Download Wireless Temperature Sensor

Wireless Temperature Sensor
Final Design Report
4/15/2007
Tye Reid
Greg Swanson
Daren Berk
James Wagoner
Executive Summary
The purpose of this project was to develop a wireless sensor system for future spacecraft
and probes. Currently, the missions that do fly with sensors have them wired into the
Thermal Protection System (TPS) of the spacecraft. Utilizing this architecture adds risk
to the system due to the process of routing wires in the TPS and the difficulty of
jettisoning the system after entry. Many current and past spacecraft engineers have
decided not to fly embedded sensors in an effort to mitigate the risk of spacecraft failure
during atmospheric entry. A wireless instrumentation system will solve these problems.
The completed design of the wireless sensor system will be a multi-stage project
spanning several years. Team ThermaSense’s focus was to research the current state of
wireless transmission, design a wireless temperature sensor and then develop a prototype.
To achieve this goal, off the self components were researched and purchased to create a
table top demonstration that confirmed the success of the new wireless thermocouple
system. Once the prototype system was operational it was then tested in the NASA X-Jet
facility to simulate atmospheric entry.
Three main wireless transmission types were researched for the system: RF, LED, and
Near Field Magnetic Communication. After analyzing each type of transmission, RF was
determined to be the best choice for the team’s needed application. Next, two main RF
technologies were considered for the wireless system, Bluetooth and Zigbee. Since
Zigbee is used for low power consumption situations, the team decided to use it for
prototyping.
After the Zigbee was chosen as the wireless transmission technology, the team designed a
circuit to collect the data from the thermocouple sensors.
Voltage signal from the
thermocouple was connected to a cold junction correction chip that also does analog to
digital signal conversion and amplification. This data was sent to a microcontroller that
attaches an identification number and the time the data was recorded.
The
microcontroller is then programmed to send the data through an inverter to the Zigbee
transmitter to be relayed to the Zigbee receiver. The receiver transfers the data to the
2
software created by the team via USB to a computer system. The software compiles the
information and exports it to an Excel spreadsheet.
For proof of design concept, testing was conducted in an X-Jet Chamber at NASA Ames
test facility. Four thermocouples where embed at different depths in the LI-900 TPS
material. Three Antennas were tested, the PCB antenna, the whip antenna, and the
external antenna. Power settings were also tested at low, medium and high power.
Testing proved that the best solution for the wireless transceivers was medium power
with the whip antenna.
The system designed by Team ThermaSense is proof that Zigbee technology is a feasible
solution to NASA’s wired sensor problem. Initial analysis shows integrating sleep mode
into the sensors design will allow for five to six years of operation from two AA
batteries. The team’s final prototype weighs 0.2 lb reducing sensor system weight by an
estimated 50%. Not only does the prototype reduce the weight of the spacecraft carrying
the system, but it also reduces space required for the system.
The next stage of this project will be to develop a flight test for the prototype, working
towards the goal of implementation into NASA missions. A successful flight test will
require a senor software update. Other areas of work needed to achieve implementation
goal would be; powering the sensor, programming the sensor for advanced power savings
capabilities, determining the transmitter’s interference tolerances, miniaturization,
transmission through RF non-transparent material research, near field magnetic
transmission research and finally a protective packaging for the wireless transmission
circuit.
3
Table of Contents
1. Background and Concept
1.1 Background
6
1.2 Problem Definition
6
1.3 Concept Considered
7
1.4 Concept Selected
10
2. Product Description and Testing
2.1 Product Description
11
2.2 Calibration
12
2.3 Analytical Analysis
13
2.4 Testing at NASA Ames
15
3. Evaluation, Conclusions, and Recommended Work
3.1 Product Evaluation
15
3.2 Conclusions
16
3.3 Recommended Work
17
Support Material
Drawings
Appendix A
Circuit Diagrams and Codes
Appendix B
Assembly and Operation Instructions
Appendix C
Math Models and Calculations
Appendix D
Testing Description and Analysis
Appendix E
DFMEA
Appendix F
Team Resumes
Appendix G
Project Timeline
Appendix H
4
List of Figures
Figure 1: Fundamental Transmitter circuit structure
8
Figure 2: Basic description of Orion Spacecraft Heat Shield
9
Figure 3: Final Transmitter Circuit Structure
10
Figure 4: Final Transmitter Prototype PCB
11
Figure 5: Top PCB Layout (Left), Bottom PCB Layout (Right)
12
Figure 6: % Error found in calibration testing
13
5
1. Background and Concept
1.1 Background
Every space craft entering a planetary atmosphere needs a Thermal Protection System
(TPS). The TPS must endure severe heat loads, which requires an understanding of
atmospheric properties, vehicle aerodynamics, TPS material properties and the physics of
the entry environment.
NASA and other space agencies would like to collect
temperature, pressure, heat flux, radiation, and recession measurements on flight tests and
flight missions in order to verify TPS design and to aid in the characterization of physical
and chemical phenomena in the entry environment. Currently the missions that do fly
with thermocouples have them wired into the TPS of the spacecraft. Utilizing this
architecture adds risk to the TPS system due to the process of routing wires in the shield
and the difficulty of jettisoning the system after entry. Many current and past spacecraft
engineers have decided not to fly embedded sensors within the TPS in an effort to
mitigate the risk of spacecraft failure during entry. A wireless instrumentation system
could collect the required measurements needed for scientists and engineers to improve
future spacecraft design while lowering the overall risk of incrementing entry vehicles.
1.2 Problem Definition
NASA would like to develop a wireless sensor system for future spacecraft and probes,
and has contracted our team, ThermaSense, to initiate the design process. The design of
the wireless sensor system is a multi-stage project spanning several years.
The
ThermaSense team’s focus was to research and understand the current state of wireless
transmission. To achieve this goal the team has been asked to design a wireless
temperature sensor and develop a prototype.
The project’s main focus was on
establishing wireless communication between the sensor and a computer for data
acquisition with a secondary focus on the actual temperature sensor and the data
acquisition. This project required the team to design a wireless system using off the self
components and to create a table top demonstration that confirmed the success of the new
6
wireless thermocouple system. To complete this goal the team researched the current
state of wireless communication architecture and decided on specific components. Once
designed and perfected the new wireless architecture could be extended to many other
different types of TPS mounted sensors such as velocity and pressure sensors.
1.3 Concepts Considered
With the basic understanding of the obstacles pertaining to the thermal wireless project,
research was completed to identify the different elements needed to complete a wireless
system that is to be embedded in the spacecraft. The team’s initial knowledge led us to a
system where a thermocouple was attached to a device for cold junction correction. The
system was attached to the transmission wireless device and the data was transferred to
the receiving wireless device and sent to the data acquisition system. After researching
thermocouples themselves, the team learned that not only does the thermocouple voltage
need to be corrected, but it should also be amplified to allow accurate temperature
readings. To amplify the voltage, the team inserted an amplifier into the system after the
cold junction correction and before the wireless transmitter.
This described system setup will work for a basic table top demonstration without any
interference, but since noise interference is a major obstacle of this project, the team
researched different ways to protect the thermocouple signal from outside interference.
One of the best ways the team found to reduce susceptibility to noise interference was by
adding an analog to digital converter to the system on the transmission side. This addition
changed the thermocouple signal to a digital signal, in which the wireless transmission
was many times less susceptible to interference.
At this point in the design, the team believed the system described above would be the
final product. Actual testing with a prototype proved the system required a
microcontroller. The microcontroller was needed to provide a clock signal for certain
7
chips and also covert between different interfaces, i.e. UART to SPI. The addition of the
microcontroller gave the team our final system.
Thermocouple
Cold Junction
Correction Chip
Amplification
Xbee Wireless
Transmitter
Analog to Digital
Conversion
Figure 1: Fundamental Transmitter circuit structure
Another main part of this project was to research different types of wireless architectures
that were available for use in the system. Research showed that there were three main
architectures: radio frequency, light emission and near field magnetic communication.
Before looking at each architecture in respect to the specific project scenario, the team
researched and developed a decision matrix for each of the types of wireless transmission
in order to develop an understanding of the capabilities and restraints. The decision
matrix (shown in Appendix 2) evaluated each architecture by; size, power consumption,
ability to transmit power, transmission distance, requiring line of sight, noise immunity
and cost.
Based on these characteristics alone, each architecture is capable of
transmitting the thermocouple signal for a table top demonstration in an interference free
facility. This will not be the environment of the final product, so the team judged the
wireless architectures mainly on noise immunity and power consumption. When looking
at the requirements of these two specifications, near field magnetic communication is the
best wireless option.
After obtaining a description of the Orion Spacecraft Heat Shield Structure (Figure 2) the
team had a new understanding of the obstacles they had to overcome for a successful
design.
8
Figure 2: Basic description of Orion Spacecraft Heat Shield
With this new knowledge the team realized the real challenge was to communicate
through the many layers of the spacecraft. Light emission was ruled out since solid
structures in between the transmitter and receiver could not have holes drilled in them.
The two options left were radio frequency and near field magnetic coupling. The team
spent some time in the lab experimenting with the magnetic communication through
aluminum, which is one of the most troublesome layers with in the Orion construction.
This rough experiment demonstrated that it would be very difficult to transmit a signal
through aluminum.
The team then researched the possibility of transmitting a radio frequency through the
aluminum layer and found the probability of success to be a little less than using the
magnetic communication process.
Due to this research, near field magnetic
communication seemed to be best choice until the team was informed that the
transmission distance was larger than the range of magnetic communication. In addition
we learned that there was no commercial availability of the wireless architecture or this
purpose. This knowledge made radio frequency, although higher in noise susceptibility,
the best choice for our system overall. The team continued to work with the near field
9
magnetic communication for the possibility of other wireless transmission within the
system that might need its specific capabilities.
1.4 Concept Selection
Taking into consideration the specifications defined during the problem definition phase
(shown in Appendix 3) and the results of our initial research, the circuit structure shown
in Figure 1, was modified to look like the structure in the following figure.
Thermocouple
Cold Junction
Correction Chip
Xbee Wireless
Reciever
Pickaxe
Microcontroller
Bit Inverter
Xbee Wireless
Transmitter
Figure 3: Final Transmitter Circuit Structure
The first change was the inclusion of four thermocouples instead of one connected to the
circuit. With four thermocouples, the circuit met our specification and also better meets
NASA’s needs by giving them the ability to imbed more thermocouples into the TPS
material.
In order to incorporate four thermocouples, the team also had to change to four CJC
chips. It was decided that the CJC chip would also incorporate the amplification, A/D
conversion, and the voltage to temperature calculations. This eliminated the amplifier
and A/D converter shown in Figure 1 and also reduced the chance that noise could enter
the system. No other changes were needed to the Structure shown in Figure 1, after this
10
change because the microcontroller has the ability to read 4 signals and pass them along
to the transmitter.
2. Product Description and Testing
2.1 Product Description
The transmitter PCB is made up of seven IC’s (integrated circuits) and four headers.
Starting from left to right there is first the terminal block.
This terminal block is
connected to four cold junction correction IC’s. These IC’s correct the voltage potential
read coming in from the thermocouples connected to the terminal block for room
temperature. The corrected temperature data is passed to the PicAxe microcontroller
using SPI protocol. The PixAxe microcontroller converts the temperature data to the
correct resolution and outputs the data to the bit inverter using UART serial protocol.
The bit inverter inverts the bits and feeds it to the X-Bee transceiver. The X-Bee
transmits the temperature data wirelessly to the receiver connected to the computer. The
receiver sends the temperature data down the USB cable and the software displays the
temperature data on the screen. This process is shown in the code for the microcontroller
itself. Refer to Appendix B for the code. The code for the software can also be seen in
Appendix B.
11
Figure 4: Final Transmitter Prototype PCB
The temperature software ThermaSense.exe was written in Visual Basic 6.0. The PicAxe
microcontroller code was written in the PicAxe Programming editor. This editor can be
downloaded from the PicAxe website, or borrowed on a disk from Joe Plummer. The
PCBs were laid out in EAGLE Layout Editor, which is a free download from the Eagle
layout website. The Layout files are available from the project website. The circuit
traces are shown below in Figure 5.
Figure 5: Top PCB Layout (Left), Bottom PCB Layout (Right)
2.2 Calibration
The CJC correction chips are factory calibrated to be accurate to with-in +/- 4 degrees
Celsius through the complete range of temperatures added to a +/- .2% reading error.
12
Once the temperature data enters the CJC, the only other possible error in data received
would be a dropped packet. To test the factory calibration a controlled test was set up
and conducted.
A voltage from a DC source inputted into the thermocouple inputs caused our circuit to
read out a temperature.
Comparing this temperature against standard type K
thermocouple tables showed the error that the circuit had. Shown below in Figure 6, are
the results of the temperature calibration test and the accuracy of the circuit in a typical
room environment.
% Error for Range of Temperatures
1.4
1.2
% Error
1
0.8
0.6
0.4
0.2
0
-0.2 0
500
1000
1500
2000
Temperature (F)
Figure 6: % Error found in calibration testing
This temperature calibration test table shows a maximum of a +/- 1.2 % error within the
range of the temperatures to be read. This error is over the manufacturers maximum
rated error of +/- .8 % error at maximum reading. The differences in max error can be
accounted by several things.
The first is the circuit layout.
Because there is a
thermocouple header between the CJC chip and the thermocouple a temperature
difference is possible and the circuit reads another thermocouple voltage. The second is a
programming error where the CJC reads off a temperature in Celsius and rounded off
decimals turn into significant errors when converted to Fahrenheit.
13
2.3 Analytical Analysis
Attached in Appendix D is a TK Solver sheet for the analytical analysis. For the
analytical analysis, a nodal method was used. The nodal analysis was one dimensional
going through the block. The one inch thick block was divided into 10 nodes. On the
heated side, the heat used was a heat flux. This method was the best assumption since a
heat flux sensor was used in the actual experiment. The only other heat transfer for the
first node, was heat transfer to the second node. Everything that was not transferred to the
second node was stored in the node, increasing temperature. All of the nodes in the
middle of the block had terms for the heat transfer between the nodes on either side of
that particular node. Anything that was not transferred was stored in that node. The final
node had convection to the surrounding environment. Since the testing occurred in a
vacuum, the convective coefficient was assumed to be low. The coefficient used was 5
W/(m2*C).
Using the heat flux calculated by the heat flux sensor, the temperatures in the block was
an order of magnitude larger than what was found in the experiment. A heat flux of 4800
W/m2 produced temperatures similar to the ones found during testing. The heat flux
sensor measured a heat flux of 580000W/m2. This value is much bigger than 4800 W/m2.
Several factors contributed to the larger measured heat flux. First a Cold Walled Gardon
gage was the measurement device. Presently a Hot Walled Gardon gage is not available
for the X-Jet. The Cold Walled Gardon gauge can produce values 20-30% higher than the
true value. Additionally, the heat flux sensor was not measured at the same distance that
the heat shield was tested in the X-Jet.
The sensor was closer to the jet than the heat
shield was during testing. The distance between the jet and the material has an
exponential effect of the heat flux and even small distances can have a large effect on the
heat flux measured.
14
Even with the known errors, the remaining difference between the two values is too
significant. Another possible error that could be affecting the values is calibration. The
X-Jet is a small facility in which only a few technicians can operate. The heat flux sensor
might not have been fully calibrated prior to testing at the site. It is also possible that the
analytical analysis is not accurate. Presently the difference between the temperature
values of the tested version and the calculated version is so vastly different that nothing
can be discerned.
The analytical analysis is correct in one aspect. The nodes response to time produced the
same graph just on a different magnitude of order. This means that our analytical analysis
set-up can accurately simulate the tests but that there is an error in the input heat flux.
The conclusion of the analytical analysis is that since the graphs of the analytical analysis
and the testing were similar the step response is accurate. The reaction of the heat shield
can be simulated. Unfortunately, since the temperatures were so different, the
temperatures cannot be verified.
2.4 Testing at NASA Ames
For proof of concept, testing was conducted in an X-Jet Chamber at NASA Ames test
facility. An X-Jet chamber consists of a vacuum chamber and plasma torch as a heat
source. Four thermocouples were embedded at different depths in the LI-900 material.
The thermocouples were then attached to the wireless prototype to transmit back to the
computer. Many tests were conducted at Ames test facility. Three Antennas were tested;
the PCB antenna, the whip antenna, and the external antenna. Power settings were also
tested along with the antennas. It was found that the best solution for the wireless
transceivers was medium power with the whip antenna. In conclusion, the wireless link
was found to be strong with system recovery under heavy EMI noise, making our
prototype a successful solution for wireless sensors.
For more information and a
complete analysis of all tests, see Appendix E.
15
3. Evaluation, Conclusions, and Recommended Work
3.1 Product Evaluation
The wireless TPS sensor system was initially specified to transmit data from the sensor to
a data collection system once a second. The wireless system was to be powered by about
9 volts and have a system life of 2 years. The size of the system was specified to be 5” x
4” x 1” and weigh 16 oz.
The system was specified to use thermocouple sensors
attached to 3 – 4 communication channels. Temperatures recorded by the system were to
be accurate to +/- 5 degrees Celsius. Finally the system was required to have medium
susceptibility to electromagnetic interference.
ThermaSense’s final product’s recorded temperature data to the data collection system
ten times a second. The system in final design required 3.3 volts for operation; the team
implemented two AA batteries to supply this energy. The current consumed by the
system was measured to be 49 mA during continuous transmission.
Given this
measurement a rough estimate for system life expectancy is 8 hours, this time span does
not meet the system life specification.
To overcome the issue the system can be
programmed into sleep when not transmitting. The system only needs to transmit for five
to six minutes and can be programmed into a sleep mode for the rest of the time, which
will only consumes a few micro amps, and means the system could roughly last five to
six years. The final system weighs only 3.2 oz and is scaled to be 3.1” x 2.9” x .75”.
This size and weight is a large improvement over the original requirements.
The final prototyped system included 4 thermocouple sensor communication channels.
Calibration testing was completed outside an electrical interference environment using a
precise mV source. Using the ASTM table given for a type K thermocouple the mV
source was used to compare measured temperature to ASTM standard. Calibration
testing on the final system resulted in a +/- 1 percent temperature reading from the
16
thermocouples. Electromagnetic interference testing was conducted at NASA Ames’ XJet facility. The system transmitter was set for medium power using a whip antenna.
These tests concluded that it has low susceptibility to electromagnetic interference; this
was the main accomplishment of our system.
3.2 Conclusions
The system designed by this team is proof Zigbee technology is a feasible solution to
NASA’s wired sensor problem. Zigbee design goals for low power and low bandwidth
were tested and achieved with ThermSense’s wireless TPS sensor system.
Initial
analysis shows integrating sleep mode into the design will allow for five to six years of
operation from two AA batteries. The designed system will allow probes to travel to
nearby planets such as Mars, Venus, and Mercury. The team’s final prototype weighs 0.2
lb reducing sensor system weight by an estimated 50%. Not only does this reduce the
weight of the spacecraft carrying the system, but it also reduces space required for the
system.
With the tested ability of the final prototype to overcome electromagnetic interference,
the system will now be miniaturized and further developed by next years senior design
team. Using the thermocouple system as proof of concept, the wireless system should be
expanded to all different types of sensors useful to embed within TPS material. Once
these steps are complete the system will be flown on a probe mission to ensure
performance and reliability.
3.2 Recommended Work
This project’s goal is a flight test and implementation into NASA missions. In order to
achieve this goal, additional areas of work would be; powering the sensor, programming
the sensor for advanced power savings capabilities, determining the transmitter’s
interference tolerances, miniaturization, transmission through RF non-transparent
17
material research, and near field magnetic transmission research and implementation
studies.
The next stage of this project would be to develop a flight test for the system as we work
to achieve NASA’s goal of implementation into missions . A successful flight test will
require a software update. Currently, if the receiver gets corrupted data the software
crashes and needs to be reset before data can be recorded. New software will need to
overcome this issue as software will not be able to reset during a flight test and will also
need to recover from lost packets and general signal disruption.
Most likely the stage of circuits may need a different sensor connected other than a
thermocouple, requiring the transmitter side of the circuit to be decoupled from the
thermocouple side. If decoupling is needed, work towards a standard transmitter circuit
and a sensor connection strategy is recommended. This change would make it easier to
connect the transmitter to any data source as long as that source transmits a standard set
of data (time, sensor ID, and reading).
Additional work needed for a successful flight test is the protective packaging of the
circuit. The environment that the circuit is exposed during testing and flight will vary
depending on the test and packaging that can resist many environmental factors would be
needed.
The environment would likely involve thermal extremes, vibrations, and
radiation which all could damage the circuit and disrupt the signal.
These are next steps and additional work recommended for next year but the list is not all
inclusive, as other issues, parameters, requirements and ideas may be uncovered as the
project progresses. Other work that fits into the scope of the project and propels the
project towards the final goal would be worthwhile and should be pursued.
18