[amsat-bb] I want this. I want that. Here comes another FM LEO sat.

[Phil Karn]

On 07/22/2014 04:06 PM, M5AKA wrote:

> Phil, the technology you describe could equally well be used in
> cross-band terrestrial transponders. Has anyone yet developed it for
> terrestrial use ?

Sure, there are several digital schemes now appearing for ham VHF/UHF
voice use, such as D*Star (championed by Kenwood) and C4FM (championed
by Yaesu).

But they have their drawbacks, including use of proprietary voice
codecs, a lack of multi-vendor support, and a general apathy among hams
towards anything invented or deployed after 1955 or so, when SSB started
to take off in the military and hams followed.

I reserve judgment on C4FM because I haven't looked at it yet, but I was
underwhelmed when I looked at D*Star some years ago. The design was
quite old and not very efficient or ambitious, and in demonstrations it
didn't seem to perform a whole lot better than FM. And that's pretty
faint praise for a digital mode.

Also, terrestrial and satellite communications are very different
problems with very different technical solutions at the physical layer.
In satellite communications power efficiency is almost always paramount,
so you try to use simple binary modulation schemes like BPSK with
coherent detection and strong forward error correction. High symbol
rates are okay because you usually have a line-of-sight path and
multipath is seldom a problem.

In terrestrial communications, including ham repeaters and mobile
phones, power is usually not much of an issue, at least on the forward
(base station to mobile) link. But unlike satellites, interference,
fading and multipath are the real problems because you almost never have
a clean line-of-sight path.

So terrestrial and satellite communications tend to use very different
and more complex modulation and error correction methods. Everybody
seems to be converging on OFDM (Orthogonal Frequency Division Multiplex)
because of its inherent resistance to multipath; it's used in everything
from WiFi to DSL to terrestrial TV broadcasting (except in North
America, which uses something else) to 4G mobile (LTE). OFDM divides a
channel into a lot of low speed channels that are inherently less
vulnerable to multipath. Each channel then uses a traditional digital
modulation like BPSK or QPSK. Higher order schemes like QAM are common
because, aside from multipath, you usually have a high SNR and can
afford to cram more bits/sec into each hertz of valuable bandwidth.

Cable TV systems are different from both satellite and terrestrial radio
as they have high SNRs and no multipath. Straight (non OFDM) QAM with
very large signal constellations are standard. 256QAM, where each symbol
carries 8 bits at once, is very common. That's something I would never
run on a satellite unless I was extremely constrained on bandwidth and
had DC power to burn.

Above the physical level there ought to be commonality between
terrestrial and satellite systems to permit interoperability between
them, but here we run into political problems. D*STAR uses a proprietary
patented digital voice codec common in public service land mobile, and I
think C4FM does too. In my opinion, proprietary technologies have no
place in ham radio, and enough people felt that way that we now have an
excellent non-proprietary alternative, CODEC2 by VK5DGR. While it has
gotten a lot of use in open source digital voice packages for HF it
doesn't seem to have gotten a lot of traction among the commercial
manufacturers of VHF/UHF mobile ham gear.

--Phil