For my sensor project coursework I will be investigating a thermistor.
Gardeners in Britain use greenhouses so that they can grow their plants in a warmer environment than that outside the greenhouse. This is to try to protect the plants from frost and animals, whilst the seeds germinate and grow, and plants are stored. After this period, plants can be put into the garden. There can be problems with greenhouses though. One example of a problem would be in the height of summer when an unventilated greenhouse can reach internal temperatures that start to kill the plants. Another problem can arise in late autumn, winter and early spring when the internal temperature of the greenhouse will decrease, also causing the photosynthesis reaction undertaken by plants to slow down, and even stop causing plants to die.
The specification that they have given me is as follows:
The temperature sensor should be able to detect when the temperature drops below 20°C and when the temperature rises above 30°C.
The optimum temperature for growing plants is about 25°C. The company have a series of heaters in the greenhouses and underneath the soil, which could be used to raise the temperature within the greenhouses, when the growing conditions become too cold. The greenhouses that the company use also have windows that could be opened when the temperature inside the greenhouse becomes too hot.
Ways to measure temperature.
Some ways to measure temperature include;
A thermocouple
A thermistor
Wheatstone bridge
A Resistance Temperature Detector (RTD)
A thermocouple uses two pieces of dissimilar metals that have a contact potential between them, and this contact potential changes as the temperature changes. The problem with using a thermocouple is that the output voltage is typically a few micro volts per °C. Normally, the output of thermocouples is amplified using an operational amplifier (op-amp) so that the output voltage is of a more useful magnitude. An application that a thermistor could be used for is to help the temperature compensation of a full Wheatstone bridge. The output voltage of a Wheatstone bridge is so small though, that it too needs amplifying.
Due to the need for amplification, I shall not use a thermocouple or Wheatstone bridge to detect the temperature although they are both viable solutions. This is because there is more to go wrong in a circuit containing an amplifier, than one without.
Resistance Temperature Detectors (RTDs) are usually wire wound components, but can be made from a thin film as well. They work on the principal that when wire becomes hotter, its resistance increases. I could use a RTD in my sensor, but I do not have easy access to them so it is unlikely that I will use one.
Instead, I shall be using a negative temperature coefficient (NTC) thermistor. The name thermistor is combined from the phrase thermally sensitive resistor. The resistance of an NTC thermistor decreases as the temperature increases, in a disproportional manner. I will use a thermistor because thermistors are reliable and cheap. They also only require a small amount of circuitry.
Potential dividers
Potential dividers, as the name suggests, divides the potential in a circuit. This is done using two or more resistors in series and works upon the principal that the current across both of the resistors is the same due to Kirchhoff's law, so the potential difference across a component must be directly proportional to the resistance of the component.
Voltage = Resistance x Current
Therefore; Voltage is directly proportional to Resistance.
Using this principal, the potential difference across a component can be 'tapped off', thus giving a divided proportion of the total potential difference. The circuit used is shown below.
Another way this circuit can be used is to sense a change in the surrounding environment. This can mean substituting one of the fixed resistors for a form of variable resistor. I will be using a thermistor, a component that's resistance will change according to the temperature of the surrounding environment.
I shall ...
This is a preview of the whole essay
Voltage = Resistance x Current
Therefore; Voltage is directly proportional to Resistance.
Using this principal, the potential difference across a component can be 'tapped off', thus giving a divided proportion of the total potential difference. The circuit used is shown below.
Another way this circuit can be used is to sense a change in the surrounding environment. This can mean substituting one of the fixed resistors for a form of variable resistor. I will be using a thermistor, a component that's resistance will change according to the temperature of the surrounding environment.
I shall use this circuit in my sensor because I think it is the one that will give good results when it is placed in the greenhouse. This circuit is probably also the cheapest to make and would only take up a small amount of space. I am going to perform a series of experiments using this circuit, to get a more accurate idea of how the sensor works and what results the setup will give.
To work out the output of a potential divider you use the formula:
Vout = Vin x R2
R1 + R2
This is the apparatus that I will use:
One lab power pack of 5v
One thermistor (resistance to be decided)
One variable resistor (resistance to be decided)
One glass thermometer 0°C to 100°C
One 200 cm3 beaker
A voltmeter/ohmmeter
Two crocodile clips
Five banana-plug wires
One kettle
Ice cubes
Diagram of setup.
When I am deciding what fixed resistor to use in my sensor I will think about the factors listed below. I have decided that I shall use a 1000? thermistor. This is because they are fairly common, and should the thermistor fail it is easily interchangeable.
Resolution
The resolution of a sensor is basically how small a change in the environment the sensor can detect. My sensor should be able to detect a 0.1°C change in temperature so that it can accurately tell the difference between 19°C and 20°C. The voltmeter I am using to take my results is digital and can read the potential difference to two decimal places. The voltmeter could be able to detect a potential difference change to three decimal places, but I think that the sensor would be more sensitive than necessary for the application that I am designing it for. I do not think that the resolution of my sensor will be affected by the resolution of the voltmeter.
Sensitivity
The sensitivity of a sensor will depend upon what fixed resistance that is used in the potential divider and what the resistance of the thermistor is, in proportion to the fixed resistor. I have worked out that when I use a 1000? thermistor (value of resistance of thermistor at 25°C) the fixed resistance that should give me the greatest range of potential difference output is as follows:
I know that the resistance of the thermistor that I am using at 20°C is 1200 ? and the resistance at 30 °C is 800?
Value of Fixed
Output Voltage at
Output Voltage at
Difference Between Output
Resistance in ?
20°C (5v input)
30°C (5v input)
Voltages at 30°C and 20°C
500
3.559
3.054
0.505
600
3.367
2.854
0.513
700
3.151
2.833
0.318
800
3.000
2.500
0.5
900
2.751
2.350
0.401
000
2.645
2.122
0.523
100
2.521
2.111
0.41
200
2.500
2.000
0.5
300
2.400
.910
0.49
400
2.315
.836
0.479
500
2.244
.778
0.466
(See Table 1)
The fixed resistance that gives me the greatest range of results is 1000 ?. Interestingly this happens to be the same resistance as the resistance of the thermistor at 25 °C
Response time
The response time of my sensor depends upon how quickly the thermistor reacts to a change in the temperature of the surrounding environment. The response time of my sensor does not need to be particularly fast under or around ten seconds would be sufficient. This is mainly because the plants in the greenhouse will not be affected be a short period of time at a slightly cooler temperature than the optimum growing temperature of 25 °C. The thermistor that I am using will be able to detect the temperature in one second or less, a period of time that will not affect the plants.
Random error
There is a chance that a random error could occur in my sensor. This however is a small chance, mainly due to the fact that the sensor will be working permanently, rather than taking results every hour for example.
Systematic errors
There is a possibility that a systematic error could occur in my sensor, these are mainly down to instances such as zero errors, and usually all of the results are affected. In my sensor, temperature could change the resistance of the fixed resistor in the potential divider. I do not think this should be a problem in my sensor though because the sensor will not become hot enough for this to be a factor.
Now that I have taken these things into consideration I am going to perform an experiment to find out what the output potential difference will give at certain temperatures. To get the different temperatures, I will place the thermistor into a beaker of water. Due to water having a high specific heat capacity, it cools down quite slowly. As a result of this, I will put my thermistor into a beaker of water. I will read the potential difference across the thermistor at different temperatures of water, starting at 50°C letting the water cool and taking readings every 5°C. When the water temperature reaches about room temperature, I will put some ice cubes into the beaker of water to lower the temperature of the water further. The potential difference output from the lab pack transformer will be 5v.
The results that I have collected are as follows:
Temperature in °C
Potential Difference in Volts
5
3.90
0
3.67
5
3.15
20
2.89
25
2.73
30
2.30
35
2.00
40
.87
45
.65
50
.45
The specification that I was given by the plant growing company stated that the sensor should be able to tell when the temperature drops below 20°C and when the temperature rises above 30°C. So I have taken results every 1°C from 18°C to 32°C so that I can give a more accurate potential difference output at 20°C and 30°C:
Temperature in °C
Potential Difference in Volts
8
3.20
9
3.16
20
3.09
21
2.90
22
2.75
23
2.53
24
2.46
25
2.33
26
2.29
27
2.27
28
2.25
29
2.21
30
2.10
31
2.07
32
2.01
This graph shows me what potential difference the potential divider will give out when the temperature output is 20°C and 30°C, information that could be used to calibrate a circuit that can open and close the windows, and turn the heaters on inside the greenhouse.
Although the potential difference output from the lab pack fluctuated, I corrected this as best as I could before every reading was taken. The potential difference was correct to one hundred millivolts of the total output. The 1000 ohm variable resistor that I used actually was 996 ohms when I measured it after the experiment. This shouldn't affect my results drastically though, because I used the same variable resistor throughout my experiment and the percentage error is only 0.4% which is quite small overall, especially given that the percentage error (tolerance) of the resistor that I used could have been up to 5%, according to the manufacturer. When I look at a graph of my results, and draw a line of best fit, all of the points fit quite well, if there are any anomalous results, they would be the results for 10°C and 25°C. I think that this is a human error.
When I took the results, I waited for the temperature of the water drop until it was the temperature that I wanted, then I adjusted that the potential difference across the lab pack, if necessary so that it was 5v, to provide a fair set of results. Next I took my reading of the potential difference across the thermistor. I suspect that if I had to correct the potential difference of the lab pack, the temperature of the water could have dropped by roughly about 1°C by the time I came to read the potential difference across the thermistor. I think that this could be the source of any anomalous results, such as the results for 10°C and 25°C. Another reason for there being anomalous results is that the sensor unit didn't have enough time to stabilise before I took my reading. This could be rectified by placing the thermistor in a bag, inside a water bath and set the temperature. This would then remain constant, and the sensor could be left for an hour or so to stabilise. This would give me a set of results that are more accurate but would take a lot longer to achieve.
I would also propose that there is a way of calibrating the circuit that my sensor controls, so that a lower temperature and upper temperature could be fine tuned in the greenhouse that the sensor will be placed. For example, the specification from the company could change so that the lowest temperature that the greenhouse is allowed to reach is 21°C, and the highest 31°C, with a slight adjustment to the circuit controlling the windows and heaters, the upper and lower potential difference boundaries of activation, which the control circuit runs from, could be changed. This would mean using a graph for the sensor showing what the potential difference across the thermistor is at certain temperatures.
Now that I have performed various calculations and an experiment, I can give the sensor and its details about the sensor the plant growing company. When they have their control circuit, they can use graph number two to calibrate at what potential difference output (from the sensor) the windows open and the heaters turn on. Should the company decide to change their specification by ±2 or 3°C, they should refer to graph number two as before.
Table 1: This is how I found out what fixed resistor would give the greatest range of potential difference output across the thermistor.
R1 in ?
R2 in ?
Rt in ?
R2/Rt
Input voltage
Output Voltage at 20°C
500
200
700
0.3559
5.0
3.559
600
200
800
0.3367
5.0
3.367
700
200
900
0.3151
5.0
3.151
800
200
2000
0.3000
5.0
3.000
900
200
2100
0.2751
5.0
2.751
000
200
2200
0.2645
5.0
2.645
100
200
2300
0.2521
5.0
2.521
200
200
2400
0.2500
5.0
2.500
300
200
2500
0.2400
5.0
2.400
400
200
2600
0.2315
5.0
2.315
500
200
2700
0.2244
5.0
2.244
R1 in ?
R2 in ?
Rt in ?
R2/Rt
Input voltage
Output Voltage at 30°C
500
800
300
0.3054
5.0
3.054
600
800
400
0.2854
5.0
2.854
700
800
500
0.2833
5.0
2.833
800
800
600
0.2500
5.0
2.500
900
800
700
0.2350
5.0
2.350
000
800
800
0.2122
5.0
2.122
100
800
900
0.2111
5.0
2.111
200
800
2000
0.2000
5.0
2.000
300
800
2100
0.1910
5.0
.910
400
800
2200
0.1836
5.0
.836
500
800
2300
0.1778
5.0
.778
Safety whilst performing the experiment
When I perform the experiment to find what the output of the potential divider across the thermistor is, these are some things that I should be aware of.
I must be careful when I use the hot water from the kettle, the water could be hot enough, if it is boiling to burn or scald me or another student. For this reason, when I use the kettle to heat the water I shall try to plug it in as close to my experiment as possible so that I don't have to carry the kettle around and risk spilling some of the hot water.
As with any experiment I should also be very careful given that I am using water and electricity. The sensor circuit will only run on 5v and therefore shouldn't be an electrocution hazard if it comes into contact with water, nonetheless there will be a short circuit. The lab pack is more of an electrocution risk because it has a 230v input, so this should be kept away from the beaker of water which will also be kept on a paper towel to try to absorb spillages.
This is a graph showing all of the results that I have obtained showing the more temperature range that I have explored further.