Joule Thief: Time Course of Events Using PhotoTransistor to Monitor LED Output
Here I’ve placed a phototransistor, NTE3037, to monitor the light output of the LED. The purpose is to determine exactly when and how much the LED is turning on, compared to the electrical signals in the circuit.
First I use a cheapo PhotoVoltaic (solar) cell to show the basic light pulse timing and the rough shape of the light curve. Then I switch to the phototransistor which has a much higher amplitude and less noisy output, but whose output wave shape depends somewhat on its supply voltage.
The deepest dips in the Input Battery Voltage trace seems to occur right at the moment of LED turn-on. The voltage can actually climb above this lowest point during the LED’s “ON” period.
I didn’t show it in this video, but I’ve hooked an identical LED to my Function Generator and used the FG to send pulsed square wave signal to the LED…. the output of the Phototransistor tracks the brightness level of the LED exactly: if I give it a square pulse, flat along the top, at the same frequency as the JT pulses, the PT responds with the same shape output. So the decreasing ramp light output that is detected in the JT circuit pulses seems to be real, not an artefact of the phototransistor. The reason it doesn’t show up as well in the second half of this video is because I used a fresher battery for the JT and the light output is more constant during the pulse… but it still ramps downward.