Order Tray | Contact Us | Home | SIG Lists

[aprssig] Recommendations for KPC-3 digis:

Robert Bruninga bruninga at usna.edu
Thu Dec 9 15:50:40 UTC 2004

Here are my recommendations for anyone with a KPC-3
WIDEn-N digi.  THis is a copy one I sent to a local SYSOP:

>>> Robert Bruninga 12/9/04 10:03:32 AM >>>
Thanks for giving me the opportunity to kibitz on your
digi settings.  I hope you will see the benefit of some of
these changes and will share them with other DIGI
operators in your area:

First the settings that are detrimental to the network.
Although some of these seem to be "beneficial" when
you look at them in the book, they are actually bad for
an APRS network.  Here is what they should be:

1) UIDWAIT should be OFF. (the default).
With it on, your digi is not doing the fundamental APRS
fratricide that is the primary mechanism for suppressing
multiple dupes.  All digis that hear the same packet are
supposed to DIGI it at the SAME time so that all those
copies do not get further digibeated between themselves.
(but outward located digs will hear it without colision
(and continue outward propogation)

2) UIFLOOD should be set to NOID (the default).
Yours is set to ID so that your digi obliterates all
prior path data in every packet by overwriting with
its own call.  This makes troubleshooting and tracing
packets to their source impossible.

3) Your BLTs and LTPaths are generating about
160 packets per hour, a tremendous load on the net.
The recommended way to still cover the same area
with decent 10 minute local and 30 miniute DX
copies is shown on:


as follows:

 BLT 1    EVERY 00:30:00 START 00:00:00
 BLT 2    EVERY 00:30:00 START 00:10:00
 BLT 3    EVERY 01:00:00 START 00:20:00
 BLT 4    EVERY 01:00:00 START 00:30:00
 LTP 1    APN383 
 LTP 2    APN383 
 LTP 4    APN383 VIA WIDE2-2

This results in a local ID once every 10 minutes, a one 
hop one every 30 minutes and a 2 hop one every hour 
for a total of only 20 total packets over all of virginina, 
but an 8 fold reduction in the QRM you had before.
(Remember that 2 hops from a digi hits all the same stations
that a 3 hop path does from a home station.)

4) Your UNPROTO should use the TOCALL of
APN383 where APN identifies it as digipeater firmware.
The "3" identifies it as a KPC-3.  and the 83 identifies
it as ROM version 8.3 (if that is what you have)..

5) Your UNPROTO path should be APN383 (no VIAs)
since it is used for yoru BText which should contain 
information of *local* value.  A good thing to put there is
an OBJECT indicating your local ECHOlink or IRLP node 
for travelers passing through your area.  Be sure to include
the node number and frequency such as

BText: ;IRLP7074 *000000zDDMM.hhN/DDMM.hhW$46760p107 in Cville

If the ECHOlink or IRLP node is say to the WEST of this
digi, then you might want to include in the UNPROTO path
a VIA diigi in that direction to make sure that anyone that is
within DIRECT range of the ECHO/IRLP node will get
the BText, but that the BText does *not* go beyond the
area where it can b e used.

OK, now for the NEW stuff under the "New n-N Paradigm"

6) Set your UIFLOOD parameter to support the I-64 corridor
This will let your digi digipeat along the entire corridor from
Norfolk to KY but even a 64LNK7-7 packet would only
generate 7 copies compared to WIDE7-7 that can generate
as many as 200!!!

7) Move your WIDEn-N support from UIFLOOD to UITRACE:
This lets your digi still support WIDEn-N for the time-being,
but helps us trace where the packets are coming from so
we can solve out-of area DX flooding problems easier...

8) We no longer need "TRACE" so you are now free to
add another alias to the UIDIGI parameter.  Now you 
Just decide what else you want to support

Other notes.  Your LT's look great but...
a) But the /A=001377 must be 6 digits to be decoded properly
b) Your digi is not a W-R digi.  It was a W-R-T with an "N: 
    Now it should be:
    64LNKn-N & Wn-N & XXXX in Cville.......
c) The overlay character should now be a "L" instead of "N"
    This is the character between the LAT and LONG.
    This tells everyone it is a LINKn-N digi

d) Put the /A=001377 at the end, not first.  Since those first
    20 bytes are what the D7 and D700 users see and the
    other stuff above is more important.

ALl of this applies to your other digi out in WV too, since it is
also more or less along the I-64 corridor.  Next, send this to
the Richmond DIGI sysop and see if he agrees.  But he
will have to drop WIDEn-N inorder to also support the
I-95 LINK of 95LNKn-N in his UITRACE parameter...

In my opinion I have no objections to droping WIDEn-N
entirely in those areas that are flooded by out of area
DX.  If you do that, then add WIDE2-2 to  your alias
list in the UIDIGI list so that passing mobiles at least
get out...

de WB4APR, bob

More information about the aprssig mailing list