Comments form Peter, G3PLX the father of PSK31:                 18 Jan 2014
---------------------------------------------------------------------------

PRIOR QRITING: I recall that I wrote-up my ideas from the PCSAT2 experiment 
somewhere - was it in an AMSAT Journal?  You might like to dig that out and 
post it.

AGC:  One of the things I didn't like about the PCSAT2 PSK31 transponder 
was the low gain of the overall signal path. I would STRONGLY urge you to 
make the overall gain such that the antenna noise level of the 28MHz 
receiver FULLY modulates the downlink baseband. AGC action should then 
ensure that the downlink baseband remains fully modulated at all times. 

Unlike a linear transponder, there is no reason to keep the downlink signal 
level low under the "no-signal" condition.  Strong AGC action should then 
keep the downlink modulation depth constant regardless of the uplink signal 
level.  I suggest a combination of a mean (or r.m.s.) level detector plus a 
deviation clipper, rather than a peak-level detector.  This strategy will 
ensure that uplink noise will be the limiting factor at all times even if 
the downlink SNR is poor. 

The strong AGC action does mean that the user will not see changes in the 
level of his own downlink baseband signal as he changes his uplink power. 
Uplink power control, based on measuring the level of one's own downlink 
signal, will therefore not work, but my experience was that this method 
would be unworkable anyway - it would be very difficult to set levels in 
the user's downlink receiver in order to ensure that all users shared the 
downlink.  

However, I did trial uplink power control based on the user 
measuring the SNR of his own downlink signal, and that worked very well 
and it only requires all users to agree on a reference SNR value to be 
set into the user's groundstation software and doesn't depend on receiver 
baseband levels.  This reference downlink SNR value could even be hard-
coded into the user software in order to force all users to the target 
value (~10dB?).  

We would need contributions from the authors of existing PSK31 software 
packages to add this feature into their software. We would in any case need
contributions from these authors to include the uplink Doppler correction, 
in which the uplink frequency is controlled to keep the frequency of one's 
own downlink signal fixed. To achieve both the uplink power and Doppler 
control, the user software needs to have an auxiliary receive channel which 
will be aligned to the frequency of the user's own transmission.  Most of 
the PSK31 software packages can do this now.  I had both these features 
working on PCSAT2, and was probably the only person to do so within the 
short lifetime of the PCSAT2 transponder (before the antenna fell off!) 
so I can contribute ideas here.

[WB4APR comment: I hope to require uplinks to use 1/4 wave vertical antennas
with 25W transmitters.  The horizon-to-zenith antenna pattern matches
inversly with the range-gain of the typical pass over any one station.  
Thus, the power received on the spacecraft is nearly level within a few 
dB independent of where a particular station is located with respect to 
the instantaneous position of the bird. (plus or minus, of course, all 
the other variables ;-)

73
Peter G3PLX