[aprssig] "Best" packet decoder solution

[Phil N6TCT]

I spent a lot of time in my basement last year trying to answer the
question of packet decode performance.

I tested the Argent Data Systems T2-135, the Byonics TinyTrak4, the
Argent Data Systems T2-301 (which has an integrated radio), and the
Coastal ChipWorks TNC-X.  In the case of the T2-135, the TT4, and the
TNC-X the radios they were connected to were either a Alinco DR-135 or a
GE MVS (both swapped back and forth).  I did not test the KPC3+ because
I didn't have one at the time.  The transmission source was track 1 of
Stephen Smith (WA8LMF)'s test CD (http://wa8lmf.net/TNCtest/index.htm)
fed into a service monitor (either a Motorola R2001 or an HP 8924C)
modulated up to 144.39 at -80 dBm where it was connected to the radio's
antenna input.  As Stephen just posted about, this is the track that has
roughly half an hour of 144.39 off the air from southern California. 
The test was to see how many packets the TNC would successfully decode.

The two best performers were (1) the TNC-X connected to the "9600 bps"
(i.e., the raw discriminator output) output of the Alinco DR-135 and (2)
the T2-301 with its integrated radio.  Both decoded between 850 and 900
packets.  The TT4 decoded between 750 and 800 packets.  The T2-135
didn't fare so well, decoding between 550 and 750 packets depending on
the particular board and test setup.  This was apparently due to some
manufacturing issues that have since been corrected.  I have not tested
newer boards from Scott Miller, N1VG, but I suspect they are better now,
if only because I know Scott uses a similar test setup and is a diligent
guy and cares about the quality of his products.  :-)

All of the above said, I would hesitate to draw too many conclusions
from my experiments due to some of the variables that Lynn Deffenbaugh,
KJ4ERJ, pointed out in his earlier post on this thread.  There are
questions of what transmit deviation you're testing with, whether you're
pre-emphasizing the transmit signal or not, etc. -- and what you expect
other people to be doing in the real world.  I.e., how closely do you
think your test setup matches what you're going to see in the real
world?  Plus it's a moving target and the hard-working ham radio
manufacturers who make these widgets for us are constantly improving
their products and addressing issues.

One thing I did learn is that radio-TNC combo matters and that some
experimentation is necessary to get the best results from a given radio
and TNC.

Another lesson is that a service monitor is a great thing to have, and I
wish I got one a long time ago.

A final thing I learned is how labor intensive it is to make these kind
of measurements -- which is partly why I stopped.  :-)

Phil, N6TCT

Greg Clark wrote:
> What's the "best" packet decoder out there?
>
> By "best", I mean the one that has the greatest ability to decode APRS
> data packets.
>
> Can be stand-alone (like the D710) or separate unit (like Kantronics TNC)
>
> Is any one in particular considered the  "gold standard" ?
>
> How do units like the new Kenwood HT, or small inexpensive ones like
> the TT4/OT/TNC-X  compare?
>
> -- Greg K7RKT
> ------------------------------------------------------------------------
>
> _______________________________________________
> aprssig mailing list
> aprssig at tapr.org
> https://www.tapr.org/cgi-bin/mailman/listinfo/aprssig
>   
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.tapr.org/pipermail/aprssig_lists.tapr.org/attachments/20101227/917f403d/attachment.html>