Experiment #3: Comparison of different tin foil sensors
Kitchen tin foil comes in different qualities. The purpose of this experiment is to assess whether there is any significant difference in between two different qualities of kitchen tin foil when used as touch sensors.
Sensors
Two sensors were created. Each has the same dimensions and means of connection to the circuit. The only difference is the type of tin foil. The sensors are labelled “Thin Foil” and “Strong Foil”. “Thin Foil” is ASDA Kitchen Foil while “Strong Foil” Tesco Strong Kitchen Foil (UK supermarket brands).
Each sensor uses a 6×4cm piece of foil glued to a 8×5cm piece of thin corrugated card. A 4cm strip of 5mm copper tape is glued to the top edge of the foil. Bare wire was stripped from a piece of jumper wire. The wire was cut into 22mm lengths and bent at right angles 6mm from the end. The longer (16mm) section was soldered to the copper wire. The 6mm end acts as a male connection that can accept a female DuPont cable connector.
The resistance between the corner of the tin foil furthest from the connector and the male end of an attached Dupont cables was measured as:
- Thin foil: 1.3Ω
- Strong foil: 1.9Ω
The difference is considered to be negligible given that a 1kΩ resistor in wired in series with the sensors.
Circuit
The circuit used was the same as that used in experiment 2, except that the lead to the sensor can be plugged into either sensor as required.
The circuit was as follows:
Once again, the Uno was supplied with 5V from the PC via it's USB port. The switch is not physically present but is implied: switching between sensors is effected by unplugging one sensor and plugging in the other. The breadboard and other components were connected in a logically similar way to the following diagram:
Pin 2 of the Uno is configured as an output pin and is connected, via a 1MΩ resistor to input pin 3 and the 1kΩ resistor that is there to protect the microcontroller from possible static discharge from a finger touching the foil.
Methodology
The SingleSensorRawDataLogger
was again used, but this time with the number of columns set to one so that a single reading is reported per line.
The Thin Foil sensor was connected to the circuit. The code was altered to use 25 samples per reading and was uploaded to the Uno. The sensor was tapped three times and serial output was captured using CoolTerm which was configured to write the data to a text file on disk. The test was run twice more, with the code modified to 50 and 100 samples per reading respectively, each time writing a new text file.
The Strong Foil sensor was now connected to the circuit in place of the Thin Foil sensor. The same tests at 25, 50 and 100 samples were run and output captured once again in three text files.
Results
The data from the text files were imported into a Libre Office Calc spreadsheet and the results for each sensor were graphed against each other. The graphs were exported as image files. They can be seen in the image carousel below. There is one graph for each sample size.
The following sensor readings are given in the table below:
- Maximum reading
- Minimum reading
- Average of readings ≥ 500 (an arbitrary value above which a key is deemed to be pressed)
- Average of readings < 500 (key deemed to not not pressed)
25 Samples | 50 Samples | 100 Samples | ||||
---|---|---|---|---|---|---|
Thin | Strong | Thin | Strong | Thin | Strong | |
Max | 1,594 | 1,901 | 3,239 | 3,802 | 6,554 | 7,648 |
Min | 2 | 2 | 0 | 0 | 0 | 0 |
Avg ≥ 500 | 1,549 | 1,842 | 3,132 | 3,624 | 6,094 | 7,283 |
Avg < 500 | 8 | 12 | 14 | 16 | 20 | 27 |
The final table gives the percentage difference between the values for the Thin sensor and those for the Strong sensor in the above table.
25 Samples | 50 Samples | 100 Samples | |
---|---|---|---|
Max | 16% | 15% | 14% |
Min | 0% | NaN | NaN |
Avg ≥ 500 | 16% | 14% | 16% |
Avg < 500 | 34% | 8% | -14% |
Download the spreadsheet (zipped)
Conclusions
As can be seen from the graphs and tables, the thicker foil sensor gives a larger reading when a finger is pressed against the sensor at all sample sizes, while the variation in readings when no finger is pressed down is sometimes lower and sometimes higher than the thin sensor.
The conclusion is therefore that the thicker foil gives a greater variation between “on” and “off” readings and will therefore be a better propostion for the instrument's keys. Future tests will only use the thicker foil.