Jim Lux wrote:
Yes. OP said "near real time," which I take to mean "OK to drop some
frames," like the satellite video phones the reporters use from the
boondocks. Thus, high-def can be confined to a lot lower bandwidth if you
don't mind seeing compression artifacts as each frame is being built on the
screen.



Here's a back of the envelope link budget for a 500E3 km link carrying 1
Mbps

Let's assume 2GHz for the working frequency (not necessarily a good
choice, but somewhere to start)

Free Space Path Loss from Moon to earth, between isotropic antennas =
32.44 + 20*log10(500E3) + 20*log10(2E3) = 32.44 + 114 + 66 = about 212 dB

Assume an antenna 2m in diameter at one end (moon end)
lambda for 2GHz is 15cm, so the antenna is 13 wavelengths in diameter
Beamwidth will be about 70/13 = 5 degrees.. OK, because Earth is 2
degrees wide from the moon, so you can just point at the middle of the
visible earth.

Assume an antenna 10m in diameter at the earth end. Beamwidth will be 1
degree, twice the visible lunar disc size, so you can just point at the
moon, generally.

Gains of antennas
2m @ 2GHz = 30dB
10m @ 2GHz = 44dB

Preceived at Earth = Ptransmitted +30 - 212 +44dB = Ptransmit -138dB.

Assume transmitting with 10 Watts or +40dBm..
Prec = -100dBm

Now, let's look at the receiver:
Power Spectral density of Noise is kT+NF.. kT is -174dBm/Hz and a decent
NF might be 2dB (allowing for some plumbing losses, etc.

-172 dBm/Hz

Eb (energy/Bit) = -100dBm -60dB (1Mbps) = -160dBmJ

So, Eb/No is about +12dB... If you allow 2dB for implementation loss,
that gets you to 10dB, which will get you a BER of 1E-6, which isn't
terrible. Coding will improve it, etc.

Take home message:

10W at 2Ghz with reasonably sized antennas at moon and earth can carry 1
Mbps.

Scales linearly with data rate.. You want 10Mbps, you need 100W. Or
bigger antennas.