Mashavu Thermometer Project

[Laura Arnold]

Contact Information

University and Department: Penn State, Department of Bioengineering

Team Members: Lisa Godiska, James Lata, Michael Mohns, Corey Latanision, Marty Linderberger, Bill Klinger

Project Information

Describe the challenge your project is trying to solve.

The goal of the BIOE 401 Thermometer Design team is to design and develop a thermometer for the Mashavu project in Kenya.  The thermometer will be designed to meet the specific needs of the Kenyan culture and economy.  The themometer must accurately work and interface with a computer via a USB, requiring little work to operate from the user.  The overall goal is for the device to be successfully implemented in the Mashavu project, and accepted in the Kenyan community.

Body temperature is a key component in a medical analysis.  Just one simple temperature measurement can provide a lot of important information about a patient.  The normal body temperature is approximately 37 degrees Celsius and as temperatures rise and fall above and below this value it may indicate medical problems.  For instance, body temperatures above 38 degrees Celsius require immediate intervention and could indicate an infection in the body.  Temperatures lower than 36 degrees Celsius are also considered dangerous and if measured on a patient require urgent attention.  As illustrated, in the medical field,  temperature is important for assessing the health conditions of a patient.

The key focuses of this project are as follows:

Context: Understanding the Kenyan culture and learning how to design to meet their specific needs.

Development: The use of engineering design tools to design the product.

Analysis: Testing to assure robustness of design.

Implementation: Implementing the device into the Mashavu project in Kenya

Describe how you addressed the challenge through your project.

Several design alternatives were evaluated in order to determine the most appropriate thermometer design for the Mashavu project.  Table 1, shows the different design ideas, including a thermistor, a thermocouple, an alcohol thermometer, and external temperature strips.  Each of the thermometer design ideas were evaluated based on how well they met the customer needs specific to the Kenyan people.  The table shows the weight, or importance, out of 100% of each customer need and then ranks each design, 0 being lowest to 4 being highest, on how well the design meets the customer need.  The design that scored the highest was the thermistor. The thermocouple, which utilizies a thermocouple to measure temperature change, and the external temperature, which are strips that change color upon temperature change, proved to be too inaccurate for body temperature measurement and the alcohol thermometer, which uses the expansion of alcohol to measure temperature, was too hard to sanitize and replace. A thermometer utilizing a thermistor temperature sensor was the most robust, inexpensive, and easy to use and will be further developed and modified.

Table 1: Pugh concept table for the thermometer design alternatives

Process Overview

The thermometer design is based upon a thermistor.  In order to meet the inexpensive cost presented by the customer needs the thermistor is enclosed in a water balloon and the leads of the thermistor are encased in the barrel of a pen.  The leads of the device are attached to wires which are connected to a quarter wheat stone bridge with resistances of 1 kohm across the elements.  The circuit is then connected to a DAQ device which is connected via USB to a computer.  An input voltage from the computer is supplied to the circuit in order to measure the change in resistance of the thermistor with varying temperatures.  This temperature change is measured as an output voltage and using the Steinhart-Hart equation this output voltage can be computed to a measured temperature.

Design Concept

The design uses generic and cheap materials so that components can be easily attained in Kenya and the design is cost effective costing approximately two dollars.  The design will produce accurate results as seen in testing analysis with over 99% accuracy and can measure small changes (0.1 degree Celsius) in temperature.  Not only will the device function properly but it is also extremely easy to use requiring only the click of a button to obtain an accurate temperature measurement.  Customer needs also focused on sanitation.  This aspect of the design is more related to the protocol for the use of the product.  In order to meet this need the device will measure body temperature through the armpit.  With the high incidence of diseases and viruses in Africa, this type of measurement will provide the most accurate and sanitary temperature measurement.

Measurement Protocol

The protocol for the thermometer demands little from both the caregiver and patient.  The patient must put the device in their underarm and place the hand of the arm which has the thermometer on their chest so that the thermometer is tightly held in the underarm.  The caregiver then must press one button on the labview program to initiate temperature measurement.  After 120 seconds the measurement is complete and the computer outputs information indicating the measurement is over and the device can be removed.  No further sanitation or work is required.

The thermometer utilizes a thermistor to measure temperature change.  The thermistor is modeled in Comsol with insulated leads.

The thermistor encasing was made out of cheap materials including a bic pen cover and a water balloon.

The thermometer is used to measure armpit temperature.

Evaluating the final design:

The final design combines the temperature measuring accuracy and cost efficiency of a thermistor with the power of labview interfacing technology to create a thermometer for implementation in the Mashavu project. 

The device temperature properties were measured using COMSOL. The software allowed for the design of proper geometries and thermal and material properties.  Using a transient analysis in COMSOL, this model generated sound-thermal results at a time of 120 seconds, thereby validating the 2 minute measurement time needed by specifications.

Figure 1: Temperature Plot of Thermistor in Comsol

Figure 2: Temperature vs. Time at Point (-.05, 0) as Seen in Figure 1

Figure 3: Comsol Temperature Plot in Kelvin at t =120s

The plastic pen housing (A) was modeled as polypropylene, with an outer boundary temperature of 310°K (98.6°F) and an inner boundary temperature of 302.44°K (85°F) representing ambient temperature. The leads (B) were excluded from the temperature analysis due to minimal thermal dependencies.  The press fit (D) between the leads and the pen tightly secure the thermistor in place and does not allow for thermistor movement. The rubber-latex balloon (C) encapsulating the device was modeled as a “soft commercial rubber” with a boundary condition temperature of 310°K (98.6°F). Finally, given thermistors are manufactured with powdered transition metal-oxides such as manganese, nickel, cobalt, and copper oxides, the thermistor (E) was modeled as a copper. Boundary temperatures were set at 302.44°K (85°F) ₅ representing air conduction at the boundary and 310°K (98.6°F) where the balloon was in contact with the thermistor. Below in table 2 the thermal and material properties of each component in the device used in the Comsol model are given.

Table 2: Properties

Material	Component	Density (kg/m3)	Thermal Conductivity (K) (W/m-K)	Heat Capacity (Cp) (J/kg-K)
Polypropylene	Plastic Pen	.09051	.122	19253
Soft-rubber	Balloon	11002	.22	20104
Copper-Oxide	Thermistor	87006	60	3856

Construction

Circuit Construction (Quarter Wheatstone Bridge)

•     Consists of: :
  

A.     Pen tube

B.     Resistor (1 kohm resistors)

C.     Sensor (thermistor)

D.     Balloon

E.     Wires

F.     Removable clip with prongs

G.    Cellophane

H.    Circuit board

•     The circuit was connected together by soldering the components B,E and F onto the circuit board
  

•      The clip allows for the sensor to be removed from circuit board therefore increasing safety and usability
  
  

Sensor

•     Thermistor soldered together with two long, threaded wires
  
  
  

•      Exposed wires were covered by shrink heating a tube
  

•      End of wires were soldered into a clip that attaches to the DAQ device
  

•      The thermistor and wires are threaded through a BIC pen tubing
  

-  Allows for easy use

-  Protects the wires

•      The thermistor is covered with a balloon and plastic wrap that provides extra protection from the environment
  

Calibration Testing

Step 1: The designed thermometer was connected to a multimeter.

Step 2: A ReliOn thermometer (model number 144-691-000) and the designed thermometer were simultaneously placed into warm water (temperatures shown below) for two minutes.

Step 3: After two minutes a resistance value was obtained from the multimeter.

Step 4: The two thermometers were then taken out of the water and set aside for future testing.

Note: The two thermometers were tested at the three temperatures important for medical analysis, above body temperature (39.39 ^oC), below body temperature (34.22 ^oC) and at body temperature (36.31 ^oC)

The temperature values, in Celcius, and their respective resistance values, in ohms, were then plugged into the Steinhart-Hart equation.  This equation input into LabVIEW allows the changing resistances to be measured by the designed thermometer.

SteinHart-hart Equation

Table 1: Temperature and resistances used in calibration of thermistor

	Calibration  1	Calibration 2	Calibration 3
Temperature (oC)	39.39	36.61	34.22
Resistance (Ohms)	616.1	678.31	746.06

Figure 1: Thermistor Steinhart-Hart Curve

Experimental Testing

Accuracy

• Multiple temperature measurements at random temperatures were gathered with the device, and compared with a reference temperature acquired from a digital thermometer
  

• The device was used first, immediately followed by a digital thermometer reading
  

• Percent error between the device and reference temperature was calculated using the equation below
  

• Average accuracy of the device was determined by calculating the mean of all individual percent errors, and equals 0.18%
  
  

Percent Error = |reference temperature - experimental temperature| x 100

       reference temperature

Precision

• Multiple temperature readings from one patient were acquired using the device
  

• A three minute wait period was implemented between each reading to provide the thermistor with identical idle times between each measurement.  The initial measurement was thrown out due to the lack of the three minute wait period beforehand.
  

• The average value of all temperatures was calculated
  

• Percent error between each individual reading and the average value was computed
  

• Average Precision of the device was determined by calculating the average of all individual percent errors, and equals 0.31%
  
  

The results of the experimentation show a small margin of error, less than the desired 1 percent.  In comparison with commercially available digital thermometers our device tested very well in both accuracy and precision.

Figure 2: Percent Accuracy of Prototype used on Sixteen Subjects

Figure 3: Percent Precision of Prototype and Reference Thermometer used on Sixteen Subjects

Conclusion:
Thermistors are extremely robust, with less than 1% error. This is necessary in order to measure the small deviations in body temperature which correlate to illness.  Additionally, the average Kenyan makes under one dollar (1 USD) a day, so the price goal needed to low; it was set at three dollars.  The thermometer also implements a cheap encasing, consisting of a water balloon wrapped in cellophane and the body of a Bic pen allowing the price goal of three dollars to be maintained.  Although cost and robustness were key features of the design, with a thirty percent incidence of AIDS in Kenya, sanitation was also a design imperative. Originally, oral and rectal points of measurement were considered, however these areas can be a point of transmission for diseases. Therefore, underarm temperature measurement was selected to minimize the contamination of the device and need for sanitation.  This point of measurement is also easy to use, and does not offer any discomfort to the patient.  The device/computer interface also allows ease of use by requiring only the press of a button to initiate and measure body temperature. The final design combines the specific customer needs of the Kenyan culture with the robustness of a thermistor temperature sensor and is appropriate for implementation in the Mashavu project in Kenya.