M20 slow decoding ?

[Patrick EGLOFF]

Hi all,

Our local Weather station has recently changed the sonde type from M10 to M20.
Since, I noticed some differences in the detection around launch time :

- no detection while the sonde is still waiting in the launch container. (my RX is line of sight)

- late detection only after some minutes after launch.

I was suspecting a problem with my remote receiver as it was not ON since a few weeks to to a PSU failure.

Being curious, I went to the launch site with a receiver ana a portable setup.

It confirmed that the behaviour is the same as with the M10.

- Power ON of the sonde at 1100

- Launch at 1115

I looked at the remote scan plot and saw that there is a signal a few seconds after the launch when the sonde is getting out of the container.

The datas are displayed much later.

It can be that the M20 have a lower output power that could explain the lack of detection while in the container.

But the late data decoding is puzzling me.

How long should it take to get a stable decoding after a signal has been detected ?

Is it slower with a M20 compared to a M10 ?

73 Patrick TK5EP

[Mark Jessop]

Not having M10s or M20s in my area, I'm not really sure if or why there would be much difference in decode times.

I am aware the M20s have a habit of failing in flight due to poor quality batteries...

I was a little curious about whether frequency offsets could be doing anything (with the fsk_demod demodulator, this is pretty unlikely), so I had a look at the data from a recent flight: https://grafana.v2.sondehub.org/d/bbaa7894-e5f4-4c0d-be96-897b4ffde43b/radiosonde-telemetry-dashboard?var-Serial=310-2-02291&from=1732360688000&to=1732368901000&orgId=1

I do see your station reporting maybe 4 kHz above a few other stations... but then there's also 'AutoRxpi2b' that is 25 kHz off frequency... so who knows what is actually real here.

If your local launchers uses a consistent frequency, you could try setting up an 'always on' decoder, which constantly listens on the known frequency:

https://github.com/projecthorus/radiosonde_auto_rx/blob/master/auto_rx/station.cfg.example#L136

(It does say this requires a multi-SDR or network-SDR decoder setup, but i *think* it might work with a single RTLSDR)

That way you skip the scan/detection time, and should be decoding packets as soon as it's ready. 

Finally, it might be worth watching your auto_rx log at launch time, to see if auto_rx is reporting anything wrong with the telemetry in the early stages of the flight.

73

Mark VK5QI

--
You received this message because you are subscribed to the Google Groups "radiosonde_auto_rx" group.
To unsubscribe from this group and stop receiving emails from it, send an email to radiosonde_auto...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/radiosonde_auto_rx/2918e1db-6ba6-4f2a-9f5c-912452da1073n%40googlegroups.com.

[Patrick EGLOFF]

Thanks Mark,

I've checked different things and came to the conclusion that there is no problem with the M20 decoding.

I found a ham working at Meteo France who has helped during the switch from M10 to M20.

He told me that the M20 have a lower output power 100 mW compared to the 200 mW of the M10.

The shelter has also been changed from a modular building to a marine container, which is completely metallic. 

The radiated signal from inside the shelter is far lower than before.

I've forced the detection on the single frequency and it worked more rapidly than the days before.

Concerning the frequency, I tried to calibrate my SDR using both methods found on your site. The LTE method didn't give me usable datas and the GSM one never compiled.

73 and thanks,

Patrick TK5EP