[Ground-station] Experimental Channels

[Phil Karn]

On 2/12/20 09:54, Wally Ritchie via Ground-Station wrote:
> At Hamcation, G4KLX and others expressed some very valid concerns that
> Digital Multiplexed Transponders will not provide the opportunities
> for experimentation with analog modes that are available through
> QO-100's analog bent-pipe transponder. While we haven't been talking
> much about this subject, the DMT can actually provide even greater
> opportunities for such analog experimentation while still taking
> advantage of all of the advantages of a pure digital design. How
> exactly can that be?
As you know, I really like this idea too. It could even be used for
*digital* experimentation with different modulation and coding schemes
not implemented on the spacecraft itself. But it's a rather inefficient
use of downlink capacity so it should be enabled only when it is
actually used.
>
> Using a standard 4K voice channel as an example, we can filter to
> 3400Hz wide, and nyquist sample with 16 bits at 8KHz, and we end up
> with a synchronous stream of 32K Bytes per second or 256kbps. Such raw
> IQ channels can then be relayed in real time through the standard
> quasi-error-free DMT downlink using GSE encapulation of a UDP or RTP
> stream.

16 bits seems a bit (!) excessive, especially if a user wants a wider
"transponder" than just 4 KHz. What's a typical uplink SNR, and what SNR
is actually needed by the user?

There's also my idea of having some number of uplink channels that can
demodulate and digitize analog FM, compress it onboard using a decent
codec (e.g., Opus) and multiplex that into the downlink stream. This
would not be as flexible as a digital uplink (e.g., the downlink
metadata could only identify the uplink channel frequency, not the
callsign of the user) but it might appeal to people who still want to
keep an analog "feel" in the system, like a squelch tail. Maybe an
analog FM uplink transmitter on 5 GHz would be easier to construct than
a digital transmitter, though it would probably require more uplink
power to close the link.

The same thing (digitization with Opus before inclusion in the downlink)
could be done for *any* analog voice mode uplink, including linear SSB
for anyone still crazy enough to insist on it for that wonderful "analog
experience". And there probably will be some challenges like frequency
stability. This would be more economical than a pure uncompressed "bent
pipe" mode, but it would be an attribute of the virtual transponder
instance when it is created.

But as far as the downlink is concerned, we're two decades into the 21st
century. It should be 100% digital, period.

Phil