Up late indeed!  News from the late-night crew:

- Pressure sensor data is now getting to the FC.  A hardware fix was
required to swap SCL and SDA pins.  Plus lots of firmware magic.
- PiCam2 stopped working a couple nights ago.  We thought it was just a
bad cable, but it turned out that the Pi had been baked with too many
volts in the wrong place.  We did all the flight mods on a new Raspberry
Pi and replaced picam2.
- Another system test, mostly to see WiFi loaded with full telemetry and
video come up from scratch.  It worked but is still finicky.  We did one
of these on Tuesday and the link bitrate was stuck at 1 MBit/s which is
too low to deliver our telemetry.  I did a bunch more testing Wednesday
night and eventually came across the right bring-up order and event
response matrix to get things working with reasonable bitrates.  It
turns out the wireless driver we are using just doesn't support being
told what bitrate to use.  And some times it gets stuck on the wrong
- We managed to crash the RNH with Theo in the room, so he pulled an
SRAM dump to sort through later.  I crashed it again after he left.  But
we have a physical reset method exposed outside the rocket now, so these
crashes won't stop a launch.
- Some of the TM4K stability issues may have been related to logs
filling the disk, so I added a 32GB SD card which should provide about
30 hours of active telemetry and video capture.  Still need to get Cacti
configured for monitoring disk and other system resources network-wide.
- PiCam2 (store only, no streaming) bitrate limit has been removed so we
will get the full 1080p experience.  At around 1 MByte/s to the local SD
card, we can record for about 3.4 hours.  Starting up the cameras
happens right before we hit the big red GO button, so this will work
- People have been putting in lots of hours this week making all the
things work.  This was just a summary of Thursday night.

The testing Wednesday night included a Venus (COTS) GPS lock from a FAB
basement window with WiFi amps blaring while streaming video around a
corner and down the hall to the FDT.

Testing Tuesday night included getting Venus lock to show up on the
telemetry page for the first time.  And post-processing the raw GPS data
collected from the jGPSv3 board we are actually flying, connected to our
flight antenna.  The data correlated and lined up with expectations
based on the satellite constellation at sample time.

Not to sound too optimistic, but this might just work.  Let's go launch
this thing!


On Thu, 2015-07-16 at 21:40 -0700, Andrew Greenberg wrote: 
> Hi everyone,
> we'll be down in the rocket room (and around FAB) testing the avionics
> system tonight. Come down and test with us if you can! We'll be up...
> late :)
> Andrew

Attachment: smime.p7s
Description: S/MIME cryptographic signature

psas-avionics mailing list

Reply via email to