[nos-bbs] More no-space

[(Skip) K8RRA]

Oh my, "No Space!!" is back.
In what follows I may tell you more than you really want to know, but I
suspect the TNC is involved in the problem.

I'm not asking for help or opinion by this mail, but if something
similar has happened in your (the reader) experience I am interested
enough to explore the similarity in search of new documentation - maybe
an FAQ entry.

Abnormal stuff surrounding incident.  I abbreviate my call K8RRA with
RRA, my south neighbor WA8RSA with RSA, and my north neighbor WB8I with
8I:

1) Two [stranded for many hours] messages in SMTP list to go to RSA
[stranger still - one has domain filled out totally. the other does
not.]  Mail delivery is suffering.

2) Two AX25 Status stations listed as connected - one is RSA, the other
8I seems idle.  RSA alternates between three states:
      CON PENDING/CONNECTED/RECOVERY.
As many as 18 packets seen to stack up in xmit buffer.

3) more than 22,000 packets between RSA and me - we have tied up the
majority of the bandwidth, the next highest is 1,400 heard on a
known-to-be-big-user.

4) When I attempt to connect to either RSA or 8I, the session is quickly
terminated and I am reconnected to RRA.

5) RRA xmit signal has an interrupted characteristic - perhaps a SWR
shutdown during mid transmission?

At this point I choose to shutdown until tomorrow morning and daylight.
BREAK FOR SLEEP.  Today I find power supply about 1 volt low but
otherwise OK.  The TNC will not leave KISS mode for testing, the
solution is to hard reset using the plug on the board.  The solution
returns the factory presets and the TNC seems OK after this point.
After setting parameters CALL, XMITLVL, and CD, then jnos restarts and
operates normally.  The signal transmitted sounds normal again.

6) In testing jnos operation today, I reviewed the JHEARD lists of my
neighbors and found RRA mutated call listed for yesterday [eg: O8RRA-1
K8RRA0-1 etc] and I conclude that the TNC/transmitter pair messed up the
jnos character string during transmission.

Now for a postmortem:
The TNC clearly had some parameter damaged, and may have been providing
a poor T/R signal, or the radio may have shut down for under voltage
while transmitting.  SWR shows OK.  The source of the damage is unknown,
this rig has been in 24/7 operation for a year without similar symptoms.

I'm going to guess the root cause is likely in the power supply.  The
power is supplied by battery, the raw source of recharge may not be
adequate for the installation.  Further, there is no low voltage defeat
implemented in the station, thus TNC & radio operation may become
erratic under low voltage.  Jnos on the other hand is not on the battery
and knows only that operation is difficult and many [! many ! many !]
retries are appropriate.  I don't understand how erratic channel text
strings can cause the malloc trouble we started with, but I'll keep
looking...  I will add some preventive maintenance steps to test the TNC
and battery condition.

73
de [George (Skip) VerDuin] K8RRA k