Lab 4: AM Detector
Introduction
For reference, here is the lab manual that was followed. Circuit diagrams given in the manual will not be reproduced here. There are three different designs that were tested.
The only metric of performance(that we care about) for a detector is the minimum input amplitude at which it works. Component number ends up being a factor but thats only because im lazy and bad at breadboarding.
What does a detector do? It takes the blue wave and turns it into the pink-wave. This seems like it would be a low pass filter but in reality its not. Recalling the spectrum analyzer from Lab 1, we can see that an AM wave isn't actually the signal plus the carrier, its the signal multiplied by the carrier making it substantially more complicated to separate.
99.9% of this JS came from here, credit is due
Diode detector
The oldest form of radio is called a crystal radio, which utilized a primitive diode in the form of a galena crystal. This design calls back to that. Basically just rectify the signal and then low-pass the signal to remove the carrier. Its performance is absolutely abysmal, not worth showing a screenshot here as there wasn't even an output signal to see. This is due to diodes having a finite turn on time and the breadboard certainly doesn't help with that. This turn on time is handled by the second detector type.
Biased Diode detector
Just take the exact same circuit as the normal diode detector, and apply a DC bias to just above the turn-on voltage of the diode.
This is done with a voltage divider. The waveform doesn't have the full amplitude of the input but it doesn't do half bad. See the waveform below. It pretty consistently extracts this intelligence with input signals from 40mV to 250mV before any clipping occurs.
Complementary-Feedback Pair detector
A CE amp basically acts as a diode detector, only letting current through when a certain voltage is present.
The problem here is the input signal is typically too small to cause a large swing in the throughput current of single transistor.
That can be fixed by setting up whats called a complementary pair. Basically as current begins to flow through the NPN the voltage will drop on the input of PNP.
This is essentially a very low input floor amplifier. The wave that comes out can be seen below.
The primary problem here is the number of components makes this very finnicky. The experience in the lab was that the system was very noisy. The acceptable voltage range was 2mVpp to 100mVpp which is a very low floor considering that we are on a breadboard.
Conclusions and reflections
While detection is a bit more complicated than just filtering out the carrier signal, its not much harder. These circuits just make sense.
Theres only two viable options for the final radio, one of which is way more finnicky and complicated than the other one, making the biased diode the clear winner.
Putting the signal from the biased diode into our audio amp of choice(lm386) makes sounds like the ones below.
It works quite well considering they are two simplest circuits we've looked at.