[aprssig] High density timeslotting

[scott at opentrac.org]

I'm thinking of trying an experiment, and I was wondering if anyone's
attempted this before.

Right now, timeslotting on the OpenTracker (and presumably the TT3 and
anything else that supports it) isn't accurate to more than a second or two
because of the way time data is presented in the incoming GPS data stream.
The practical limit would seem to be about 2-3 seconds separation between
trackers.

With an accurate time reference, though, you could do a lot better than
that.  Deluo's pin out info shows (or did the last time I was able to find
it) a 1 pulse per second output, but it turns out that at least for the
current generation of receivers it's actually a TTL version of the data
output.  I'm not sure if that applies to the SiRF units as well.  The Garmin
GPS18 does provide a real 1PPS output, and better yet, it's well-documented
- something you can't say for ANY of Deluo's offerings.

Assuming you're not running something with a horribly slow TXD (like a
D700), getting two beacons per second should be pretty easy.  Three looks
doable with decent transmitters, and four might be possible with good
equipment and very short packets (200 ms for a 30-byte frame and < 50 ms
TXD).

The only hardware modification required for an OpenTracker should be an
inverter circuit for the 1PPS pulse to make it into a negative edge to
trigger the interrupt request.  Actually, even that might be optional, as
long as you know exactly how long the pulse is.

The firmware would need a bit of tweaking to synchronize the internal clock
to the 1PPS signal, and some more to make sure that it stops acquiring fixes
a second or two before it's due to transmit so it's not caught in the middle
of an incoming fix.

Digipeating would present an interesting challenge.  Obviously if the
channel utilization is over 50%, you can't digipeat in-channel.  Are there
any dual-port TNCs out there that can handle full duplex digipeating, and
what transmitters would be suitable for 100% transmit duty cycle for
cross-band operation?

It might seem like an exotic application, but I've got a few users who have
expressed interest in this sort of thing.  I think they're all for races or
competitions of one sort or another - organized events where everything's
coordinated in advance, with appropriate infrastructure.  Higher
timeslotting density means more tracker users and/or a faster update rate
without the added expense of multiple channels.

Anyway, has this been tried before?  Should be interesting to experiment
with.

Scott
N1VG