Checking deviation

[Sven Ruin]

Hi

I was wondering if there is any simple way to check FM deviation using SDR, expecially for packet/APRS? For example, the Yaesu FT-818ND manual states that the optimum deviation is between ±2.5 kHz and ±3.5 kHz for packet at 1200 bps. Would it for example be possible with the Gqrx packet decoder to get some information about the packet's deviation?

73 de Sven SM5PMF

[Mike Waters]

Oh yes there is! I've done that myself on an narrow band FM signal.

Zoom way in on the signal, and it'll get wide enough on the waterfall to see its bandwidth. It's hard for me to explain it step by step.

73 Mike 

W0BTU 

[Sven Ruin]

Thank you for the helpful reply Mike! So I suppose that during the short burst of a 1200 bps packet one should see a bandwidth of about 10-11 kHz in the waterfall. As I understand, Carson's bandwidth rule (https://en.wikipedia.org/wiki/Carson_bandwidth_rule) says that 2.5-3.5 kHz deviation and a highest tone of 2.2 kHz corresponds to 9.4-11.4 kHz bandwidth. And if the packet decoder is running at the same time, it can show the contents of each packet, which should provide information about where the packet comes from.

[David Ranch]

Not so much GQRX but check out the devcal tool from Svx which uses a RTL dongle to check FM deviation:

   https://www.svxlink.org/doc/man/man1/devcal.1.html

--David

KI6ZHD

[Robin Gape]

Sven,

Bessel Zero
a very accurate method of setting FM deviation, which can be done using GQRX, is the Bessel Zero method. In practical terms:

a) set a particular audio modulation frequency

b) increase the deviation from zero until the carrier frequency level is reduced to zero

c) a narrow bandwidth spectrum analyser can be used for this, or a receiver in CW mode, set to narrow bandwidth

There is a reasonably clear explanation of the method at https://www.qsl.net/g3oou/fmdeviation.html, a description of its use in practice at https://ramdor.co.uk/2021/03/18/fm-deviation-using-bessel-zero-method/, and another in-practice article at https://www.rtl-sdr.com/measuring-frequency-deviation-fm-transmitter-rtl-sdr/.

More adventurous
As one might hope, a brave soul has designed a direct FM deviation meter using GNU Radio. It’s on GitHub at https://github.com/mpvano/FMDeviance.

Carson’s Rule
The problem with Carson’s rule is that it is very much an approximation. It was pulled out of thin air, probably after a pub lunch.

HTH, 73,

Robin, G8DQX

[David Ranch]

Great set of URLs Robin!  Thank you for those and I'll have to give them a try.

--David

KI6ZHD

[Sven Ruin]

Thank you for all replies. I began with the Bessel Zero method, which seems to have now worked for me to check deviation, at least approximately. I used a 1200 Hz tone, which as I understand corresponds to a deviation of ±2.9 kHz (within the optimum deviation window stated by Yaesu for 1200 bps).

On the transmitter side, I set up a computer to generate a mark calibration tone of 1200 Hz, using DireWolf 1.7 (command: direwolf -x m). On that computer, running arcOS, I in AlsaMixer gradually increased the Speaker setting for DigiRig Mobile, which is my USB interface between computer and transceiver, from a very low initial value. I also gradually increased the PKT MIC setting on my Yaesu FT-818 transceiver, from a very low value. When Speaker was set to 23 and PKT MIC to 12, I could on another computer running Gqrx (connected to a SDR receiver) see that the carrier frequency level was clearly reduced, which I had not seen with any other setting. In Gqrx, I had zoomed in on the spectrum using the mouse wheel. However, it did not seem possible to deterimne the optimal settings from only looking at the bandwidth shown in the waterfall. At the same time, the deviation meter function (MTRmod) on the FT-818 showed three bars. Yaesu's manual says that for FM transmission, the proper setting of the FM Mik Gain will produce five bars of indication on voice peaks, slightly less on lower levels of speech input.