A LOW-COST LINEAR-RESPONSE WIRELESS TEMPERATURE SENSOR FOR EXTREME ENVIRONMENTS

eggplantcinnabarΚινητά – Ασύρματες Τεχνολογίες

21 Νοε 2013 (πριν από 3 χρόνια και 6 μήνες)

151 εμφανίσεις


A LOW
-
COST LINEAR
-
RESPONSE
WIRELESS
TEMPERATURE SENSOR
FOR EXTREME ENVIRONMENTS



A Thesis Presented

by

Michael Fortney

to


The Faculty of the Graduate College


of

The University of Vermont



In Partial Fulfillment of the Requirements

For the Degree of
Master of Science

Specializing in Electrical Engineering




February
, 2007









1

Accepted by the Faculty of the Graduate College, The University of Vermont, in partial
fulfillment of the requirements for the degree of Master of Science, specializing in
E
lectrical Engineering.






Thesis Examination Committee:




______________________________


Jeff Frolik, Ph. D.



Advisor






______________________________



Steve Titcomb
, Ph. D.





______________________________



Jun Yu
, Ph. D.



Chairperson






__
____________________________



Frances Carr, Ph. D.



Vice President of Research



Date:
December
7,

2006


Abstract





This Thesis contributes to the area of Wireless Sensor Networks (WSN) by
presenting a system level evaluation of design strategies and r
outing algorithms.
Specifically, power consumption (and consequently, network lifetime) and cost of
implementation are considered. To date, these two facets of WSN design have inhibited
the deployment of sensor networks on a large scale envisioned by man
y researchers.
Thus, this research strives to minimize these parameters whenever the opportunity
presents itself.

Existing hardware implementations were deemed too complex and too
costly
……………………….
.

Specifically, the objectives of this research include
:

1)
Maintain a simple
, low
-
cost
, wireless

design
for educational purposes
.

2)

Provide wide
-
range
atmospheric
temperature measurements

from
-
90 ºC to +60 ºC
.

3)
Provide a linear response circuit for improved resolution and simplified calibration
.

4) Evalua
te the design, testing accuracy and linearity across the specified temperature
range.


The platform developed for this research is inherently open ended. Through the
use of a simple I2C communications protocol
…. Results and costs…..






Acknowledgements


Frolik, Keller


Vermont EPSCoR EPS 0236976

Vermont Space Grant

HELiX








4

Table of Contents

Acknowledgements

3

List of Tables

6

CHAPTER 1: INTRODUCTION

................................
................................
................................
..................

10

1.1.

Thesis Motivation

................................
................................
................................
.......................

10

1.2.

Thesis O
bjective

................................
................................
................................
.........................

10

1.3.

Contributions

................................
................................
................................
..............................

13

1.4.

Thesis Organization

................................
................................
................................
....................

13

CH
APTER 2: ADAPTATION OF A LOW
-
COST WIRELESS SENSOR FOR FRESHMAN AND
OUTREACH PROGRAMS

................................
................................
................................
.......

15

2.1.

Intro to Paper

................................
................................
................................
..............................

15

2.2.

Abstract

15

2.3.

Introduction

................................
................................
................................
................................

15

2.4.

The CricketSat System

................................
................................
................................
...............

16

2.5.

UVM Cri
cketSat Development and Testing

................................
................................
...............

18

2.6.

High School Outreach

................................
................................
................................
................

20

2.6.1.

2003


2004 HELiX Team

................................
................................
................................
.........

21

2.6.2.

2004


2005 HELiX Teams
................................
................................
................................
........

22

2.7.

College Freshman Engineering Course

................................
................................
......................

27

2.8.

Conclus
ions

................................
................................
................................
................................

28

2.9.

Acknowledgements

................................
................................
................................
....................

29

CHAPTER 3: CRICKETSAT SENSOR AND SYSTEM DEVELOPMENT

................................
..............

29

3.1.

Introduction

................................
................................
................................
................................

29

3.2.

The Original Stanford CricketSat Design (1999)

................................
................................
.......

31

3.3.

UVM CricketSat Worksho
p Planning (Spring 2003)

................................
................................
.

33

3.4.

UVM CricketSat Revision A (Spring 2003)

................................
................................
..............

34

3.5.

2003 CricketSat Workshop and Test Flights (June

16


June 20)

................................
..............

38



5

3.6.

UVM CricketSat Test Flight (08/08/2003)

................................
................................
.................

41

3.7.

GIV CricketSat Workshop and Test Flight (08/09/2003)

................................
...........................

43

3.8.

HELiX CricketSat Weather Stations (12/2003


02/2004)

................................
........................

44

3.9.

UVM CricketSat Revision C (January 2004)

................................
................................
.............

48

3.10

UVM CricketSat Revision D (May 2004)

................................
................................
..................

49

3.11.

2004 CricketSat Workshop and Test Flights (June 14


June 18)

................................
..............

51

3.12.

First Collaborative BalloonSat Flight (July 17, 2004)

................................
...............................

54

3.13.

Second Collaborative BalloonSat Flight (July 30, 2004)

................................
...........................

55

3.14.

UVM CricketSat Revision E (December 2004)

................................
................................
.........

61

3.15.

UVM CricketSat Revision F (May 2005)

................................
................................
..................

63

3.16.

Conclusions

................................
................................
................................
................................

65

CHAPTER 4: A LOW
-
COST LINEAR
-
RESPONSE TEMPERATURE SENSOR FOR EXTREME
ENVIRONMENTS

................................
................................
................................
....................

68

4.1.

Intro to Paper

................................
................................
................................
..............................

68

4.2.

Abstract

68

4.3.


Introduction

................................
................................
................................
...............................

68

4.4.


555
-
Timer Astabl
e Oscillator

................................
................................
................................
....

72

4.5.


Linear Frequency Control Methods
................................
................................
...........................

76

4.5.1.

Threshold Control Voltage Method
................................
................................
............................

76

4.5.2.

Ladder Voltage Control Voltage Method

................................
................................
...................

81

4.5.3.

Current Source Method

................................
................................
................................
..............

84

4.5.4

Compa
rison of alternative methods

................................
................................
............................

87

4.6

Implementation and Test

................................
................................
................................
............

88

4.61

Implementation of a linear sensor

................................
................................
..............................

88

4.62

Test Scenarios

................................
................................
................................
............................

89

4.63

Test Results

................................
................................
................................
................................

91

4.7

Conclusions

................................
................................
................................
................................

96



6

4.8

Acknowledgement

................................
................................
................................
......................

96

CHAPTER 5: SUMMARY AND CONCLUSIONS

................................
................................
.....................

98

5.1.

Conclusions

................................
................................
................................
................................

98

5.2.

Future Research

................................
................................
................................
........................

100

5.3.

Final Thoughts

................................
................................
................................
.........................

101

Bibliography

101


List of Tables

Table 2.1:

HELiX CricketSat Workshop Schedule

................................
.........................

20

Table 3.1:

UVM CricketSat sen
sor Revsions A
-
E and proposed Revision G.

................

65

Table 3.2:

UVM CricketSat significant milestones.

................................
........................

66

Table 4.1:

Comparison of simul
ation results for various the oscillator control methods.

87

Table 4.2:

Calibration methods and sources

................................
................................
....

93

Table 4.3:

Mean error
/standard deviation (ºC) produced using various calibration
methods. “Z” calibration includes absolute zero data point, “NZ” does not.

..................

94


List

of Figures


Figure 2.1:

CricketSat Wireless Temperature System

................................
................................
..................

17

Figure 2.2:

UVM CricketSat Wireless Temperature Sensor

................................
................................
........

19

Figu
re 2.3:
The John D. O'Bryant School CricketSat presentation of the MHS
-
1 flights. The team received
a 2
nd
-
place finish at the 2005 Boston Regional Science Fair

................................
................................
.........

23

Figure 2.4:
Medgar Evers Col
lege (MEC) team preparing for a BalloonSat launch at the Milton High
School. The CUNY school provided flight support for the MHS
-
2 payload.

................................
..............

24



7

Figure 2.5:

The Milton High School CricketSat Ar
ray System flown on the MHS
-
3 flight. This project
allows the measurement of several CricketSat sensors over 50 miles away.

................................
................

25

Figure 2.6:

Properly segregated data collected from the Milton H
igh School developed CricketSat Array
System. Raw frequency results are shown.

................................
................................
................................
..

26

Figure 2.7:
Example student design projects: Wireless Wind
-
Chill Instrument (left) and Wireless Door
Alarm (r
ight)

................................
................................
................................
................................
.................

28

Figure 3.1:
The original Stanford CricketSat schematic and completed circuit board.

................................

32

Figure 3.2:
UVM initial redesign

(RevA) of the Stanford University CricketSat.

................................
.......

35

Figure 3.3:
UVM RevA CricketSat temperature, pressure and humidity sensors. Pressure and humidity
sensors are inserted in the prototype area.

................................
................................
................................
.....

37

Figure 3.4:
CricketSat flights conducted from the bridge near the Sandbar State Park. Students are shown
preparing, launching, and tracking CricketSat sensors.

................................
................................
.................

39

Figure 3.5:
Results from a pulse
-
mode CricketSat launched from the UVM campus experience a balloon
burst.

................................
................................
................................
................................
..............................

42

Figure 3.6:
CricketSat weather station loca
ted at the Waldorf High School fitted with a solar panel for
long
-
term use.

................................
................................
................................
................................
................

45

Figure 3.7:
The Waldorf CricketSat weather station. Plastic
-
plate and threaded
-
rod construction (left),
stacked
CricketSat sensors (center), sensor selection switch and battery (right).

................................
..........

46

Figure 3.8:
CricketSat receiver designed to work with the CricketSat weather station and was later used for
all Cri
cketSat applications.

................................
................................
................................
............................

47

Figure 3.9:
Schematic diagram of the UVM RevC CricketSat.

................................
................................
...

48

Figure 3.10:
CricketSat RevC temperature s
ensor circuit board. Green solder mask, white labeling and
strain
-
relief holes were added to this design.

................................
................................
................................

49

Figure 3.11:
CricketSat RevD schematic diagram. A voltage regulator was added fo
r oscillator stability
and active sensors.

................................
................................
................................
................................
.........

50

Figure 3.12:
CricketSat RevD temperature and pressure sensors.

................................
................................

51

Figure 3.13:
Stu
dents using the Spectra RTA software to perform CricketSat calibration measurements.

.

52



8

Figure 3.14:
CricketSat daytime temperature sensor flight. The sensor experienced a 114 ºF temperature
change during the flight.

................................
................................
................................
................................

53

Figure 3.15:
Night
-
flight data collection from a CricketSat pressure sensor used as an altimeter showing a
linear ascent rate.

................................
................................
................................
................................
...........

53

Figure 3.16:
Students prepare for a BalloonSat flight from the Milton High School. Temperature data (red
trace) in the figure to the right shows the flight bag temperature during flight.

................................
............

54

Figure 3.17:
CricketSat sensor array timing diagram.

................................
................................
..................

56

Figure 3.18:
CricketSat power sequencing circuit using a BASIC Stamp II controller.

..............................

57

Figure 3.19:
CricketSat Sensor Array System

................................
................................
.............................

57

Figure 3.20:
BalloonSat flight path (left) and raw segregated CricketSat temperature and pressure d
ata
(right) received during the flight.

................................
................................
................................
..................

58

Figure 3.21:
Converted CricketSat temperature data (left) and altimeter data (right). CricketSat data is
compared to known data for sensor validatio
n.

................................
................................
.............................

60

Figure 3.22:
CricketSat RevE circuit board.

................................
................................
................................

61

Figure 3.23:
CricketSat RevE schematic.

................................
................................
................................
....

62

Figure 3.24:
CricketSat RevF schematic.

................................
................................
................................
....

64

Figure 3.25:
CricketSat RevF circuit board.

................................
................................
................................

64

Fig
ure 4.1:

CricketSat non
-
linear frequency response to temperature

................................
.........................

71

Figure 4.2:

555
-
timer astable oscillator CricketSat design

................................
................................
...........

73

Figure 4.3:

555
-
timer internal circuitry composed of the threshold voltage ladder, comparators, RD latch
and discharge transistor. Image for Texas Instruments NE555, SA555, SE555 Precision Timers data sheet.

................................
................................
................................
................................
................................
.......

74

Figure 4.4:

PSpice astable timing waveforms of the 555
-
timer circuit. The top trace shws the voltage on
the timing capacitor. The bottom trace represents the digital output of the timer.

................................
.......

75

Figure 4.5:

PSpice simulation circuit for using a voltage source (
V
2
) to override the timer’s internal
threshold compare voltages. The 5
-
Volt source is used to represent the regulated voltage, which is now
necessary for this mode of
operation.

................................
................................
................................
............

77



9

Figure 4.6:

Simulation waveforms showing the effect of control voltage on the charging cycle of the
timing capacitor. The top and bottom waveform shows the capacitor voltage usi
ng a control voltage of 4.0
Volts and 2.0 Volts, respectively. Note that the lower limit is always one
-
half the control threshold value.

................................
................................
................................
................................
................................
.......

78

Figure 4.7:

Simulation results demonstrating 555
-
t
imer frequency response to varying threshold control
voltage.

................................
................................
................................
................................
..........................

81

Figure 4.8:

Simulation circuit used to investigate the effect of RC ladder voltage (
V
2
) on timer frequency.
In comparison to F
ig. 4.5, the value of R
2

has been reduced to increase duty cycle and minimize the non
-
linearity.

................................
................................
................................
................................
........................

82

Figure 4.9:

Simulated response of timer frequency to control of RC ladder voltage.

................................
..

84

Figure 4.10:

PSpice simulation circuit used to investigate the use of a current source method for 555
-
timer
frequency control.

................................
................................
................................
................................
..........

85

Figu
re 4.11:

Schematic showing implementation of the LM234 current source device as a temperature
sensor. The value of R
1

(i.e. R
SET
) is selected to provide a suitable sensitivity and frequency range.
..........

89

Figure 4.12:

CricketSat module modified using the LM234 current source for use as a temperature sensor.

................................
................................
................................
................................
................................
.......

90

Figure 4.13:

Test results for the thermistor
-
based and current
-
source
-
base C
ricketSat temperature sensors.

................................
................................
................................
................................
................................
.......

92

Figure 4.14:

Figure to the right demonstrates the linear response of the CricketSat oscillator circuit over a
wide current range. The highlighted region near

the origin is shown in the left figure, showing the trend
line passing near the origin.

................................
................................
................................
...........................

95

Figure 5.1:
Hopper

node and
ALOHA
node photo

................................
................................
........................

99









10


















CHAPTER 1
: INTRODUCTION

1.1.

Thesis Motivation




Importance of wireless sensors

o

General

o

Interesting apps

o

Problems with these

o

CricketSat

o

Stepping stone forward…


1.2
.

Thesis Objective



The work presented herein seeks to address
….





11












12


Specifically,
t
he main objectives of this thesis are as follows:

1

Develop
low
-
cost, wireless, sensor
hardware that can

measure extreme atmospheric
temperatures from
-
90

º
C to +60

º
C
.

This range encompasses the recorded limits of atmospheric temperatures measured on the
s
urface of the earth. It also accommodates
most
atmospheric measurements within 30 km
of the surface, typically experienced with weather balloon flights.

2

Provide a linear response design
to for improved resolution and to
simplify
calibration procedures
.

A linear response sensor provides improved resolution over a
wide
dynamic temperature
range than a thermistor
-
based solution. The linear design also simplifies sensor
calibration since the
sensor

response
may

directly propo
rtional to absolute zero, allowi
ng

minimal calibration to be performed at conveniently warmer temperatures.

3

Evaluate the performance of the linear sensor design calibration procedures.

Testing of the linear design to assess its linearity and accuracy using a calibrated
temperature sens
or
and comparing the results to prior art.

4

Considering 1
-
3 above, maintain
a
simple

circuit

design for middle school, high
school and college applications.

A primary application for this sensor is for use in the classroom for engineering outreach
and to pe
rform atmospheric measurements for earth science studies. The design should
use common off
-
the
-
shelf components, demonstrating basic electronic principles and
avoiding the use of components requiring special programming. Component cost should
be kept low
, affording schools to provide kitted sensors for each student or small teams to


13

assemble.



1.3
.

Contributions



Linear temperature sensor



Low
-
cost receivers



Assembly and usage documentation



Non
-
linear pressure, humidity, light



Multi
-
sensor arrays

o

Stamp,
C
S
onde
, LP apps



Custom
application designs
for
first
-
year

d
esign class



HELiX outreach


To meet the objecti
ves outlined above…



The key contributions of this work

.


To evaluate
the
CricketSat

design, both simulation and laboratory exper
iments..
.



System l
evel development

o

Spectrogram

o

Stamp
-
based multi
-
sensor


1.4
.

Thesis Organization


This chapter presented an introduction to
..
.





14






The differences between the existing designs and the proposed
CricketSat

were
highlighted. The remainder of this thesis
is organized as follows. Chapter 2
addresses the
educational aspects of the CricketSat
.

Chapter 3 describes the
development of
the UVM
CricketSat sensor revisions A through F
. Chapter 4 describes

the design and analysis of
the linear response CricketSat
design
. Finally, Chapter 5 summarizes the key results of
this research
, addresses present ongoing activity

and proposes further improvements and
avenues for future
research.





15

CHAPTER 2
:

ADAPTATION OF A LOW
-
COST WIRELESS SENSOR FOR
FRESHMAN AND OUTREACH

PROGRAMS


Mike Fortney and Jeff Frolik

University of Vermont

Underrepresented Groups in Engineering



2.1
.

Intro to Paper

This paper was
published in
the ASEE Journal
, Vol #, date

...

and was presented at the
p
roceedings conference, Connecticut, date.

I
t describes the use of the CricketSat sensor
at the Univ
ersity of Vermont for a first
-
year

engineering design course and high school
engineering outreach.


2.2
.


Abstract

This paper details the development of new CricketSat designs and education programs a
t
the University of Vermont (UVM). UVM first explored the use of this wireless sensor in
Summer 2002 after attending a NASA Starting Student Space Programs workshop. Work
at the university has since involved improvements to the design to expand functionali
ty
and facilitate successful student circuit assembly. High school, undergraduate and
graduate level students are involved with CricketSat sensors and systems, design and
testing. Collaborative and outreach programs involve other institutions.


2.3
.

In
troduction

The CricketSat wireless temperature sensor was originally designed in 1999 at
Stanford University's Space System Development Laboratory as part of the NASA Space
Grant "Crawl, Walk, Run, Fly" student satellite program
i
. The purpose of this NAS
A


16

program is to instruct students into methods of space hardware development. Student
satellites range from the simple balloon
-
borne CricketSat to the more complex earth
-
orbiting CubeSat. To assist colleges and universities in developing their own progra
ms,
"Starting Student Space Hardware Programs" workshops are held frequently at the
University of Colorado campus in Boulder
ii
. The workshop covers the range of student
satellite designs, with emphasis on the BalloonSat program.

Representatives from UVM at
tended workshops and a CricketSat program was
implemented at the University in 2002. UVM CricketSat objectives involve
improvements to the original design, its use as an educational tool, and outreach
activities. This paper discusses the following activi
ties which take place within the
program:

1.

CricketSat development and testing

2.

Collaborative work with other colleges and universities

3.

High school outreach

4.

Freshman introduction to engineering course


2.4
.

The CricketSat System

The CricketSat system (Fig.
2.
1) is composed of a single wireless sensor and a
receiving station. The CricketSat transmitter contains a simple, 555 timer
-
based circuit
that produces an audio tone that changes frequency in response to changing temperature.
This tone amplitude modulat
es a 434 MHz carrier. Calibration of the sensor is performed


17

by measuring the tone frequency taken at various temperatures. From the calibration,
graphs are produced for converting frequency to temperature during use.


Figure
2.1:

CricketSat Wireless T
emperature System

For flight, the CricketSat device can be attached to a helium balloon as small as 2
feet in diameter. During flights, the 434 MHz signal is received by the ground station.
The ground station consists of a Yagi antenna, a UHF radio rece
iver and an audio
frequency measurement device. The frequency of the tone is measured with a frequency
counter or audio
-
spectrum software. Flights have been tracked for 90 minutes before the
signal becomes too weak to measure reliably. During this time,
the balloon may travel a
distance over 150 km, reach an altitude of 10 km and experience temperatures less than

70


C. The sensor and balloon are seldom recovered.

BalloonSat is a much larger system, toting a payload of several pounds, and a
price in exc
ess of ~$500. This system contains a GPS device and a radio transmitter used
to broadcast the position coordinates for tracking during flight. Sensor data is usually
collected and stored during the flight. Recovery of the payload is necessary for the
ex
pensive equipment and data. Complexity, cost, and logistics for flight preparation and


18

tracking may make this system undesirable. Flights may achieve altitudes of 30 km in
100 minutes before the balloon bursts and the payload parachutes back to earth.

In comparison to BalloonSat, the CricketSat system has benefits of low cost
($10), low weight, and live data telemetry. It is also simple to understand, easy to
assemble, and simple to use. Drawbacks include single
-
sensor operation, and
requirements for c
alibration and frequency conversion. Work at the UVM is involved
with improving the performance and flexibility of this device.

2.5
.

UVM CricketSat Development and Testing

Development work at UVM includes improvements to the original design,
adaptations
for new sensing capabilities, and the design of multi
-
sensor systems.
Improvements have also been made to better the electrical performance, system
reliability, and the likelihood of successful assembly. Concerning the latter, component
outlines and desi
gnations have been added to the board, and a protective layer to
minimize electrical shorts. A prototype area has been expanded to support student
adaptations to the CricketSat design. The most recent UVM CricketSat design is shown
in Fig. 2
.2
.



19


Figure 2.
2
:

UVM
CricketSat Wireless Temperature S
ensor

One objective of the UVM CricketSat work is to extend the circuit’s capability
beyond simply measuring temperature. Common sensors for air pressure, humidity,
light level, and accel
eration can be now accommodated in the design. In addition, easily
built sensor circuits
iii

will also interface with this design. With this added flexibility, the
CricketSat may used as a platform for a wide variety of wireless sensor applications. To
date,

multi
-
sensor CricketSat designs have also been developed by both university and
high school students. Eventually, the improved circuitry will lead to the development of
a low
-
cost student radiosonde (CricketSonde) containing meteorological and other
scie
ntific instruments. Such a design may enable community
-
based, meteorological
measurements at a much finer spatial resolution than those currently available using the
current network of National Weather Service radiosonde stations. This work toward this
e
nd is detailed in the following sections.




20

2.6
.


High School Outreach

The HELiX (Hughes Endeavor for Life Science Excellence) Program at UVM


HELiX/EPSCoR
iv

is a NSF funded outreach program supporting area college and
high school students. Among the HELiX

activities is a summer workshop for high
school students that is designed to provide students insight into the "real world" of
science. Teams, consisting of a teacher and a few students, conduct a research project,
assisted by scientists at the universit
y. At least one of the students must be female. A
HELiX
-
sponsored CricketSat workshop titled "Building and Launching Cricket Satellites
to Measure Various Atmospheric Conditions" was conducted during the summers of
2003 and 2004 (one is also planned for Ju
ne 2005). The one
-
week session, outlined in
Table
2.
1, involves lectures and hands
-
on activities for the students and teachers.
Classroom instruction includes an introduction to the earth's atmosphere and operation of
the CricketSat sensors. Hands
-
on ac
tivities involved the assembly, soldering, calibration,
and flight of these sensors. Students fly balloons, collect data and analyze the results.
School teams must then conduct a related research project to be conducted over the
following year.

Table
2.
1:

HELiX CricketSat Workshop Schedule

Day

Activity

Monday

Temperature profile of the atmosphere

Introduction to the CricketSat sensors

Practice soldering

Tuesday

CricketSat assembly

CricketSat testing

Wednesday

CricketSat calibration

Thursday

Balloon
flights and data collection



21

Friday

Spreadsheet data entry

Analysis and results


The high school teams have provided an important role of testing and evaluating the
CricketSat sensor designs. These teams have also developed and evaluated more
complex mul
ti
-
sensor CricketSat designs. Each balloon flight strives to surpass previous
results, contributing towards advancing the system. Parameters analyzed for each flight
are duration, distance, minimum temperature and maximum altitude.

2.6
.
1
.

2003


2004 HE
LiX Team


This initial workshop (2003) involved a team from the Waldorf High School in
Charlotte, Vermont, consisting of a female science teacher and three female students.
The June flights involved the initial testing of the newly developed pressure and

humidity
sensors and ground station receiving system. Balloons were released over the water from
a causeway on Lake Champlain. The results were not very encouraging. Frequency
measurements using a meter became unstable less than 10 minutes into the fli
ght. The
meter did a poor job of measuring the signal in the presence of background radio noise.
Expected qualitative variations were observed for the temperature, pressure and humidity
sensors.

In short, the system worked satisfactory for close
-
range
work, but not for balloon
flights. As such, for their long term research project, the team decided to build a wireless
weather station consisting of temperature, pressure and humidity sensors. The goal was
to make automatic measurements at periodic interv
als. A frequency measurement meter
was connected to a computer for data collection. Since all of the CricketSat sensors share


22

the same radio frequency, a rotary switch was added to the station to provide power to the
CricketSat sensor to be measured. Th
e system worked for three
-
hour intervals before the
meter would turn itself off. This appeared to be caused by the meter’s inability to
properly track the changing signal.

A new method was devised for performing the frequency measurements during
balloon
flights. Several audio spectrum analyzer programs were investigated. These
programs allow for individual frequencies in the audio signal to be “seen” on the
computer and measured. The CricketSat signal was easily identifiable, even when far
away, allowi
ng it to be measured reliably. The SpectraRTA
v

software was selected at the
time due to its data logging capability. Spectrogram
vi

is now recommended due to its low
cost.

2.6
.
2
.


2004


2005 HELiX Teams

This improved platform was utilized for the second H
ELiX workshop (2004) in
which two high schools participated. The first team was from the Milton High School
located in Milton, Vermont. The team was composed of a female science teacher and
two female students. The second team was from the John D. O'Bry
ant (JDOB) School of
Mathematics and Science (Boston Public Schools) located in Roxbury, Massachusetts.
This team consisted of a female science teacher and two students: one female and one
male.

Day and evening CricketSat flights (MHS
-
1) to monitor atmosp
heric temperature
profiles were conducted. A CricketSat humidity sensor was also flown along with an
experimental audio alarm device attached. Collectively, the flights were a remarkable


23

success. The SpectraRTA software performed well, allowing measurem
ent of the tone
signals for a much longer period of time than the previous method using a frequency
meter. The shortest flight was tracked for 45 minutes and the longest for 91 minutes.
This allowed for measurements much higher in the atmosphere. Accordi
ngly, the lowest
measured temperature was
-
41


F (
-
40


C), and the highest recorded altitude was 26,732
feet (8.1 km). Factoring in the velocities of the upper
-
air winds obtained from the
National Weather Service (NWS), the longest CricketSat altimeter fl
ight was 144 km.

For their follow
-
on project, the JDOB team prepared a detailed presentation of the
MHS
-
1 flights, placing second at the 2005 Boston Regional Science Fair in March 2005
(Fig.
2.
3). They will go on to compete at the Massachusetts State Sci
ence Fair to be held
in May at MIT.



Figure

2.3:
The John D. O'Bryant School CricketSat presentation of the MHS
-
1 flights. The team received
a 2
nd
-
place finish at the 2005 Boston Regional Science Fair


For the Milton High School team, their work was jus
t beginning. The school
hosted two BalloonSat flights in July 2005 for students from Medgar Evers College


24

(MEC) of the City University of New York (CUNY). CricketSat sensors were flown as
payload to provide real
-
time flight support data for the MEC team
. These flights also
allowed CricketSat sensors to achieve altitudes and conditions not normally experienced
with the smaller balloon flights.


Figure
2.
4
:
Medgar Evers College (MEC) team preparing for a BalloonSat launch at the Milton High
School. Th
e CUNY school provided flight support for the MHS
-
2 payload.

For the first BalloonSat flight (MHS
-
2), the MHS team monitored the temperature
inside the BalloonSat instrument flight bag (Fig.
2.
4). The temperature was measured for
125 minutes and never dro
pped below 63


F (17


C). For the MEC team, this validated
the use of the insulated lunch bag for holding instruments during BalloonSat flights. This
experiment also demonstrated the compatibility between CricketSat and BalloonSat
payloads concerning rad
io co
-
interference.

With the successful results of the single CricketSat sensor on the MHS
-
2 flight,
the team was now presented a challenge of measuring data using several CricketSat
sensors. Unlike the earlier weather station, this system would need to s
equence through
the sensors automatically. The problem was presented to the Milton team, and with a


25

little guidance, they devised a sequential timing algorithm for segregating and identifying
sensors during flight. In addition, the design needed to be li
ght weight for the balloon
application. A circuit was designed for the students using a BASIC Stamp controller.
One ambitious student assembled the circuit, wrote a PBASIC program employing the
timing algorithm, and tested the controller.



Figure
2.
5
:

The Milton High School CricketSat Array System flown on the MHS
-
3 flight. This project
allows the measurement of several CricketSat sensors over 50 miles away
.

The CricketSat Array System (CAS) assembled for flight is shown in Fig.
2.
5.
During the Ballo
onSat flight (MHS
-
3), the CricketSat flight bag and external
temperatures were measured, as well as altitude (air pressure). The system worked very
well, properly segregating the data from the various CricketSat sensors, as seen in Fig.
2.
6.



26


Figure
2.
6
:

Properly segregated data collected from the Milton High School developed CricketSat Array
System. Raw frequency results are shown.

New levels of performance were achieved. The flight was tracked for 134
minutes, to an altitude of 85,781 feet (26 km), a
nd with a bone
-
chilling external
temperature of

92


F (
-
69


C). The CricketSat altimeter worked properly below 32,000
feet (10 km), meeting expectations. External temperature versus altitude data correlated
with NWS sounding balloon data. The CricketSa
t altimeter data agreed well with altitude
data provided by the onboard GPS. The results were presented at the Northeast Regional
Space Grant Conference held in October 2004 in South Burlington, Vermont by the
author and the two Milton High School student
s.

Design changes to the CricketSat are necessary for improvement to the
temperature and pressure measurements. As can be seen in Fig.
2.
6 (
Ext Temp 1 & 2
),
for very low temperatures, the CricketSat frequency is very low and difficult to measure.
In addi
tion, the CricketSat pressure sensor only works properly up to altitudes of 10 km.
To completely characterize the environment experience during these large balloon
launches, the sensors need to perform measurements to altitudes of 30 km with


27

temperatures
as low as

90


C. As such, thesis work is in progress by the author towards
the development of a linear frequency response CricketSat temperature sensor for use in
extreme cold environments. This may lead to the development of a family of linear
response

CricketSat sensors for radiosonde experiments (i.e., CricketSonde). The linear
response provides benefits of uniform sensitivity, and simplified calibration and
conversion methods. The sensors will be evaluated on future HELiX and BalloonSat
flights.


2
.7
.


College Freshman Engineering Course

As a result of the above successful programs, the CricketSat was chosen in Spring
2004 as a project platform for UVM’s freshman design course for electrical and
mechanical students (instructed by the co
-
author)
vii
. I
n this course, students first
fabricate, test and calibrate the basic wireless temperature sensor. Then, working in
teams, the 60 students adapted over a six week period this sensor for an application of
their own choosing. The project requires electrica
l modification of the circuit and
mechanical design requisite of the application. Student projects from the first offering
included a wireless wind
-
chill instrument (Fig.
2.
7
-
left), a wireless synthesizer and a
wireless alarm system (Fig.
2.
7
-
right). We
view the breadth of these designs as being
indicative of the flexibility and simplicity of the CricketSat platform to accommodate a
variety of introductory
-
level student projects. The course is currently in its second
offering to 70 students and utilizing
the improved CricketSat design illustrated in Fig.
2.
2.



28




Figure
2.7:
Example student design projects: Wireless Wind
-
Chill Instrument (left) and Wireless Door
Alarm (right)

2.8
.


Conclusions

UVM’s CricketSat activities have in addition enabled collabora
tive
meteorological studies with the University of Alaska along with the aforementioned work
with Medgar Evers College. Like UVM, the University of Alaska is involved with
CricketSat development and testing. Designs and methods are being shared between t
he
two schools towards a common goal of improving this design. With this wide range of
“customer” input, we view the UVM CricketSat design as rapidly migrating towards a
simple, flexible and yet a powerful platform upon which meaningful projects can be
de
veloped for a wide range of wireless monitoring applications. We hope that with this
paper along with additional material (schematics, project ideas and kit information)
available online
viii

will provide a resource that other institutions may utilize to

deve
lop their
own entry level sensor programs. The author encourages interested institutions to contact him
ix

should they
have any questions.



29

2.9
.


Acknowledgements

The authors would like to acknowledge the Vermont and Colorado Space Grant
Consortiums, and the

UVM HELiX outreach program for their support in the
development of the CricketSat program at UVM. The authors would also like to
acknowledge Dr. Shermane Austin of Medgar Evers College and Dr. Neal Brown of the
University of Alaska for their flight and t
echnical contributions, respectively.


CHAPTER 3:

CRICKETSAT SENSOR AND SYSTEM DEVELOPMENT


3.1.

Introduction

In this chapter we discuss the development of the UVM CricketSat sensor and
systems beginning with the original Stanford CricketSat design and co
ncluding with the
current UVM Revision F (RevF) design. System development is also discussed relating
to CricketSat sensor arrays and a reliable ground
-
based measurement system.
Educational applications, discussed in Chapter 2, provided the motivating fa
ctors driving
the development work. These factors were primarily related to performance, simplicity,
and reliability. In addition, we considered the likelihood of successful assembly, testing
and adaptability. Limitations of the RevF design discussed a
t the end of this chapter, lead
to the development of a linear CricketSat sensor, discussed in Chapter 4.

CricketSat performance relates to sensor accuracy and resolution. For example,
of primary concern is that the oscillator change frequency based on s
ensed parameter.


30

Resolution may suffer due to a non
-
linear frequency response to the measured parameter
and this is addressed in detail in Chapter 4.



Performance



Simplicity



Reliability


A simple design is important to allow students to understand the oper
ation with
little or no electronics experience. The 555
-
timer based design was maintained due to its
simplicity and wide
-
spread use. Operation of the 555
-
timer oscillator is discussed in
Chapter 4. A circuit with fewer components could have been develop
ed using a small
microcontroller, but the “black box” circuit would not demonstrated simple electrical
principles and would have required a custom pre
-
programmed microcontroller which
may not be available in the future.

Likelihood of successful assembly is

important for students untrained in soldering
skills and circuit board assembly. A quality assembly manual, in addition to a clearly
labeled and coated circuit board, helps guarantee proper insertion of components and
minimization of soldering shorts.

Re
liability is mostly concerned with mechanical issues relating to wires and
components breaking from exposure and repetitious use. The use of strain
-
relief holes
helps alleviate wires from breaking. Components should be mounted tight to the circuit
board
to avoid flexing. Vertically mounted components may be placed in proximity of
other components for physical protection or mounted horizontally if room permits.



31

Testability is important for debugging newly assembled circuit boards. Providing
clearly lab
eled test points on the electrical schematic and the printed circuit board allows
for confirmation of proper signals in debugging a non
-
functional circuit. The test points
also aid in instructional use for demonstrating 555
-
timer operation with the use of

test
instruments.

Adaptability allows the CricketSat circuit to interface with additional types of
sensors primarily for student
-
based designs. Test points provide access to power and
timing signals to interface with custom circuitry. For small circu
its, the on
-
board
prototype area can be used. Larger circuits may be constructed on external prototype
circuit boards and wired to the CricketSat circuit board at the test points.



3.2.

The Original Stanford CricketSat Design (1999)

The original CricketS
at was developed in 1999 for the Space and Systems
Development Laboratory (SSDL) located at Stanford University. David Joseph
(W7AMX), a student mentor, designed the CricketSat at the suggestion of Professor Bob
Twiggs, director of the laboratory. The ci
rcuit (Fig.
3.
1a) is composed of a simple 555
-
timer based oscillator, a thermistor, an LED and an RF transmitter module. The circuit
produced tones or clicks in a receiver (or flashes on the LED), dependent on the size of
the timing capacitor, C1. A prot
otype area (Fig.
3.
1b) is provided on the circuit board
supporting adaptations.



32

(a)


(b)

Figure
3.
1
:
The original Stanford CricketSat schematic and completed circuit board.

The printed circuit board was designed using Express PCB software. This was a
go
od choice, as the software is free to download and is widely used. The circuit board
design files were provided at the first “Starting Student Satellite Hardware Programs”
workshop held by the Colorado Space Grant Consortium at the University of Colorado
in
2002.




33

3.
3.

UVM
CricketSat Workshop

Planning

(Spring 2003)

Dr. Mark Miller (UVM), attended the 2002 Colorado workshop and decided to
use the CricketSat for student outreach, sponsored through the HELiX program at the
University. A workshop entitled “
Building and Launching Cricket Satellites to Measure
Various Atmospheric Conditions” was planned for Summer 2003. The goal was to
perform balloon
-
borne atmospheric profile measurements similar to those made by the
U.S. National Weather Service. These dat
a would be compared to NOAA radiosonde
sounding data and used to validate know atmospheric relationships taught in the
classroom.

A UHF ham radio transceiver (Kenwood THD
-
7A), tuned to 433.92 MHz, was
designated to receive the remote signal. Methods use

to measure the received data vary
dependent on the CricketSat mode of operation. For a CricketSat operating in pulse
mode, a stopwatch is used to count clicks heard in the speaker over a specified time
interval (i.e. 15 seconds) or used to measure indivi
dual click intervals. A CricketSat
operating in the tone mode requires a frequency measurement device to measure the
audio tone produced.

For allowing students to duplicate measurements made by NOAA, the CricketSat
sensor would be required to support th
e use of pressure, humidity and various other
sensors. Fortunately, due to the versatility of the 555 timer circuit, interfacing with a
variety of sensors types is simple. Passive (resistive and capacitive) and active (voltage
and current output) sensor
types are easily interfaced with the timer oscillator circuit.


34

Various methods of interfacing these sensor types for control of the timer are analyzed
and discussed in Chapter 4.

The immediate concerned required changes to the printed in support of a
rep
lacement for the TWS
-
434 radio transmitter module. The original six
-
pin module was
replaced by the manufacturer with a four
-
pin version, with no equivalent substitute
available. Since the circuit board required modification, the opportunity was taken to
make a few additional changes. Unknown was which CricketSat operational mode (pulse
or tone) would prove superior relating to long
-
distance reception of the signal. Features
allowing the selection between these CricketSat modes of operation would be usefu
l for
initial evaluation.


3.4.

UVM CricketSat Revision A (
Spring
2003)

For this initial release of the UVM CricketSat, changes were made to support the
new transmitter module, along with those to provide safeguards and flexibility. To
protect the Cricket
Sat circuitry (Fig.
3.
2) in the case of a reverse
-
battery connection,
diode (D2), in series with the power, allows current to flow if the battery is properly
connected. Three shorting
-
block jumpers (JP1
-
JP3) provide for flexibility of circuit
operation an
d provide connectivity for other sensor types.



35


Figure
3.
2
:
UVM initial redesign (RevA) of the Stanford University CricketSat.

In support of the two modes (pulse and tone) of CricketSat operation, JP2 allows
mode selection without replacement of the timi
ng capacitor. Pads were provided on the
RevA circuit board supporting two timing capacitors, C1 and C3. C1 is intended to be
the smaller capacitor (0.1uF) and C3 the larger (100uF) electrolytic capacitor. JP2
connects the larger, C3, to the circuit, pla
cing it in parallel with C1, which is always
enabled. Since the electrolytic capacitor is typically hundreds of times larger than the
disc capacitor, its value dominates the oscillator timing intervals while connected.
Therefore, tone mode is provided wi
th JP2 disconnected and pulse mode with it installed.

Two additional jumpers were added to support two other features. Jumper JP1 is
used to disable the LED, conserving power for extended use. Jumper JP3 allows for the
connection of active sensors to the

threshold control (Pin5) of the 555 timer. Varying the
threshold voltage may be used to control the oscillator frequency as described in Chapter


36

4.

One goal of the redesign was to simplify the circuit by eliminating the need for
the two inductors L1 and
L2 (Fig.
3.
1
a
). These inductors appear to serve the primary
purpose of noise decoupling, typically accomplished using capacitors. Two decoupling
capacitors were added to the design as inductor replacements, while the inductors were
retained until their p
urpose was understood. Inductor L1, associated with the 555 timer
circuit, appears to serve a decoupling purpose and its replacement more straightforward.
The use of L2 is more complicated, appearing to serve a dual purpose of noise decoupling
and aidin
g modulation.

Pins 1, 2, and 6 of the TWS
-
434 transmitter module (Fig.
3.
1
a
) are tied together,
driven by the logic output of the 555 timer IC and coupled through L2 to the battery
power. The transmitter power pins (1 and 2) are connected to the data pi
n (6) on the
circuit board using a pinched metal trace, intending it to be cut. Indications are that this
connected arrangement may provide a stronger modulation of the transmitter. For this
RevA design, the data and power signals to the new transmitter m
odule were separated.
As a precaution, an unnamed jumper was added to the circuit board to provide
reconnection if needed. The L2 inductor was configured to couple the transmitter power
pins to the supply voltage.

To make the prototype area more usable,
the bottom row was freed from the
power rail, providing four rows of unconnected pads (Fig.
3.
3a). This change allows for
inclusion of an 8
-
pin DIP IC, such as another timer, or an op amp to be used in the space.


37

Power and ground connections were added to

additional prototype pads to provide access
to those signals.

(a)


(b)



(c)

Figure
3.3
:
UVM RevA CricketSat temperature, pressure and humidity sensors. Pressure and humidity
sensors are inserted in the prototype area.

Active pressure and passive hum
idity versions of the CricketSat sensor were
created. A Motorola MPX4115AP pressure sensor (Fig.
3.
3b) was used as an altitude
sensor for balloon flights and as barometric pressure sensor for ground
-
based
measurements. The active device produces a voltag
e linearly proportional to


38

temperature. This voltage was used to vary the threshold control voltage on the 555
timer, affecting the frequency of oscillation.

For the CricketSat humidity sensor (Fig.
3.
3c), a passive humidity sensor
(Humirel HS1101) repl
aced the timing capacitor (C1) in the oscillator circuit. The
HS1101 data sheet includes a 555
-
timer based circuit that works well with the layout of
CricketSat circuit board. Unfortunately, the center frequency of the circuit is greater than
6000 Hz, ex
ceeding the bandwidth of the RF transmitter. A doubling of the timing
resistors reduced the frequency in half, allowing adequate operation.

Two CricketSat sensors of each type were assembled and calibrated (tone mode)
in preparation for the first HELiX
CricketSat workshop. The tone mode was thought to
provide a higher resolution and allow automated data logging to a laptop computer. A
Radio Shack multi
-
meter with and RS
-
232 interface and accompanying software was
selected to perform the sensor frequenc
y measurements.


3.5.

2003 CricketSat Workshop and Test Flights (June 16


June 20)

Three female students and their science teacher representing the Waldorf High
School (Charlotte, Vermont), participated in this first workshop, whose format is outlined
in Chapter 2. Students assembled and calibrated four CricketSat temperature sensors.
Due to the wiring complexity of the pressure and humidity sensors, sensors assembled
and calibrated in the prior section were used.



39

Test flights were conducted Wednesd
ay (18
th
) from the Sandbar Bridge

(Fig.
3.4
a
)

in Colchester, Vermont. Since no flight procedures were available from the
Colorado workshop, this was a flight of firsts. The flight objectives were four
-
fold.



Develop flight procedures relating to preparati
on, launch and data collection



Observe qualitative and quantitative atmospheric effects on sensor types



Evaluate the data collection system



Develop benchmarks to evaluate CricketSat and system performance

o

Tracking time

o

Estimated range

o

Coldest temperature

o

Highest altitude

o

Lowest humidity

(a)

(
b)

(c)

Figure
3.4
:
CricketSat flights conducted from the bridge near the Sandbar State Park. Students are shown
preparing, launching, and tracking CricketSat sensors.

Four sensors were flown

(Fig. 3.4b)

and track
ed

(Fig. 3.4c)

in sequence and
students recorded measurements at 20
-
second intervals. It was obvious from the first


40

balloon flight that the data collection system using the frequency meter had a problem.
After only a couple minutes of flight, the meter c
ould not discriminate the tone signal
from the background noise. Little temperature or humidity variation was observed over
the water surface and therefore no quality quantitative measurements were obtained in the
short span.

Qualitative temperature resul
ts were better since the audio signal was audible for
several minutes and a reduction in the tone frequencies were noticeable in
correspondence with decreasing temperature and altitude. Results were even more
impressive for the pressure sensor. This sens
or produced a noticeable decrease in
frequency after a couple minutes of ascent. Unfortunately, the antenna of wildly
swinging sensor popped the balloon. The descent to the lake surface produced a tone
variation sound similar to that of incoming artiller
y; that is u
p until impact with the water.
Several thing
s were learned from this flight:



Qualitative results were observed for the temperature and pressure sensors,
demonstrating sensor response to the stimulus



The frequency meter was inadequate for makin
g distant measurements



Future balloon flights would use CricketSat sensors configured for pulse
-
mode
operation, using a stopwatch for measurement until a better frequency
measurement method is adopted



The CricketSat sensors require a longer balloon attachm
ent string to stabilize the
flight system and avoid popping balloons



41



Measurements over water do not provide adequate temperature or humidity
variation versus altitude


3.
6
.

UVM
CricketSat
Test Flight (08/08/2003)

This test flight was conducted to evaluate

feasibility of using stopwatch
measurements relating to pulse
-
mode CricketSat balloon flights. This was in preparation
of a CricketSat workshop and flight to occur the following day. A CricketSat sensor was
prepared and released on the green behind the
UVM Cook Science physics building.
Instead of counting CricketSat “clicks” in a 15
-
second period, a more accurate method
was adopted which measured the time interval of five consecutive clicks. The resolution
of the stopwatch provided a much improved res
olution for the measurements. A second
watch was used to provide elapsed time since launch for each measurement.

This flight and results and were extremely encouraging. The temperature dropped
from 83 ºF to 40 ºF in 25 minutes (Fig.
3.
5a), at which point
it quickly began warming.
We realized that the balloon had burst and was rapidly descending. The signal was
tracked to landing, only six minutes later, indicated by a return to the original
temperature. Using a directional antenna and a RF spectrum anal
yzer, the balloon was
located in the top of a maple tree located two miles away in Winooski.



42

(a)


(b)

Figure
3.5
:
Pulse
-
mode CricketSat temperature (top) and altitude (bottom) fight profiles.


Using an adiabatic cooling rate of 4.5 ºF per 1000 feet of
altitude, the second plot
(Fig.
3.
5b) provides an altitude profile for the flight, indicating a burst altitude of 11,000
feet. This graph can be used to estimate ascent and descent rates due to the nearly linear
characteristics. From the data, the balloo
n ascended at a rate of 440 fpm (5.00 mph) and
struck the maple tree at 1833 fpm (20.8 mph).



43

This flight clearly demonstrated the successful operation of the CricketSat
temperature sensor and the pulse
-
mode stopwatch measurement system. This flight also
d
emonstrated use of the CricketSat to demonstrate atmospheric characteristics in a short
time interval. The linear temperature profile, similar to those produced by BalloonSat
flights indicates a linear temperature change over time. Since these balloons (a
nd
BalloonSat weather balloons) rise at a nearly linear rate, a constant adiabatic rate in the
lower atmosphere is demonstrated.


3.7.

GIV
CricketSat Workshop and Test Flight (08/09/2003)

This workshop was conducted at 20
th

anniversary celebration for t
he Governor’s
Institute of Vermont (GIV) held at Shelburne Farms in Shelburne, Vermont. A small
group of six students assembled pulse
-
mode CricketSat temperature sensors which they
were allowed to take home. Temperature measurements could be made using t
he
flashing LED and a stopwatch to perform measurements.

A calibrated pulse
-
mode CricketSat sensor was used for the demonstration flight
on a cold day and occurring during a thunderstorm. The CricketSat sensor was placed in
a zip
-
lock bag to protect it fr
om the rain. In an attempt to achieve a higher flight than the
previous day (11,000 feet), less helium was placed in today’s balloon. The under
-
inflation of the balloon and the additional weight of the rain
-
laden bag and balloon
resulted in a slow ascent

rate.

The flight was successful in relation to benchmarks relating to flight duration,
minimal temperature and estimated terrestrial range. After 90 minutes of tracking the


44

flight, the minimum temperature experienced was +20 ºF. After factoring in an ad
iabatic
cooling rate, an estimate for an altitude was determined to be approximately 8000 feet.
Factoring in flight time, strength of the winds and the low angle of the Yagi antenna, the
terrestrial flight range was estimated to be in excess of 15 miles.

As in the previous
flight, the pulse
-
mode CricketSat using a stopwatch proved to be a very successful
system for data collection and measurement.


3.8.

HELiX CricketSat Weather Stations (12/2003


02/2004)

HELiX teams and their UVM sponsors have a one
-
yea
r commitment to develop a
project in response to knowledge and skills obtained during the summer workshop. The
Waldorf team decided to use CricketSat sensors to develop two weather stations to
investigate lake
-
effect on temperature. One station was locat
ed at the Waldorf High
School (Fig.
3.
6), 1.5 miles inland from Lake Champlain, at an altitude of approximately
150 feet above sea level. The second station was located in Williston, Vermont, six miles
from the lake at an altitude of approximately 400 fee
t. The weather stations were fitted
with solar panels and rechargeable batteries for long
-
term operation.



45


Figure
3.6
:
CricketSat wea
ther station located at the Wald
orf High School fitted with a solar panel for
long
-
term use.

The mechanical construction
of the weather stations consisted of plastic plates
(Fig.
3.
7a), sheet aluminum, shelf brackets, threaded rods and nuts. Along with
temperature, the pressure and humidity sensors developed during the workshop were
stacked (Fig.
3.
7b) and mounted center in

the weather station. Autonomous data logging
capability was required, requiring the CricketSat sensors to operate in tone
-
mode. Again,
the Radio Shack meter was used, but since the receiver was in close proximity to the
weather station, the signal noise

issues, experienced in the first flight, were not replicated.



46

(a)

(b)

(c)

Figure
3.7
:
The Waldorf CricketSat weather station. Plastic
-
plate and threaded
-
rod construction (left),
stacked CricketSat sensors (center), sensor selection switch and ba
ttery (right).

Since CricketSat sensors are always transmitting data, only one is allowed to be
powered on at a time to avoid interference. Consequently, the weather stations were
fitted with a switch box (Fig.
3.
7c), allowing the selection of one of the
three sensors.
Since temperature was the measurement of interest in this application, this restriction did
not cause a serious problem. The station would remain in the temperature mode for
autonomous data logging and only briefly switched to pressure or
humidity to make those
measurements manually.

To eliminate the dependency of the Kenwood receiver, costing hundreds of
dollars, a CricketSat receiver (Fig.
3.
8a) was designed to work with the weather station.
The circuit (Fig.
3.
8b) uses the accompanying
receiver module to the transmitter module
used on the CricketSat. The signal weak produced by the receiver module is amplified by
a LM386 amplifier, driving the speaker and the Radio Shack frequency meter.



47

(a)

(b)


Figure
3.8
:

CricketSat receiver d
esigned to work with the CricketSat weather station and was later used for
all CricketSat applications.


The station functioned with frequent problems mostly relating to loss of data
acquisition after a few hours. Apparently the meter was again the culpri
t, now failing to
take frequency measurements, even in the presence of a strong signal. Turning the meter
off briefly returned it to proper operation. Apparently, the meter assesses the initial
signal and makes a determination of a threshold voltage leve
l used to convert the analog
input signal into a digital one for counting. Changes in the CricketSat signal strength
resulting form battery level or duty cycle variations would render the initial threshold
level in error, resulting in the loss of data.

Results from this work demonstrate the need to find or develop a reliable method
for measurement of CricketSat data in the presence of noise. It also demonstrated the
need for an automatic multi
-
sensor array system sequencing through the sensors. Such a
system could be used for ground based or balloon
-
borne applications. Progress is this
area is achieved later during the year,
demonstrated

in
Section
3.
12
. Meanwhile, during
the course of this weather station project, a new CricketSat sensor was develope
d

for use
in the newly created first
-
year

design class at UVM.



48


3.9.

UVM CricketSat Revision C (January 2004)

The CricketSat sensor development
jumped quickly

from RevA to RevC.
CricketSat RevB was not manufactured and only exists in prototypes. Revision

C was
designed to address previous problems with RevA and to meet the needs of the new
freshman design class. The most significant change for this version was the addition of
strain
-
reli
ef holes for the battery wires.
These connections were frequently b
reaking at
the board surface due to flexing, and were dif
ficult for students to repair. All but one of
the jumpers (JP1) was removed from the circuit (Fig. 3.9) to simplify assembly.


Figure
3.9
:
Schematic diagram of the UVM RevC CricketSat.

This is th
e first CricketSat version to be used for the freshman design course
(Chapter 2), requiring a quality PCB (Fig.
3.
10) that would be easy to assemble and
tolerant of soldering errors. Production quality boards were ordered which provided
solder
-
mask and si
lkscreen layers. The solder mask, covering all metal traces, except for
the solder pads, minimized short circuits caused by newly developed soldering skills.


49

The silkscreen layer provided reference labeling to help guide students relating to
component lo
cation and orientation on the circuit board.


Figure
3.10
:
CricketSat RevC temperature sensor circuit board. Green solder mask, white labeling and
strain
-
relief holes were added to this design.

To reduce complexity and assembly for the students, only o
ne timing capacitor
location was provided on the board. However, to accommodate multiple size capacitors,
three pads provided a choice of 0.1” or 0.2” pin spacing. This feature supported disc,
electrolytic and a capacitive humidity sensor as well. A sim
ilar three
-
pad feature was
provided for R1 in support resistor, thermistor or photocell devices.


3.10

UVM
CricketSat
Revision
D

(
May

2004)

CricketSat Revision D (RevD) was designed to directly support pressure and
humidity sensors and address problems wit
h the earlier designs. Prior testing revealed a
sensitivity of the oscillator circuit to battery supply voltage, later evident due to self
-
heating of the thermistor. The addition of a 5
-
Volt regulator (Fig.
3.
11) resolved the


50

problem and provided the spe
cified voltage for the active pressure and humidity
(Honeywell HIH3610) sensors.


Figure
3.11
:
CricketSat RevD schematic diagram.
A

voltage
regulator
was

added for oscillator
stability

and active sensors.

The unregulated voltage remained connected to t
he RF transmitter module
(through D2) to provide maximum power and reception range. A self
-
protection, low
-
dropout regulator (Texas Instrument TL750L05C) allowed stable circuit operation down
to 5.3 Volts, extending effective battery life in comparison to

a traditional 2
-
Volt dropout
regulator (LM7805). The addition of the regulator provided isolation from the RF
module, allowing the removal of the L2 (Fig.
3.
9
) inductor associated with the timer.

A mechanical improvement was the enlargement of the four m
ounting holes,
allowing the use of common #4
-
40 mounting hardware. This is necessary for stacking
the CricketSat sensors in the weather station and for use in student projects.



51

(a)


(b)

Figure
3.12
:
CricketSat RevD temperature

(top)

and pressure

(bottom
)

sensors.

Pre
-
wired pads and component outlines were provided to directly support the
Motorola pressure and Honeywell humidity sensors. This would simplify the assembly
and allow the students to install the sensors themselves, unlike the RevA version.

The
prototype area was sacrificed to retain the original dimensions of the CricketSat. Ground
(G) and data (D) test points were added to allow testing of the oscillator and allow off
-
board interconnections.


3.11
.

2004 CricketSat Workshop and Test Fligh
ts (June 14


June 18)

A series of CricketSat balloon flights during this workshop produced results
exceeding all benchmarks established on previous flights. More importantly, a reliable


52

tone
-
measurement system was successfully demonstrated. During the
search for viable
audio spectrum analyzer software, a team of students from the UVM freshman design
class discovered the Spectra RTA program. This software, shown for use during
CricketSat calibration (Fig.
3.
13), allowed CricketSat tones to be visualized

and
measured even in the presence of noise. This allowed for flight measurements to be
collected over a longer duration, higher altitudes and terrestrial distance than by using a
frequency counter.


Figure
3.13
:
Students
using the Spectra RTA software
to perform CricketSat calibration measurements.

Three flights were conducted during the daytime and three in the evening
involving temperature, pressure and humidity sensors. One of the daytime flights (Fig.
3.
14a) involved a CricketSat temperature sensor
. Data (Fig.
3.
14b) was collected nearly
continuously for 73 minutes resulting in a minimum temperature of
-
41 ºF (114 ºF drop)
in during the flight. Eventually, measurements could not continue due to the complete
loss of the signal. Prior to that final
ity, even after loss of the signal discernable to the ear,


53

for a period of time, the Spectra RTA software continued to display a tone peak and
allow measurements to continue.

(a)

(b)

Figure
3.14
:
CricketSat daytime temperature sensor flight. The senso
r experienced a 114 ºF temperature
change during the flight.

The highlight of the evening was the CricketSat altimeter flight (Fig.
3.
15a). This
flight was tracked for 91 minutes, reaching an altitude of 26,732 feet (Fig.
3.
15b) after 90
minutes. The fla
ttening of the data after 80 minutes is due to the limited pressure range
of the MPX4115 sensor resulting in an artificially low result. Wind velocities from
Albany sounding balloon data along with the CricketSat rate
-
of
-
ascent altitude slope (Fig
3.
15b)
was used to generate a flight path and estimate the terrestrial distance (51 miles).

(a)

(
b
)

Figure
3.15
:
Night
-
flight data collection from a CricketSat pressure sensor used as an altimeter

showing a
linear ascent rate.



54

The collective sets of flights

throughout the day resulted in all previous
measurement benchmarks to be surpassed due to the success of the measurement system
based on the Spectra RTA software. This system now proved viable for performing long
-
range atmospheric measurements for single
-
sensor flights, meeting the first of two goals
outlined at the end of the Section
3.10
.


3.12
.

First Collaborative BalloonSat Flight (July 17, 2004)

As described in Chapter 2, collaboration was established between Vermont and
New York Space Grant Consor
tia (Medgar Evers College), the UVM HELiX high school
outreach program and the Milton High School. The Medgar Evers College had launched
their first BalloonSat (Condor
-
1) in June, but had lost communication with it in mid
-
flight. One of several possibili
ties related to temperature of the flight bag, housing the
GPS and communications equipment. Too cold of a temperature might cause the
equipment to fail.

(a)

(b)


Figure
3.16
:
Students prepare for a BalloonSat flight from the Milton High School. Tem
perature data (red
trace) in the figu
re to the right shows the

flight bag temperature during flight.



55

Since the June CricketSat flights demonstrated that measurements could be
collected as far as 51 miles away, a CricketSat sensor could be used to provide r
eal
-
time
BalloonSat flight
-
bag temperature measurements. This experiment would also
demonstrate CricketSat compatibility with the BalloonSat system, allowing the
measurement of various sensor data.

The flight was conducted at the Milton High School (Fig.

3.
16a) and the
temperature data was collected by three HELiX students. The flight bag temperature is
shown as the red trace in the graph (Fig.
3.
16b). The temperature was collected 125
minutes and only dropped to +66 ºF, indicating that low temperature
was not likely to
have been the cause for communication loss in the Condor
-
1flight. Additional results of
the flight include a GPS validation of altitude exceeding 45,000 feet and a terrestrial
range of 18 miles. Results were probably greater than these,

but cannot be confirmed due
to loss of GPS data on this flight as well (green trace).

This flight demonstrated the CricketSat could provide real