Forum: GNU Radio Filterbank stuff

Announcement (2017-05-07): www.ruby-forum.com is now read-only since I unfortunately do not have the time to support and maintain the forum any more. Please see rubyonrails.org/community and ruby-lang.org/en/community for other Rails- und Ruby-related community platforms.
Marcus D. Leech (Guest)
on 2009-02-13 01:32
(Received via mailing list)
I'm toying with some insane ideas for EME (Moonbounce) for a friend of
mine.

I'm toying with taking a modulated baseband from a QAM16 modulator, and
replicating it across several spectral pickets
  over a wide band, then on reception doing "something" with these
redundant copies to improve SNR/BER.

I'm currently thinking Analysis Filterbank/Synthesis Filterbank to take
the baseband signals, and spread them across
  a wider spectrum, with replication.

Ignoring whether this approach is sane or not is the filterbank the
right block to use to take multiple narrow baseband signals and
  "map" them onto a wider baseband?

--
Marcus L.
Principal Investigator, Shirleys Bay Radio Astronomy Consortium
http://www.sbrac.org
Matt E. (Guest)
on 2009-02-13 01:52
(Received via mailing list)
Marcus D. Leech wrote:
>
> Ignoring whether this approach is sane or not is the filterbank the
> right block to use to take multiple narrow baseband signals and
>   "map" them onto a wider baseband?
>


The filterbank is the way to do what you are asking.  But the overall
idea is not sane :)  QAM-16 makes no sense on such a weak signal.  The
purpose of QAM16 is to get more bits per second per hertz.  What you
want is more bits per second per watt.  In any case, if you just do
OFDM, then you don't need to worry about filterbanks.


I would suggest looking up and reading the paper that Phil Karn (KA9Q)
wrote a long time ago, called EME-2000.

Matt
Johnathan C. (Guest)
on 2009-02-13 02:29
(Received via mailing list)
On Thu, Feb 12, 2009 at 3:31 PM, Marcus D. Leech 
<removed_email_address@domain.invalid>
wrote:

> I'm currently thinking Analysis Filterbank/Synthesis Filterbank to take
> the baseband signals, and spread them across
>  a wider spectrum, with replication.

Aside from what Matt says regarding QAM and OFDM, if you truly want to
exactly replicate the baseband signal, and not apply some form of
coding across each of them to make them different, why not just zero
stuff your baseband up to the wider sample rate?

On the sanity of the original idea, though, I agree with Matt--OFDM
gets you the protection against frequency selective fading I think
you're going for here.

Another thought is to use spread spectrum, similar to GPS, which would
trade off bit rate for bandwidth.

Johnathan
Bob McGwier (Guest)
on 2009-02-13 02:31
(Received via mailing list)
Marcus D. Leech wrote:
>
> Ignoring whether this approach is sane or not is the filterbank the
> right block to use to take multiple narrow baseband signals and
>   "map" them onto a wider baseband?
>
>
PLEASE do yourself a favor and look at WSJT by Joe Taylor.  This guy has
years and years and years of experience doing this.  He has studied the
problem absolutely to the n-th degree.  His system is really beautifully
done.  And,  you know he isn't a nut case because he won the Nobel prize
for physics proving observational evidence for General Relativity by
discovering and observing certain types of radio emission pulsars.  The
only thing that has ever caused me to doubt his sanity was his taking
the job of Dean of Faculty at Princeton for seven years!

http://www.physics.princeton.edu/pulsar/K1JT/

Bob

--
(Co)Author: DttSP, Quiktrak, PowerSDR, GnuRadio
Member: ARRL, AMSAT, AMSAT-DL, TAPR, Packrats,
NJQRP, QRP ARCI, QCWA, FRC.
"It is human nature to think wisely and act in
an absurd fashion.", Anatole France.
Marcus D. Leech (Guest)
on 2009-02-13 03:01
(Received via mailing list)
Bob McGwier wrote:
> http://www.physics.princeton.edu/pulsar/K1JT/
>
> Bob
>
Thanks Bob.

I'd actually looked at JT65 a few years back, and had forgotten about
it.  Doh!

I'm looking into this for a friend who wants to use his USRP2 for EME.
I'll see if I can do something JT65-like (or exactly that!)
  in Gnu Radio.


--
Marcus L.
Principal Investigator, Shirleys Bay Radio Astronomy Consortium
http://www.sbrac.org
Frank B. (Guest)
on 2009-02-13 03:33
(Received via mailing list)
On Thu, Feb 12, 2009 at 7:59 PM, Marcus D. Leech 
<removed_email_address@domain.invalid>
wrote:


> I'm looking into this for a friend who wants to use his USRP2 for EME.
> I'll see if I can do something JT65-like (or exactly that!)
>  in Gnu Radio.


You should be able to just use it via portaudio on top of JACK, talking
to
the USRP through GNU Radio. There may be some sample rate issues that
need
to be solved in GNU Radio -- last time I looked, WSJT only pretended to
handle variable sample rates.

Frank
Bob McGwier (Guest)
on 2009-02-13 06:17
(Received via mailing list)
WSJT uses portaudio directly and that can talk to jack through its
portaudio host if jack has been compiled with that support in it.

Bob


Frank B. wrote:
> talking to the USRP through GNU Radio. There may be some sample rate
> issues that need to be solved in GNU Radio -- last time I looked, WSJT
> only pretended to handle variable sample rates.
>
> Frank
>
>
> --
> ...we ain't going to hell, we're going to the rebel side of heaven. --
> Langhorne Slim
>


--
(Co)Author: DttSP, Quiktrak, PowerSDR, GnuRadio
Member: ARRL, AMSAT, AMSAT-DL, TAPR, Packrats,
NJQRP, QRP ARCI, QCWA, FRC.
"It is human nature to think wisely and act in
an absurd fashion.", Anatole France.
Frank B. (Guest)
on 2009-02-13 07:17
(Received via mailing list)
On Thu, Feb 12, 2009 at 11:15 PM, Bob McGwier 
<removed_email_address@domain.invalid>
wrote:


> WSJT uses portaudio directly and that can talk to jack through its
> portaudio host if jack has been compiled with that support in it.


Ideally, yes. However, unless some of the WSJT code has been rewritten,
despite appearances, you're limited to using the WSJT default sample
rate.
Sample rate choices made in the audio subsystem aren't actually
propagated
correctly through the entire DSP chain.

This causes problems with JACK in a number of situations, depending on
which
driver (ALSA, FFADO, etc.) JACK itself is using.

Frank
Marcus D. Leech (Guest)
on 2009-02-13 17:31
(Received via mailing list)
Frank B. wrote:
> talking to the USRP through GNU Radio. There may be some sample rate
> issues that need to be solved in GNU Radio -- last time I looked, WSJT
> only pretended to handle variable sample rates.
>
> Frank
>
>
> --
> ...we ain't going to hell, we're going to the rebel side of heaven. --
> Langhorne Slim
>
So, just feed the audio as if it's a baseband signal right into the
USRP?   Or do I need to put it through an SSB modulator
  first?

--
Marcus L.
Principal Investigator, Shirleys Bay Radio Astronomy Consortium
http://www.sbrac.org
Frank B. (Guest)
on 2009-02-13 20:40
(Received via mailing list)
On Fri, Feb 13, 2009 at 10:28 AM, Marcus D. Leech 
<removed_email_address@domain.invalid>
wrote:


> So, just feed the audio as if it's a baseband signal right into the
> USRP?   Or do I need to put it through an SSB modulator
>  first?


You'll want to modulate it. WSJT gives real output. Of course
"modulating"
here means not much more than complexifying and filtering, since WSJT
will
do the tuning on its own.

BTW if sample rates do turn out to be an issue with JACK, you can always
use
the dummy driver simply for connecting your GR program and WSJT. Dummy
will
take whatever sample rate you tell it, within reason. You can't monitor
the
signal by ear that way, but in this case, big deal.

Frank
This topic is locked and can not be replied to.