Call for GNU Radio Expertise

[Gary Mikitin, AF8A]

HamSCI Community - This is a call for expertise on GNU Radio software, particularly with spectrograms and waterfalls.

Short story:  the Grape 1 Low IF receiver system, set to either 2.5, 5, 10 or 15 MHz,  has been deployed to dozens of locations, and will be going to dozens more.  HamSCI plans on using GNU radio to process the received audio stream, format it in Digital_RF (DRF*) and then upload it to the Personal Space Weather Station (PSWS) server at the University of Alabama.  That part of the project is working, thanks to the efforts of Bill Engelke, AB4EJ (and perhaps others - sorry if I omitted anyone).

One critical element is missing:  A user friendly, easily adapted/modified display of the received spectrum.  Bottom line, users need to know what signals their Grapes are receiving, have some idea of signal to noise ratio, etc.  The most easily understood method (especially for citizen scientists) is a spectrum display, aka a waterfall.  (Example att’d from a completely different package, fldigi).  Who has experience with spectrum displays in GNU Radio, and some time to devote to implementing a solution (it will be running on a Raspberry Pi, 4gB memory)?

If you can help with this need, feel free to reply here or to contact me at gmikitinaf8a <at> gmail <dot> com

Thanks, all de Gary, AF8A

*DRF is a standardized HDF5 format for reading and writing of radio frequency data and the software for doing so. The format is designed to be self-documenting for data archive and to allow rapid random access for data processing. https://github.com/MITHaystack/digital_rf

[Art Botterell]

Gary, 

I’ll hold short of claiming expertise, but I’ve done a bit of fiddling with GNU Radio and displays in the past.

73,

Art KD6O

<Spectrum Display.png>

--
Please follow the HamSCI Community Participation Guidelines at http://hamsci.org/hamsci-community-participation-guidelines.
---
You received this message because you are subscribed to the Google Groups "HamSCI" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hamsci+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/df5c06a2-a726-4e00-b742-6b00aa24f2c7n%40googlegroups.com.
<Spectrum Display.png>

[David G. McGaw]

To be clear, it would be desirable to have a real-time spectral display in addition to the waterfall, to be able to see the noise floor and any interfering signals.

As I mentioned on the call this morning, we have Gnuradio flowgraphs for setting up our our research radios that include both.  We use USRPs, RTL-SDRs and RSPduos.

73,

David N1HAC
Dartmouth College

On 8/10/23 4:18 PM, Art Botterell wrote:

You don't often get email from a...@botterell.net. Learn why this is important

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/5847FE06-9503-4BDA-B18A-CA231440D735%40botterell.net.

[Art Botterell]

David,

Am I correct in understanding that what comes out of the Grape is a 1KHz IF, filtered rather narrow?  SNR and interference could be eyeballed even in that narrow window, I suppose.  And I imagine the HF antenna feed could also be split and sent to a separate wider-band SDR if a wider view was desired.

I also wonder whether a tool for visualizing a longer period, say, the past 24 hours, of Doppler variability might be helpful, or at least reassuring to the local operator.  Although maybe that should be a utility run on a laptop or whatever to leave the Pi’s workload as steady as possible.  Am I just describing a DRF file viewer?  Sorry, I’m new here.

Art KD6O

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/24a19042-139f-6f9a-a3b1-643b5f80fd5c%40dartmouth.edu.

[David G. McGaw]

I believe the sample rate is 8 kHz, so 4 kHz Nyquist, i.e. audio band.  John N8OBJ should confirm, but the output is filtered to 160 Hz to 1600 Hz.  It is not a sharp filter, so signals would be seen to either side.  The signal of interest is indeed at 1 kHz.

David N1HAC

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/36BCEA33-F236-4019-A30F-34E10DAA918F%40botterell.net.

[Bill Owens]

Is there a setup for GNU Radio to gather and log the DRF data, or is it premature to ask that? I happened to get a modern Pi4 which was said to not work with the old Grape OS image, so I put Ubuntu and GNU Radio on it instead, but haven't had the time to dive into the documentation and figure out how to construct a config yet. I'd like to put it in parallel with one of my original Grape machines.

Thanks,

Bill N2RKL

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/a787cecb-9298-c8d7-d94c-d5e452fc99c5%40dartmouth.edu.

[Marcus D. Leech]

My dance card is utterly overflowing, but I did build this GR- based application a couple of years ago, and we ran it at our
  field observatory site until an ice-storm destroyed our antenna, and then we had to move from that site...

https://github.com/ccera-astro/wwv_doppler

[Gary Mikitin, AF8A]

Bill - You've hit the proverbial nail on the head.  The HamSCI Grape Team is working towards just what you describe - a distributable image for a Grape 1/Raspberry Pi, running GNU radio creating DRF data.  Not there yet, progress is a bit slow, but it's always darkest before the dawn <grin>. Dawn WILL be here before October!

You are in grid FN13?  You are in a very good place to generate data during the eclipses, especially the total eclipse in 2024.  Please keep an eye on this list.  We'll publish details as they become available.  Also, running 'dual Grapes' is most helpful, as there are uses for both the legacy fldigi generated data and the soon to come DRF data.

73 de Gary, AF8A

[Michael Fene]

This is Michael Fene’; I would like to offer my services re your GNU Radio Request.

Kind Regards,

Michael Fene’

KI6HEF

Get Outlook for iOS

---

From: ham...@googlegroups.com <ham...@googlegroups.com> on behalf of Gary Mikitin, AF8A <gmikit...@gmail.com>
Sent: Thursday, August 10, 2023 4:37:01 PM
To: HamSCI <ham...@googlegroups.com>
Subject: Re: [HamSCI] Call for GNU Radio Expertise

 

--
Please follow the HamSCI Community Participation Guidelines at http://hamsci.org/hamsci-community-participation-guidelines.
---
You received this message because you are subscribed to the Google Groups "HamSCI" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hamsci+un...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/bd8f9a8b-f72b-48cb-a786-bb8c6d521e8fn%40googlegroups.com.

[Art Botterell]

Again, please forgive my newly questions.  What’s the life cycle of the DRF file / data stream?  How long a duration is represented in the upload files?  And can the files be read while they’re still partial?

73

Art

--
Please follow the HamSCI Community Participation Guidelines at http://hamsci.org/hamsci-community-participation-guidelines.
---
You received this message because you are subscribed to the Google Groups "HamSCI" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hamsci+un...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/bd8f9a8b-f72b-48cb-a786-bb8c6d521e8fn%40googlegroups.com.

[Jonathan]

Why can't fldigi be installed for diagnostic purposes? When setting up the hardware, you use fldigi to verify you have the proper signal, then start up the GNU Radio graph and set it up to start automatically on boot?

Can someone post the graph file so people can start investigating?

Michael KI6HEF,

The goal for this request is to generate a useful waterfall and spectrum display from a GNU Radio pipeline, like using the QT GUI waterfall sink and QT GUI frequency sink, for diagnostic and monitoring purposes. The pipeline takes an audio input from the Grape mixer board output, downmixes it to baseband, and produces carrier baseband data. The waterfall and spectrum display is used to verify the correct signal is being produced by all of the hardware components.

Another option is to use an external waterfall and spectrum application and have the data from GNU Radio feed it. I'm not familiar with any applications like this, but they might be out there.

Jonathan

KC3EEY

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/E4B7B774-F543-4843-9105-E3B64887922A%40botterell.net.

[Jonathan]

I found two good references to get started, here and here. 

Jonathan

KC3EEY

[Dr. Nathaniel A. Frissell Ph.D.]

Hi Jonathan,

 

That *might* be an ok short-term solution, but I think we really need to be visualizing the data as it goes through the actual data processing chain. I think the use of FLDigi could lead to misleading results.

 

All, I’ll be out until Monday. Have a great weekend!

 

73 Nathaniel W2NAF

 

From: ham...@googlegroups.com <ham...@googlegroups.com> On Behalf Of Jonathan
Sent: Friday, August 11, 2023 6:27 AM
To: ham...@googlegroups.com
Subject: Re: [HamSCI] Call for GNU Radio Expertise

 

Why can't fldigi be installed for diagnostic purposes? When setting up the hardware, you use fldigi to verify you have the proper signal, then start up the GNU Radio graph and set it up to start automatically on boot?

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/CAOY0kB2U3yXOumsXv78s%3DPcvxaJ149vVB9kDyZareMGjGr10UQ%40mail.gmail.com.

[Black Michael]

Does FLDigi's FMT mode work for this purpose?  That will give you real-time Doppler.

Mike W9MDB

[Jonathan]

Hi Nathaniel, All,

I would start with the mentioned QT GUI sinks. I remember others mentioning that some waterfall display plotting sink wasn't "capable", but there would be no details as to what would be missing or not working. Does anyone have more on that?

Jonathan

KC3EEY

On Fri, Aug 11, 2023 at 8:56 AM 'Black Michael' via HamSCI <ham...@googlegroups.com> wrote:

Does FLDigi's FMT mode work for this purpose?  That will give you real-time Doppler.

Mike W9MDB

--

Please follow the HamSCI Community Participation Guidelines at http://hamsci.org/hamsci-community-participation-guidelines.
---
You received this message because you are subscribed to the Google Groups "HamSCI" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hamsci+un...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/107139066.1683508.1691758564758%40mail.yahoo.com.

[Steve Cerwin]

Hi Mike,

The problem with FLDIGI is that it reports only a single answer for frequency per data point, based on strongest amplitude. During multipath propagation during times of varying solar flux, such as with simultaneous 1-hop, 2-hop, or more modes, the Doppler signature exhibits mode splitting where the Doppler trace splits into multiple frequency tracks. A true spectrogram records all the tracks which contains valuable information. Any frequency tracker that records only one frequency value per point loses this information. 

Attached are some plots that illustrate the point. The time of interest around 1200 UTC in these 5 MHz WWV recordings is the morning transition between night and day when the F2 ionization layer lowers from more than 300 km down to ~ 220 km. Because the different modes have different path lengths yet all are shortening by the same lowering ionization layer, the longer paths have a faster rate of change (velocity) and produce more Doppler. This is the working theory behind the observed mode splitting. It has been supported by analyses of the different Doppler tracks that produced results consistent with hmF2 data reported by ionosonde.The spectrogram recorded by Spectrum Lab clearly shows the mode splitting but FLDIGI and another frequency tracker based on an FM demodulator smeared the separate Doppler tracks into a wide noise-like vertical band. So that is why FLDIGI is not the best choice for Doppler analyses. It gives an incomplete picture of some interesting Doppler physics.

73, Steve WA5FRF

On Aug 11, 2023, at 7:56 AM, 'Black Michael' via HamSCI <ham...@googlegroups.com> wrote:

Does FLDigi's FMT mode work for this purpose?  That will give you real-time Doppler.

Mike W9MDB

--
Please follow the HamSCI Community Participation Guidelines at http://hamsci.org/hamsci-community-participation-guidelines.
---
You received this message because you are subscribed to the Google Groups "HamSCI" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hamsci+un...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/107139066.1683508.1691758564758%40mail.yahoo.com.

[Black Michael]

Gotchya'.

So do we need real-time display of such a waterfall or just recording of the audio for post-analysis or both?

Would real-time serve any purpose other than entertainment?

Mike W9MDB

[Phil Erickson]

Hi Mike,

  It has been our long experience at MIT Haystack that if you do not have some real time diagnostics available on an instrument, it quickly (and often silently) goes out of calibration or stops feeding data or other types of problems - and you find these problems out only later when it is too late to remedy them.  A real-time display dashboard is essential and is not pure entertainment.

73

Phil W1PJE

--

Please follow the HamSCI Community Participation Guidelines at http://hamsci.org/hamsci-community-participation-guidelines.
---
You received this message because you are subscribed to the Google Groups "HamSCI" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hamsci+un...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/568382172.1713313.1691765138884%40mail.yahoo.com.

--

----
Phil Erickson
phil.e...@gmail.com

[Black Michael]

Good point...

So is QT installed on the grape image?

Sounds like FLTK is  installed.

It is Linux, right?

Mike W9MDB

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/CAAZaqEvK36XrNpwsXfhH2b_jZAULhFuk6CpSrRK4-ZxhbZisCg%40mail.gmail.com.

[Jonathan]

Mike, 

I don't know if QT is installed on the image, but you can easily install it with apt-get and install the appropriate packages. The call out is for someone to do the development of the method to display a usable waterfall and possibly a spectrum plot.

Jonathan

KC3EEY

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/688735875.1705962.1691766257866%40mail.yahoo.com.

[Black Michael]

This has some good starting points

https://sourceforge.net/projects/qwt/

cpuplot is real-time with multiple items  plotted.

spectrogram is non-realtime but shows a spectrogram plot.

Mike W9MDB

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/CAOY0kB2SufbrE9cSNRM-V-xfTxhAvLr0G9uhs8MMcdUmd8xwYw%40mail.gmail.com.

[Michael Fene]

Jonathan,

Thanks for your prompt reply.

To-date I have only monitored the google-groups communications regarding this Hamsci effort.

I am an experienced SW / Sensor integration engineer...but will need to come up to speed to meet your requirements.

I would be happy to discuss over a telephone call.  

Kind Regards,  Michae Fene'

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/CAOY0kB2U3yXOumsXv78s%3DPcvxaJ149vVB9kDyZareMGjGr10UQ%40mail.gmail.com.

[Jonathan]

Michael,

Do you have any experience specifically working with GNU Radio? That is primarily what this call is looking for. The goal is to create a waterfall display and possibly a spectrum display for diagnostics and live, active monitoring of the signal on the Grape PSWS. It uses a GNU Radio graph that takes input from the audio input of the soundcard, resamples it to 8 kHz, then downmixes it to produce a baseband signal, then finally encodes the data in Digital RF format. Within the graph, they are looking to implement a waterfall display similar to the picture in the initial post of this thread.

Jonathan

KC3EEY 

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/CAPioHAbGT7aCwErtQievmiSOQKuVNKycephzYXGs3PDjs66Z1A%40mail.gmail.com.

[Derek Kozel]

Hi Gary,

I'd recommend looking at the CyberEther waterfall and spectrogram displays. They're available for use in GNU Radio through the gr-cyber module. Running on a Raspberry Pi is an explicit goal of the project. There's a chance that these displays will become available by default in a future GNU Radio release as well.

https://github.com/luigifcruz/CyberEther

https://github.com/luigifcruz/gr-cyber

Luigi Cruz will be at the GNU Radio Conference next month. Will anyone from the Ham Sci community be going? Sorry I don't have time to work on this myself at the moment, but if it's still in-progress in November I'd be happy to jump in and help out.

Cheers,

Derek

MW0LNA / K0ZEL

To view this discussion on the web visit https://groups.google.com/d/msgid/hamsci/CAOY0kB3KFSwu3rgMZ4d7YPBAQYn6nYEwWEUb4rOGodLXzKDzcg%40mail.gmail.com.

[Gary Mikitin, AF8A]

Hello All - In the interest of circling back on this topic, I'd like to offer a link to the new HamSCI/DRF_processing GitHub site.  This site hosts the Python code that reads and plots the Grape 1 DRF files.    https://github.com/HamSCI/DRF_processing

If you haven't seen the announcement or results of Bill Engelke's efforts in this area, take a look through the recent posts on the HamSCI-Grape group. The results are quite something.   https://groups.google.com/g/hamsci-grape/c/5eyE8arseZE (top most post)

73 de Gary, AF8A

On Saturday, August 12, 2023 at 2:04:36 PM UTC-4 Derek Kozel wrote:

Hi Gary,

I'd recommend looking at the CyberEther waterfall and spectrogram displays. They're available for use in GNU Radio through the gr-cyber module. Running on a Raspberry Pi is an explicit goal of the project. There's a chance that these displays will become available by default in a future GNU Radio release as well.

https://github.com/luigifcruz/CyberEther

https://github.com/luigifcruz/gr-cyber

Luigi Cruz will be at the GNU Radio Conference next month. Will anyone from the Ham Sci community be going? Sorry I don't have time to work on this myself at the moment, but if it's still in-progress in November I'd be happy to jump in and help out.

Cheers,

Derek

MW0LNA / K0ZEL

On Sat, Aug 12, 2023 at 4:21 PM Jonathan <emum...@gmail.com> wrote:

Michael,

Do you have any experience specifically working with GNU Radio? That is primarily what this call is looking for. The goal is to create a waterfall display and possibly a spectrum display for diagnostics and live, active monitoring of the signal on the Grape PSWS. It uses a GNU Radio graph that takes input from the audio input of the soundcard, resamples it to 8 kHz, then downmixes it to produce a baseband signal, then finally encodes the data in Digital RF format. Within the graph, they are looking to implement a waterfall display similar to the picture in the initial post of this thread.

Jonathan

KC3EEY 

The goal for this request is to generate a useful waterfall and spectrum display from a GNU Radio pipeline, like using the QT GUI waterfall sink and QT GUI frequency sink, for diagnostic and monitoring purposes. The pipeline takes an audio input from the Grape mixer board output, downmixes it to baseband, and produces carrier baseband data. The waterfall and spectrum display is used to verify the correct signal is being produced by all of the hardware components.