Gnu Radio for RA

[Stephen Arbogast]

I am  a dummy ......    I  tried  randomly  switching  between  collecting  samples  for  reference and    antenna......     this  does  not  work.....       I  got  errors  about    sizes  of  arrays   in python ......

[Robert Meade]

Ted Cline was dealing with some shenanigans related to this. He had to "clear the pipes" so to speak in the GNU radio flowgraph for this to work. You almost want to call the flowgraph externally on one frequency, stop it, and run it with another. I think he used some messaging block functions to purge it before switching sources or frequencies without it being unclear where in the stream the switch occurred. You may consider looking at GNU radio scanner implementations for an idea of how to implement a switching for reference sample collection, or reach out to Ted and he's probably willing to share his draft flowgraphs.

On Sat, Feb 14, 2026 at 5:12 AM 'Stephen Arbogast' via Society of Amateur Radio Astronomers <sara...@googlegroups.com> wrote:

I am  a dummy ......    I  tried  randomly  switching  between  collecting  samples  for  reference and    antenna......     this  does  not  work.....       I  got  errors  about    sizes  of  arrays   in python ......

--
--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to
sara-list-...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
---
You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sara-list+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/sara-list/9d2e9c7c-97ce-4e7e-b886-1c33adba5578n%40googlegroups.com.

[Marcus D. Leech]

On 2026-02-14 17:35, Robert Meade wrote:

Ted Cline was dealing with some shenanigans related to this. He had to "clear the pipes" so to speak in the GNU radio flowgraph for this to work. You almost want to call the flowgraph externally on one frequency, stop it, and run it with another. I think he used some messaging block functions to purge it before switching sources or frequencies without it being unclear where in the stream the switch occurred. You may consider looking at GNU radio scanner implementations for an idea of how to implement a switching for reference sample collection, or reach out to Ted and he's probably willing to share his draft flowgraphs.

You can look at stream tagging.    But unless the *hardware* inserts tags at an appropriate moment, there will ALWAYS be samples-in-flight that lead to ambiguity.   Gnu Radio is
  inherently a streaming architecture, with ring buffers between each block.     You can lower latency (and thus ambiguity windows) by reducing the sizes of those ring buffers,
  but at the risk of dropping samples, depending on sample rates.

To view this discussion visit https://groups.google.com/d/msgid/sara-list/CAMZzgQeTw6mHs9gy0Nt2N-j5FuMXx2NRuGmrLydyPTwBk2iAiw%40mail.gmail.com.

[Robert Meade]

Last I checked Ted avoided the stream tagging approach. I might give it a whirl, but I think I'm more interested in trying it when I have a noise diode reference source implementation vice retuning, so may be a while.

To view this discussion visit https://groups.google.com/d/msgid/sara-list/0b50c636-814e-40f8-a52c-5d48fab4f7fd%40gmail.com.

[Marcus D. Leech]

On 2026-02-14 17:42, Robert Meade wrote:

Last I checked Ted avoided the stream tagging approach. I might give it a whirl, but I think I'm more interested in trying it when I have a noise diode reference source implementation vice retuning, so may be a while.

There's also an approach I used a few years back called "oblivious slicing".   You gather up a bunch of samples, and if the reference is significantly "brighter" than the average sky,
  then samples fall into two populations, so you know which ones are from the reference and which ones are "sky".   Ken Tapping (of DRAO) taught me this one.

To view this discussion visit https://groups.google.com/d/msgid/sara-list/CAMZzgQcuC6aMMCtHX-6252JYm6N9%3DVwBCxB0Cq%2B2LrFbNVV%3Dsg%40mail.gmail.com.

[Robert Meade]

I like this one because it works in post processing. This will probably the first thing I try if/when I built the SRT dish vertex calibrator.

To view this discussion visit https://groups.google.com/d/msgid/sara-list/b6d38a8b-7d02-42a6-955d-f23e0f2eee0b%40gmail.com.

[andrew....@googlemail.com]

Ted has successfully overcome this problem by introducing a delay on changing over from signal to reference and visa versa – any concern about the delay’s effect on signal collection is mitigated by Ted also providing the choice of how often a reference sample is collected – from every other signal sample to every 10 or more.

 

I have successfully tested the beta versions of ezColGR (GNR Radio ezCol software).

 

The bottom line is so far we have found everything can be overcome – it is about learning how to do it in GNU Radio, which is often not stated in any documentation.

 

In the BAA GNU Radio group that meets monthly, we are working through learning to use the software and create our own radio astronomy implementations. Please join us!

 

Andy

To view this discussion visit https://groups.google.com/d/msgid/sara-list/CAMZzgQcEw6YGbbYQ-f2Vii-GN66esjSffwMVQxf9Qn8s0-zqHw%40mail.gmail.com.

[Stephen Arbogast]

Some  thoughts......

1.    Stream tagging  approach.......   I have been  playing  with  this  but  since  Ted Cline gave up on  this approach  I am wondering...

2.    Marcus   Leech   mentioned  the problems  with the  streaming architecture  of  gnu radio...   Yes ...  I have a little  experience  with this  in other   application  domains not easy  to deal with  streams.

3.   Not   sure   I understand  "oblivious slicing"     did this work?

4.  I will    be there  in next GNU  Radio  Group meeting.

Final  thought ...

I watched the latest   Drake Lounge  on  You  Tube  this  eve ....  wow ...  really  enjoyed  the  discussion about  RFI    from  satellites.

Stephen

[Stephen Arbogast]

Very  sad to see funding   cut   for NCAR  and  NOAA in  Boulder,   Colorado..     I worked   at both  places as an undergrad   E.E.     back in the  1970's...  Very cool  places to work..

[Marcus D. Leech]

On 2026-02-15 23:23, 'Stephen Arbogast' via Society of Amateur Radio Astronomers wrote:

Some  thoughts......

1.    Stream tagging  approach.......   I have been  playing  with  this  but  since  Ted Cline gave up on  this approach  I am wondering...

2.    Marcus   Leech   mentioned  the problems  with the  streaming architecture  of  gnu radio...   Yes ...  I have a little  experience  with this  in other   application  domains not easy  to deal with  streams.

If you have a stream of samples that contain both "sky" and "reference" values, and the reference value is distinct enough in amplitude from "sky", then you take those samples,
  sort them, and use the mid-point as a "slice" to determine which samples are in the set "sky", and which are in the set "reference", then just subtract one from another.

To view this discussion visit https://groups.google.com/d/msgid/sara-list/0b03e424-55fa-43e3-903d-088fc3115fdfn%40googlegroups.com.