AFSK packet investigation

Raw AFSK data output from a local APRS transmission

Raw AFSK data output from a local APRS transmission

To further my investigation as to why the soundmodem TNC appears to be flaky, I decided to scope the output of some local APRS transmissions. After setting the squelch to the point where no background noise was seen / heard, I set my DS2202 oscilloscope to single-shot trigger the AFSK data.

AFSK data output from local APRS transmission showing 0dB variation on 1200 and 2200Hz tones

AFSK data output from local APRS transmission showing 0dB variation on 1200 and 2200Hz tones

To see the audio spectral content of the transmission, I engaged the FFT math function of the DS2202 and took a number of “snapshots” with the single-shot triggering.

I collected a few samples where there was a 0dBmv difference between the amplitude of the 1200Hz and 2200Hz tones. I also saw quite a few samples where there was a significant (6dBmv or more) difference. None of the packets / samples were successfully decoded by soundmodem.

AFSK data output from APRS recordings

AFSK data output from APRS recordings

When I scoped the APRS recordings that I’ve been using to test with, I captured a packet which showed a 17.6dBmv difference (which by decibel math for voltage corresponds to an 8-times) difference in amplitude. This packet for whatever reason was successfully decoded (as usual).

Now to be completely honest, I am not sure of the accuracy of my method, nor am I even sure that I can simply use FFT functions to measure the amplitude differences to establish a meaningful baseline. What I do know is that if the AFSK passes through my FT-857D it appears to be meaningless on receive.

Suspecting that it is something to do with my transceiver (perhaps alignment?) I used the APRS test CD to make a test transmission (on a vacant frequency, after announcing my callsign and “testing” while using minimal power) and attempted to decode the AFSK packets via the radio and soundmodem.

Damnit! It worked. So this suggests that the incoming APRS data that I’ve been trying to decode may very well be swamped by noise or some other characteristics, lending to it not being decoded by soundmodem.

I am very open to suggestions at this point – I am nearing my wits end with this! I think that my next course of action will be to bring the test bench into town where the APRS transmissions will tend to be stronger and try decoding the packets.

If it is a matter of the signals being too weak, I would like to investigate solutions to this problem. Some possible remedies include: receiver with better sensitivity, better antenna system (I am using a mobile 5/8λ whip indoors), preamplifier for RX.

More to follow …


This entry was posted in General, Operating and tagged , , , , . Bookmark the permalink.

4 Responses to AFSK packet investigation

  1. Pravin says:

    Hi James, I had noise issues with my Ft 857D when using the rear ACC socket with isolation transformers for digital modes it was a weird noise from the radio, I put a 10nf capacitor across the signal line and gnd. Noise gone. The noise was evident on the waterfall on HRD which i was using to monitor signals. Now this rig works well.

    • James says:

      Interesting result, Pravin. I’ll have a look at the output of the ACC jack with the scope to see if adding a cap. will help kill any noise. Truthfully, I haven’t seen anything tell-tale on the waterfall display in Fldigi .. it is worth investigating however!

  2. Bob VE3MPG says:

    Try leaving the squelch open (noise). I last had a packet station in CZ land in 1990 and setup was to leave the squelch open especially important for weak sigs. I was connecting to Budapest from the Slovak city of Kosice at the time.

    Good luck getting it all working,

    Bob VE3MPG

    • James says:

      Hi Bob,

      I did try with the squelch wide open too – I was merely using the squelch to probe the audio with my scope (avoiding all the noise). Still no real luck to speak of!

Leave a Reply

Your email address will not be published. Required fields are marked *