The 8.4GHz band may be quite an interesting band to make some RX experiments of very far away signals not just for the pleasure of receiving very far away transmissions but also for the challange of making a very low noise and stable receiver at microwave frequencies.
On the right you can see the 8.4GHz converter, that works at the focal point of my 5.6m dish. It employs a 8.3-8.5GHz LNA of my design with about 45K of noise temperature (at room temperature). It is followed by a band pass filter, also my design. After I use a surplus 8GHz image rejection mixer operates reasonably well up to 8.5GHz. In practice the band pass filter is a bit of a luxury as the IR mixer alone would supress the image frequency by more than 15dB. This IR mixer unit has also amplification in both RF and IF ports so no more gain is nessessary at the frontend converter. For LO I used a MACOM brick PLO that was originally operating at 7.5GHz and required extensive retunning to work properly at 8.360GHz. My IF is centered at 70MHz working well from 50 to 90MHz. Final rig is usually a NRD545 0-2GHz wide coverage receiver, and recently have experimented a bit with several home made SDRs.

Converter assembly at the antenna focal point.

Mars Reconaissance Orbiter
Received at 8.438444 GHz while mars was at a distance of about 1.63AU aprox. 243 million Km from earth.
(first detection 30/03/2006 at 19:09)

Main Carrier has about 30dB [Hz] S/N, picture of FFT with 12Hz resolution and 32x average.

Mars Express
Received at 8.420432 GHz while mars was at a distance of about 1.63AU aprox. 243 million Km from earth.
(first detection 30/03/2006 at 19:16)

Signal has about 20dB [Hz] S/N, picture of FFTwith 5Hz resolution and 8x average.

Mars Odyssey
Received at 8.406852 GHz while mars was at a distance of about 1.63AU aprox. 243 million Km from earth.
(first detection 31/03/2006 at 19:36)

Signal has about 15dB [Hz] S/N, picture of FFT with 5Hz resolution and 4x average.

Rosetta
Received at 8.421790 GHz 'en route' to the comet 67P, at a distance of about 2.61AU aprox. 389 million Km
from earth.
(first detection 2/04/2006 at 13:30)

Signal has about 20dB [Hz] S/N, picture of FFT with 5Hz resolution and 8x average.

New Horizon Pluto Charon
Received at 8.4378947 GHz while en-route to Pluto but still relatively close to earth about 0.7AU at the present
time, aprox. 104 million Km from earth.
(first detection 13/04/2006 at 00:10)

Signal has about 25dB [Hz] S/N, picture of FFT with 12Hz resolution and 32x average.
Note the excess noise like signal from the spacecraft transmitter, something I could not observe in any other spacecraft signal,
it was probably transmitting data at a reasonable high bitrate.

Spitzer, Space telescope
Received at 8.4166265 GHz while obiting the sun trailing earth at about 0.3AU earth distance, aprox 48 million Km.
(first detection 28/04/2006 at 21:10)

Signal has about 26dB [Hz] S/N, picture of FFT with 5Hz resolution and 16x average.

Signal has about 21dB [Hz] S/N, picture of FFT with 5Hz resolution and 4x average.

Stereo A & B
Received at 8443.518520 and 8446.234570 GHz orbiting the sun trailing earth at about 0.5 AU distance from earth,
aprox. 152 million Km.
(first detection 08/05/2006 at 10:10, 10:20)

Signal has about 21dB [Hz] S/N, picture of FFT with 5Hz resolution and 8x average.

GAIA
Received at 8464.990304 GHz on its way to L2 point, still quite close to earth at about 899477.901 Km away.
(first detection 30/12/2013 at 23:30)

Signal has a S/N of about 52dB [1Hz], picture of FFT with 2.93Hz.

The signal was marginal on a water-fall spectrogram but clearly measurable with just one minute integration time.

Signal spectum with doppler corrected 10 minute integration time. Identification of the Cassini signal was based on the fact that signal has the right doppler variation as predicted by the relative velocity calculation (19Hz). Signal is 4.9 dB [Hz] above noise.

To detect this signal that is expected to be 13dB/Hz below the noise floor I had to aquire and integrate spectrograms for a long time. I did several aquisition periods of 15 minutes (900s) the minimum I would expect to see something. More than 15min in each chunk is also undesirable because of the doppler change correction scheme used. The receiver is operated at fixed frequency and the doppler variation was corrected by skewing sucessive spectrograms in software while accumulating. Positive identification of the V1 signal arises from the fact that signal is only visible for the right skew ammount that corresponds to the doppler variation as predicted by the relative velocity calculation.

Combining four of the best integrated spectrograms, we could obtain this 1 hour (3600s) integrated spectrum where all the other peaks nearly disapear. The blue trace, being the best one, has about 0.2dB of (S+N)/N that corresponds to -13.3dB of S/N [4Hz] or -7.2 dB [Hz] (Using 1.3Hz resolution we could see that my rx signal was spread over 4Hz).
The recent V1 reception with the 20m dish at Bochum's (by the AMSAT DL Team) produced 2dB of (S+N)/N [10Hz] that is -2.3dB of S/N [10Hz] or 7.7dB [Hz]. As I should have about 12dB less antenna gain and 3dB of polarizarion loss (as I was running linear polarization) therefore I would expect to receive -7.3dB S/N [Hz]. I consider my result in very good agreement with this particular experiment.