workexperience.ms (5870B)
1 .TL 2 Notes on work done 3 .AU 4 Lucas Standen + Nigel Standen 5 .AI 6 29/5/24 7 .NH 1 8 How the current of an LED can effect the wavelength of the output 9 .NH 2 10 The experiment 11 .LP 12 In the experiment we measured how the wavelength of an LED is effected by the current passed through 13 it. The results lead one to believe that as current increases the wavelength does too. 14 15 We made sure to measure using a pulse, rather than a continuous current, this is to ensure that the 16 thermal energy that the LED gives off, does not effect its results; however we believe our pulse 17 was still to long and that it did effect the results. 18 19 The pulse is generated by an Arduino with a simple script running, allowing me to press a button 20 that triggers the power supply and the spectrometer (tool used to measure wavelengths). Before 21 starting the experiment, we took a reading of ambient temperature, as we were looking at 22 current (which effects temperature) against wavelength. 23 24 We found the temperature to be 25 .B "20 deg C". 26 27 .NH 3 28 Triggering circuit 29 .LP 30 The triggering circuit was kept simple, with most things being done in code, one thing of note 31 however is that the push button used as a trigger is in a push to make configuration. 32 33 To wire the circuit follow these steps: 34 35 1) Attach a button between pin 2, and ground. 36 37 2) Attach pin 7 to the spectroscope's live input, and attach its ground to a ground rail on the 38 bread board. 39 40 3) Attach pin 8 to the PSU's live input and its ground to a shared ground. 41 42 4) Using the Arduino ide, flash the code (can be found on the USB you gave me). 43 44 5) Set up the program on the PSU to have a pulse duration as long as you need, and then press the 45 button. 46 47 .LP 48 The code for this can be seen here as well: 49 .LP 50 .B1 51 const int instrument1Pin = 7; // Pin for instrument 1 (0V to 5V) Spectroscope 52 53 const int instrument2Pin = 8; // Pin for instrument 2 (5V to 0V) PSU 54 55 const int triggerPin = 2; // the pin the button is connected too (active high) 56 57 // Configure pulses 58 const int pulseDurationInMs = 100; 59 const int psuDelayInMs = 30; 60 61 void setup() { 62 // Configure pins 63 pinMode(instrument1Pin, OUTPUT); 64 pinMode(instrument2Pin, OUTPUT); 65 pinMode(triggerPin, INPUT_PULLUP); 66 67 // Initialize pins 68 digitalWrite(instrument1Pin, LOW); // LOW corresponds to 0V (min value) 69 digitalWrite(instrument2Pin, HIGH); // HIGH corresponds to 5V (max value) 70 71 } 72 73 74 void loop() { 75 if (digitalRead(triggerPin) == LOW){ 76 // Create pulses 77 digitalWrite(instrument2Pin, LOW); 78 delay(psuDelayInMs); 79 digitalWrite(instrument1Pin, HIGH); 80 81 delay(pulseDurationInMs - psuDelayInMs); 82 83 84 digitalWrite(instrument2Pin, HIGH); 85 delay(psuDelayInMs); 86 digitalWrite(instrument1Pin, LOW); 87 delay(1000); 88 } 89 90 } 91 .B2 92 .LP 93 It is relatively simple in function, when the switch on pin 2 goes high, it will send a low to the 94 PSU (the PSU is active low in this case) and a high to the spectrometer (active high). The setup 95 function (defined in the first set of {}) is used to initialise the pins, and the loop function 96 will be run as fast as the microcontroller can. It constantly checks if the trigger pin goes low 97 (the button has shorted pin 2) and if it does it sends a pulse of pulseDurationInMS down 98 each wire. The variable psuDelayInMs is used to set how much sooner the PSU will trigger compared 99 to the spectrometer. 100 101 .NH 2 102 The results 103 .LP 104 We measured results using a green LED, and our pulse's had a width of 450ms 105 106 A graph of results can be found below: 107 .PSPIC graph.ps 108 .LP 109 .I "y axis = wavelength (nm), x axis = current (mA)" 110 .LP 111 I modelled a line of best fit for this graph to be y = 1/9x + 568.4, with X being the current in mA, 112 and Y being the wavelength in nm; however this is just by eye. From this one can assume the 113 temperature coefficient to be 1/9 nm/mA. 114 115 The results show that as the LED had more current flowing through it, the colour of the light it 116 produced changed to a higher wavelength (towards red). 117 118 It is worth noting that when we reduced the pulse duration to 300ms instead of 450ms, at 30mA 119 the wavelength fell to 570.61nm, which suggests that even on a pulse as small as 450ms the heating 120 of the LED has effected its wavelength. 121 122 Of note we found that there was a tiny delay between when the when the spectrometer went high, and 123 the PSU went low, we found it to be around 6us, we believe this to be because the spectrometer is 124 set to go high first. 125 126 This effect can be seen in this oscilloscope: 127 .PSPIC scope.ps 128 .I "The yellow traces is the spectrometer, the blue is the PSU" 129 130 .NH 2 131 Takeaways 132 .LP 133 From this experience I think I should takeaway that, first LED's are effected by temperature very 134 slightly, and that the equipment resolution/accuracy is important to note, my results may be wrong, 135 however the trend shown by that graph is correct, I find this interesting and important. 136 137 I've believe this work will help me in further life, it really has been my first glimpse into a real 138 working world with electronics, I look forward to working more on it. 139 140 .NH 1 141 Building LED's 142 .LP 143 The process consists of taking the die which is made from silicon and a handful of other metals 144 that decide the wavelength of the light emitted. The die is placed on a package with apoxy, that 145 contains silver, then a wire is bonded to the top (see pictures), to the other part of the package. 146 147 Here is some images of the process or wire bonding LED's: 148 .PSPIC wirebonds.ps 149 .I "A view of the die on the package, and the machine" 150 .PSPIC wirebonds2.ps 151 .I "A view through the lens" 152 .LP 153 This was a very interesting process to see, I always imagined an LED as one bit of metal that did 154 the job, I had never through how much precision work went into it. The LED's we worked on 155 were 280um x 280um.