FW: FLL 90 UV Filter Data results

[Pei-lin Chiang]

---

From: ache...@hotmail.com
To: chiang...@hotmail.com
Subject: FW: FLL 90 UV Filter Data results
Date: Sun, 17 Aug 2008 18:00:44 -0400

---

From: ache...@hotmail.com
To: hale-...@googlegroups.com
Subject: FLL 90 UV Filter Data results
Date: Sun, 17 Aug 2008 17:55:07 -0400

      We uploaded the files from the NXT. The files were "FLL90_Time" and "FLL90_Filter." The pinwheel spun every five seconds, taking readings, then it slept for 55 seconds. So we did about twenty five readings per minute.
      We noticed after 144 minutes, there was a half hour without any readings. The next time it started taking readings was at 176 minutes (a half hour later.) We thought the payload was already on the ground, so we will ignore the data after 144 minutes.
      Most of our filter readings were in the 20's, 30's, or 40's. Occasionally in the 50's and the 100's. Our highest reading was 195, second was 188, and third was 185. Our lowest was 27. In general, the readings were very low, and it was  impossible to tell which readings matched which filters. We think that was because our payload went up at sunrise, but the probe was pointing straight up.
     We did get some high readings, but the highest reading was not at the highest altitude. We think that was because the box may have been tumbling a bit on the way down. Therefore, the probe sometimes faced the sun.
     We graphed the filter and the time readings. We used the Excel Chart Wizard. We chose to use a smooth XY (scatter) graph. It's a scatter with data points connected by smooth lines, without markers. The x-axis is minutes. The y-axis is UV readings.
    You can see our chart and raw data in the file section of the HALE group. The chart is called FLL90_uv_chart.xls.

Arianna Chen and Pooja Rathnashyam
Rising 6th Graders
Reston, VA    

EMAILING FOR THE GREATER GOOD Join me

---

Get thousands of games on your PC, your mobile phone, and the web with Windows®. Game with Windows

[Brian Davis]

On Aug 17, 6:03 pm, Pei-lin Chiang <chiangpei_...@hotmail.com> wrote:

Thanks for the update and the data team! Very cool.

> We noticed after 144 minutes, there was a half hour without any readings.
> The next time it started taking readings was at 176 minutes (a half hour
> later.) We thought the payload was already on the ground, so we will ignore
> the data after 144 minutes.

Yes, I think it was on the ground. I tried looking at your data in a
couple of other ways, and drew up some graphs of my own. According to
the GPS data, it landed at MET (Mission Elapsed Time) of 111 minutes.
Curiously, there are some high readings after that - I guess maybe the
payload got tumbled a bit for 5-10 minutes after landing. On the graph
"UVvsMET" (in the files section), I marked the when the balloon
popped, and when it landed. More curious, is how the lowest minimum
readings happened at the highest altitudes...

> In general, the readings were very low, and it was  impossible to tell which
> readings matched which filters. We think that was because our payload went
> up at sunrise, but the probe was pointing straight up.

Yes, but there does seem to be *some* regular variation. In each
"window" of readings, the highest readings almost always happen near
the beginning (look at "UVFilter" in the files section). The red data
is from a few series that seemed to be above background during the
descent - the blue data is from some of the records during ascent. I
"cherry picked" these records - I went through the data to find series
of records that weren't all low, but showed some high points -
evidence that for one reason or another, the UV sensor was getting at
least a little direct exposure.

Based on this, I'd guess that the "no filter" portion of the filter
wheel was always the first sector over the sensor. Is that right?

> We did get some high readings, but the highest reading was not at
> the highest altitude.

I also tried graphing the readings vs. the *altitude*, not just the
MET - this is a lot more complicated (no, I didn't arrange everything
by hand), but Excel is able to do it with some work. The result is
"UVvsAlt" in the files section. You can see that on the way up (blue
data), everything is fairly evenly spaced... but on the way down (red
data), there are far fewer data points "up high" than "down low"...
because really high, the payload didn't spend much time (it was
falling fast). Much more interesting is the position of the "no low
readings" section of your data - there's a point where the low
readings were in the very high 30's and low 40's, not the 20's and
30's like everwhere else. Curiously, this occurs only right above an
apparent "cloud deck" that I see in the data from my Gypsy payload.
I'm not sure what this means... but it's a rather close correlation.
It *could*, of course, be an accident, but I can think of at least two
other possibilities. First, maybe at that point the Sun had finally
risen high enough to start hitting the sensor, only to be cut off as
the payload dropped below the clouds. Second, maybe the clouds had
something to do with the higher "lowest" readings - maybe the cloud
deck was higher on your mission than my estimate (remember, yours was
in a different place and time), and it scattered light into the
sensor.

In any event, I'm sorry the data isn't exactly as you expected... but
there's a lot of interesting things in their I *didn't* expect. And I
always learn a lot more from all the stuff I *didn't* know, rather
than just having the stuff I *thought* I knew confirmed again. Good
job!

--
Brian Davis

[David Levy]

> Based on this, I'd guess that the "no filter" portion of the filter
> wheel was always the first sector over the sensor. Is that right?
>

Unfortunately no based the following observations

1) . The launch team would have to align the filter that way but
could never be exact because once on the axle they were instructed not
to turn the pinwheel by hand.

2) The pinwheel motor was not programmed to turn exactly 90 degrees for
each quadrant. This was partly due to #1 above. It was also thought
that a slight error on rotation would compound over time. We did have
a special design to try to come up with a mechanism using a touch
sensor to ensure an accurate rotation over each quadrant. But
unfortunately that design never materialized. In the end the team
opted to merely spin the pinwheel for 5 seconds. This was a time
observed to accomplish a 360+ degree rotation consistently over a 3
hour period.

The assumption was that for any given 1 minute segment in the data, the
highest values would be attributed to the zero filter, followed by the
sunglasses, cotton shirt, and then denim. Again this was a big
assumption because even at noon on the ground in a hot summer day in
Virginia, the three filters showed the same readings which were
interpreted as "no UV detection" at the time. The hope as to show that
more UV would come through the filters at higher altitudes and even
denim would prove ineffective at 100k feet.

There is a pretty good picture of the pinwheel here:
http://hale-teams.googlegroups.com/web/FLL_Team_90_HALE_Payload.pdf .
I believe it spin clockwise.

David

[Eric]

We did align it so that the "no filter" position would have been
first. Also, remember that Gypsy and your payload were on the same
balloon (first one launched) - along with th SLR camera. So you can
"see" what the view looked like by looking through the photos posted.
Of course, your payload won't be facing the same direction as the
camera, but you can get an idea in terms of cloud deck, etc.

Also remember, that we filled the payload with the foam peanuts to
help cushion the contents. Your students might want to think about
whether that helped or hurt in terms of keeping the payload warm (I
have my theories, but it should be an interesting exercise for them).

Eric