Просмотр сообщений

В этом разделе можно просмотреть все сообщения, сделанные этим пользователем.


Сообщения - df6nm

Страницы: [1] 2 3 4
1
Markus, email to QRP Lab about that problem, pse.
I did mention this to G0UPL a few years ago but have not received a response from him. Unfortunately I do not own an U3S kit myself, so I can't reproduce the problem here. However I have read that there are different "park mode" options which control the synthesizer frequency during key-up periods. Maybe you can tell the firmware to maintain a constant frequency, so that the phase will not change during a pause.

73, Markus

 

2
What can I do?
If you want you could use the original Opera software from EA5HVK.  My preference is to transmit with the SpecLab digimode terminal, set to 8.192 second on-off-keying. You could also enable DTR from the serial port for keying. You will need to generate a 11011001101001... message for your callsign, eg. using my opgen2b utility: http://df6nm.bplaced.net/opera/opgen2b.zip

The best option would be to ask Hans Summers for a U3S firmware which implements coherent keying ;-)

73, Markus

3
Еще спектр Оперы становится шире, если сигнал передается некогерентно, то есть каждая посылка имеет какую-то свою случайную начальную фазу относительно непрерывной несущей. Не знаю, принимаются ли в алгоритме U3S меры по обеспечению когерентности, если нет, то в причина может быть в этом.
Yes, unfortunately the U3S does not send Opera dashes with coherent phase. Opds can also detect incoherent signals, but sensitivity is about 3 to 4 dB worse.
73, Markus (DF6NM)

4
Hi again Alex,

I loaded your "My_VLF_TX_1pps_PLL.USR", and it worked right away. Coincidentally the samplerate on my computer happens to be similar to yours, 47999.98 Hz. However playing with different enries in the calibration table, and turning output resampling on or off, made no difference. It always worked correctly...

But there may be a problem with the DAC output channel selection. In this configuration, the wanted signal from the digimode terminal comes out at L5, connected to the left speaker output. However R5 gets the signal from the signal generator directly, without correction. So if you have connected the amplifier to the right speaker, the PLL would have no effect, and the original unstabilized output frequency would be transmitted. This could explain the problem! The solution is simply to physically connect the amplifier to the left speaker, or flip the lower DAC-switch in the upper position so that both outputs carry the L5 signal.

73, Markus

5
Hmm, normally R0 and R1 should both work the same way. It would only makes a difference if you have "Sync'd Resample" activated, but in this application we have selected "don't resample".

73, Markus

6
Hi Alex,
 
this is strange. Even if the true samperate is slightly lower than the samplerate in the calibration table, the SR detector should be able to measure and correct the difference. Then the PLL should steer the digimode frequeny a little bit higher than 8270, such that the frequency at the analog output (and the feedback) input is correct.
 
Are you sure that the 1pps SR detector in the first (TX) instance is working properly? "Enabled" and "GPS phase lock" should be ticked, "measure only" not ticked (!), the status box at the bottom should be green. Also make sure that analog input level at the feedback input (L1) is greater than the PLL threshold in conditional actions (-40 dB).
 
Best 73,
Markus

7
Hi Alex,
 
this is all very interesting!
 
> I still cannot explain this, why the signal is perfectly visible in the transition zone and is completely absent at the beginning of the far zone.
 
Well once again we see that the hardest part is getting out of the nearfield ;-) Normally we would expect a r^-3 dependence up to 5.7 km (18 dB drop at double distance). So yes, if you could **hear** the signal at 3.7 km, it should be detectable in some mHz bandwidth at 7.8 km.
 
Note that the pattern and orientation of the quasistatic magnetic field lines are quite different from the far field. There is also an axial component which is twice as strong as the transversal component, but disappears completely in the far field. This is probably what you picked up at the second location, providing another 6 dB signal strength and unusual RX-loop orientation. It is also possible that the signal was locally enhanced by reradiation from underground wires or metal pipes.
 
>  the frequency of the signal continuously and irregularly changes within small limits
 
This is normal behaviour of the PLL. Depending on the measured phase and small samplerate errors, it steers di.freq, such that on average the phase is held for the correct frequency (8270.00625). This was confirmed by the green trace, so I believe you were transmitting on the intended frequency.
 
> another laptop with rx-loop + DHO38-calibration.
 
Samplerate calibration based on DHO38 also has some difficulties. First it can be affected by soundcard glitches, wich cannot be corrected and can cause a residual phase change which will disturb the frequency. Using 1pps + GPS-phaselock is the only method available in SpecLab which can correct such glitches, because it knows absolute time and can calculate the phase which the signal should have, even after an interruption.
In addition, the DHO38 transmitter is not exactly on 23.4 kHz. When Wolf and I measured it ~ 12 years ago it was approximately 0.03 ppm high, which is why it is announced as 23400.00006 Hz in the table. But I do not know if this is still correct today.
 
> - should Input Calib and Output Calib be equal?
Normally yes. I know that there are some soundcards which have different ADC and DAC samplerates, but these seem to be rare.
 
> - do I need to enable the "resample to nominal output SR" checkbox?
No, this is not necessary for this application.
 
73 and good luck,
Markus

8
Hi Alex,

1. Stefan is right. SpecLab only shows the rounded number in the signal generator entry field, but it internally keeps the precise frequency (8270.00625). To check it, you can type
print(str("#####.######",gen[1].freq))
into the command interpreter window and press the "F2" key.

2. The sign of the correction also depends on the "Rotation of antenna" selection in the "Options - Radio Direction Finder" menue, which can be CW or CCW. So maybe that has been changed some time ago. Now I have changed the expression from ".001*(azim(8260,8280))" to ".1*sin(azim(8260,8280)*pi/180)", which can lock on positive or negative slopes like an analog phase detector.

The PLL function has also been improved by using the SpecLab mixer instead of the digimode terminal, and allowing faster PSK switching for EbNaut:
https://klubnl.pl/rsgb_lf_group-archives/html/rsgb_lf_group/2016-12/msg00206.html
http://df6nm.bplaced.net/VLF/fec_tests/vlf_tx_pll_fasterpsk.zip

73, Markus

9
PS (off topic): Obninsk - the location of the world's first nuclear power plant, delivering electricity to the grid since 1954. Did you go to the museum? 73, Markus

10
Hi Alex, DL0AO is watching with antennas configured as an east cardioid (424 uHz and 47 uHz), unfortunately with no result yet. I hope we together will succeed some day, but this is really the most difficult time in the year. Do you think it is possible to extend the length of the TX antenna even further?

73, Markus

11
Today is the second series. Moreover, the daily session
Sergey, do you want to try a daytime transmission now? Noise at DL0AO is very low, and the loop antenna is pointed in your direction.
73, Markus

12
Indeed a good signal at DL0AO before UT midnight. Chayka lines from Slonim were also strong.

73, Markus

13
Thank you very much for watching.
Watching is our pleasure. Thank you for transmitting!

14
A trace also appeared at DL0AO...

15
Launched the lighthouse EbNaut
Hi Alex,

this is a strong signal, but I still cannot decode it.

73, Markus

Страницы: [1] 2 3 4