Spectrogram settings

[Andre' Kesteloot]

M. Bruno wrote:

> At 09.15 13/02/99 -0500, Bob K3DJC wrote:
> >I have been playing with this program and wondered what the best settings
> >are for maximum signal sensitivity. If / when we get the new bands am
> >looking to the fun you fellows are having. I also have FFTDSP program
> >running under DOS and have been trying to evaluate the difference. The
> >DOS program doesn't have the ability to look in real time as does
> >Spectrogram.
> >73   Bob    K3DJC
> **************************************
> Hello Bob and all,
> In October I posted the following message:
> --------------
> I did some tests to found the best setting of Audiogram for weak signal
> detection.
> I used a signal generator modulated by SLOWCW. The signal was radiated by a
> short wire and then received by the main aerial, to include the 'true' band
> noise.
> After several tests, I found the following:
>
> - The best setting of Audiogram is 5.5 kHz, 16 bit sampling; 16384 points FFT;
> 200 msec dwell time; 4 times spectrum average. This gives a nominal BW of
> 0.36 Hz.
>
> - The best dot interval is (as expected) 3 seconds. Even 2 seconds is
> useable if the S/N is a bit better. I  found no significant improvements
> going to 10 sec dots and 16 times average.
>
> During the tests the S/(S+N) was roughly 6 dB, but since the noise is
> highly impulsive it is difficult to assert a true ratio.
>
> -------------------------
>
> This was the original posting. Now, the majority of stations here use
> 3 or 4 second dots. The 0.36 Hz bandwidth gives an improvement
> of at least 10-15 dB over the ear copy with a 100 Hz filter. I extimated
> that the ear-brain system (!) has an approximative BW of 10 Hz.
>
> Sometimes I use a larger BW (like 4096 samples or 0.8 Hz) for band
> watching. If a signal is present it is usually recognizable with this BW,
> and then I switch to 16384 samples and 0.36 Hz.
>
> The averaging is a matter of personal preference. I leave it a 1 (no AVG).
> Some operator prefers to have a longer one. A very long one (128)
> allows detection of steady carriers buried into statics.
>
> I found a very good excercise to receive aero NDBs in 200-400 kHz
> band with Spectrogram. I found many beacons that I supposed not to
> be receivable ! Distances of 1500 km are not unusual on a calm night,
> and you have repetitive signals to play with.
>
> A good thing for our North American friends should be to identify
> the carrier of DBF39 on 138.830 kHz. Bob, have you a quiet listening
> position?
> 73 de IK1ODO Marco
> Rivalta, ITALY JN35SA (N 45 01' 25.6", E 7 31' 09.4")