First BAA GNU Radio for Radio Astronomy meeting video now available online

[andrew....@googlemail.com]

For anyone who is interested, this video can be accessed at:

 

BAA RA Section GNU Radio meeting #1 - 7th Oct. 2025

 

If you would like to receive a meeting link for the next meeting please send me an email on andrew (at) thornett (dot) net

 

Andy

[Stephen Arbogast]

Andy,

I   watched  this  You Tube   video...  very  cool.   Totally  slipped my mind... we are flooding  here in South  West  Colorado......   I have  two  comments....

1. DISPIRA  while  very cool,   I have been  warned by  Marcus  Leech  the  three stage  low pass  filter  is  over  kill,    costs processing  time.  A  single   stage IIR  Low  Pass Filter  is sufficient.   One must  be careful of  low  value of  alpha in the  the filter .... can  lead  to   errors   in  numerical  processing.. must adjust  it  for  good  results.

2.  Ted  Cline   mentioned    he  is  investigating Tags in  GNU    so I   started playing with  them.   I think  this might  work..  here is a  link   https://wiki.gnuradio.org/index.php/Stream_Tags

I am  using    this  software   with success  but  it  does not  do  what  Ted  Cline's ezRACol  does  ..  collect samples for background  reference...     https://github.com/Radio-Source/GSR

If  anyone  decides to try  it  I was  having  success  with a  split_ratio of  10k.  Here is my  .grc  file    Modify it for  your  SDR   My  observations   have been corrupted  by lightning and  thunder...

Stephen

[Stephen Arbogast]

I am  running   Ubuntu 24.04.3  on  mini pc  Beelink  I bought  from Amazon...  nothing  exotic.

[Marcus D. Leech]

On 2025-10-14 23:03, 'Stephen Arbogast' via Society of Amateur Radio

Astronomers wrote:
> Andy,
>
> I   watched  this  You Tube   video...  very  cool.  Totally  slipped
> my mind... we are flooding  here in South West  Colorado......   I
> have  two  comments....
>
> 1. DISPIRA  while  very cool,   I have been  warned by Marcus  Leech 
> the  three stage  low pass  filter  is  over kill,    costs
> processing  time.  A  single   stage IIR  Low Pass Filter  is
> sufficient.   One must  be careful of  low value of  alpha in the  the
> filter .... can  lead  to   errors  in  numerical  processing.. must
> adjust  it  for  good results.

Numerical instability.  It shows up in various types of mathematical
algorithms implemented on finite-precision hardware.  Like, you know,
computers :) :)

>
> 2.  Ted  Cline   mentioned    he  is  investigating Tags in GNU    so
> I   started playing with  them.   I think  this might work..  here is
> a  link  https://wiki.gnuradio.org/index.php/Stream_Tags
>
> I am  using    this  software   with success  but  it  does not  do 
> what  Ted  Cline's ezRACol  does  ..  collect samples for background 
> reference...  https://github.com/Radio-Source/GSR

I'll note that baa_seminar has a "subtract background" feature. But,
really, this can be done post-facto, which is how many observatories do
this in the real world.

[Stephen Arbogast]

Yes..   very  cool......   I  am  experimenting   with    curve fitting .....       hope Discovery  Dish   releases   there az/el  control soon   would  love  to  play  with it  ...

[Marcus D. Leech]

On 2025-10-14 23:03, 'Stephen Arbogast' via Society of Amateur Radio Astronomers wrote:

Andy,

I   watched  this  You Tube   video...  very  cool.   Totally  slipped my mind... we are flooding  here in South  West  Colorado......   I have  two  comments....

1. DISPIRA  while  very cool,   I have been  warned by  Marcus  Leech  the  three stage  low pass  filter  is  over  kill,    costs processing  time.  A  single   stage IIR  Low  Pass Filter  is sufficient.   One must  be careful of  low  value of  alpha in the  the filter .... can  lead  to   errors   in  numerical  processing.. must adjust  it  for  good  results.

2.  Ted  Cline   mentioned    he  is  investigating Tags in  GNU    so I   started playing with  them.   I think  this might  work..  here is a  link   https://wiki.gnuradio.org/index.php/Stream_Tags

I am  using    this  software   with success  but  it  does not  do  what  Ted  Cline's ezRACol  does  ..  collect samples for background  reference...     https://github.com/Radio-Source/GSR

If  anyone  decides to try  it  I was  having  success  with a  split_ratio of  10k.  Here is my  .grc  file    Modify it for  your  SDR   My  observations   have been corrupted  by lightning and  thunder...

Stephen

Stephen:

I'm not sure I understand why you're changing "split_ratio", which is used both in BAA seminar and GSR to determine the "split" in the integration function between the
  "Integrate" function (which is a compute hog) and the "Single Pole IIR" which has numerical stability issues when processing very very small inputs with very very small
  values of 'alpha'.   Unless you're asking for stupidly-long integration times (10s of hours), 'alpha' should always stay reasonable.

On Monday, October 13, 2025 at 6:19:52 AM UTC-6 andrew....@googlemail.com wrote:

For anyone who is interested, this video can be accessed at:

 

BAA RA Section GNU Radio meeting #1 - 7th Oct. 2025

 

If you would like to receive a meeting link for the next meeting please send me an email on andrew (at) thornett (dot) net

 

Andy

--
--
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/b1626bfc-0207-4573-84e5-e3f0b80e4ee6n%40googlegroups.com.

[Stephen Arbogast]

Marcus,

No  particular reason  for  setting  split_ratio = 10k   other  than  experimenting.   However,  I have  found that I need a value  greater than the  default  of  100  to  avoid  numerical errors in the  IIR  Single  Pole Low  Pass  Filter.  

Some  info if others are  interested,  includes  filter design   software....      

 

https://tomroelandts.com/articles/low-pass-single-pole-iir-filter

[Marcus D. Leech]

On 2025-10-16 19:13, 'Stephen Arbogast' via Society of Amateur Radio Astronomers wrote:

Marcus,

No  particular reason  for  setting  split_ratio = 10k   other  than  experimenting.   However,  I have  found that I need a value  greater than the  default  of  100  to  avoid  numerical errors in the  IIR  Single  Pole Low  Pass  Filter. 

Interesting.  So some questions:

   what receiver?
   what sample rate?
   what integration time?

Also, what type of PC?   x86?  ARM?

To view this discussion visit https://groups.google.com/d/msgid/sara-list/d5efb9cd-463a-4447-a09a-b90ddf4b2062n%40googlegroups.com.

[Stephen Arbogast]

Receiver is AirSpy  R2

Sample  Rate is  2.5 Mhz

Integration  time is  I don't  know.   I am still experimenting with  integration  count  of  10K    I don't  believe  the  60 seconds shown in the  gui..

My  PC is  https://www.amazon.com/dp/B0C2P486GQ?ref_=ppx_hzsearch_conn_dt_b_fed_asin_title_1&th=1

I will  try and post  some   screen shots but  at the  moment experiencing    RFI  from  neighbors' heat   pumps  and  inverters.  Tonight  is first night  we  will  drop  below  freezing  here.

[Marcus D. Leech]

On 2025-10-16 20:44, 'Stephen Arbogast' via Society of Amateur Radio Astronomers wrote:

Receiver is AirSpy  R2

Sample  Rate is  2.5 Mhz

Integration  time is  I don't  know.   I am still experimenting with  integration  count  of  10K    I don't  believe  the  60 seconds shown in the  gui..

My  PC is  https://www.amazon.com/dp/B0C2P486GQ?ref_=ppx_hzsearch_conn_dt_b_fed_asin_title_1&th=1

I will  try and post  some   screen shots but  at the  moment experiencing    RFI  from  neighbors' heat   pumps  and  inverters.  Tonight  is first night  we  will  drop  below  freezing  here.

How does the "numerical issue" manifest in your case?

To view this discussion visit https://groups.google.com/d/msgid/sara-list/41fed8a3-e54b-40b5-8245-58974f2eea66n%40googlegroups.com.

[Stephen Arbogast]

From Computational Physics   by Steven E. Koonin.......    see chapter  1.  This  is  a  very old  text   but  the principles are the same. I have   a Jupyter   notebook  on another  computer  that  demonstrates this.

I have  observed  that  total power jumps from  0 to 1.5  randomly  but  when  I  increase the   split_ratio  these  jumps   decrease  and smooth  out.   I have  compared to  some  test  software  (not  for  release)  that   Ted  Cline  has  sent  me. His preliminary software  uses an AirSpy R2 and  works   fine  at  2.5  Mhz   sample  rate  with integration  count   of 10 k

And by the  way...    Very cool  work to Eduard  Mol.....  I am  73  with  Ankylosing   Spondylitus   ..  so  won't be climbing   ladders soon.

[Stephen Arbogast]

Here   is  a  screen   shot    at  1:00 am  from Durango. Colorado, USA   33 degrees   F........

[Stephen Arbogast]

Cool   stuff