[ao16aprs] Uplink packet processing

[Jim White <jim@coloradosatellite.com>]

The Microsat software listens to all four uplink frequencies simultaneously
and will queue to the downlink packets heard at the same time.  That's why
they can support multiple ground stations.
It is possible, even likely, that the receivers are no longer exactly on
their original frequence.  Each frequency is set independently.  It would
be very interesting to know what the exact uplink centers are now, after 9
1/2 in space.  If testing for this, just remember to change only one thing
at once - the frequency up the uplink.

I sense some frustration as folks make changes and test the results.  You
change something and get results that don't make sense, change it back and
it gets worse yet, making even less sense.  A few of us spent a couple of
years attempting to optimize low power uplinks and I well remember that
frustration.  

The unknown factor that causes most of these unreliable test results is the
uplink interference situation.  It changes all the time, minute to minute,
day to night, by the day of the week, and as Tom Clark said - by the phase
of the moon.  As long as you are running low EIRP that factor will be there
and will tend to confuse other test results.

I think there are two ways to set up a low power station for optimum operation
1) Beg or barrow an amplifier.  This minimizes the interference problem.
Tune all other parameters for best throughput.  Takes a few dozen passes.
Can be instructive and interesting.  Return the amplifier knowing you are
optimized.  Further variation in results will be from interference (other
users, or unintentional non-satellite).
2) Use appropriate test equipment to measure and set the below parameters
as close as possible.  Following are from memory, I'm sure I'm forgetting
some things.
For the uplink:
- Deviation 3.2 KHz
- Frequency right on.  Check to assure connecting the modem to the
transmitter doesn't pull it off frequency (can be a problem if directly
connecting to the varactor without enough isolation)
- Tones balanced, one not any stronger in amplitude than the other.  Assure
balanced deviation
- No hum or RFI induced into transmitted audio
- Short low loss coax 
- Tuned antenna for max RF out, best match.
- If using a amp assure its output is clean, no hum, no RF feedback
- Set TXD to 80ms.  This is a safe value for just about any
transmitter/amplifier.  After other parameters are tested 'good', back this
down till the link goes bad then increase it till good again.  This gives
adequate time for the tx to lock up without excessive channel time spent on
flags.
- Assure the amp is pulling in on transmit without contact bounce or
excessive delay, or just lock it on.
- If using tracking antennas: Assure time in PC is within a couple of
seconds, elements are less than a month old, antennas direction is calibrated.
- Use circular polarization if possible.  Sense doesn't matter.  Uplink
antenna on the satellite is linear but it's orientation will change with
satellite motion, Faraday rotation, reflections and just about everything
else.

On the downlink:
- Use a mast preamp
- Low loss coax
- Assure receiver tuning 'beeps' don't feed into the demod and interfere
with the received signal.
- All the downlink antenna stuff
- Assure demod or TNC isn't inducing hum into the received audio
- Assure level from receiver into demod is adequate and not marginal
- Make receiver tuning steps as small as possible.  On a 1200 bps psk
signal into a typical TAPR demod a 100 Hz tuning step will kill a bit and
cause a packet to be lost.  Use a cheap audio O'scope to observe the eye
pattern for extra wave forms, noise, multipath, and loss of sync on tuning
steps, other stuff that will cause errors.
- If capable, switch polarization sense on signal fades.  Reflections,
multipath, Faraday rotation and other effects with change the sense during
a pass.  (Satellite motion won't, in the northern hemisphere.  But it will
for stations near the magnetic equator).

jw
wd0e@amat.org
jim@coloradosatellite.com

>
>Hummh...  Maybe the satlelite is not exactly on 145.88.  Lets say its
>receiver was really on 145.883, that would explain why I always have
>better success late in the pass and not early.  THus, it might not be QRM
>in central America, but just the effect of dopler and a few KHz off
>frequency,.
>
>This is another data point we must thoroughly test.  When we are all
>working correctly, someone needs to sweep +/1 5 KHz in 1 KHz steps and
>assure ourselves exactly where the center of the uplink channels are.
>
>And sicne each one has its own XTAL, then the results will be on a per
>channel basis.. 

---
You are currently subscribed to ao16aprs as: lyris.ao16aprs@tapr.org
To unsubscribe send a blank email to leave-ao16aprs-833L@lists.tapr.org