[aprssig] Delayed packet issue caused by wx station
Mark Conner
mconner at aer.com
Fri Aug 1 08:52:38 EDT 2008
Hi all,
I thought I'd relay a recent experience I had on this issue........
We got word that one of the local digipeaters (a KPC-3) was 'holding' all traffic, then digipeating it every few minutes as a long data burst. This in turn would cause the other area digipeaters to also retransmit in a long burst. At first we thought it was just a busy channel, as the propagation in the Central Plains has been better than usual with a strong inversion. However, once we found the interval was exactly five minutes we dismissed that idea. We rechecked the settings for something like UIDWAIT being turned on and other grasping at straws with no luck.
Because the co-located Peet weather station had recently been reconnected after a long inactive period, we started to look at that. When in packet mode, the Peet sends serial data (one line) to the KPC-3 every 5 minutes. This was hooked to the KPC-3 second serial port. Turning off the weather station resulted in normal digipeater function immediately, so we left it that way as having the digipeater back to normal was the most critical.
After getting home I looked some more at the KPC-3 manual. The second serial port input on the KPC-3 DB9 connector is an XCD ('external carrier detect') pin when not in use as a second serial port (turned on by GPSPORT command). I think the GPSPORT command was not set, or this particular KPC-3 is prior to 8.3. Our next trip will confirm both these issues. If the input pin is not acting as a second serial port, connecting the weather station to it probably caused the pin to tell the KPC-3 it had an external carrier detected and to hold transmission. When the Peet hit the 5-minute timer and send the packet, the pin went to 'no carrier' during the weather data transmission and the KPC-3 released its held traffic.
This is a somewhat improbable set of circumstances leading to this situation, but perhaps our experience will help someone else in the future with diagnosing a similar problem on their network.
73 de Mark N9XTN
More information about the aprssig
mailing list