[amsat-bb] PSK-31 Auto Doppler Correction

[Bob Bruninga <bruninga@xxxxxxxx>]

In response to the discussions about a PSK-31 Satellite Transponder,
Peter Martinez has now worked out the details of a second order AFC
correction that will allow PSK-31 to stay locked onto a worstcase 16 Hz
per second per second Doppler rate of change  on 10m meters.

As we also noted in that disucssion, these most of the time AO-40 is well
within these low rates..  Thus making PSK-31 quite usable on AO-40.
His Email below is theoretical and will need to be applied in the more
popular PSK-31 porgrams, but it shows that it can be easily done.

de WB4APR@amsat.org, Bob

PSK-31 Transponder: http://www.ew.usna.edu/~bruninga/psk31uplink.html

---------- Forwarded message ----------
Date: Wed, 27 Nov 2002 19:55:28 +0000
From: Peter Martinez <Peter.Martinez@btinternet.com>
To: PSK31 <psk31@bipt106.bi.ehu.es>
Subject: [psk31] 28MHz PSK31 satellite uplink

>From Peter G3PLX:

After discussion with Bob, I have worked out that the 28MHz uplink
frequency will go from 711Hz high to 711Hz low during an overhead pass,
which is the worst case for Doppler. As the bird goes overhead it will
be shifting by 16Hz/sec. This is way outside the capability of the AFC
on my original PSK31 program, and I suspect most others will be unable
to follow at this rate.

My suggestion - and this is aimed at PSK31 programmers - is to go for a
second-order AFC loop. I have simulated it here using the worst-case
scenario and it will comfortably hold a PSK31 signal within 2Hz (plenty
enough to copy QPSK) right through the absolute worst-case overhead
pass.

To explain how it's done, let me first explain how the basic
(first-order) AFC works. Every time the program decodes a 31.25 baud PSK
symbol, it can calculate the frequency error. Rather than subtract ALL
of this error and be spot-on-tune ready for the next symbol (which would
be too sensitive to QRM and QSB), the idea is to remove only a small
part of the error. The effect is that the error is removed slowly and
gently over a period of ten seconds or so. Such a system will stay on
frequency through QSB and QRM, but if the signal was drifting at a
steady rate, there would be a steady residual frequency error. If the
signal was drifting too fast, the residual frequency error could exeed
the limit for the PSK decoder (7.8Hz for BPSK, 3.9Hz for QPSK). Not only
that, but even if the AFC feedback was increased to reduce the residual
error, it would only take a 0.5 sec drop-out (0.25 sec on QPSK) to throw
the AFC off lock as the satellite went overhead.

The second-order AFC which solves this problem has an additional tuning
rate correction. As well as using the frequency error to correct the
receive frequency, it also uses this error to adjust the tuning rate.
The effect is that a steady drift will eventually build up a tuning rate
correction which will cancel out the residual frequency error
completely. In the satellite situation, it means that the residual error
is almost completely removed, and even if the satellite signal drops out
as it goes overhead, the AFC will stay locked.

The second-order AFC system could be simply added to the existing AFC
system for satellite use. The tuning rate value needs to be reset to
zero before a pass. Some experiment will be needed to choose how much
correction to apply. It won't work if there isn't enough, and it could
go unstable if there is too much.

This scheme could just as easily be applied to the idea suggested for
correcting the 28MHz uplink frequency of an FM-downlink satellite by
monitoring your own transmission through the satellite.

I hope the PSK31 programmers will be able to use this idea for
"satellite capable" PSK31 programs.

73
Peter
_______________________________________________
Keep in mind that only registered email addresses can post in the list.
Psk31 WWW Site at http://aintel.bi.ehu.es/psk31.html
Psk31 list info at: http://aintel.bi.ehu.es/mailman/listinfo/psk31

----
Via the amsat-bb mailing list at AMSAT.ORG courtesy of AMSAT-NA.
To unsubscribe, send "unsubscribe amsat-bb" to Majordomo@amsat.org