radiosonde_auto_rx v1.8.2 - Bugfixes, DFM17 & RS41-SGM improvements.

[Mark Jessop]

Hi all,

I've just released radiosonde_auto_rx v1.8.2!

This release brings the following major changes:

• Fixes a bug where on Debian/Raspbian Trixie (and probably any OS with GCC v14) the reported SNR was nonsensical and unchanging.
• KA9Q-radio support: The radiosonde detection step is now performed in parallel by default, with 4 parallel detectors running. Thanks @cartpauj for this addition!
• Vaisala RS41: Added a close_on_encrypted setting which if set to False allows auto_rx to continue decoding and logging telemetry (serial number, frame count, and SNR) from encrypted RS41-SGMs. By default this is set to True to avoid using up a receiver.
• Vaisala RS41: Tweaks to the demodulator settings, which seems to improve performance in weak signal situations.
• Graw DFM: Added a realtime filter to help reduce the amount of corrupt Graw DFM17 positions. Thanks @DB8LE for the addition!
• Intermet iMet: A new 'IMETWIDE' sonde type has been added (usable in the always_decode setting and web interface), which forces wide-band decoding for iMet sondes. May be useful to lock on to the increasingly rare wide-band iMet-1 and iMet-4 sondes.
• Web Interface: Various small updates and fixes (Thanks @danieldegroot2, @cartpauj, @penfold42 and @ke5gdb )
• Log Files: Fixed a bug where some log files would fail to be read in Numpy versions >2.3. Info on this bug here.
• RD41 Dropsonde: Some very initial support for RD41 dropsondes is in this version, but it doesn't appear to be working. Without a RD41 sonde on-hand this is very difficult to test. If anyone is willing to mail me a RD41 please contact me!

You can find out all about the changes in this release here:

https://github.com/projecthorus/radiosonde_auto_rx/releases/tag/v1.8.2

A big thanks to all the contributors that helped out with this version!

73

Mark VK5QI

[Bert Thuiswerkplek]

Thanks Mark!

I have been looking out for this release for the last couple of weeks.
Thank you for your hard work and this amazing software radiosonde_auto_rx!

73,

Bert PE1KZU

[Mark Jessop]

Just a note that I'm getting reports that M20 sondes are sometimes being mis-identified as M10s in version 1.8.2, resulting in a SDR being 'consumed' trying to decode a M10 sonde until the timeout kicks in. It looks like it will eventually detect the sonde as a M20 and decode correctly.

This looks to be a result of a change in dft_detect behaviour which we can hopefully get fixed soon.

73

Mark VK5QI