[Ground-station] Balloon Launch - experiments?

[Phil Karn]

On 4/22/18 21:10, Zach Leffke via Ground-Station wrote:
> Last note for today....I promise.
> 
> We've flown the Adafruit Ultimate GPS on multiple flights in the recent
> past and they are good up to at least 100k ft.  Breakout available
> here:  https://www.adafruit.com/product/746.  It also has a 1 PPS
> output. (don't forget the external antenna)

I've got this one too. Great little unit.

1pps is now becoming common on low-end GPSes. But getting it into a
computer in a way that it can be used is still tougher than it should be.

> In our case we kept things super simple and married the GPS to 'gpsd'
> and then wrote a simple little standalone python script that launched on
> boot that logged the NMEA strings to a file.  Allowed for 1 Hz
> timestamped balloon positions.

That's the way to go. gpsd is everywhere.

> As a byproduct......having the GPS based timestamp in a log that also
> had the 'OS' time stamp (from the python 'datetime' module) allowed us
> to figure out the time offset between the computer clock and true time
> (OS time was wildy wrong without NTP or an RTC to keep it on track
> through power cycles and when disconnected from internet).  We're
> experimenting with updating the computer clock automatically by tying
> gpsd and ntpd together...along with the 1 PPS reference from the above
> GPS.....but not there yet.

It seems to be taking forever for this stuff (automatic host clock
setting from GPS) to "just work". It's especially important with a Pi
since it has no battery-backed clock. You tend to forget this when you
run them on networks with reliable NTP service.

> Above is relevant to reversing out the filename timestamps for the IQ
> recordings (that would come from the OS and be wrong) so that you could
> get the true UTC time for the recordings........

Depending on how accurate you want to be, it might be worth looking at
how to get (some) of the 1pps pulses directly into the radio and
recorded so you can establish a precise time scale at the sample level
and determine the error in the sampling clock so you can correct it later.

Phil


-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0xF14E4354ED463446.asc
Type: application/pgp-keys
Size: 3301 bytes
Desc: not available
URL: <http://lists.openresearch.institute/pipermail/ground-station-openresearch.institute/attachments/20180422/e3c7e668/attachment.key>