The M17 specification is being updated with the following changes, most of them breaking:
We have spawned a separate dev branch for the draft: https://github.com/M17-Project/M17_spec/tree/dev
The document is available for review below. Comments are welcomed!
We’ve just translated OpenRTX into Ukrainian and Polish.
Join the effort: https://hosted.weblate.org/projects/openrtx/openrtx/
Update: German, Czech, Portugese (Brasil), Spanish, and Swedish followed really fast. Thank you!
A simple DSP trick added into the most recent version of N1ADJ’s Go M17 gateway reduced MER down from 3.5% to under 2% (tested with a CS7000). A short IIR filter equalizes the frequency response of the CC1200 demodulator, improving the eye diagram. Additionally, the demodulator’s sensitivity has been adjusted for optimal M17 reception.
GitHub: https://github.com/jancona/m17
Silvano IU2KWO:
Demodulator fixes merged into master, next round of improvements for M17 decode available here: https://github.com/OpenRTX/OpenRTX/tree/m17-ber-threshold
The new demodulator improves M17 symbol timing and slicing. The next planned change will put a threshold on the decoded VIterbi output, based on its metric. That’s where the Codec 2 payload is. Erroneous data will not be played back anymore. This should improve overall user experience with the M17 mode.
Great job so far!
Vlastimil OK5VAS will be presenting LinHT at OpenAlt 2025 in Brno on Sunday, November 2, at noon.
Update:
Vlastimil’s talk was streamed live:
After receiving a new 10W 40dB attenuator, I decided to give the GRF5604 a spin again. I connected it to my LinHT prototype (Revision A hardware, for the record) and supplied with 5.1V. I’ve set the SX1255’s DAC gain to 0dB and mixer gain to -13.5dB, to prevent the power amplifier from saturating.
Here’s a spectrum plot showing harmonics with the fundamental at 36dBm (M17 signal at 433.475MHz):
Harmonics are below -60dBc, which I would call a good result (the output filter is a MiniCircuits LFCN-490+). The peak to the left is the PLL reference leakthrough (32MHz).
Stay tuned for more news.
We are making headlines again, this time with the LinHT – our revolutionary, open-source handheld radio running Linux. Make sure to grab CQ DL issue 11-2025 🙂
Please donate to keep our projects alive: https://www.paypal.com/donate/?hosted_button_id=4HTHZCS8UYPU6
Thank you!
This release adds a new function, sq_eucl_norm, and fixes the unterminated-string-initialization warning.
The squared Euclidean norm (fancy name for a simple sum of squared differences) offers some more robustness, as it doesn’t use sqrtf, which might be computationally expensive. Calculating pure L2 norm is not always required, as the similarity between two vectors can be estimated without using square roots.
Of course it does. We knew that a while ago!
While I’m trying to marry M17 text message decoder with SQLite database and Vlastimil OK5VAS is preparing a test PCB for the upcoming Revision B, Andreas OE3ANC prepared a nice fbDOOM demo. This is an important step, demonstrating the SoM’s capabilities. Stay tuned for more news regarding M17 support in our amazing device.
Extra: quick teaser – M17 message preview. This is not the final design.
Changes include GPS and RTC improvements along with some code clean-up.
See the GitHub release page for full changelog: https://github.com/OpenRTX/OpenRTX/releases/tag/v0.4.2