heat transfer with wind
Developing and testing an Earth Scaled Atmospheric
Temperature Sensor
Philipp Wittmann
Space Engineering, masters level 2016
Luleå University of Technology
heat transfer with wind
Developing and testing an Earth Scaled Atmospheric
Temperature Sensor
Philipp Wittmann
A thesis presented for the degree of
Master of Space Science and Technology
Department of Computer Science, Electrical and Space Engineering Lule˚a University of Technology
The HAbitability, Brine Irradiation and Temperature Package (HABIT) instrument of the ExoMars Surface Platform will investigate the present day habitability of Mars at the near surface environment. This instrument includes three Atmospheric Temperature Sensor’s (ATS’s) which are similar to the ones previously used on the Rover Environmental Monitoring Station (REMS) of the Mars Science Laboratory (MSL) rover, that has now been operating on Mars for more than four years. The ATS of REMS is only used to provide the air temperature, however on HABIT it will be used furthermore to provide information about winds and heat transfer at the surface of Mars. The retrieval method needs to be further investigated and vali-dated. This master thesis is aimed at three goals: 1) the development and testing of an Earth Scaled Atmospheric Temperature Sensor (ESATS) to test the retrieval con-cept; 2) the validation with other Earth-based standard wind sensing technologies under outdoors uncontrolled conditions; and 3) the analysis of the existing observa-tions of the ATS of REMS on Mars to get a better understanding of its expected future performance on HABIT once it operates on Mars.
The ESATS is a up-scaled semi-autonomous prototype version of an ATS which consists of a rod of different size and material to those that are used on REMS and will be used on HABIT. The rod shall be heated from the base where it is attached to. The temperature profile shall be measured at three different measurement points. All these temperatures are different from the one of the atmosphere to which the rod is exposed to. The temperature profile along the rod changes depending on the air temperature, air density and the wind speed because of the convective heat transfer. A preliminar analysis is used to define what is the ideal length of the rod, and what is the material that is best adapted for this experimental prototype. Since the air density is needed to retrieve the wind speed, the pressure will be monitored as well. In parallel, a second wind measuring technique based on the dynamic pressure changes detected in a Pitot tube is used as control.
under static conditions. 4) The second part of the measurement campaign will take place outdoors, where the ESATS is exposed to forced convection due to wind. In this setting first the influence of the Sun on the system is measured, as it is important to know, if the measurement can be performed when the illumination conditions change. 5) Next, the system is tested with the 50 cm rod in long term tests with the reference measurement of a commercial weather station (HOBO) next to it. With the data obtained the convective heat transfer method is used and the retrieved wind speed is compared to the one received from the HOBO.
Finally, to get a better understanding of its expected future performance on HABIT once it operates on Mars, the data of the ATS of REMS is used to perform the wind speed retrieval for Mars and to compare it with the data received from the REMS wind sensor. It is only operating during daytime and has still difficulties to retrieve a precise wind speed.
The measurement campaign has given several information about ATS in general. First it was decided to place the temperature sensor in the middle at 14 of the rod length, which is optimal for the retrieval process and which is also coincident with the one chosen for HABIT and REMS. The measurements in the laboratory are providing good and constant temperature profiles with the chosen setup which correspond with the one expected from the equations that describe the heat transfer problem in a long rod. On the other hand, it is not possible to calculate a valid ambient temperature for the short rods, which is because of an overheated boundary layer around the rods due to the heating. For this reason, it is recommended to use the longest rod in the lab.
For outside testing the influence of the Sun could be confirmed and was affecting the measurements of the copper rod. During the time span where the prototype was in the Sun, it was not possible to get any reasonable results. The next measurement campaign was defined in a shadowed area with diffuse light only. Finally, the tests of exposure to dynamic changes over time are in excellent agreement with the ones provided by the HOBO station and can even give a better resolution and sensitivity to small changes of wind magnitude. This prototype has confirmed experimentally, that under Earth conditions, this method can be used to retrieve the wind speed. Finally, the Martian data of the REMS ATS are analyzed and the comparisons suggest that the method is sensitive to wind changes on Mars as well, and shows better time and magnitude resolution than the existing REMS wind sensor. This confirms that this method can be successfully used for the HABIT sensor.
This project has been funded with support from the European Commission. This publication [communication] reflects the views only of the author, and the Commis-sion cannot be held responsible for any use which may be made of the information contained therein.
1 Introduction 1
2 Instrument 3
2.1 General Physics and Formulation . . . 3
2.1.1 Properties of Earth Atmosphere . . . 3
2.1.2 Rod Material and Length . . . 4
2.1.3 Temperature Sensors . . . 6
2.1.4 Required Heating Power . . . 8
2.1.5 Tripod . . . 11
2.1.6 Microcontroller and Other Electronics . . . 11
2.1.7 Mathematical Model of the Prototype . . . 12
2.2 Prototype Requirements and Implementation . . . 13
2.3 Pitot-Tubes . . . 15
2.4 Electronic Calculations . . . 16
2.5 Software . . . 19
3 Evolution of the ESATS 21 4 Measurement 23 4.1 Measurement Protocol . . . 23 4.2 Laboratory Measurements . . . 24 4.3 Outdoors Measurements . . . 25 5 Data Analysis 26 5.1 Data Formats . . . 26 5.1.1 File Names . . . 26 5.1.2 Raw-Data-File . . . 27 5.1.3 Processed-Data-File . . . 28 5.2 Earth Data . . . 29 5.2.1 Laboratory Experiments . . . 29
5.2.2 Field Site Experiment . . . 47
5.3 Martian Data . . . 58
6 Results 62 6.1 Laboratory . . . 62
6.2 Outside . . . 63
6.3 Mars Data . . . 65
6.4 Comparison Between Mars and Earth . . . 65
8.2 Static Condition Test . . . V 8.2.1 Static Condition Test for 16 cm Rod . . . V 8.2.2 Static Condition Test for 33 cm Rod . . . .VIII 8.2.3 Static Condition Test for 50 cm Rod . . . XI 8.3 Reasoning for Rod Length . . . .XIV 8.4 HOBO Measurements for 50 cm with 20◦C Heating . . . XV 8.5 HOBO Measurements for 50 cm with 6◦C Heating . . . .XVI 8.6 Arduino Code . . . .XVII 8.6.1 Debugcode with Serial Communication . . . .XVII
List of Figures
2.1 Length of rods . . . 6
2.2 Power level of the heating resistor against case temperature . . . 8
2.3 The aluminum plate with heating resistors attached . . . 10
2.4 Sketch of the prototype . . . 13
2.5 Mounting for the rods . . . 14
2.6 Sketch of a commercial Pitot-tube . . . 15
2.7 Sketch of the basic electronics setup . . . 16
2.8 Sketch of the electronic circuit . . . 18
2.9 Serial to txt GUI sketch . . . 20
3.1 Preliminary tests of the Pt1000 version 2 . . . 22
3.2 Pitot tubes front view . . . 22
3.3 Preliminary tests of the Pt1000 version 3 . . . 22
3.4 The ESATS in its final form . . . 22
5.1 Effects of noise on temperature readout for uA741 . . . 31
5.2 Pattern of noise measured at the output in mV for uA741 . . . 32
5.3 Effects of noise on temperature readout for TL081 . . . 32
5.4 Pattern of noise measured at the output in mV for TL081 . . . 32
5.5 Effects of noise on temperature readout for OP113 . . . 33
5.6 Pattern of noise measured at the output in mV for OP113 . . . 33
5.7 Noise in the temperature profile without using mean . . . 34
5.8 Calibration curve for 16cm rod . . . 34
5.9 Calibration curve for 33cm rod . . . 35
5.10 Calibration curve for 50cm rod . . . 35
5.11 Calibration curve for 75cm rod . . . 36
5.12 Calibration data for the two pressure sensors . . . 37
5.13 Comparison between measured temperature and calculated tempera-ture profile over the rod length for 16 cm rod . . . 38
5.14 Comparison between measured temperature and calculated tempera-ture profile over the rod length for 33 cm rod . . . 38
5.15 Comparison between measured temperature and calculated tempera-ture profile over the rod length for 50 cm rod . . . 39
5.16 Comparison between measured temperature and calculated tempera-ture profile over the rod length for 75 cm rod . . . 40
5.17 Reasoning of the precision for detecting a valid ambient temperature. Up-scaled figure in 8.22 . . . 41
5.18 Temperature profile in lab for 10◦C heating . . . 42
5.19 m and h-values in lab for 10◦C heating . . . 43
5.23 Wind profile in lab for 20◦C heating . . . 45
5.24 Temperature profile in lab for 30◦C heating . . . 46
5.25 m and h-values in lab for 30◦C heating . . . 47
5.26 Wind profile in lab for 30◦C heating . . . 47
5.27 Outside test setup with a reference anemometer . . . 48
5.28 Influence of the wind on position of an Anemometer . . . 48
5.29 Influence of the Sun on the temperature measurement . . . 49
5.30 Random Influence of the Sun on the temperature measurement . . . . 50
5.31 Temperature profile of the 50 cm rod for a 20◦C heating . . . 51
5.32 m-h profile of the 50 cm rod for a 20◦C heating . . . 51
5.33 Wind profile of the 50 cm rod for a 20◦C heating . . . 52
5.34 Temperature profile of the 50 cm rod for a 20◦C heating . . . 53
5.35 m-h profile of the 50 cm rod for a 20◦C heating . . . 54
5.36 Wind profile of the 50 cm rod for a 20◦C heating . . . 54
5.37 Direct comparison of the HOBO and the ESATS wind speed values . 55 5.38 Temperature profile of the 50 cm rod for a 6◦C heating . . . 56
5.39 m-h profile of the 50 cm rod for a 6◦C heating . . . 56
5.40 Wind profile of the 50 cm rod for a 6◦C heating . . . 57
5.41 Direct comparison of the HOBO(left) and the ESATS(right) wind speed values . . . 58
5.42 m and Tf values . . . 59
5.43 Wind speed comparison . . . 60
5.44 Wind speed day . . . 60
List of Tables
2.1 Material suggestions for the rod . . . 4
2.2 Properties for sensor comparison . . . 6
2.3 REMS ATS Properties (P1k0.161.7W.A.010) compared with ESATS (C420) . . . 8
2.4 Heating of the copper rods above ambient temperature . . . 8
2.5 Properties of used materials for the volume and mass of a 33 cm cop-per rod . . . 9
2.6 Energy needed for heating up the system for different rods up to a specified temperature ∆T . . . 9
2.7 Power needed for heating up the different rods up to a specified tem-perature in a given time . . . 9
2.8 Voltage needed for heating up the rod within 5 minutes . . . 11
2.9 Different available rod lengths . . . 13
2.10 Constant values for temperature calculation . . . 19
α Angle of incidence
∆p Pressure difference
∆T Temperature change
∆t0.5 Time in which 50% of the temperature change are reached
∆t0.9 Time in which 90% of the temperature change are reached
µ Dynamic viscosity of gas
ρ Density
A Calibration value 2.10
a Length of the longer side of a cuboid
AC Cross section
amu Atomic Mass Unit
AT S Atmospheric Temperature Sensor
B Calibration value 2.10
b Length of the shorter side of a cuboid
BM P Barometric pressure sensor
C Calibration value 2.10
c Specific heat capacity of a material
C420 Pt1000 temperature sensor
Cp Temperature dependent heat capacity
CW 004A Copper alloy with high purity
D Diameter
Dh Diameter of the cross section
EDM Entry Decent Module
ESAT S Earth Scaled Atmospheric Temperature Sensor
F R4 Flame Retardant
GT S Ground Temperature Sensor
GU I Graphical User Interface
H Height
h Heat transfer coefficient
HABIT HAbitability, Brine Irradiation and Temperature Package
HOBO Commercial Weather Station
I Current
IR Infrared
ISA International Standard Atmosphere
k Thermal conductivity of rod
kf Thermal conductivity of fluid
L Length of the rod
LM ST Local Mean Solar Time
M Mass
m Dimensionless fin/rod parameter
MEarth Molar mass of Earth atmosphere
M D REMS data file
M ER Mars Exploration Rover
M SL Mars Science Laboratory
n Position of the middle temperature = 4
N ASA National Aeronautics and Space Administration
N u Nusselt number
N V REMS data file
OP 113 Operational Amplifier OP113
OpAmp Operational Amplifier
P Standard atmospheric pressure
P0 Power
p1 Pressure read out sensor 1
p2 Pressure read out sensor 2
PP er Perimeter of the rod
P DS Planetary Data System
P r Prandtl number
P S Pressure Sensor
P t1000 Temperature sensor consisting out of Platinum
Q Heating energy
Qf Rate of heat transfer
R Resistance
r Radius
R0 Resistance at 0◦C
Rt Temperature dependent resistor value
RGas Universal gas constant
Re Reynolds number
REM S Rover Environmental Monitoring Station
RHS Relative Humidity Sensor
RT D Resistance temperature detector
SCL Clock port of I2C-Bus
SDA Data port of I2C-Bus
T Temperature
t Temperature in ◦C
T0 Temperature at the tip of the rod equal ambient temperature
T∞ Ambient temperature
Ta/T a Temperature at tip of rod
Tb/T b Temperature at base
Tf/T f Ambient temperature
Tref/T ref Reference temperature
T GO Trace Gas Orbiter
T L081 Operational Amplifier TL081
U Voltage
uA741 Operational Amplifier uA741
U V Ultraviolet Sensor
V Volume of a body
Vcc Supply voltage = 5V
Vout Output voltage
vW ind Wind speed
value Value read out from one of the analog ports of the Arduino
Introduction
As weather or climate is influencing our every days life on Earth, it will also affect life on other planets. For this reason, for exploring another planet, like Mars, the goal is to obtain data, which is describing the parameters, which can improve the development of climate models. Therefore, meteorological instruments, which can resist high accelerations during the start and landing phase of a spacecraft have to be developed.
In the 50 years of history of planetary exploration on Mars, the improvement of these measurements can be seen. The first probes sent to Mars were four soviet spacecrafts and one of NASA. These were followed by Viking 1 and 2. Both of them included landers, named Viking I and II, which successfully landed on Mars. These landers had already meteorological probes on board, including three hot-wire anemometers and the same amount of fine hot-wire thermocouples [1]. The data obtained from these landers included information about the temperature range and wind speed, which is needed for the development of a climate model. With help of this the first steps in the search for traces of organic life on another planet were made.
These two successful landers were followed by several other mission to Mars, which included the rover Soujourner on the Mars Pathfinder mission and the two Mars Exploration Rovers (MER), Spirit and Opportunity. These missions were followed by the Mars Phoenix Lander and concluded until now with the MSL or Curiosity in November 2011 [2].
The hot wire anemometer used on the viking probes, was consisting of two hot wires, with a mounting perpendicular to one another. In between these two hot wires a reference air temperature was measured, with an identical hot wire. To figure out the direction of the wind a quadrant sensor was used, which was mounted on the top. A quadrant sensor is a positioning sensor, which can detect the direction of a force, in this case wind, that is acting on the system. The Mars Pathfinder mission had a meteorological boom, where the atmospheric wind-profile of the Martian atmosphere should be measured. The MERs did not have a meteorological experiment on board, but with the Phoenix lander a dynamic pressure instrument in a thin kapton tube was brought to Mars[3].
mounted on a silicon chip. For the temperature measurements several devices were used. These are a Wind Sensor (WS), an Air Temperature Sensor (ATS), a Pressure Sensor (PS), a Relative Humidity Sensor (RHS), an Ultraviolet Sensor (UVS) and a Ground Temperature Sensor (GTS). The ATS is used in this instrument package for the first time and is described as a 35 mm long fin consisting of multilayer FR4, a composite material. On it two thermistors are mounted at the tip and at 14 of the length of the rod, starting from the base. The thermistors used are Pt1000 which are measuring in a range of 150 K to 300 K. With an ATS the atmospheric temperature or air temperature should be retrieved with the convective heat transfer method, which will be described later in this document and in more detail in [4] for the case of the HAbitability, Brines, Irradiation and Temperature (HABIT) instrument. This instrument package is monitoring all necessary atmospheric properties.[3]
In the future there will be several other missions to Mars, with two of them will be named here. First ExoMars 2016, consisting of the Trace Gas Orbiter (TGO) and the Entry, Descent and Landing Demonstrator Module (EDM), which will be a technology demonstrator and second, ExoMars 2018. This mission, including a surface platform and a rover, is now called ExoMars 2020 due to a delay. Both of them will be send up to Mars in an ESA-Roscosmos collaboration. This mission will include a follow up instrument of REMS, HABIT. This instrument package includes a GTS, to measure the ground temperature, photodiodes, for the sunlight penetrating the atmosphere of Mars, a conductivity measurement of perchlorates, which are present in the Martian soil and a similar model of the ATS, compared to the one of REMS. In this case there will be three Pt1000 mounted to each of the three FR4 rods, which have the same dimension as the ones used on REMS. The temperature sensors will be mounted at the same positions as on REMS and the third one will be mounted at the base. With this setup and a mounting of the rods in x, -x and y direction, if one is looking on the instrument in front, the aim of the ATS is to retrieve the wind speed and direction on Mars and to calculate the ambient temperature.
Therefore, it is necessary to validate the model, which should be applied on the instrument in the end. For this reason, the Earth Scaled Atmospheric Temperature Sensor (ESATS) is developed. It should verify under Earth conditions that the principle can be used on Earth, that the assumed model is working in the expected region and in the end that it is working on Mars as well, with data received from REMS, included in the developed model.
Instrument
This chapter contains some general information about why the parts used in the prototype were used, how the circuit is built, how the prototype is designed, and which assumptions and theories were used.
2.1
General Physics and Formulation
2.1.1
Properties of Earth Atmosphere
This section contains some general information about Earth atmosphere, like pres-sure, temperature and molar mass, which are used in calculations several times during this project. For this reason, these information are provided right away.
Firstly, the molar mass of Earths atmosphere is calculated from the three most common atoms or atom pairs, Nitrogen, Oxygen and Argon. Their distribution, rounded to full percent, in the atmosphere are as follows [5]:
• Nitrogen 78% with a mass of 28 amu • Oxygen 21% with a mass of 32 amu • Argon 1% with a mass of 40 amu
Here, one can calculate with the following formula MEarth = 0.78 · 28 + 0.21 · 32 +
0.01 · 40 the molar mass of Earths atmosphere. For this calculation all other atoms and molecules are not taken into account, as they have a probability of less than 1% corresponding to less than 1000 particles per million, because of too small influence. So the result of this equation is MEarth = 28.96 u.
The temperature T = 288.15 K is given as the absolute temperature under stan-dard atmospheric conditions [6] at sea level equivalent to 15◦C. The pressure of P = 1013.25 hPa under the same conditions, is needed to calculate the density of Earth atmosphere, as well as the universal gas constant, which has a value of RGas =
0.082L·atmK·mol. With the molar mass of Earth’s atmosphere one can then calculate the density ρ = P ·MEarth
T ·Rgas = 1.225
kg
m3. If one compares this value to the one of the international standard atmosphere (ISA) [6], it can be seen that they fit well.
kf = 0.02534m·KW [7] is the thermal conductivity of air, which is in general
tem-perature dependent, but here as a simplification the thermal conductivity for 15◦C is used.
As for this project the velocity of wind vW ind is an important factor, as it should
be calculated with help of the temperature profile received. Therefore, different velocities, 1, 5, 10 and 14ms where chosen. One has to mention that vW ind 6= 0, as
otherwise the equation Re = ρ·vW ind·Dµh, which will be explained in 2.2, would result
in zero and therefore, the length, L, of the rod calculated with L = 4.0 · q
Dh· 4·hk ,
explained in 2.7, would increase to infinity. As a maximum velocity 14ms[8] was chosen, as the instrument has to withstand the wind without being mounted on a fixed place.
2.1.2
Rod Material and Length
To get first hints about how to perform the experiment, a conference report from H.I. Abu-Mulaweh [9] and a paper of D.W.Mueller Jr. and H.I. Abu-Mulaweh [10] were studied. In the conference report three different materials were suggested. These were copper, steel and aluminum, with their properties given in table 2.1. For steel the alloy stainless steel 304 was used. As one can see, the melting temperature of all of the suggested materials is way higher than the maximum temperature, which will be reached in the measurement campaign on Earth. So all three materials can be used in this point of view, but it is also important to have a good thermal conductivity through the material. Copper has the best properties, with a value of 4.01cmKW , which is nearly twice as high as the one of the next best material, Aluminum. Compared to these both, steel has a very small thermal conductivity. So copper seems to be the best material for such a setup. The downside of copper is that it is quite heavy and expensive compared to the other two recommended materials.
Steel Copper Aluminum
Density 7.90cmg3 8.96 g cm3 2.70 g cm3 Melting Point 1425.00◦C 1084.62◦C 660.32◦C Thermal Conductivity 0.15cmKW 4.01cmKW 2.37cmKW
Table 2.1: Material suggestions for the rod [11]
As untreated copper has impurities, which can be spread unevenly along the rod, it is hard to get a stable thermal conductivity, beside increasing the purity of the copper. For this reason, CW004A [12] copper rods were used. This copper alloy has a purity of more than 99.90% of a mixture of copper and silver and less than 0.04% Oxygen enclosed. With a thermal conductivity of 3.88cmKW they are very close to the value of pure copper which was used in table 2.1. Whereas the thermal conductivity of untreated copper alloys is between 2.40m·KW and 3.80m·KW . The density of these pure copper rods is at 8.890cmg3.
diameter means a higher mass, which is increasing by the factor of four, if the radius is doubled. For simplicity one can assume that only the maximum of the wind speed should be detected, which means that equation 2.1 has to be at its
maximum. This happens if the cosine becomes maximum, so for α = 0◦, which
implies that Dh = D = 1cm, where Dh is the cross-section diameter and D the
diameter. This results from the way α, the so called angle of incidence, is pointing, because the angle is giving the direction from which the wind is flowing around the rod. This means, if the rod points in the same direction as the wind flows, the angle is 90◦ and it is hard to measure anything. If the wind instead hits the rod perpendicular, the maximum effectiveness is reached.
Dh =
D
cos(α) (2.1)
Inserting Dh into 2.2 results in the Reynolds number, which is calculated for the
different wind speeds, which are given in ??, as well as the density and the dynamic viscosity. The Reynolds number is an a-dimensional number, with which one can point out similarities in flow patterns of different flow situations.
Re = ρ · vW ind·
Dh
µ (2.2)
The next step is to calculate the Prandtl number 2.3, which is another dimensionless number, resulting in the ratio of viscous and thermal diffusion effects [13].
P r = Cp·
µ kf
(2.3) With both previous equations the Nusselt number 2.4 can be calculated. It is again an a-dimensional number, as only the Prandtl - and Reynolds number are used to calculate it. It is quantifying the ratio between the two heat transfers, the convective and conductive one.
N u = 0.3 + 0.62 · Re 1 2 · P r 1 3 1 + (0.4/P r23) 1 4 · (1 + ( Re 282000) 5 8) 4 5[13] (2.4)
This equation is applicable, as long the factor P r · Re > 0.2 is valid. With the resulting value, one is now calculating the heat transfer coefficient 2.5, h, with the unit mW2K.
h = kf ·
N u Dh
[13] (2.5)
To determine the rod length, which will be used, one needs to receive valid results for Qf =
√
hPP erkAC·(T0−T∞) [9], where Qf is the value by which the rods are heated,
Acthe crosssection and PP er the perimeter of the rod, which is cylindrical for Earth
conditions. This implies, that the temperature in infinity and the temperature at the tip of the rod have to be close to each other. So one has to make sure, that mL > 4[9] is valid for this equation 2.6.
m = L r
4 · h k · Dh
For this reason, the factor was set to 4.0 as seen in 2.7, to receive the minimum rod length needed. L = 4.0 · r Dh· k 4 · h[9] (2.7)
This results in a rod length of 47.7 cm, for a wind speed of 10m
s. To test the effects
of a too short or too long rod length, the sizes of the rods were designed around this value, with a rod length of 50.0 cm as the closest. These values can be seen in table 2.1.2. The short rod lengths were chosen to fit in a future test in a wind tunnel, where the system should be tested and the longest rod was limited by weight, as the stability of the system has to be increased, which results in a higher mass the longer the rod gets.
• 0.75 m • 0.50 m • 0.33 m
• 0.16 m Figure 2.1: Length of rods
2.1.3
Temperature Sensors
There is a wide range of different atmospheric temperature sensors available, so it is important to choose the right type and specifications of the temperature sensor. For this reason, one has to have a closer look on the different types of temperature sensors. For this work, thermistors, resistance temperature detectors (RTD) and thermocouples were taken into account.
As one wants to recalculate the wind speed out of the measured data even small changes in the temperature measurements are important, but a fast thermal response is also recommended. Furthermore, it should run outside in a cold climate with
temperatures which can be below −40◦C and on the other hand, the maximum
temperature will not exceed 100◦C so that the temperature range has to be in between these limits. Another criteria for the decision process was to have a sensor, which can deliver similar specifications as the ones used in REMS, so it may be used in a future mission. For such a mission it is important to have a high long term stability, as the sensor will be in a remote place, where it is not possible to re-calibrate it.
Thermocouple RTD Thermistor
Range −200◦C - +2000◦C −250◦C - +850◦C −100◦C - +300◦C
Accuracy >1◦C 0.03◦C 0.1◦C
Thermal Response Fast Slow Medium
Long term stability Low High Medium
Table 2.2: Properties for sensor comparison [14]
and with a possibility to detect temperatures above 300◦C the maximum temper-ature is high enough to never be reached on Earth under a normal experimental setup.
The measuring accuracy presents big differences, which means that RTD’s are the most precise temperature sensors and thermocouples have the lowest precision. The low precision of the thermocouples results in the wide temperature range they operate in, as in such a range it does not matter anymore if the result variates by one degree. These sensors are normally used in industry in high temperature production chains.
More precise are thermistors. They are normally operated in a temperature range around room temperature. For high sensitive operations, like calibration, RTDs are used with the highest precision for the temperature measurement with 0.03◦C.
The thermal response of the temperature sensors is reversed compared to the accuracy. This means, that the thermocouple has the fastest response, because of the material used for measurement. These sensors are consisting of two different materials within a voltage difference in which the same temperature occurs. The thermistor has a medium response, resulting from bigger changes in the size of the thermal resistance compared to the resistor changes of an RTD. Such resistance temperature detectors are normally consisting of platinum.
For the long term stability of the sensors, the thermocouple has the lowest one, followed by the thermistor and with the highest stability the RTDs are at the end. Additionally, all 3 types of sensors need some sort of calibration. Also some errors could occur by measuring the temperatures with different methods.
Despite that the thermocouple is the sensor with the fastest thermal response, it has a too low accuracy and a low stability and can not be used for measuring. For the other two sensor types the main difference is that the accuracy is a bit better for the RTD and the thermal response is better for the thermistor. But for an RTD the thermal response is differing if the surrounding material is changing. If one compares the values for other materials, the RTD has a lower response time, which means it is reacting faster to environmental changes. So the sensors used will be RTD-Type sensors. Commonly RTDs are consisting of Platinum, but there are also some other materials possible. As for REMS a Pt1000 [15] was used, which is an RTD the focus of the search was on other Pt1000s as the other sensor types discussed beforehand did not behave better than them. RTDs are the preferred type of sensors, as they have a high acceleration resistivity, which is needed to survive the launch and landing of a rocket, where harsh forces interfere with the system during both phases.
For the ESATS the size did not matter, as the system is staying on Earth and therefore, has only the limitations which are chosen in the design phase. There-fore, as the chosen copper rods are already bigger than the ATS on both Martian prototypes, the size of the temperature sensors 2.3 is not an important criteria for comparison. The most important point which has to be taken into account is the op-erating temperature, which should be in a range of at least −40◦C to 100◦C. This
range includes the maximum temperature on Earth of 60◦C, with the maximum
on Mars the minimum temperature should be below the −90◦C mentioned in [16] as lowest temperature for the ATS of REMS. For Earth operation the temperature range of the Pt1000 was set to −40◦C to 140◦C, as the sensor should be capable of the Martian conditions in theory, but for the ESATS it is not needed to cover this range. The reason why the temperature range is set to the range given is explained later in this document. An important criteria was the thermal response time of the Pt1000 which is for the chosen C420 [17] in the same range as for the ones used on REMS. This means even if they are not better than the ones of REMS, these temperature sensors have similar properties and can be working in the same environments. For the response time ∆t0.5 stands for the time in which 50% of the
temperature change can be adapted and ∆t0.9 for a change of 90%. Even if there
may be better sensors available for other temperature regions, it was one of the few who had a temperature range fitting for both environments and one of the best documentations found during the search.
REMS ATS ESATS
Operating Temperature Range −200◦C - +750◦C −196◦C - +150◦C
Measuring Temperature Range −90◦C - +40◦C −40◦C - +140◦C
Temperature Coefficient 3850ppmK 3850ppmK
Resistance at 0◦C 1000 Ω 1000 Ω
Class A (DIN EN 60751 F0.1) B DIN EN 60751
Size 1.6 mm x1.2 mm 3.9 mm x1.9 mm Response Time Water (v = 0.4ms) ∆t0.5 = 0.05s ∆t0.5 = 0.08s ∆t0.9 = 0.18s ∆t0.9 = 0.25s Air(v = 1ms) ∆t0.5 = 1s ∆t0.5 = 3.5s ∆t0.9 = 2.5s ∆t0.9 = 15.0s
Table 2.3: REMS ATS Properties (P1k0.161.7W.A.010) compared with ESATS (C420)
2.1.4
Required Heating Power
To specify the type of heating for the copper rods one needs to know how much power the system will need to reach the given temperatures, ∆T , above the room temperature, seen in 2.4.
• +0◦C
• +10◦C
• +20◦C
Table 2.4: Heating of the copper rods
above ambient temperature
With these temperatures in mind and the length of the rod known, one is able to calculate the power consumption. For this one needs the volume of one of the rods, calculated with V = π · r2 · H, where r is giving the radius of a cylinder and
H the height. The volume also has to be calculated for the aluminum plates, which are cuboids, so the volume V is a2· b, with a the length of the sides and b the height.
With these two volumes one has to calculate the mass of each with help of M = V ·ρ, where the density ρ of copper [12] and aluminum [19] is given in table 2.5. In the same table also the volume of one of the rods, with length of 33 cm is given, as well as the mass and the specific heat capacity, which is needed for the calculation 2.8 of the heating energy for the specified rod.
Q = M · c · ∆T (2.8)
In this calculation ∆T is one of the temperatures given in 2.4. Also the aluminum plate has to be taken into account twice, as the ESATS consists of two of them.
Material Volume Density Mass Specific heat capacity
Copper 25.918 cm3 8.90 g cm3 230.67 g 380 J kg·K Aluminum 28.125 cm3 2.80 g cm3 78.75 g 862 J kg·K
Table 2.5: Properties of used materials for the volume and mass of a 33 cm copper rod
After solving this equation for all four rods, for a ∆T , the temperature difference, of 10◦C and 20◦C, seen in table 2.6, one has to specify the time frame in which one wants to increase the temperature to receive the power needed for heating up the system. This power can be seen in table 2.7, where the values for the different rod length are displayed for times of one and five minutes.
∆T 16 cm 33 cm 50 cm 75 cm
+10◦C 1756.1 J 2234.2 J 2685.8 J 3349.8 J
+20◦C 3512.2 J 4468.4 J 5371.5 J 6699.6 J
Table 2.6: Energy needed for heating up the system for different rods up to a specified temperature ∆T
∆T 16 cm 33 cm 50 cm 75 cm
1 min 5 min 1 min 5 min 1 min 5 min 1 min 5 min +10◦C 29.27 W 5.85 W 37.24 W 7.45 W 44.76 W 8.95 W 55.83 W 11.17 W +20◦C 58.54 W 11.71 W 74.47 W 14.90 W 89.53 W 17.91 W 111.66 W 22.33 W
Table 2.7: Power needed for heating up the different rods up to a specified temper-ature in a given time
through the surrounding air would be bigger than the possible heating power and it is impossible to heat up the system fast enough. For this reason, there are no values calculated for more than five minutes. In this way one will choose the five minute heating, as it is still a fast way to heat up the system, with a possibility to control the heating manually.
As heating foils for this power region are large compared to heating resistors which can handle the same power with less size, the heating resistors are preferred. As resistors a 47 Ω [18] heating resistor was chosen, which can take up to 50 W if its temperature stays constant at around 20◦C. For this setup this is not the case, because the used materials should heat up. This means, that the power which can be used is limited, as seen in figure 2.2. Here the rated power in % is showing the power which can be used for heating if the temperature stays within the given limits. For this reason, two resistors of the same type are used to stay below the given limits. As for the long rods with a temperature difference of 20◦C two resistors are already close to their limit, a third resistor was added to the instrument. These resistors are connected with kapton and heating paste to the aluminum plate and no screw is used, as the surface of the aluminum plate should be as stable and as isolating as possible, which changes with screws. The setup of the heating resistors is shown in 2.3.
Figure 2.3: The aluminum plate with heating resistors attached
For the calculated power values some approximations were applied. First in between the two aluminum plates the space filled with air is excluded, in order to have simpler equations. Furthermore, a perfectly isolated body, where no heat is dissipating in any direction is assumed, which is also not feasible. So with the calculated power applied to the system, it will take a longer time span to heat up the prototype. On the other hand, no other problematic side effects are influencing the system because of this assumptions.
Temperature 16 cm 33 cm 50 cm 75 cm
+10◦C 11.73 V 13.23 V 14.50 V 16.20 V
+20◦C 16.59 V 18.71 V 20.51 V 22.91 V
Table 2.8: Voltage needed for heating up the rod within 5 minutes
2.1.5
Tripod
For the tripod one has to define how much force it has to take so it is still stable and will not fall apart. For this reason, the mass of the longest rod was calculated, with the help of the density 8.890 cmg3 and the volume of a cylindrical body, V = π · r
2· H.
This results in 0.524 kg mass of the rod, where the mass of the mounting has to be added, so that the mass of the ground plate can be calculated. For this reason, the mounting was built out of aluminum as it has a low density and is still stable. Furthermore, the heat transfer coefficient is low, so it will be used as an isolation as well. This mounting results in a mass which is adding to the copper rod, but it is comparable small. If one assumes that all the weight is on a lever at the tip of the rod, where the other end is mounted, one can calculate the counter weight which is needed to keep the rod in place. The assumption is not precise and the weight of the plate will be too high, but as the force of the wind is not included, it will have enough mass to withstand the attacking wind. As the foot of the tripod should be in a reasonable size, the material used to manufacture it, was aluminum. Therefore, the weight of the plate was chosen to be 1.8 kg after calculating the force it has to withstand. To get such a weight, the dimension of the aluminum plate is 150 mm x 200 mm x 20 mm.
2.1.6
Microcontroller and Other Electronics
To read out the data received from the temperature sensors and the pressure sensor, one needs a device to handle and store the data. For this an Arduino was chosen, as it is easy to access and easy to program. Also it is a cheap alternative compared to other more advanced microcontrollers. The program which has to be written for this is not too computational expensive, so that it is possible to run it on such a simple architecture. As model type the Arduino Uno Revision 3 was chosen, as it was not clear in the beginning, how many pins will be in use and if there will be more sensors added to the circuit. So it was a compromise between space needed and the possibility to increase the setup. If the prototype will be rebuilt, it is possible to use an even smaller Arduino microcontroller, as only three analog inputs, the 5 V, 3.3 V, ground, SDA, SCL and four digital pins are needed.
The system has connected a micro-SD-card-shield to be able to store data if it is run without connection to a computer. Therefore, a micro-SD-shield from Adafruit, which is a company providing parts for Arduinos, was used. This shield has a well documented example code, which leads with small adaptions to a simple plug and play device.
pressures, two standard pressure sensors are enough and no too high requirements had to be fulfilled.
As one can see, it is easy to look for premade parts for Arduinos and one can assemble and program them well. For this reason, an Arduino is preferred compared to other microcontrollers.
2.1.7
Mathematical Model of the Prototype
The prototype is consisting of a rod out of copper with three temperature sensors mounted on this rod. One of these temperature sensors is mounted at the base, where Tb, the base temperature is measured. The second one is placed at the tip
of the rod, where Ta, the infinite temperature or also the temperature at the tip is
obtained and the third one is placed somewhere in between, at a specified length of the rod, which is given by L ·n1. There the temperature Tln is measured.
With these temperatures given, one can calculate the m-value, which is the dimensionless fin parameter. This value is calculated with help of
Ta− Tf = (Tb− Tf) · 1 cosh(m)[13] (2.9) and Tln− Tf = (Tb− Tf) · cosh(m · (1 − n1)) cosh(m) [13] (2.10)
Here Tf, the ambient temperature of the environment, and m are the values one has
to calculate. To be able to solve these equations, one has to cancel out one of these values. Therefore, one is subtracting equation 2.9 from equation 2.10.
Tln− Ta= (Tb− Tf) ·
cosh(m · (1 − 1 n)) − 1
cosh(m) (2.11)
Furthermore, equation 2.9 has to be reformulated a bit, seen in 2.12, Tb− Ta = (Tb− Tf) · (1 −
1
cosh(m)) (2.12)
so one can cancel out Tb− Tf by division of 2.11 through 2.12.
Tln− Ta Tb− Ta = cosh(m · (1 − 1 n)) − 1 cosh(m) − 1 (2.13)
With a simple reformulation one can compare the two sides of the equation, as they are growing with different speeds, to calculate the correct m-value, as seen in equation 2.14.
Tln− Ta
Tb− Ta
· (cosh(m) − 1) = cosh(m · (1 − 1
n)) − 1 (2.14)
With the calculated m-value one can now calculate Tf from equation 2.9.
Tf =
Ta− cosh(m)Tb
The resulting parameters can be used to calculate the heat transfer coefficient, h, with help of m = LqD4·h
h·k2.6.
With the before mentioned equation, h = kf · N uD
h 2.5, the velocity of the wind can be determined. As all values have to be known, also the pressure has to be measured with the same periodicity as the temperatures are measured. It is hard to solve this equation for the wind speed by inserting the equations 2.1 to 2.4 into 2.5, so the wind speed is varied until h from equation 2.5 equals h calculated in equation 2.6.
Figure 2.4: Sketch of the prototype
2.2
Prototype Requirements and Implementation
The ESATS will consist of two aluminum plates with 7.5 x 7.5 x 0.5 cm. These two plates are connected with screws and will have a distance in between the two nearest surfaces of approximately 0.6 cm, where the heating resistors are attached with kapton tape. The mounting can be seen in figure 2.5. On the left plate one of the rods will be mounted in the center. The rod will be screwed in, as it has a thread of length 1 cm at one side. The rods have different lengths, as seen in 2.9 and are consisting of copper with the properties specified in 2.1.2.
• 0.16 m • 0.33 m • 0.50 m • 0.75 m
Table 2.9: Different available rod lengths
Figure 2.5: Mounting for the rods
The plate at the bottom is consisting of aluminum, and is only for having a counterweight to the longest copper rod, so the construction can not fall over. The x-profile from the base plate to the heating plate is available in three different lengths, 10 cm, 20 cm and 30 cm and the 30 cm profile is also available with two of them screwed together. To one of the profiles the plates with the heating resistors are mounted, with a total height from ground of 22.5 cm to the lowest corner of these plates. It is also possible to increase the stability with connecting even more x-profiles together. In addition, the height of the system can be increased due to a connection of two or more rods on top of each other, but this is not recommended before stabilizing the base further. The heating resistors are glued to the left and the right of the screwing which is sticking through the front plate. To power these resistors a maximum of 250 V is allowed, if the temperature stays constant at 20◦C on the resistor itself. As this is not the case, one has to check how much power is allowed. For the construction used, a maximum voltage of 20 V is recommended and should not be exceeded. If one is getting close to it, one should better connect another heating resistor below the mounted rod. Each of the heating resistors is powered by one channel of the standard laboratory power supply.
The data will be gathered with help of a voltage divider, where the resistance of the constant resistor will be similar to the resistance of the temperature at 0◦C, so 1 kΩ. The sensitivity with the 8 bit analog input of the Arduino is in this setup at 1.22 ◦C
step. The possible precision with this temperature range can be calculated by
the maximum temperature.
2.3
Pitot-Tubes
For the calculation of the wind speed a pressure measurement is required, which will be done by two BMP180 sensors. These sensors have the possibility to measure the pressure, the temperature and to recalculate the altitude above sea-level. One of the temperature measurements will be used as reference temperature value for the ambient temperature, so that one can compare these value with the calculated one later on. The altitude calculation is not needed, but one can add it if the prototype will be used in different regions. The most important part of these sensors is the pressure sensor, where one will be used to get the static atmospheric pressure and the other one will be used to calculate the dynamic atmospheric pressure, which is influenced by wind. This means that there will be a pressure difference ∆p between the two sensors under dynamic conditions. With this difference one can recalculate the wind speed vwind =
q
2 ·∆pρ [20], which will be compared with the wind speed calculated with the copper rod.
To acquire this, one has to build up something like a Pitot-tube. This kind of tube is in a commercial product consisting of one tube, which has a big entrance hole in front and some smaller ones at the sides. These holes are connected to tubes of the size of the holes and lead to the pressure sensors which are placed after a 90-degree bend of the tubes. The big tube is connected to one of the sensors to detect the dynamic pressure, the smaller ones are merged together and connected to the other sensor to measure the atmospheric pressure. This setup can be seen in figure 2.6.
Figure 2.6: Sketch of a commercial Pitot-tube [20]
the wind. So one has to decide where the experiment should lead to. A preliminary experiment has shown that the cross section of the tubes may be too small and has to be increased, which has to be confirmed with more measurements.
With the small cross section, it is important to point precisely in the direction of the wind, as already small changes make it hard to detect the pressure changes, as the device only has a small field of view due to the small inlet. The Pitot tubes were 3D-printed with an Ultimaker 2. For this reason, it is easy to copy the setup and build up more devices like this in a broader region to detect the wind speed on a bigger region.
Figure 2.7: Sketch of the basic electronics setup
2.4
Electronic Calculations
An Arduino Uno has six analog pins, where three will be used, as every temperature sensor needs one. The data received from these pins can have a range between 0 and 1023, which is a range of 1024 values.
An analog value is received by reading in the voltage in millivolt along a voltage divider. For a very basic setup only a reference resistor and a Pt1000 are needed, to build up a voltage divider as seen in figure 2.7. With this one can now calculate a resistance value Rt from the first readouts, by applying
Rt=
R
1023 value − 1
(2.16) to the system, where R = 1 kΩ . With this setup it is possible to achieve a precision of 1.2◦C per step. To increase this, as mentioned before, the measurement range has to be reduced, as one is measuring in the simple setup from 0 V to Vcc, which
is represented by 5 V. The lowest resistance, occurring at −196◦C, has a value of R =202 Ω. So with equation 2.16 rearranged and applied for voltages
Vout =
Vcc· Rt
Rt+ R
(2.17) one obtains a voltage of 0.84 V for the mentioned values and R0 =1000 Ω. This
reference voltage is obtained has to be implemented. If this reference voltage is fed in an OpAmp at the inverting end and the measured voltage at the non-inverting input, one has set the 0 V to the lowest measurable temperature, which is represented by the lowest resistance value of the system. If the system will be tested now it will result in the same precision, because the only thing done so far is a shift of the full supply voltage to the higher resistance values. To change this, one has to define the upper limit of the system, which can be done with equation 2.17. Therefore, the upper resistance value of 1586 Ω, which can be found at 150◦C, has to be inserted for R. This results in a voltage of 3.07 V as upper border voltage. So one can use the 3.3 V voltage output of the Arduino as reference voltage to decrease the voltage span in which the measured values can be. With this method applied, the whole temperature range can still be measured, but with a higher sensitivity of the single sensors.
If one is now measuring the different voltages with a voltmeter, one can see for the 0.84 V, a drop of a couple of millivolts. This phenomenon is occurring, due to a non stable setup. To handle this, one has to include a voltage follower for the low voltage border and for the value from the Pt1000s to stabilize the current level, so that the voltage is not changing anymore by drawing some current.
For this project the full range of the temperature sensors is not needed, so the lower border voltage is set to 1.506 V, which corresponds with a temperature of −40◦C, by using the system described above. Also the upper part of the measure-ment range is not needed in total. This will result in a cut off due to an OpAmp with an amplification of a factor around 2.2. The imprecision occurs due to the resistors, which have a default deviation around the value given for them. In addition the internal reference voltage, which is set to 1.1 V, is changed so that there will be a saturation of the amplifying OpAmps for voltages over 2.006 V. These changes are reducing the upper border of the measurement range from 150◦C to 138◦C, which is equal to 2.0 V. This adaptation of the circuit results in an increase in sensitivity up to 0.17step◦C , which is four times the sensitivity than in the beginning.
The described circuit for three temperature sensors is shown as an electronic schematics in figure 2.8, where the values for all resistors are inserted. The basic principles of this circuit are described in [21] with some improvements received from [22], which are the additional voltage followers for the reference voltage and after the voltage divider of the temperature measurement. The resistor values used are adapted to the needs of the circuit. The capacitors are removed as there is no im-provement gained by them. The part of the multiplexer was removed, as a time synchronous measurement is preferred and there is no difference in the dimension-ing of the circuit. For more temperature sensors in use a multiplexer is strongly recommended. The 220 kΩ resistor at the output of the circuit is used to decrease the noise level of the circuit. With this the voltage is lowered by a specified amount, but it is not changing the measurements. The last improvement, of adding the re-sistor, is the only one copied directly from [22], as all other parts are necessary and therefore, both circuits have them.
Then from the determined resistor value the temperature can be calculated with help of two different formulas, specified in DIN EN 60751 [23], 2.18 and 2.19, for calibration.
Rt = R0· (1 + A · t + B · t2) (2.18)
• A = 3.9083 · 10−3 1 ◦C • B = −5.7750 · 10−7 1 ◦C2 • C = −4.1830 · 10−12 1 ◦C4
Table 2.10: Constant values for temperature calculation
One of these formulas is used for positive temperatures 2.18, the other one for negative ones 2.19. This can be decided by checking if the calculated resistance is bigger or smaller than 1 kΩ and then the corresponding formula has to be applied. In these equations A, B, C are constants, given in table 2.10, R0 the resistance of
the temperature at 0◦C, 1 kΩ, Rt the resistance of the measured temperature and
t the measured temperature in ◦C. After rearranging and solving the calculation of the temperature for the different temperature sensors the results will be transmitted to a computer or stored on a Micro-SD card. If they are sent to the computer, a time stamp will be added after displaying them in a GUI, if the data is stored on a SD-card no time stamp is added. Both methods are storing the data in a txt-file so that the data can easily be recalled and processed further. The reason for storing the data and not doing a direct processing on the board is, that it is easier to access the data again on a computer and if some values are changing, one can use the same measurement data for comparison.
2.5
Software
The project is consisting of different software. Two are implemented for the Arduino, one is for serial communication with a computer and the fourth one is for processing the data.
The two software for the Arduino are basically the same with the same function-ality. The only difference is, that one is used for communication with the computer and therefore, has all the debug messages included and the other one is used with an external power supply without any connection to a computer, writing the data to files on a micro-SD-Card. Both software products are listed in the appendix and are public for further use. In the software the equations from chapter 2.4 are implemented to recalculate the temperature out of all values read in at the analog pins. Additionally, the values from the pressure sensor are read out, but only the temperature and the pressure are used further. For one program they are sent to the serial port of the Arduino in a specified format, for the other one they are written to the micro-SD-card and stored there in the format explained below.
To describe the two interfaces shortly, in the first software one can choose in a drop-down menu, to which port a con-nection has to be established and then one has to press the start button, where the transmission process will be started, even if the Arduino is already sending not displayed data. Then the Graphical User Interface (GUI) will tell how many trans-missions where already successfully pro-cessed and underneath it will tell the time in seconds, the raw value of each sensor, the resistance value of each sensor and also each temperature value in ◦C. In the bottom there will be written to which Com-Port the program is connected and the current status of the program.
The other GUI has a directory but-ton, where one can choose the data set which has to be read in and processed to recalculate the wind speed with help of the calculations given above.
Evolution of the ESATS
The ESATS evolved stepwise. The first prototype was built for testing the Pt1000 temperature sensors. Therefore, these sensors were attached to a plastic spoon, which can be found for example, in every cafeteria. This setup was used to prevent the temperature sensors to shorten, because if metal would have been used the cables could have been connected directly. This issue could be solved with some non conducting foil, like kapton, under the cable connection, but at this time, it was not accessible. The sensors were fixed to the spoon with some wire around one of their legs and were placed 1 cm from each side and somewhere in the middle, which was later changed to 1
3 from one side of the spoon. The explained setup can be seen
in figure 8.1. Later the setup was enhanced again in form of a better connection between the cables and the temperature sensors. The setup was designed in a way, so that the Pt1000 can be easily exchanged but the fixture of them is still well enough so they do not fall off. Another positive side effect was the higher distance to the spoon, which lead to a static shortcut prevention. On the other hand, as one can see in figure 3.1 the mounting is not ideal due to a wake, which is resulting from the height of a plastic mounting from at least one direction, if the wind is blowing parallel to the long side of this mounting. But for preliminary tests under static laboratory conditions the setup could be used. For this reason, after pre-calibrating the sensors and enhancing the circuit to a higher precision, the spoon was placed close to a heater, to see if a heat profile could be observed. Because of a bad heat conductivity of plastic and a too low temperature difference, there was no such profile visible.
The next version of the ESATS-prototype was a copper rod with three tempera-ture sensors mounted with kapton to it, as seen in figure 3.3. The mounting was only on the sensor itself, which turned out to be a bad idea, as in this way the sensors were holding the whole weight. In this setup, the 33 cm rod was used and the sensors were mounted at the base, the tip and 13 of the rod. The first test was done again with the heater, but later the rod was mounted with one side to a table and was heated on the mounted side with a heating gun. With this it was possible to heat the rod approximately to 30◦C and to 50◦C. As a disadvantage the temperature profile disappeared over time after a while in consequence of an even temperature distribution along the rod.
the aluminum plates the exchangeable copper rod is screwed. With the temperature sensors glued with kapton on the known distances to the rod. Therefore, the gluing was adapted to a more stable version, where the plastic cases to which the cables were attached to, were glued as well. As the heating plate was not strong enough to heat the rod to the specified temperatures, it was changed into two heating resistors with a power of 50 W each.
For the Pitot tubes there were two versions. In the early test phases, they were made out of old pens and cut and glued together again in a way where they could be used. Later these tubes were 3D printed, which resulted in a smaller and more compact design with similar properties. One of the tubes is closed, the other one is open in front, so for one tube the wind pressure is measured and for the other one the atmospheric pressure. Later a mounting for the tubes was added to the setup. The final setup of the Pitot tubes can be seen in figure 3.2 and 8.2.
This evolution lead to the final setup of the ESATS, which can be seen in figure 3.4 and explained in section 2.2.
Figure 3.1: Preliminary tests of the
Pt1000 version 2 Figure 3.2: Pitot tubes front view
Figure 3.3: Preliminary tests of the Pt1000 version 3
Measurement
In this section a short overview of the planned activities and measurement campaign is given. For all the measurements in the lab and outside at least 3600 values were taken into account, for the operational amplifier comparison 1200 values were taken into account. The measured time should be similar to 20 minutes for the 1200 values and 1 hour for the 3600 values, but due to a non exact 1 s delay between each measurement, which takes a bit longer because of the measurement and the transmission itself, the measurement time takes longer. The heating of the rod was controlled manually which means, that in some cases the heating is not exactly matching the required temperatures 2.4, but the temperature error was kept as low as possible around them.
4.1
Measurement Protocol
The campaign for measuring the temperature with the prototype consists of two main parts. First there will be a measurement in a laboratory environment with controlled settings and the second environment will be outside, with uncontrolled settings.
• Laboratory
The laboratory conditions are ambient conditions without wind. Here because of the absence of wind a stabilization in the thermal profile can occur. In this case stable conditions are defined as conditions, where the temperature change of the Pt1000 will be less than 0.5◦ in both directions within the next mea-surement. The following steps will be taken into account in this environment:
1. The OpAmps are compared to each other. 2. Calibration of the sensors against each other.
• Outside
Outside will be uncontrolled wind applied so one does not know how strong the wind is and how good the weather conditions are.
1. The influence of the Sun without heating is measured.
2. Steps 3 through 5 from the laboratory setup will be repeated with winds from different directions in dynamic conditions.
3. Data will be used to calculate the m-parameter and try to convert it to equivalent wind speed with the data gathered from the temperature and pressure sensors, ignoring IR loses.
4. The wind speeds are compared with the ones taken by a commercial weather station. The comparison is done by looking for frequencies and changes in the speed, as well as comparing the order of magnitude. • Further Information
The prototype itself will have a weight of approximately 5 kg, to compensate the leverage of the longest rod. As a heating plate in the prototype is used, only touch the ESATS if it is cooled down as otherwise you may get burned.
4.2
Laboratory Measurements
As mentioned in the measurement protocol, there were three different temperatures used to provide different measurements in the laboratory environment. The first measurement was to figure out how the system behaves with different operational amplifiers under the same conditions. The second one was to figure out what hap-pens, when one uses the same rod and places the middle temperature sensor at three different positions for n. In this case n = 2; 3; 4, which means that the middle temperature sensor was placed at 14;13 and 12 of the rod length starting from the base. The last measurement was done with the rods of different length, to figure out which one behaves best under the conditions given in the different environments.
For the first measurement the rod with a length of 50 cm was used to determine the noise level and the different effects of the noise on the measured temperature. For this the unheated system was used as an example. Therefore, one of the outputs of the temperature sensors was connected with an oscilloscope and the noise was compared to the output at the end of the circuit, shortly before the Arduino. In this way, one is able to monitor the difference in the noise level of the signal induced by the OpAmps. Furthermore, the Arduino was used to monitor the raw temperature values, which were plotted in GNU-Plot, to obtain the differences in the tempera-tures measured. There all three outputs were observed and plotted instead of only one, as it was done with the oscilloscope. But in this case there is no reference how much the different measured values are differing from the ones which were really observed by the system.
To compare these measurements, one has to compare the thermal profiles and the Tf and m-value which can be calculated with it. With help of these values it is possible to see if the static laboratory conditions are fulfilled or if the results are differing too much from the expected values. This setup was done with the 33 cm rod and gave no clear result beside a changing temperature profile, depending on the location where the temperature sensor was placed. For this reason, it was decided, that the placement of the middle sensor will be similar to the placement of HABIT and REMS.
The third measurement is used to observe how different rods are changing the precision and results of the measurement under stable conditions. The longer the rod is, the closer the temperature at the tip is to the temperature in infinity, which is equal to the ambient temperature Tf. So for this measurement one wants to figure out which length is necessary and which is too much, because the length of the rod is too long and so the last part is not heated anymore. Therefore, the temperature at the tip is equal to the ambient temperature. The measurement plan for this type of measurement is to use all different rods with three temperature sensors each. These sensors are mounted at the best position detected with the test before and will use the best OpAmps, so that the noise level is as low as possible. For each of the different rods a temperature profile of room temperature, 10◦C and 20◦C will be used.
4.3
Outdoors Measurements
For the outside measurement there are two tests planned. The first one is to measure the temperature profile on a unheated rod and to see the difference in measuring in a shadowed area and in the Sun. For this, the rod will be placed in the shadow and due to cloud movement and the movement of the Sun across the sky the settings will change steadily. With this type of profile one is able to observe, if the prototype will be able to detect the wind speed, even if it is in the Sun or if it is heating up too quickly. For the case of an influence due to the Sun it has to be decided, if all continuing measurements have to be done in the shadow, in the Sun or if the thermal load needs to be increased.
Data Analysis
5.1
Data Formats
The data of the measurements on Earth have to be stored in a specified way to make them easily accessible and to be able to see the differences between the various sets of data. For this a special naming and file format was introduced. The file format used is oriented at the Planetary Data System (PDS) file structure, with some adaptations to differentiate each file from one another, which can be found mainly in the name of the files.
5.1.1
File Names
shows when it was calibrated the last time. This can be avoided by recalibrating the system with every start of the software by the system itself. At the end the most important part is shown: the information about what kind of test was run. The word ’static’ indicates that the test was run in laboratory under static conditions whereas ’outside’ shows that the test took place outside. Outside one has to bear in mind that the relatively nearby ground reflects the heat. Two examples of this naming regulation can be seen in figure 8.3.
5.1.2
Raw-Data-File
Earth Data:The raw-data-file consists of eight columns with multiple lines. In the first column the time in seconds from the start of the Arduino is written. This means it would be possible to leave it running continuously without any interruption and measure only every nth iteration for m seconds. But for the measurement needed, it is not necessary. The second column includes the time in the normal local time frame on Earth in the standard 24h format HH:MM:SS, which is not working if one is using the setup with the SD-Card-Shield, as a continuously running clock is not included in the prototype. The next three columns are including the temperature of the three measurement points along the rod. The order is Tb, the base temperature, Tln, the temperature at point n, and Ta, the temperature at the tip of the rod. These temperatures are given in ◦C. Continuing the format with the two pressure values from the Pitot tubes, which are given in hPa. The first one is the stable one, the second one is the one which is influenced by wind. The last column of this data format is filled with the ambient temperature values of one of the temperature sensors, which is mounted on one of the pressure sensors. This value is again given in ◦C.
This file corresponds to the environmental data file of the PDS, where the un-processed data in SI-units is stored. If the measurements of this prototype has to be enhanced or the order in the output file is changed, the calculations in the processing part of the program will lead to different results, which will not be reasonable. Planetary Data System (PDS) for Martian Data:
The PDS is an archive for science data provided from NASA for planetary mis-sions, which is globally used by scientists. This type of archive provides access via standardized data sets, distributed in different fields of science. Which are over-watched by scientists, who are specialists in the corresponding field all across the US. There are two standard formats which are provided. One is the ”PDS Stan-dards Reference”[24] and the other is the ”Planetary Science Data Dictionary”[25]. Both of them are defining the data allowed to be labeled as ”Conform to PDS Stan-dards” and ”funded by PDS”. The standardization was introduced to store the data centrally as a long term archive file system, which is providing access to future scientists. Therefore, it is important to provide the data in a simple way with al-ready processed data, so that even without knowing the experiment it is possible to interpret the provided data and process it further.
