Joe & Gypsy - Preliminary Results

[Brian Davis]

Well, I'm been busy with other things so it's taken me a while, but
I've finally got the first pass at the results from Lil' Joe and
Gypsy. I uploaded a single 2+ Mb zip file (JoeAndGypResults.zip) with
screenshots of all of the semi-cleaned up graphs I've produced with
the data (the Excel data files are currently topping about 4 Mb, and
aren't well documented, so I've not put them up yet... but certainly
will, and anybody who wants a portion of the data just ask & I'll try
to oblige!). Quick descriptions (Gyp means on Gypsy, Joe means from
Lil' Joe)

GypAmbSounds: Average and maximum sound levels during the mission.
Very quiet with some "creaks" or "pops", and then you can "hear" the
descent, even at that altitude. No way to know if this is actual
sound, or vibration through the structure (I suspect largely the
second).

GypBat: NXT battery level during mission. You can clearly see the
voltage drops when the platform was being moved, but other than that
there's just a slow variation in voltage largely (it seems) associated
with temperature.

GypCDS: The CDS cell readings - a large number indicates "bright",
while a small number indicates "dark". The orange trace along the
bottom shows the platform pitch angle, because sometimes the CDS cell
was facing up (I've colored those CDS datapoint red) and sometimes is
was facing down (colored green). You can find a cloud deck in the
data, and you can see the effect of the constant spinning around of
the platform - probably when the CDS sensor was facing the rising sun
it registered bright, while when it was facing away it would have been
dark.

GypMotorSounds: any time the platform used the pitch motor (the motor
that controlled the platform angle), it recorded the ambient sound
(black bars), the motor sound (big red dots), and the sound level of
an NXT "beep" at 100%. No big surprise, it gets very quite at high
altitude (very low pressure), but the motor noise still comes through
fairly well, almost certainly due to transmission through the
structure.

GypPitch: The measured platform angle over the course of the mission.
All of these pitch excursion are planned and executed correctly,
except the "drift" near the end. I'm not sure what happened here, i'll
have to check exactly what was programmed in the script files.

GypPres1: The reported pressure over the course of the mission.

GypPres2: The reported pressure as a function of altitude, which looks
very good. Since the sensor was at two very different temperatures
during ascent and descent, it seems the temperature correction I'm
applying to the sensor is fairly close.

GypProfile: The elevation profile for Gypsy & Lil' Joe from the GPS
data. Also here is the "ressure altitude" derived from the pressure
data, based on an assumption of a constant scale hieght (NOT a good
assumption, but simple). This also has all the elevations and some
velocities.

GypScript: The elevation profile "decorated" with colors to show where
(what elevation) Gypsy was trying to take pictures and video. I still
don't know what was up with the camera, but this is exactly the
profile of pictures & video I was expecting, so the NXT-side of the
payload was working flawlessly here.

GypTemp: Internal temperatures for Gypsy. It got cold :)

Joe3Axis1: The acceleration data from Lil' Joe, showing all three axis
separately. The "Z" axis maxes out the sensor for a good part of the
deceleration, but probably not by too much.

Joe3Axis2: A small part of the above graph, magnified to show in more
detail what's going on.

Joe3Axis3: A detailed 3 axis view of the acceleration chaos during the
peak accelerations.

JoeCutdown: High-resolution view of the actual cut-down of Lil' Joe
and the accelerations immediately following. True "Zero G" only lasts
for about 1.5 seconds, after which the payload starts swinging back
and forth before flipping over (remember Lil' Joe was dropped with the
low-density, high-drag tail down, so it should have flipped at some
point, and it looks like this is the first thing it did before 10
seconds had elapsed.).

JoeNetAccel1: The net accelerations (what you would "feel", ignoring
direction) during the first two minute of free-fall. Note there is a
strong general acceleration that stops after about 28 seconds, and
then a long turbulent "tail" of accelerations. The parachute
deployment is there... but doesn't look like it has any real effect,
implying at least at this point it either deployed early (hard to say)
or had almost no effect visible in the data.

JoeNetAccel2: The net accelerations in more detail for the first 40
seconds.

JoeNetAll: The net accelerations for the entire descent. Two minutes
of detailed data (30 Hz sampling), followed by 13 minutes of "window"
sampling (detailed logging of a 2 second window once a minute). Allows
the landing time to be deduced to within a minute independent of the
GPS data.

JoeProfile: The actual GPS data, showing the descent elevation vs
time, and also where the GPS logger did not record any data. Some
artifacts are clearly visible, but the velocity estimates here have
worked around these and not included anything obvious.

JoeWindow9, 10, 11, & 12: Detailed views of several of the 2-second-
wide "windows" logged during descent. It seems the payload/parachute
combination spent a lot of time descending like a dumbell, with the
parachute well out to the side of the payload and the two spinning
around held together by the parachute tether.

Let me know what in this doesn't make sense, or where anybody wants
more detail or actual data.

--
Brian Davis