Wednesday, January 14, 2009

An AC to DC Converter

In one of my classes, E E 331 Devices and Circuits I, our final project was to build an AC to DC converter. Our design specifications were to take a 10 Vpp 60 Hz DC input from a center-tap transformer and have an adjustable 10 V to 20 V DC output. It was expected to have output noise of maximum 100 mV, and able to deliver 1 mA current for all voltage settings. Also, we were graded on how cheap the circuit was to build.



Above is a block diagram of our design. The circuit works by rectifying an AC input, creating a high frequency square wave, and using that wave to drive a boost converter to amplify the rectified input signal to a level dictated by a differential amplifier you could control using a potentiometer.



And above is the circuit schematic as viewed in PSPICE schematic and simulation software. The circuit was designed mostly on computer, then built later when the simulated circuit met specifications. The square wave was provided by a 555 timer, and the differential amplifier was built using a LM741 operational amplifier. The amplifier compares the voltage at the anode of a 5 V zener diode with the output voltage, and increases its output voltage when the output voltage falls too low, and decreases its output when the output voltage gets too high. This feedback system maintains a constant output voltage, which is calibrated using a network of resisters to operate within the specified range.



Above is the circuit built on a breadboard. The total cost of all the electrical components is a little under $8. Our design had barely any noise in the output, and it met all performance requirements within a narrow error margin.

2 comments:

Anonymous said...

Jonny;
Du hast wirklich gut gemacht!!!! Strong work.
Vater

Anonymous said...

741s are op-amps and do not work as well as hysteresis switches as comparators. An LM393 might be better. You've designed a burst-regulating converter. The next step is to model the loop dynamics in the complex-frequency domain. You will need to first linearize the switch (2N7000). It is really possible to have a linear model of a switch! If interested, I'll send you a weblink to my article on it at www.en-genius.net

I recently spent 8 weeks trying to refine the unified theory of current-loop converters. After 35 years of switching converter theory development it is still unfinished business. One of the leading contributors to switching converter theory lived with Dottie and me for 5 years in Portland. He and his wife will be visiting us in FEB10.

Uncle Dennis