FYI: Old pcb (RSM421)versus new ones issues.

[py2uep]

Hi Nevvman18,

First of all, thank you for your work on the RS41-NFW firmware.

I'm reporting an issue I encountered when flashing RS41-SGP sondes with the newer RSM424 board revision. When running RS41-NFW on these units, the Horus telemetry is not decoded, and it appears there are timing problems during transmission. The sondes do power up and transmit, but the signal is unstable and does not follow the expected timing used by the decoder.

When I use the exact same firmware on older RS41-SGP sondes equipped with the RSM421 board, everything works perfectly — Horus packets decode normally, and the timing is correct. This suggests the issue is specific to the newer RSM424 hardware revision.

Both sonde models I tested are RS41-SGP (with the pressure sensor). I have repeated the flashing process several times, trying different configurations, and the results are consistent. Additionally, a friend of mine performed the same tests independently in another location and obtained identical results.

Please let me know if I can assist further by providing recordings, logs, PCB photos, or performing any additional tests that may help identify what differs on the RSM424 boards.

Thank you in advance,

Best regards,
Demilson, PY2UEP

[obi7zik]

I’m also seeing a similar issue, but on a sonde with the RSM414 board revision.

This unit was recovered in very poor condition (heavy rain exposure, possibly run over, then left in direct sun for about 74 days), but it still powered up and I was able to flash RS41-NFW successfully. After flashing, it does transmit, but the timing seems unstable and decoding is unreliable, similar to the behavior reported on the RSM424 boards.

[py2uep]

My friend PU2SWX compiled the firmware using the Fastest (-O3) optimization with LTO, and the Horus v3 transmissions ran smoothly on an RSM424 board. However, a new problem was discovered: the onboard pressure sensor isn’t working. It only worked on the older board.

[darksidelemm]

This may be related to #47

RS41-nfw appears to use the internal RC oscillator, not the crystal. If the RC oscillator is off-frequency, it could result in degraded decode performance for Horus and APRS modes.

I also note that the 4FSK modulation is done using delays - if there's anything else interrupting the code, that could mess with symbol timing. That might explain why compiling with a different optimization level improves things. Ideally 4FSK modulation should be done using a timer interrupt for more accurate symbol timing.