I’m trying to understand the interference environment for my application
(radio astronomy), using the DBS_RX daughtercard.
I’ve had an ongoing problem with a pulse source getting in via the feed
spillover of my 12’ dish. I initially thought this
was due to the local 5.6Ghz doppler radar de-sensing my LNA. The
radar is only 10km north of here, and has
peak power of 250KW. Since my LNA is a GaAsFET device, it has gain up
into the several Ghz. But I put
a commercial 11-pole interdigital filter in front of the LNA, which
covers from 995Mhz to 1600Mhz, with
out-of-band rejection of at least 85dB, and likely much more. My
pulsey little friend is still there.
If it weren’t for the fact that it’s only weakly directional, and the
pulse structure is rather “elastic”, it’d
be a good candidate as a pulsar–the main pulses show up every five
seconds, with sub-pulses that have
variable timing with respect to the main pulse. The pulses are 30dB
out of the noise, which makes them
pretty strong. Their duty cycle is quite small (another
characteristic of a natural pulsar).
It doesn’t appear to show up in my spectral display, at 4Mhz bandwidth,
but the short duty cycle would probably
preclude it, unless I turn the averaging way down, which I might try.
I’ve just modified my code to turn on the baseband filtering in the
DBS_RX, but that seems to have made the total
detected power less stable, rather than more stable.
My understanding of the way things are processed in the FPGA is that the
baseband signal (possibly low-pass filtered by
the DBS_RX) coming out of the DBS_RX is sampled at 64Mhz, then filtered
and decimated down to my desired bandwidth
(4Mhz in this case). How effective would the filtering in the FPGA be
at keeping intermod products out of my
passband, etc?
In the band that I’m observing, 1418-1422Mhz, there should be no
deliberate transmission sources at all, but I’m starting
to suspect that my pulsey little friend is in-band, rather than
out-of-band. But the duty cycle makes it hard to observe
on the spectral display.