[aprssig] Soundcard Packet Analyzer

Bob Bruninga bruninga at usna.edu
Sat Oct 22 18:20:05 CDT 2011

What we really need is a PACKET ANALYZER software!

Something to start pointing fingers at all the bad apples on the air.  Every packet expert I know will admit that Most if not 90% of signals on the air are poorly adjusted.  Levels, Skew, TXD, distortion, pre-emphasis, etc.

Performance could be SO much better!

Trying to get everyone to look at a scope is hard to do.  Or ven interpreting what they see.  But if they could just download a sound-card PACKET ANALYZER then at least we would have a baseline.  Here is what it would do on EVERY packet:

1) note the modulation level
2) Note distortion content
3) Note skew
4) Note pre-emphasis
5) Note the TXD
6) Note SNR
7) Measure length
8) Comment on PATH

etc...  Of course, most of the above analysis depend completely on  the receiver used to receive the signals, so this has always been a difficult and subjective process.

HOWEVER, a large number of signals on the air are from three rather popular factory-aligned packet radios, the D7, D700 and D710.  They might not be perfect, they might not be best, they might not be perfectly set, but having come from a factory, they should be consistent.

SO, my suggested calibration of this PACKET ANALYZER would be to monitor the channel for D7's, D700's and D710's and benchmark their performance.  Then compare that to what ever we think PERFECT is, and then from that, we can benchmark what the ideal parameters should be.  Of course just listening to empty channel nooise can also tell lots about the receiver used.

Then this program measures all signals against that benchmark.

Then we can immediatleyl point out the bad apples.

Not that we are going after individual stations.  But we can go after the DIGIPEATERS because they represent the 95% of all packets on the air.  If we can get those digipeatetrs adjusted, then we can really clean up the network!

Mike Chesena, KA2ZEV suggested something like this.
I think it is a GREAT idea!

Do we have any DSP jockies that can write this code?  Or probably most of it alreayd exists somewhere.  But the self-calibration will be important so we are all comparing to the same orange.


More information about the aprssig mailing list