UPDATE July 2018
Next phase of testing
After receipt of a new batch of SCS Tracker DSP TNCs, I loaned one to a local tester along with a TS2000 cable for testing against the VAPN prototype. Things worked reasonable well with good decodes both directions. The volunteer visited the BBS, downloaded SP mail, etc. with relative ease. All in all this was a great verification of the ability of the SCS Tracker’s to get the job done.
Received spectrum didn’t sound right
During the testing when the VAPN prototype was receiving, I kept hearing a higher pitch note during transmission later confirmed to be approximately 1,944 Hz… well above the approximately 500 Hz Robust Packet (aka RPR) signal. Here is a spectrum plot of the RPR signal as heard by a little microphone near the receiver’s speaker, fed into Audacity and processed by its spectrum tool. Click to enlarge into new tab.
Key observations:
- The ~500 Hz wide RPR signal pedestal is quite obvious.
- There is a relatively weak signal spur below the main RPR signal.
- Above there is a strong ~1944 Hz note.
- Both the low and the high notes are symmetrical around the main pedestal.
This note did not seem to interfere with the operation of the successful packet tests in any way. Still this modulation method’s ability to function in a nice tight 500 Hz bandwidth is compromised a bit if the Tracker TNC really generates this extra energy outside these limits.
The astute reader will notice the 500 Hz pedestal isn’t quite centered on the expected 1500 Hz. Frequency inaccuracies of both transmitter and receiver can account for this slight error. It’s interesting to note the pedestal, the lower frequency spike and the higher frequency spike seem to all be about 56 Hz too low.
Next step – testing without radios
Further tests are in order to see if the SCS Tracker is actually generating this spectral oddity or if something about the way it interacts with the transmitter’s signal chain is to blame. The obvious next step is to feed the Tracker’s transmit signal directly to a sound card… preferably a decent studio grade model, but almost anything should be good enough.
Conclusion
Robust Packet’s most noble feature, IMO, is good performance using minimal bandwidth. Whatever we learn from our additional tests will be shown here in amendments to this post. Perhaps we are simply driving the transmitter a bit too hard. We will find out.