The RTL SDR sticks as mentioned seem to have delays when changing frequency that cause
lost calls in trunking mode. I've added a new feature (updated in git branch
max-trunking-update2 and in the wiki) which uses a software Local Oscillator instead of
hardware retuning which essentially makes tuning instantaneous. There is a restriction
that all of the channels to be tuned including the trunk CC must fall inside the tuning
passband (whose width is equal to the sampling rate). See the SignalScope wiki page for
more information.
Outside of this, LSM trunking works very well now with RTL SDR
Max
Show replies by date
Hi Max, thanks for the update.
Unfortunately, the higher sample rate is chocking my CPU. It isn't usable for me at
this point. I gotta figure something else out here...
-Scott
I got to thinking about this last night, and found that the freq_xlating_fir_filter is the
bottle-neck I'm having the most trouble with.
This is going to be a silly question because I don't know any better. Is there a way
to pull 25k out of the spectrum and filter that, rather then the whole spectrum buffer? It
seems rather silly to me to process all that data when only a maximum of 12.5k is needed.
We'd obviously need to have the offset tuning capabilities too. Is there a block to do
that??
-Scott
Not at all a silly question (there is no such thing, for the most part - there are silly
questioners, but this is not an example of that).
So I think there's another method that uses a PFB (polyphase filter bank) instead of
the frequency translating FIR filter. I'll experiment with that ... All in all
there's no substitute for a nice beefy CPU : )
Max