- a light bulb connected to a VARIAC to vary the light intensity
- the potential divider circuit I had devised
- the portable lux-meter
They were arranged as shown below:
There were many problems with the first calibration attempt and they are listed below along with the chosen correction for the second calibration attempt:
-
The greatest problem to overcome was the quality of the lux-meter. The first calibration attempt used a portable lux meter which was not that accurate, meaning that the first calibration curve was not at all linear. To fix this problem an Xlogger lux meter was used, connected to a laptop to record the results in real time. This meant that not only were the results more accurate, but they were also plotted into Microsoft Excel.
- The second problem was the placement of the LDR in relation to the lux meter. They were placed next to each other, but the distance from the lightbulb wasn’t fixed which meant that results could vary. The second calibration design incorporated the LDR and the lux meter into the top of the cardboard box where they were sealed with gaffer tape to ensure no ambient light leaked in.
-
The cardboard box itself was a problem as it had gaps in its structure. These were covered up with gaffer tape to ensure the most reliable results. Another precaution was that light may leak in around the base of the cardboard box. To overcome this problem I would have had to tape the box down to the workbench which would have been problematic when changing the bulb. I instead recorded the lux of the highest and lowest voltages across the bulb and found that there was no difference in the recorded light intensity regardless of whether the ambient lights were on or off. This meant that there would be no need to tape the box to the workbench.
Second Calibration:
The second method again consisted of three main circuits:
- The potential divider circuit
- The bulb and the VARIAC
- The Xlogger Lux Meter connected to the laptop
They were arranged in the following setup:
This improved method incorporates the new Xlogger lux meter which is integrated in the top of the cardboard box, right next to the LDR. This means that they will be both recording the same light intensity thus resulting in the most accurate calibration possible. The gaps in the cardboard box have been sealed and the bulb stand is also secured to the desk to prevent movement and a bias towards either sensor. To ensure a fair test, the same VARIAC was used for all experiments as actual figure of voltage through the light bulb may depend on the accuracy of the VARIAC. Of course this is not a problem so long as the same apparatus is used each time.
The above graph shows a much more regular set of results when compared with the first calibration. This is excellent as it ensures that when the real experiment is done, the light intensity figures collected by the LDR will be as close as possible to the true values. If the graph was more curved, then any values recorded at the lower light intensity (voltages) would be less reliable because the difference between the potential differences would be too small to reliably calculate lux readings.
All the points lie on the curve which increases the reliability of any recordings in the final experiment. This is due to the revised method changes such as the more accurate Xlogger Lux meter and the placement of the LDR and Lux meter in relation to the bulb. If the plot points were outside the curve, it would be difficult to be certain of light intensity values recorded after the lux meter was removed during the final experiment.
When using the revised method for the final experiment it is important to be weary of the temperature of the bulb, especially the non-energy efficient ones. To overcome this problem heat protective gloves will be used when changing the bulb.
Using the Calibrated Sensor:
The sensor was used to collect recordings of the light intensity of various light bulbs as different voltages. The apparatus used for the final experiment was identical to the equipment used during the second calibration except for the lux meter no longer being connected to the laptop. Note that the lux meter was not removed from the apparatus as this might affect the performance of the LDR.
The recordings were taken across three different light bulbs:
- A 100W normal filament light bulb across a range of voltages ranging from 80-100V with increments of 10 volts.
- A 100W energy efficient light bulb across the same range of voltages
A 60W normal filament bulb again across the same voltage
The light bulbs will be subjected to each voltage with recordings of the voltage across the LDR being taken at each point. The recorded LDR voltages will then be compared to the calibration curve to work out the lux figure.
Results:
The 60W filament bulb produced an error when trying to calculate the lux in comparison to the output voltage. This was due to the second calibration graph not allowing for output voltages below 0.85V. This meant that it was impossible to convert the output voltage of 0.63V into a lux reading. If I was to repeat the experiment, I would make sure that I had a calibration curve that had a range of values that exceeded the range of the output voltages in the final experiment.
The results proved a high level of certainty in the recordings taken:
- The overall trend of the two filament light bulbs is fairly linear which is what is to be expected as there is no optimum voltage to operate at. However the 100W energy efficient bulb has a slight curve showing a relatively lower intensity of light at a lower voltage compared with higher voltages. This is to be expected of an energy efficient bulb as they do have an optimum voltage operation point. Upon looking at the graph this appears to be at approximately, to be sure more frequent recordings would be necessary, 100 volts. This is close to the 120 volt power supply that exists in UK households and therefore is to be the performance expected of the bulb.
- Looking at the two graphs it is possible to see that the bulb that emitted the highest lux level across the range of potential difference values across the light bulb was the 100W energy efficient bulb. This was to be expected as it requires less energy to deliver the same amount of light as a 100W filament bulb. This can be seen when looking at the graph as the 100W filament bulb produced lower lux values for the same potential difference values across the light bulb. The 60W filament bulb produced the lowest lux readings for the range of potential differences across the light bulb. This again was to be expected as it needs the most energy to produce a lux figure when compared with the two other bulbs. The fact that the graph matches the science of the bulbs reinforces reliability in the results.
However there were also some levels of uncertainty:
- The calibration curve was drawn up as a result of 7 readings taken using one bulb. More recordings taken during calibration would increase the accuracy of the calibration curve and therefore the certainty of the final results.
- When calibrating the sensor, the Xlogger lux meter was difficult to use as the light intensity appeared to fluctuate with a degree of 40-50 lux either side of a modal value. This meant it was difficult to record what the light intensity was when compared to the sensor potential difference. This could have been corrected by using multiple Xlogger lux meters to generate an average, more accurate reading.
- Readings were only taken once every 10 volts across the bulb. Both with the calibration curve and the curves on the final results graph would have been more reliable had there been more readings across the same voltage spectrum. For example if there had been more readings on the 100W energy efficient bulb, it would be possible to deduce whether the curve at the bottom of the trend line was due to the nature of the bulb or a problem with the recordings.
- The lowest value for the fixed resistor was 250 Ω. This was because it was the lowest value that the multiple resistor box I was using could offer. There could have been a resistor with a lower resistance that would have been better suited to the experiment however this was not investigated. However, the difference between the potential difference value across the fixed resistor at high and low light intensities was high enough to draw up a calibration curve.
Sensitivity:
The sensitivity of my sensor can be worked out by investigating the equation for a linear line of best fit through the calibration curve points. I worked out the equation to be y = 1477x - 1573. This would mean that my sensitivity would be 1477 lux per volt across my fixed resistor. However by using a linear line of best fit, you cannot take an accurate reading for each voltage because the line represents an average.
Behind the Scenes Physics:
The Lightbulb side:
In a filament bulb as the potential difference across the filament increases, the atoms in the thin wire become increasingly excited. This increased excitement causes collisions between atoms which causes an electron to jump to a higher energy level. The electron falls back to its original energy level thus releasing the extra energy in the form of a light photon. The more often the electron moves out of orbit the more photons are released every second thus varying the light intensity of the bulb. The longer this goes on for, the less economical the reaction. This is due to the atoms vibrating and releasing not only photons but also heat. This is where the energy saving bulb comes in. The special dimmable energy saving bulbs still generate light by passing current through mercury vapor to produce light. However they also incorporate a ballast: a silicone chip that regulates the current through the vapor allowing it to be dimmable. The fact that no energy is wasted on heating up the bulb means that more is used to actually generate light which is what makes energy saving bulbs more efficient than their filament counterparts. The results show this: the 100W filament bulb generated the highest lux reading; 2394 lux compared to the 100W filament bulb’s 1390 lux and the 60W filament bulb’s 1295 lux.
The Potential Divider side:
As explained at the beginning a potential divider circuit works by using a variable resistor (the centre of the sensor) in conjunction with a fixed resistor along with a power supply. By measuring the potential difference across the fixed resistor, the LDR, a sensor reading can be recorded. The LDR achieves this by being a semi-conductor, that is to say that it is a material with a conduction rating between that of a conductor an an insulator. The atoms inside the LDR have bound electrons. These electrons require different amount of energy to escape from the atom and pass a current through the semiconductor. Light, which is made up of packets of photons, determines how many of these electrons are free to flow and pass a current. Therefore the higher the light intensity, the lower the resistance of the LDR and thus the lower the potential difference across the LDR.
Limitations to the method:
The method can be used to measure only a certain range of light intensities for which a) the LDR is calibrated for and b) the range that the LDR can accurately interpret. This is due to the nature of the calibration curve: there is only a specific range range of input potential differences for which the LDR can operate when used with a specific fixed resistor. This is why the experiment only measured light intensity between 80 and 140 volts because outside of that spectrum, the ratio between the 250Ω fixed resistor and the LDR is too small or too great to give an accurate reading. When presented on a graph it is at the extremities: when the gradient is close to 0 and where the gradient is close to 1.