[SARA] AGC Control With Command Line rtl-sdr

[Jim Abshier]

I have been experimenting with using a NooElec R820T dongle as a
back-end for my 1420 MHz interferometer.  I am using the command line
software rtl-sdr running under Linux to record the 45 MHz IF signal from
my interferometer.  I can't find any documentation that shows how to
control the hardware AGC using the rtl-sdr software.  Is that even
possible?

I have made some runs that show fringes, but I don't know if the data is
being affected by any AGC action in the dongle.  I have attached some
images showing results with this rig, and the fringes look reasonable,
but I don't know if they have been affected by AGC.  The first attached
image shows a plot of the raw output for my interferometer operating in
simple adding mode.  The large change in baseline at the beginning
appears to be a warm-up gain drift.  The fringes from Taurus A appear at
about 5.5 hours local mean sidereal time.  The second image shows plots
of the fringe data after processing to remove the large total power
offset and baseline drift.  The lower plot is from high-pass filtering
and the upper plot is from additional low-pass filtering to remove the
high frequency fuzz.  These results seem reasonable, but I would like to
know if they have been affected by AGC.  It would be difficult to
calibrate this system if the gain is being affected by signal level. 
Can anyone on the list provide me with information on how to control
(turn off) AGC of these dongles?

Jim Abshier

[Michiel Klaassen]

Hi Jim,

Nice fringes; have you done some calculations on it already?

I do not see anyone answering, so I give it a try.

When you define the gain (g 24) then the AGC is switched off, and when you do not define the gain (no g xx)  then AGC is switched on.

I have used this for pulsar capturing, but now I use SHARP# for all the settings.

I that way I can check the result immediately and also check for RFI.

If everything looks ok, then SHARP# is terminated and I start cfrad2.exe.

Now 289 files are written to the disk each stacking 100000 spectra automatically.

With a python script the files are processed.

regards

Michiel

Op do 20 sep. 2018 om 22:45 schreef Jim Abshier <abs...@ameritech.net>:

--
--
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.
For more options, visit https://groups.google.com/d/optout.

[Jim Abshier]

Hi Michiel,

Thank you. Your answer is just what I was looking for.  I have been setting the gain to 0 (-g 0) so I guess the AGC was off.  Because I was digitizing the IF from my receiver, which is at about -20 dBm, I had to attenuate it by 60 db to avoid overloading the dongle.  That is why I set the gain to 0.  I didn't need anymore gain in the dongle. 

I have been using octave to control the recording and processing of the IF signal.  I call rtl_sdr from octave and only record one integration period of data at a time.  I process that chunk of data to generate one sample of fringe data, then repeat the process for a couple of hours.  This avoids generating very large IF data files.  I have tried generating large single files, but have encountered software failure problems when trying to record data over more than about 2 hours time.

I have very little experience in using these SDR dongles, so I am not familiar with all the software options available.  Thank you for the information on what you do for pulsars.  Perhaps I could try using SHARP#.

Jim Abshier