One of my RF digis is shown in wrong spot, and shows off air ONLY on aprs.fi

[jeep2liberty]

The station shown in the wrong spot *ONLY* on aprs.fi is W8VFR-1. Adjacent APRS station logs and a trip to the site confirm it is beaconing with the correct (hard coded) coordinates by two mobile aprs stations. aprs.fi also shows it with no active beacons since ~> 2days now. Again, adjacent stations and 'other' APRS-IS services show it in the correct spot, and with recent activity. This is not a browser issue. Shows incorrect on Android based phone also. The 'raw' data page on aprs.fi shows no activity from the station. It has been in service for several weeks with no local operational issues.

http://aprs.fi/?c=raw&call=W8VFR-1

I do note a repeated message of "[Delayed or out-of-order packet (timestamp)]" in the raw data. This station is RPi based and no internet connection is available to update ntpd. I run the beacon with out "timefix" in aprx program and aprx program uses all ones (1111111z) as time. Even still why do other IS services show it in the correct position and operating normally?

Thanks for any help!

73s
Curtis
W8VFR

[Heikki Hannikainen]

Hi,

On Fri, 1 Apr 2016, jeep2liberty wrote:

> http://aprs.fi/?c=raw&call=W8VFR-1
>
> I do note a repeated message of "[Delayed or out-of-order packet (timestamp)]" in the raw data. This station is RPi
> based and no internet connection is available to update ntpd. I run the beacon with out "timefix" in aprx program and
> aprx program uses all ones (1111111z) as time.

That's either a problem with the configuration of your aprx, or a bug in
it. It should not send an "all ones" timestamp. If there's no timestamp to
send, there's no need to send a timestamp at all.

Maybe you just need to remove the "111111z" from the config file if it's
inserted there manually in the packet contents?

For further debugging on the aprx side, please get in touch with the aprx
support forum. Please post your beacon config from aprx.conf when posting
there, it'll make it easier to start the debugging.

https://groups.google.com/forum/#!forum/aprx-software

> Even still why do other IS services show it in the correct position and
> operating normally?

aprs.fi parses the timestamp and uses it to detect old, duplicate and
out-of-order packets. This functionality reduces the amount of stations
jumping back and forth when packets get delayed. This also implies that
you need to transmit a correct and valid timestamp if you send one at all.

Most other IS services that I know of do not do anything with the
timestamp. Some display it, most don't do anything at all.

- Hessu

[jeep2liberty]

Hessu,

I appreciate the response. I'd again like to point out that aprs.fi was the only APRS-IS type site on the net that had the station position and data on the station wrong. That is what I'd like to fix if possible, I think there may be room for improvement in how aprs.fi processes the packets. Again, all other -IS services showed correct data.

I made *no changes* to the station at all, not even a reboot. I eventually had to 'trick' the packet ingestion on aprs.fi by spoofing the W8VFR-1 packets using another station's hardware to send a few position packets in order to break the 'cycle' you will see posted below. As for aprx and it's configuration, I believe that 111111z is the default beacon style if a precise timing source is not available such as ntpd or a GPS feed. Timefix is optional and not the default, nothing I am aware can be done to stop the 111111z output. Other stations continue to beacon this format with no issues (including W8VFR-1 after the spoof).  

Here is the raw data (annotated with appended #comment) from the aprs.fi page. Please have a look at it. Could a small change in the way we process the packets prevent this?

Initial issue:

2016-03-30 23:13:41 EDT: W8VFR-1>APRX28,qAR,KD8LXB-10:/111111z3936.32NI08440.24W#40KM digi #normal beacon for reference
2016-03-30 23:14:33 EDT: W8VFR-1>APRX28,qAR,KD8LXB-10:T#565,37.6,0.9,190.0,5.0,5.0,00000000 #routine telemetry packet
2016-03-30 23:21:51 EDT: W8VFR-1>APRX28,qAR,W8VFR-4:/111111z3936.32NI08440.24W#40KM digi #Last good position packet
2016-03-30 23:31:26 EDT: W8VFR-1>MPRX28,qAR,KS9N-10:/171111z3906.02NI08440.24W#40KM digi #Packet decoded with unfortunate FCS result. 111111z wound up as 171111z and LAT wound up as 3906.02N vice 3936.02N
2016-03-30 23:31:26 EDT: W8VFR-1>APRX28,qAR,W8VFR-4:/111111z3936.32NI08440.24W#40KM digi [Rate limited (< 5 sec)] #Not sure why position did not update upon receipt of this packet?
2016-03-30 23:34:31 EDT: W8VFR-1>APRX28,qAR,KD8LXB-10:T#566,37.5,0.7,176.0,8.0,4.0,00000000 #routine telemetry packet
2016-03-30 23:39:35 EDT: W8VFR-1>APRX28,qAR,N8VKX-10:/111111z3936.32NI08440.24W#40KM digi [Delayed or out-of-order packet (timestamp)] #Delayed from what reference point?
2016-03-30 23:47:55 EDT: W8VFR-1>APRX28,qAR,N8VKX-10:/111111z3936.32NI08440.24W#40KM digi [Delayed or out-of-order packet (timestamp)]
2016-03-30 23:54:35 EDT: W8VFR-1>APRX28,qAR,KS9N-10:T#567,45.5,0.9,187.0,17.0,5.0,00000000 #routine telemetry packet
2016-03-30 23:56:48 EDT: W8VFR-1>APRX28,qAR,W8VFR-5:/111111z3936.32NI08440.24W#40KM digi [Delayed or out-of-order packet (timestamp)]

Now the 'fix' was again to spoof the W8VFR-1 station with another piece of hardware. This broke the "Delayed or out-of-order packet (timestamp)" cycle.

2016-04-05 07:37:33 EDT: W8VFR-1>APRX28,qAR,W8VFR-4:/111111z3936.32NI08440.24W#40KM digi [Delayed or out-of-order packet (timestamp)]
2016-04-05 07:44:56 EDT: W8VFR-1>APRX28,TCPIP*,qAC,THIRD:!3943.50N/08436.79W#Portable RPi digi or digi-gate temp=26.1'C  #This was the packet that broke the cycle from an RPi station with W8VFR-1 temporarily as callsign.
2016-04-05 07:46:24 EDT: W8VFR-1>APRX28,qAR,W8VFR-5:/111111z3936.32NI08440.24W#40KM digi [Location changes too fast (adaptive limit)]  #As expected this message due to ~15KM difference in distance/time.
2016-04-05 07:54:29 EDT: W8VFR-1>APRX28,qAR,W8VFR-4:T#299,44.6,2.9,268.0,6.0,18.0,00000000
2016-04-05 07:56:02 EDT: W8VFR-1>APRX28,qAR,N8VKX-10:/111111z3936.32NI08440.24W#40KM digi #Station now back showing on map, with correct fix, time since last report/position and good now for many days. Again, no changes to the station at all !!

73s

[Heikki Hannikainen]

Hi,

On Wed, 6 Apr 2016, jeep2liberty wrote:

> I appreciate the response. I'd again like to point out that aprs.fi was
> the only APRS-IS type site on the net that had the station position and
> data on the station wrong. That is what I'd like to fix if possible, I
> think there may be room for improvement in how aprs.fi processes the
> packets. Again, all other -IS services showed correct data.

Of course, some improvements can be done, but I am not going to make
aprs.fi *not* use the timestamp any more. Other -IS services don't use the
timestamp like aprs.fi does. Adding timestamp processing improved data
quality nicely, and reduced complaints about stations jumping around due
to old/delayed packets, so it's going to stay - it's a good feature.

> I made *no changes* to the station at all, not even a reboot.

Right, but you're still sending invalid timestamps, and that needs to be
fixed to make things work right.

> As for aprx and it's configuration, I believe that 111111z is the
> default beacon style if a precise timing source is not available such as
> ntpd or a GPS feed. Timefix is optional and not the default, nothing I
> am aware can be done to stop the 111111z output. Other stations continue
> to beacon this format with no issues (including W8VFR-1 after the
> spoof).  

Ok, then aprx needs to be fixed, that'd be clear and simple bug in aprx.
Those other stations may have some trouble too later on.

Alternatively, if you're using a "raw" beacon configuration in aprx, and
you have the "111111z" written out in your aprx.conf, it may be a
configuration issue and then 111111z could be simply removed from
aprx.conf.

In any case, please contact the aprx support group about this issue:

https://groups.google.com/forum/#!forum/aprx-software

Thanks!

- Hessu

[Chuck Kimball]

Good Day:

I've been unable to look at specific dates for on the telemetry page for
sites that return a choice.

The specific one I was trying earlier was Kendal
http://aprs.fi/telemetry/a/KENDAL
If I open the telemetry page using the link above and then set some
specific dates (in this case in September), and hit enter, it then
returns the pick list, but clicking on the Kendal digipeater and not the
ship returns only the 48 hour display.

Thanks for all you do Heikki

Chuck Kimball n0nhj
Glenwood Springs, CO USA