Request: Command Line INI/CFG Options for multiple TCP_SMEM instances

[azwirko]

Description of the setup:

• Device: STEMLAB 125-14 (two or more simultaneously)
• SD card image: 20241222
• Application: sdr_receiver
• Other relevant information: smem_tcp with intf_smem and SkimSrv integration

Description of the problem/request:

I currently run CWSL_Tee https://github.com/HrochL/CWSL with two instances of SkimSrv where each one has its own install directory (SkimSrv, SkimSrv2) and each directory has its own copies of CWSL_Tee.dll/CWSL_Tee2.dll and CWSL_Tee.cfg/CWSL_Tee2.cfg file that allow pointing to two different Redpitaya MAC addresses. A Windows batch file starts the entire system running from the command line including Aggregator without any other user interaction.

First issue is TCP_SMEM seems to require a GUI interaction (Connect button press). Second it not does seem to allow for running two instances with two different IP addresses. I copied the EXE and DLL files each to unique folders and names. I start the EXEs and enter different IP addresses. I start two different instances of SkimSrv but the last SkimSrv app seems to cause the frequencies in both TCP_SMEM UIs to be the same. In the SkimSrv apps the drop down only shows a single SMEM 0 while in CWSL_Tee the dropdown options names are unique.

Is it possible to have an INI or CFG file added to TCP_SMEM to provide for the configuration of two or more instances and to start without GUI/user interaction and such that each has its own memory space addressable by two different SkimSrv instances?

Thansk & 73!

andyz - K1RA

Single instance of TCP_SMEM (showing second directory for SkimSrv2 and second TCP_SMEM, not yet running)

Trying to run two instances of TCP_SMEM where first instance is overridden by values of second running SkimSrv instance

[pavel-demin]

Hello Andy,

Thank you for testing the SDR receiver application. This is still a relatively new application that is used by very few people and I appreciate all the feedback.

It is already possible to use multiple SMEM instances. There is an example showing how this can be achieved at this link. By default, the SMEM index is 0. To use another index, it must be appended to the file names of programs and libraries. For example, tcp_smem_1.exe, intf_smem_1.dll, etc.

I will check later if it is possible to start the connection automatically when the tcp_smem.exe program starts. Hopefully it is as simple as calling Connect right after Init. Otherwise it would be easier for me to add a non-GUI program with all parameters passed via command line arguments.

All settings are stored in the registry and at the moment I would prefer not to have configuration files.

Best regards,

Pavel

[azwirko]

Thanks Pavel - I'd been following these instructions https://pavel-demin.github.io/red-pitaya-notes/sdr-receiver/ as I didn't have an 122-16, but I tried them and it works as expected with two 125-14s. I'm now trying to understand the cat_smem, file_smem and wave_smem apps you've written to see how they might be used to emulate something similar to CWSL_Digi https://github.com/alexranaldi/CWSL_DIGI that I'd used in conjunction with CWSL_Tee. Your SMEM system seems to have fewer requirements than CWSL and is seems more streamlined. Thanks for your efforts!

andyz - K1RA

[pavel-demin]

I think that it would be quite difficult to have CWSL_DIGI functionality with the current state of this project.

I can think of the following approach to reimplement the CWSL_DIGI functionality:

• add dump_smem.exe that would write I/Q samples to a file
• write a program in Python that would
  • periodically run dump_smem.exe
  • preprocess I/Q samples using numpy
  • call the decoders (wsprd and jt9)
  • upload results to wsprnet and pskreporter

[azwirko]

Thanks for the ideas. Can you provide any high level help or example usage for your file_smem, cat_smem and wave_smem apps? I was wondering if any of those could be leveraged with the WSJT-X jt9 command line decoder. I was thinking if any of your apps produced WAV files that might be another means to perform the decoding with a Python wrapper and then parse the text decode output and upload to PSKr and WSPRnet which I've done in the past.

[pavel-demin]

Can you provide any high level help or example usage for your file_smem, cat_smem and wave_smem apps?

wave_smem.exe is used to send I/Q samples from one of the SMEM I/Q channels to CW Skimmer via a virtual audio device and cat_smem.exe is used to control CW Skimmer from Thetis. These two programs can be used as shown in this video.

file_smem.exe reads I/Q samples from a file and sends them to a virtual audio device and a SMEM I/Q channel. I used it to simultaneously test the dynamic ranges of CW Skimmer and CW Skimmer Server.

I was wondering if any of those could be leveraged with the WSJT-X jt9 command line decoder.

Implementing the CWSL_DIGI functionality was not the goal of this project when I worked on it last year.

I could add dump_smem.exe and Python code to convert the I/Q samples to WAV files for all the sub-bands shown in the table at this link.

[azwirko]

Appreciate the explanation. I didn't expect CWSL_Digi functionality, but was curious to what extent any of your other tools provided mirrored the CWSL_Wave, CWSL_File, CWSL_Net, capabilities.

So now that I have a working SMEM, I guess the only request is to auto connect on startup if possible as I understand configuration is stored in the registry. That would allow me to script starting the system without any other user interaction.

Thanks again & 73.

andyz - K1RA

[pavel-demin]

Calling Connect right after Init seems to work and starts the connection automatically when the tcp_smem.exe program starts.

Here is a link to the pre-built SMEM binaries with this modification:

https://www.dropbox.com/scl/fi/kto4dgsn1dh58saohc5r6/sdr-smem-20241229.zip?rlkey=8effibpq4r582wepjee7cs8d0&dl=1

[azwirko]

Thanks for the quick turn around. Works great. Happy New Year.