no call.  Earlier NBS
people in the meteor business were Franklin Montgomery, W3FQB, Peter Sulzer,
x-W3HFW, and George Sugar, x-W3KQS -never on the air. You San Diego folks may
know my boss during most of the MB work, Ken Bowles, x-W0???, also of UCSD
Pascal fame.

73, Bob W3OTC  (now mostly retired from NIST)

ps - I'll be away a couple of weeks.
--
  - Postings to rec.radio.info:            [e-mail]
  - rec.radio.info administrivia:        [e-mail]