Lock in Amplifier. Introduction. Motivation. Liz Schell and Allan Sadun Project Proposal

Size: px

Start display at page:

Download "Lock in Amplifier. Introduction. Motivation. Liz Schell and Allan Sadun Project Proposal"

Alan Williamson
6 years ago
Views:

1 Liz Schell and Allan Sadun Project Proposal Lock in Amplifier Introduction A lock in amplifier is an analog circuit that picks out and amplifies a particular frequency of oscillation and rejects the other frequencies. Lock in amplifiers are useful because they can extract very weak signals from noisy environments as long as the signal is at a precise and known frequency. We plan to build a system that transmits audio by modulating light at a particular frequency, and receives audio using a lock in amplifier. We plan to first make a working model with AM, which will be simpler and more robust. The FM implementation will have better performance and be more complex, because we ll have to add a phase locked loop and a voltage controlled oscillator. Because we plan to continue adapting our design, we will build the different modules with the same function to be interchangeable. Depending on how much time we have, we may make our frequency or bandwidth tunable, make our demo more impressive, or simply iterate our design to achieve higher levels of precision. Motivation Lock in amplifiers are useful for taking measurements in scientific research. For example, part of an experimental set up could be modulated at a given frequency so that the lock in amplifier can filter to that frequency to separate signal from noise. Learning about lock in amplifiers could be relevant to research that either of us does in the future. We also find this project interesting, because finding a signal which may appear to not exist when looking at the full, unfiltered signal is really cool. The interchangeable modularity of this project is also interesting. We ll be able to compare different techniques for similar processes and see which approaches work better and why.

2 Block Diagrams Figure 1: Block Diagram of AM System with shared reference oscillator In the AM topology, the output of the microphone is mixed with a reference tone and used to drive an LED, and the output of the photodetector is mixed with the same reference tone for demodulation in order to drive the speaker. A low pass filter is required on the input of the modulating mixer in order to prevent noise at higher harmonics of the reference frequency from interfering with the transmission, and a low pass filter is required on the output of the demodulating mixer in order to extract the audio signal. Figure 2: Block Diagram of FM System with Phase Locked Loop

3 In the FM topology, there is no need to cheat by having both the transmitter and the receiver synced up to the same oscillator, because the phase feedback in the receiver dynamically matches to the frequency of the transmitter. The microphone output is not mixed with anything, but is used to drive a voltage controlled oscillator, just as the speaker input is read from the input to the voltage controlled oscillator in a phase locked loop. Module Specifications (Note: All modules will be drawing power from a +15/ 15/+5/0 source such as a Protoboard.) Microphone and Amplifier: We have a microphone that will fit right into the breadboard. We will use the interferencing circuit that is suggested in the datasheet. The microphone will produce a signal with small amplitude, perhaps tens of millivolts. The amplifier will boost the signal to an amplitude of around half a volt. Figure 3: Microphone Circuit Suggested in CUI Inc. Datasheet Figure 4: Input Amplifier and Sallen Key Low pass Filter

Low pass filters: We will build two identical low pass filters. One will clean up the microphone output before using it for modulation, ensuring that we only have signals in the audio band.

4 Low pass filters: We will build two identical low pass filters. One will clean up the microphone output before using it for modulation, ensuring that we only have signals in the audio band. The other will be used to clean up the demodulated signal before sending it to the speaker. The corner frequencies will be around 10 khz. Because this is not too many decades away from our modulating frequency, our filter may need a relatively sharp roll off in order to avoid harmonic distortion. These low pass filters will be built using a Sallen Key topology. Reference oscillator (AM mode only): We will build an oscillator capable of producing square waves at frequencies between 100 kthz and 1 MHz, which is the range we allow our modulation frequency to lie in. The peak to peak voltage output of this oscillator will be at least 3 V, so as to overcome any diode drops in the mixing process. Because of its built in 50% duty cycle, the Schmitt trigger topology lends itself nicely to this application. Figure 5: Schmitt Trigger Topology for Reference Oscillator Mixer (AM mode only): In order to modulate our audio signal, we don t need an analog multiplier. A Gilbert cell mixer, driven with a square wave at the LO input and driven with the microphone output at the RF input, will do the job fine. In order to reproduce the audio signal faithfully, the RF input transistors will need to be operating linearly, which means that the RF signal must be small (on the order of of 50 mv peak to peak). In addition, the RF and LO signals must be biased properly, requiring a few coupling capacitors.

5 Figure 6: Gilbert Cell Mixer VCO (FM mode only): In the FM topology, the microphone signal goes to the oscillator, not a mixer. Consequently we will need a voltage controlled oscillator whose frequency can fluctuate up or down by approximately 20 khz as it receives an input that fluctuates up or down by approximately half a volt. The center frequency should be somewhere in the hundreds of kilohertz. We have not decided upon a topology for our VCO s, but some options we are exploring include a voltage controlled current source and a 555, or possibly a 555 in astable operation with a variable voltage threshold. LED and Photodetector: We will need an LED driving circuit to take the modulated voltage signal for transmission and convert it to a current signal. We will also need a circuit after the photodiode to convert the current signal back to a voltage signal which can be fed into the demodulator. The channel between the LED and the photodetector is where a lot of noise can be added to the system depending on how much we shield it from extraneous light sources, such as the room s lighting or the sun. The LED will need to be chosen so that its response time allows it to keep up with the modulated input signal, which could be as high as 1MHz. The photodetector will be chosen so that its response time allows it to keep up with the signal from the LED and so that its spectral sensitivity includes the wavelength at which the LED operates. If we want to place our system near a window and use sunlight as the noise, we ll have to ensure that sunlight alone does not cause our photodiode to max out, because then we won t be able to retrieve the data signal.

6 Figure 7: Transconductance Amplifier used to drive LED Figure 8: Transimpedance Amplifier used after photodiode Phase locked loop (FM mode only): We will require some sort of feedback circuit which takes in the noisy signal from the photodiode, and uses it to adjust the frequency of a VCO to match. Most implementations of a phase locked loop rely on a phase detector topology which is similar to the Gilbert cell mixer we will build for the AM topology, but a challenge will be ensuring that the output of the photodiode is well biased and well sized to the input of the Gilbert cell. An additional constraint is that the output of the phase locked loop must be of the right magnitude to drive the VCO, i.e. about equal to the output of the microphone amplifier, i.e. about a 1 Volt swing. Amplifier and Speaker: This is the output stage, which amplifies the signal before feeding it into a speaker. We plan to use the class D amplifier that we made in class, so that we can devote more time to different parts of our project. For a speaker, we plan to use one of the speakers in lab. Task Breakdown and Schedule

7 Our initial goal is to build the AM system and demonstrate its operation as quickly as possible. To that end, Liz will be responsible for the low frequency pipeline (the microphone, its amplifier, the LED and photodiode, the speaker, and its amplifier), while Sadun will be responsible for the high frequency pipeline (the mixers, the reference oscillator, and the low pass filters). We plan to have the AM system completely working by April 16th. This will give us the bulk of Patriot s Day weekend to assess design flaws with the AM system and to specify a design plan for the phase detector and the VCO. Once our design plan for the FM system is finalized, we will split up the responsibility of modules between the two of us. Our task deadlines are as follows. April 8: Order all parts that cannot be found in the lab. April 16: Complete AM system. April 19: Assess design flaws of AM system. Based on these flaws, finalize design plan, including labor breakdown, for FM system. April 26: Complete building and debugging FM modules. April 30: Complete integration of FM modules into full system. References and Resources Figure 3 is from the CUI Inc. datasheet for the microphone. Figure 5 is from Hyperphysics. ( astr.gsu.edu/hbase/electronic/ietron/square.gif ) Figure 6 is from Radio electronics.com. ( electronics.com/info/rf technology design/mixers/mixer gilbert cell.gif )

Exercise 2: FM Detection With a PLL

Phase-Locked Loop Analog Communications Exercise 2: FM Detection With a PLL EXERCISE OBJECTIVE When you have completed this exercise, you will be able to explain how the phase detector s input frequencies