Radiosonde_Auto_RX

[John Smith]

Today I dug out an old pi B rev 2.1, an RTL stick and managed to install Auto_RX.

All seems to work and in the afternoon here in the UK, I managed to lock onto a sonde.

I stayed up late for the midnight launches, and was in a position to try and track 4 different Radiosondes, however my setup seems to pick up one of them and stay locked onto it, is this normal?

To word my question another way, is it possible for auto_rx to lock on to a sonde and upload for say 10 seconds, then scan to the next sonde and upload for 10 seconds etc etc?

Regards

Bozzey

[Mark Jessop]

In short, no. auto_rx was written with the intent of locking onto a sonde, and tracking it for as long as possible, giving the tracker the highest chance of getting a landing position and being able to recover the sonde.

The feature you describe is noted in this github issue here: https://github.com/projecthorus/radiosonde_auto_rx/issues/72  However, I haven't had time to work on this. 

You can track multiple sondes simultaneously by adding more RTLSDRs. There is also work in progress to point auto_rx at a 'SpyServer', which will allow multiple sondes to be tracked with a single RTLSDR or airspy, to the limit of the SDRs receive bandwidth. This will likely not work reliably on anything short of a Pi 4 though.

73

Mark VK5QI

--
You received this message because you are subscribed to the Google Groups "UKHAS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ukhas+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ukhas/b3ef5d22-b282-43b5-ab0a-15feafd02512%40googlegroups.com.

[John Smith]

Thanks for the reply.

Problem I am having today is that I am being bombarded with RS92's that are all coming my way (In York UK), and all of them are transmitting on the same frequency, as they are so close to each other they interfere.

Result is auto_rx locks onto them, cannot decode due to interference, so missing any other balloons that could be decoded.

I was interested in the dirty hack posted by pit711  (hacked code --> https://pastebin.com/raw/YGBhAvMX )

But looks like this was based on code from June 2018 so I am failing to manually patch this into 1.2.0  as a lot seems to have changed under "if autorx.scan_results.qsize() > 0:" and I am not a real coder.

Example below picture, I admit this doesn't normally happen here though, but is doing today.

Regards

Bozzey

[Ben Z en de rest]

The good hack would be to add 405.100 to your blacklist. That way it's skipped after the scan.

Regards,

Ben

Op ma 28 okt. 2019 om 16:36 schreef John Smith <satsea...@gmail.com>

--
You received this message because you are subscribed to the Google Groups "UKHAS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ukhas+un...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/ukhas/15c516b4-b782-4b47-809b-af77b11bc293%40googlegroups.com.

[John Smith]

Thanks for the reply Ben

I think I may have another issue now. They have stopped the serial launches now, so a few minutes ago, there was just 1 RS92 in reception and still failed to decode.

INFO	Decoder #0 RS92 405.099 - Closed decoder subprocess.
2019-10-28T16:39:18Z	ERROR	Decoder #0 RS92 405.099 - RX Timed out.
2019-10-28T16:36:18Z	INFO	Decoder #0 RS92 405.099 - Starting decoder subprocess.
2019-10-28T16:36:17Z	INFO	GPS Grabber - Ephemeris downloaded to ephemeris.dat successfuly!
2019-10-28T16:36:16Z	INFO	Decoder #0 RS92 405.099 - Using experimental decoder chain.
2019-10-28T16:36:13Z	INFO	SDR #0 has been allocated to Decoder (RS92, 405.099 MHz)

So not sure what my problem is, could my old pi B rev 2.1 not be up to it for RS92's? Seems to work fine for RS41's

Still, would be nice to see something like the dirty hack implemented.

Regards

Bozzey

[Mark Jessop]

Should be no difference in CPU usage between the different sonde types I think. A Pi 2B should still be OK, to decode on, with a single RTLSDR anyway. 

It is worth noting that auto_rx will usually *detect* the presence of a sonde at a lower SNR than a valid decode is possible, so it's possible the signal was just a little bit too weak.

Let me know if you keep having problems, and I can help debug what's going on. Probably best to take that off-list though!

73

Mark VK5QI

--
You received this message because you are subscribed to the Google Groups "UKHAS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ukhas+un...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/ukhas/0f52e4a6-0a68-4ae4-b8f1-5098fd407fb4%40googlegroups.com.

[John Smith]

Thanks for the reply.

It just so happens that the same place is sending up RS41's again today, so I have tried again and still failing to decode with pi2.

So, just setup a pi4 with a clean install the same way I did the pi2, booted up and it decoded the RS41's without any issues.

Can only assume an issue with RS41's on a pi2?

Regards

Bozzey

[David Brooke]

On Tue, Oct 29, 2019 at 10:51:57PM +1030, Mark Jessop wrote:
> Should be no difference in CPU usage between the different sonde types I
> think. A Pi 2B should still be OK, to decode on, with a single RTLSDR
> anyway.

I'd have expected a 2B to cope, I have one doing ADS-B and Open Glider
Network with 2 SDRs.

My auto_rx runs on a 3B with one SDR and I notice that the new auto_rx
release is using more CPU, albeit still well within its capability. You
can see when I upgraded on this graph:

https://stuff.dbrooke.me.uk/sp-iPFXYE.png

David

[John Smith]

Hi David

I did try it again today, also updated the pi2 to version 1.2.2 of auto_rx, but still fails to decode RS92's, works fine with RS41's

I ran top on the pi2 and it peaks at around 82% processor when it try's to decode an RS92.

Switched back to my pi4 and again works perfectly.

Regards

Bozzey

[Mark Jessop]

It's most likely the new FSK demodulation chain which is using more CPU. You could try disabling that by setting the 'experimental decoder' settings near the bottom of the config file to False.

If this does fix it, let me know and I'll make a note in the instructions for people running on lower powered boards.

73

Mark VK5QI

--
You received this message because you are subscribed to the Google Groups "UKHAS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ukhas+un...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/ukhas/05380a09-c483-435b-bb9e-eb2d0a59d8d8%40googlegroups.com.

[John Smith]

Mark

To be honest, I would rather take the hit that the pi2 is not up to it.

Your software is a cracking bit of kit, which I fully appreciate, worked great today on the seldom launches of RS'92's in the UK

Would rather you could look at the option (if possible) to allow scanning of other sondes and to not lock onto a single one.

I have read this is more or less possible with the revised web interface changes, but I must be missing something, as I have enabled the extra web interface and cannot see anything different?

Most probably me doing something wrong, as I am very new to Hab.

Regards

Bozzey

[Mark Jessop]

Adding in the ability to jump between sondes is a fairly big change to how auto_rx operates at the moment, and will require a lot of work and testing.

It would also result in gaps in tracking data. I would prefer to put effort into ways of allowing tracking of multiple sondes with a single SDR, which is progressing at the moment (though it will take a lot more CPU).

73

Mark VK5QI

--
You received this message because you are subscribed to the Google Groups "UKHAS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ukhas+un...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/ukhas/422b59cd-d097-4e6a-b2cd-baf69a46afbd%40googlegroups.com.

[PE2BZ]

John,

I think what Mark suggest would be a good solution, with one Pi and RTL device you can run a spyserver with 2 MHz bandwidth, in which all common UK sonde frequencies fit. Center at 404 MHz will catch 404.2, 404.4, 404.800 and 405.100/ 

With another Pi you can run the version (which is in development, not released yet) which detects radiosondes in the 2 MHz spectrum and can decode multiple sondes at once.

73'

Ben

To unsubscribe from this group and stop receiving emails from it, send an email to uk...@googlegroups.com.

[John Smith]

Hi Mark

Just for feedback

I have disabled all the experimental decoders as suggested, CPU usage has significantly dropped to around 35% and my pi2 is now decoding the RS92's with no problems. 

Regards

Bozzey

On Tuesday, 29 October 2019 20:58:48 UTC, Mark Jessop wrote:

To unsubscribe from this group and stop receiving emails from it, send an email to uk...@googlegroups.com.