[aprssig] Fixing older KPC-3 Duping Digis! (rev 1)

Bob Bruninga bruninga at usna.edu
Wed Aug 31 17:00:24 EDT 2011


> Still better, I think, for old digipeater Sysops 
> to just add UIDigi WIDE1-1 in the Kantronics version 
> KPC3-3F427265-8.2 TNC as previously recommended.
> http://www.aprs.org/kpc3/kpc3+82WIDEn.txt
> In practice, this certainly works well enough.

Yes, for the rare 8.2 digi, but if you have more than one, or as in the case
of New Mexico, if they are all that way, then it is a disaster.

The Version 8.2 fix of putting WIDE1-1 in the UIDIGI Alias list fixes the
one problem of the path being broken and going no further than one hop, but
it does nothing to solve the lack of DUPE suppression.  The above
recommendation at least lets the TNC be used, though at the expense of
multiple dupes if other verision 8.2's are around.

Tom has made an excellent point, that the multiplication of dupes makes this
8.2 fix an exponential dupe problem in areas where many 8.2 KPC3's can hear
the mobile.  I wonder if other areas have noticed this?

Anyway, the result of this revelation is to cause me to go back and REVISE
the above recomemendation to INCLUDE the  "use only WIDE2-2" (or WIDE3-3) in
this area, (meaning WIDE1-1 is not supported).

RECOMMENDATION TO NEW MEXICO?

The goal is to RECEOMMEND to all New Mexico permanent users to avoid using
WIDE1-1,WIDEn-N path.  But to in the mean time, support the occasional
visitor that drives into town with the above common path.  How about this:

1) EDUCATE everyone permanently in New Mexico to use the WIDEn-N path only.
2) REMOVE UIDIGI WIDE1-1 in all Albuquerque digis that cover the metro area,
3) Change ALL New Mexico digis to "NoW1 NM5-5 comments..."

* The result is that all the digi's with UIDIGI WIDE1-1 removed will
digipeat the W1-1,W2-1 packet once and only once using the normal (but
broken/truncating) WIDEn-N algorithm.  But since this is a metro area
everyone in Albuquerque will get it as well as the internet.  So the
visiting WIDE1-1,WIDE2-1 mobile is heard (locally).

* Distant selective digis in wilderness areas can still support WIDE1-1 as
an alias because there are fewer mobiles out there, and hopefully they will
hit fewer simultaneous digis?  We don't want a visitor lost in the toolies
to not be heard.

* Wherever significant overlaps occur, try to make sure that no more than
one digi supports the WIDE1-1 uidigi alias.

* The "NoW1 NM5-5" recommendations inform all the locals (and visitors) of
proper operations in that area. (I am assuming that NM5-5 anywhere in the
state will hit all digis in the state(once and only once)..)

Would that work?  Notice, I have changed the recommended indicator to
"NoW1..." as a better way to get the message across that WIDE1-1 is not
supported (because of the 8.2 dupe problem).

Bob, Wb4APR






More information about the aprssig mailing list