Does gr-difi work with the Kratos SpectralNet Narrowband (VITA-49) as a source?

[evba0101]

Hi all,

I hope this is the correct place to ask a question!

Can I use the gr-difi source block to ingest VITA-49 from a Kratos Narrowband SpectralNet? Or do I need to request that the SpectralNet be configured for DIFI rather than VITA-49?

Thanks!

[777arc]

I'm pretty sure what Kratos calls VITA-49 is DIFI, it's just the older DIFI 1.0 version.

[evba0101]

Interesting! Thanks for your reply. In my hardware quote, they state the following:

Please note, this device supports a VITA-49 data stream, if a DIFI compatible device is required a
different product will need to be quoted.

I will ask them to clarify which exact standard is being used. If it is indeed DIFI 1.0 as you suggest, would it then be compatible with your gr-difi block? Thanks!

[777arc]

Pretty sure it works with both kratos narrowband and wideband, that's why there are parameters like this - https://github.com/DIFI-Consortium/gr-difi/blob/main/grc/difi_sink_cpp.block.yml#L73 but I personally can't verify, they would have to be the ones that verify it will work. It would be nice to have Kratos involved in this out of tree module :)

[ChristianSchindler]

Hi @evba0101,

I am also currently working with the Kratos SpectralNet Narrowband. At the moment, I can get data from the devices (so if I get data into GNU-Radio with DIFI Source), but I have not yet found a valid configuration for DIFI Sink. Did You find something that worked?

@777arc I also trayed context_pack_size with ['72', '108'], but both did not yet work for me (I am currently only getting lots of 'Ignored Packets` but now real data)

This Is My current Configuration for the Sink:

My working configuration for the Source:

What could be interesting is that I am getting for the Source the following Warning:

source_cpp :warning: The context packet size is not 108 bits per DIFI spec. The context packet received size is: 84

[777arc]

The ignored packets issue is the classic symptom of the MTU not being large enough, so it is blocking data packets, see if that's the problem

[ChristianSchindler]

Hi @777arc:

I tried the samples_per_packet of 2 , 10, ..., 500 and non if it worked

GNU Radio -> SpectralNet:

When I communicate with 2 dives, it looks like this as reverence (SpectralNet -> SpectralNet):

And I am not creating so mutch data:

[777arc]

Regarding the source_cpp :warning: The context packet size is not 108 bits per DIFI spec. The context packet received size is: 84 I think things should still be fine, the IQ samples will definitely be fine but the important part of the context packets should also be OK and that info isnt really used in GNU Radio anyway.

For the Sink issue, your params look good, I would leave it at 128 samples per packet. The only things I can think of to try is 1) in the SNNB RF Output tab I think there are settings on the left related to the buffer, it's possible there's too much jitter or delay and the SNNB is ignoring things or 2) instead of standalone mode you can try the other mode, but that's more of a hail mary.