Fwd: LF: Auto-EbNaut reception on LF in JN80

[Markus Vester <markusvester@aol.com>]

...guess this was intended to go to the reflector. Best 73, Markus


-----Ursprüngliche Mitteilung-----
Von: Domenico IZ7SLZ <iz7slz.domenico@gmail.com>
An: Markus Vester <markusvester@aol.com>
Verschickt: Do, 13. Jul 2017 1:24
Betreff: Re: LF: Auto-EbNaut reception on LF in JN80

Hello  LF,

EbNaut auto-decoder is still online at http://qsl.net/iz7slz/
but with some modifications.
Thanks to Stefan and Markus for their kindly suggestions.

For impatient people   there are some links where  the results of last EbNaut decoding (maybe still in progress) and the squared spectrum plot, are updated every five minutes.
(tnx to Markus's 'show_rawsyms3b.exe' program).

A third frequency has been added. The decoder is also scheduled to start at XX:30 with 6 char, 2s 8K19A on 137.485 kHz. This new decoding process starts about at XX:53.

Thanks to all operators whom have tested this system. If someone has tested without having a decode can contact me for a deep investigation in the raw audio recordered in these days.

I want attach the linux shell command's file that is running here. I'm not an expert in Linux. Maybe some skilled operator  can check and improve the operations.

73 to all

Domenico, IZ7SLZ

 

On 9 July 2017 at 11:09, Markus Vester <markusvester@aol.com> wrote:

Hi Domenico,

many thanks for installing your auto-decoder! Due to frequent thunderstorms I don't transmit much around this time of the year, but this weekend looked ok so I decided to push out the TX antenna for some tests. I started experimenting at 11 UT yesterday but initially had some problems with the phase feedback and a new compensation for the 0.3 s soundcard delay. The first "good" transmission went out at 12:59:57.1, which was 2.1 s early due to an NTP error (you might look for that one in your recordings). The next one at 14:00 was decoded successfully. Repeats at 15, 16 and 19 UT should have gone out ok but were apparently lost in the QRN. A different message was sent at 20, 21 and 22 UT (as you know ;-).

I am using a different phase lock scheme now, omitting the need for an Rubidium controlled LO. The audio signal for the SSB transmitter is generated at 1990 Hz as usual. But phase feedback is done directly from the RF signal, using soundcard subsampling to 137.49 - 96 = 41.49 kHz. This is possible because my Delock USB soundcard-dongle has only ~ 30 dB alias rejection at LF, and easily picks up the strong transmit field. Thus 1pps on the other channel allows directly sensing and phase-locking the radiated RF signal. I am actually running the TX instance on the VLF grabber PC (a cheap ODYS Cosmo Win-X9 tablet), reusing the soundcard connected to the VLF-receiver.

During three transmissions (Sat 16 UT and Sun 8 and 9 UT), I have simply disconnected the Rb, so the LF TRX crystal is free-running about 0.08 Hz low. Your recent decode shows that the RF-to-audio PLL successfully managed to compensat the LO offset and drift. 

Stefan if you like you could perhaps send some rawsyms plots, as I would be interested in the performance of the new PLL.

Perhaps one suggestion for the autodecoder setup: I am an impatient character and find it hard to wait half an hour for the upload ;-) With stable signals, the first decode usually happens already after a couple of seconds during the flat-phase trials. Would it be possible to upload a preliminary first result soon after startup? You could then let the decode process continue and re-upload the complete results after the decoder has terminated.

Thanks again for the fine work!

Best 73,
Markus
 


-----Ursprüngliche Mitteilung-----
Von: Domenico IZ7SLZ <iz7slz.domenico@gmail.com>
An: DK7FC <selberdenken@posteo.de>; Markus Vester <markusvester@aol.com>
Verschickt: Sa, 8. Jul 2017 19:57
Betreff: Re: LF: Auto-EbNaut reception on LF in JN80

Hallo Markus,

Glad to see your call in my EbNaut system.

Thanks you for testing it.

Me and Stefan have some idea how to improve it. What do you think?

73 Domenico

On 8 Jul 2017 19:37, "Domenico IZ7SLZ" <iz7slz.domenico@gmail.com> wrote:

Yes Stefan,

The advantage of vlfrx tools is that you can easly record the raw audio with the needed bw and pick up the interested  time period whenever you want, every time you test different settings.

So i'm recording, on 160 Gb Storage Hard disk , audio from receiver on L channel (centered on 1500 Hz) after applying vttime and downsample it at 9600 samples/s.

WTWRITE is now using abt 7 Gb / day of my HD. 

Seems there is no degrading of EbNaut performarce, as you see from my 'auto' system.

About vtblank settings, i have just guessed them after some attempts in improving ebnaut Eb/N0 of Riccardo signals.

For sure they can be more optimized for different situations (time of the day, qrn level,..).

But i want to do this by using a new version of vtblank that has now  different parameter's ranges.

I have planned to do this upgrade on the next week.

So actually i can analize signals in the bw of my receiver (1500 +- 500 Hz) since july 6th with some interruptions.

If you have some transmissions to be checked, let me known.

Have fun.

Domenico IZ7SLZ.

On 8 Jul 2017 17:12, "DK7FC" <selberdenken@posteo.de> wrote:

Ciao Domenico,

Your LF RX works very well with vlfrx tools. I still didn't monitor real radio signals with a QRN background, just test signals.
Do you now record many days in vt data files to post process if you like? I am planning to do that :-)
And do you optimise the blanker settings for LF? There are a lot of possibilities to analyse signals and optimise the results :-)
I'm looking forward to that :-)

73, Stefan

Am 08.07.2017 17:06, schrieb Domenico IZ7SLZ:

Ciao Stefan,

Seems that gpsd automatically sets its baudrate and protocol to accept nmea.

1pps is accepted on rs232 pin 1 Data Carrier Detect.

For best results i have separatly setted my gps Jupiter  with Binary protocol at 9600 baud

Thanks for the signal. I'm sure that experience done on LF will help me on VLF.

Have a good week end

Domenico

On 8 Jul 2017 16:47, "DK7FC" <selberdenken@posteo.de> wrote:

Hi Domenico,

Am 04.07.2017 11:13, schrieb Domenico IZ7SLZ:

These are the programs running on PC:

- gpsd, ntpd for getting utc synchro from GPS (that has nmea + 1pps)

Somehow i don't get my GPS module running with gpsd. It is installed on the Raspi. The NMEA messages arrive at the serial port. How did you manage the clock syncing with the GPS module and gpsd?
How do i need to modify /etc/ntp.conf to set the clock with gpsd alternatively? It would be fine that the clock is set either by data from the web or by data from the GPS module...

73, Stefan

PS: We have massive QRN again but you can decode me at Eb/N0 = 15 dB :-)

upload2
Description: Binary data

auto
Description: Binary data