Variable Frequency Power Inverter

Introducing…

            The wait is over!  Finally, A reliable variable frequency power inverter is yours to build!  Boasting a wide operating frequency from 20 to 80 hertz, this device can operate even the most discriminating frequency dependent devices.  Output current is rated at 500 mA at 120 Vac, delivering 60 watts to your load, and a highly accurate frequency measurement and indication circuit let you know when you are providing that power at 60 hertz (a Green LED), above 60 hertz (a Red LED) or even below 60 hertz (a Yellow LED) to within 1%!  This inverter is also capable of driving inductive, capacitive and resistive loads with no change in output frequency.  As an added benefit of this design, by improving the current handling capabilities of the Power Mosfets and the Transformer by replacing them with devices rated at higher capacities, you can output higher currents and/or voltages.

            Why build it you ask?  To provide a source of 120 Vac anytime, anywhere from a 12 volt battery!  But not just any 120 Vac, variable frequency AC which gives you the ability to vary the speed of synchronous AC motors!  A definition: 

Synchronous Motors:
Synchronous motors provide rotation at a fixed speed in synchronization with the frequency of the power source, regardless of fluctuation of the load or line voltage.

So as you can clearly see, A variable frequency power inverter can not only power your favorite lava lamp, but it can also be used to control the speed of Synchronous Motors!

            The requirements for the entire project are as follows:

1.      Components: student lab kit or instructor approved devices

2.      Power: Tek PS280 power supply set to 12 volts or a 12 volt battery

3.      Inverter Output: 120 Vac 0 – 70 Hz.

4.      Visual Indication: Green LED – 60 Hz. +- 1%; Yellow LED - >10 Hz < 59 Hz; Red LED-> 61 Hz < 70 Hz.

5.      Accuracy: < +- 5 Volts output; frequency +- 1 Hz.

6.      Special:  Either the Green, Yellow or the Red LED may be on, but only one may be on at a time.

Equipment:

1.      Normal Lab Equipment

2.      Lab design Kits

3.      Department Surplus Components

Abstract

            Included is the complete process and design of a VFPI.  When successfully constructed, this project provides 120 VAC power from a single 12 volt source and allows you to vary the frequency over a range of 30 to 80 Hz.  My results indicate that this circuit would be useful for driving small loads like fluorescent light bulbs and small motors, but is not suited to drive more demanding loads like a television.

Theory

            The basic theory of operation flows much like my block diagram.  An oscillator at the heart of the circuit will provide a square wave signal to the rest of the circuit to drive the frequency detection circuit and the power driver stage.  The frequency detection circuit extrapolates the actual frequency of operation by some method and presents a logic signal to the decoding logic stage.  The decoding logic will then ensure that only one LED is on at any given time and inform the user of the frequency of operation.  The power amplifier stage, or power driver stage, will amplify the weak signal provided by the core oscillator and drive current through the primary windings of a transformer.  The transformer will step the voltage up to 120 Vac to drive a load of the user’s choice.  This encompasses the basic theory of operation.  I have included a theory heading under each block to simplify and enhance readability.

Methodology

            The description of each part and its operation is given in the breakdown of the block diagram.  What I will be discussing here is the building, testing and evaluation of the circuit.

The circuit was built one block at a time.  The core oscillator was bread boarded and tested.  During its testing phase we concluded that to vary the frequency over its useable range, a 10k potentiometer in series with a 6.8 K ohm resistor was needed.  This gave us a frequency range of 30 hz to 80 hz.  A stock 2N3904 was used, along with the diodes are kit supplied.  We chose offhand to use a 1 microfarad capacitor and it worked quite nicely.  After the core oscillator circuit was built and tested, the schematic was recorded with all the chosen values and we moved on to the frequency detection circuit.  The frequency detection circuit went through many stages before it reached its final form that you see now.  Each iteration of the frequency detection circuit was built and tested.  The problems were identified and analyzed, and then systematically worked on until the design was stable and reliable.  In building and testing the circuit, it was determined that a 741 was simply not capable if accomplishing the required tasks due to its slow slew rate and inability to swing from rail to rail (the 741’s output swing is limited to within 2 volts of each rail) The circuit was changed to allow the use of an LM324, which operated from one voltage supply and allows voltage all the way down to its negative rail (which in our case is ground).  The circuit was tested by applying a square wave of varying frequency to the input, and after fine tuning the output, making sure that it accurately measured the frequency repeatedly.  I even went as far as sticking the circuit into the freezer for half an hour and then running it through the input frequencies to see if temperature had large effects on the accuracy of the circuit.  Due to the fact that it depends upon the RC charging curve, I expected it to be highly temperature sensitive.  I was shocked to discover that when the circuit was chilled to approximately 15 degrees, the accuracy was only off by about 1 Hz.  The logic for the circuit was built with standard 2N2222 resistors.  It was tested by applying the three logical conditions to the inputs and observing the LED’s to ensure that they came on at the right times.  The power inverter stage was tested by assembling the entire circuit and driving the power inverter stage using the core oscillator.  The circuit was tested by measuring the voltage and current going into the primary of the transformer, and the voltage waveform on the secondary of the transformer.  The purpose of the “extra” two transistors in the power inverter stage is to protect the rest of the circuit in the case of a catastrophic failure.  If for some reason the gate were to become shorted to the drain (which could happen) then the gate would be at +12v all the time.  If this were to happen, the transistor will get destroyed, but not the rest of the circuit.  The “extra” transistors are there to protect the rest of the circuit from a failure in the power driver stage.  Once every part of the circuit was assembled and tested, the entire circuit was tested on various 120 volt appliances.  We tested the circuit on a Compact Fluorescent light bulb, a small fan and a motor.  All of these tests were conducted over the entire frequency range to ensure correct operation of all parts of the circuit.  It was noted that our transformer saturated at frequencies approaching 25 Hz.  Attempting to make useable power below these frequencies would result in high current and destruction of the power driver stage.

Structure and Design

        The Variable Frequency Power Inverter (or from now on the VFPI) can be divided into many functional sub circuits.  A diagram of how the blocks are placed together to form a functional VFPI is given.  A brief overview of each block is provided, and then a more detailed explanation will be given.

1.)    “Raw 12 volt Power” is your choice of a 12-volt supply, but must be capable of supplying 12 volts at a minimum of 4 amps with minimal voltage ripple and/or sag.

2.)    “12 volt Glitch Filter” provides 11.3 volts to all parts of the circuit that require a relatively constant voltage supply with minimal dropouts and ripple.  The Glitch filter maintains voltage to the circuitry in case of a momentary lapse in the raw 12-volt power.

3.)    “Core oscillator” is the ‘heartbeat’ of the circuit.  This circuit provides a variable frequency square wave with a fixed 50% duty cycle over its entire frequency range of 20 to 80 hertz.  The frequency is adjustable via a user accessible potentiometer.

4.)    “Power Driver Stage” is driven by the core oscillator and switches the current through the transformer.  This stage is designed to handle the high currents required for the power inversion process.

5.)    “Transformer” is a center tapped 24-volt to 120-volt AC transformer capable of operating at the high currents demanded by your particular application.

6.)    “5-volt Regulator” provides a constant 5 volts to the frequency detection circuit regardless of load conditions.  This constant voltage source is paramount in the accuracy of the frequency detection circuit.

7.)    “Frequency Detection Circuit” monitors the frequency of the square wave provided by the core oscillator, and provides a logic signal to the decoding logic circuit related to the actual frequency being generated by the core oscillator.

8.)    “Decoding Logic” interprets the logic signal from the Frequency Detection Circuit and indicates the frequency on a color-coded LED display.

The 12-volt Glitch Filter

            Theory:

A sharp drop in the voltage from the raw 12-volt power (due to a high current demand or supply sag) could spell disaster for any circuits that take their power directly from the raw power supply.  This Glitch Filter will supply power to the circuit for a short period of time while the supply is temporarily unable to cope with the demands of the inversion process.

Methodology:

The Operation of this circuit is rather simple; When the supply is at its full potential, the diode is forward biased and supplies the circuit with 11.3 volts and simultaneously charges the large capacitor.  When the supply drops below 12 volts, the diode is reverse biased, and the capacitor supplies energy to the circuit for a short period of time.  A worst-case load of 100 ohms was used to design the glitch filter.  This provides approx. 42 milliseconds of power without the voltage falling below 10 volts.  In practice, I do not expect the circuit connected to the glitch filter to require the 113 milliamps consumed by the 100 ohm resistor, so the glitch filter should perform well.  Equations are provided in the appendix under the heading “12 volt glitch filter”. 

I chose 4000 microfarads simply because I felt that it should be sufficient to supply the demands of the circuit with plenty of reserves during a small glitch.  On the following pages, I give a schematic of the glitch filter and its output while connected to a 100-ohm load with a faltering input voltage is provided in the appendix.

Testing and evaluation:

Once the circuit was powered by the glitch filter, and no longer powered by the raw power, the circuit stabilized and provided much more accurate measurements of the frequency of operation.  This was tested by applying many different types of loads to the VFPI and monitoring its stability and accuracy when the conditions were changed suddenly or high demands were placed on the power supply.

Core Oscillator

Theory:

            The Core Oscillator of the circuit could be thought of as the “heartbeat” of the circuit.  We had a few design requirements that needed to be met.  These requirements are not dictated by the project requirements; they are our own requirements to allow this system to be easily integrated into the project.

1.)    The oscillator must output a square wave with a 50% duty cycle over its entire frequency range

2.)    The oscillator frequency must be adjustable

3.)    The output frequency must be stable within 10 to 80 hertz.

4.)    The oscillator must be simple

Methodology:

After carefully evaluating many oscillator designs with numerous advantages and disadvantages, we discovered this simple 555 timer design which boasted an adjustable frequency range set by a single potentiometer, a 50% duty cycle over its entire frequency range, and a low parts count.  A normal 555-timer circuit may have an adjustable frequency range, but its duty cycle varies over the entire frequency range because the 555 timer will charge a capacitor through two resistors, and discharge the capacitor through only one of those resistors.  The frequency is set by altering the value of ONE of those resistors resulting in a different charge or discharge time, but the charge and discharge time is asymmetrical which creates an asymmetrical duty cycle. 

This clever circuit allows the capacitor to charge AND discharge through the same variable resistor which creates a symmetrical 50% duty cycle over its ENTIRE frequency range.  The circuit operates like so:

1.)    Pin 7 (discharge) goes high, which allows the base of Q1 to be pulled high by the 1.5K ohm resistor.  This puts Q1 into hard conduction and begins to charge C1 through the potentiometer and series resistor.

2.)    The voltage at pin 6 (threshold) reaches 2/3 of VCC which switches Pin 3 (output) low, and simultaneously switches Pin 7 (discharge) low as well.

3.)    Since Pin 7 (discharge) is now low, the diode on the base of Q1 is forward biased and pulls the base to within 0.7 volts of ground.  Since the emitter is well above 0.7 volts (the emitter will not go below 1/3 VCC while the 555 timer is operating) the transistor is forced off.  The diode D3 is now also forward biased and allows the capacitor to discharge through the potentiometer and series resistor into Pin 7 (discharge).

4.)    The voltage on Pin 4 (trigger) goes below 1/3 VCC and switches Pin 3 (output) high and simultaneously switches Pin 7 (discharge) high as well, which begins the process all over again.

So as the potentiometer is varied over its range, it alters the RC time constant of the charge and discharge curve of the capacitor, thus increasing or decreasing the amount of time required to reach 2/3 and 1/3 of VCC.  This results in a change of frequency.

Testing and evaluation:

A value of 1 microfarad was chosen for C1, and the value of the series resistor was chosen so that when the potentiometer is at its lowest resistance, a ceiling was set on the maximum frequency (approx. 80 hertz).  10K ohms was chosen for the potentiometer because it gave us a low frequency of about 32 hertz, (slightly higher than the requested 10 hertz) but allowed us to tune the frequency with much greater accuracy within the range of 60 hertz.  If a higher value of potentiometer is used, then the range of resistances over which the 555 timer produces 60 hertz +- 1% is made small, which results in great difficulty tuning the circuit to 60 hertz. 

To rephrase, when the potentiometer’s resistance is made higher than 10K ohms, the 555 timer output frequency becomes more and more sensitive to small changes in position, and soon it becomes impossible to achieve 60 hertz with any degree of accuracy because the movement required to change to frequency a small amount is so small that a human hand is not capable of doing it.

The circuit was connected to the oscilloscope and its frequency and duty cycle were measure over its entire operating range.  We were pleased to find that the circuit did indeed operate as we had designed it to.

Recommendations: 

In the later design, we may add a switch to allow frequency tuning over a high and low range to span the required 10 to 80 hertz.

 The following pages contain our Core Oscillator schematic. Pspice file and outputs with the potentiometer adjusted to both extremes to show the output frequencies and duty cycles achieved are included in the appendix.

Power Driver Stage

            Theory:

It is the duty of the Power Driver Stage to supply the transformer with the current and voltage it demands.  The Power Driver Stage is driven by the core oscillator and alternately switches the direction of the current through the transformer at a rate determined by the core oscillator. 

Methodology:

            In order to create AC, current must be constantly switched from one direction through the transformer windings to the other.  There are many ways to accomplish this, (H-Bridges are a fantastic way to achieve this, but they require considerably more complex circuitry than I wished to use) but I was looking for a simple and effective way to alternate the current through the windings of the transformer.  It dawned upon me that if I used a center tapped transformer, then I could supply 12 volts to the center tap, and alternately ground out each leg of the transformer causing current to first travel one way through the transformer, and then the other.  If I could locate a transformer that was rated at 12 volts from each leg to the center tap of the secondary and 120 volts on the primary, then the inversion process would be possible with a very small amount of circuitry.  With a bit of searching I came upon the right transformer, and now the circuitry became trivial.  I wished to use Enhancement Mode N-Channel Power Mosfets to alternately ground each leg of the secondary due to the extremely low on resistance (around 0.1 ohms) resulting in low power dissipation through the mosfet when it is on and conducting high current, and also because turning on an N-channel mosfet with its source connected to ground simply requires supplying its gate with a positive voltage above its threshold voltage.

            I examined the specs of a few mosfets and settled with using an IRF510 N-channel enhancement type power mosfet.  The IRF510 is capable of dissipating 20 Watts and sinking 4 amps of current continuously, or 16 amps pulsed.  The on resistance of the IRF510 is 0.6 ohms and its threshold voltage is 2 volts.  At a current draw of 4 amps and a Ron of 0.6 ohms, the mosfet will be dissipating 9.6 watts, well within its limits.  Another reason I chose the IRF510 was its availability (radioshack).

            With my Mosfet chosen, it was time to design the drive buffer.  The drive buffer serves two purposes.

1.)    To ensure that only ONE mosfet will be on at any one time

2.)    To protect the rest of the circuit in case of a catastrophic failure

I ensured that only one mosfet would be on at any one time by feeding one mosfet with a non-inverted drive signal from the core oscillator, and feeding the other mosfet and inverted signal from the core oscillator.  This provided 12 volts to only ONE mosfet at a time, and also allowed me to drive the Power Driver Stage with a single square wave.

      The circuit operates like so:

1.)    The input is driven high by the core oscillator

2.)    The base of Q1 and Q3 are driven high, causing them to go into hard conduction.

3.)    Current flowing from the Collector to the Emitter of Q1 and Q3 cause a voltage drop across R3 and R5 effectively pulling their outputs to ground

4.)    Q3’s signal ( ground ) is sent straight through to the mosfet Q5’s gate which switches that mosfet off.

5.)    Q1’s signal is sent to another inverting transistor circuit which effectively restores the original signal and presents Q4’s gate with 12 volts, turning that transistor on and allowing current to flow from the center tap to the outside leg of the transformer.

6.)    When the input to the power driver stage is driven low, Q4’s gate is presented with ground, turning that mosfet off, and Q5’s gate is presented with 12 volts, turning that transistor on, and allowing current to flow in the other direction through the transformer.

Testing and evaluation:

            The circuit was driven by a square wave created by the HP3312A function generator.  The HP 54600A oscope was connected to the secondary, and the voltage was monitored as the frequency was increased and decreased on the function generator.  It was noted that the transformer ceased operating properly when the frequency was driven below about 30 Hz.  This was due to the transformer saturating.  When the transformer is allowed to have current flowing in one direction for too long (and in the case of our transformer too long was 1/30 seconds) then the transformer begins to look more and more like a wire, and less and less like a transformer.  When this happens, enormous amounts of current begin to flow through the primary windings, but this results in little to no power being created in the secondary.  There is no way to circumvent this problem except to get a transformer with much more core material which results in a greater capability for not saturating at higher flux densities, but also increases weight and cost significantly.

            Recommendations:

            A larger transformer would decrease core losses and also lower the lowest possible operating frequency.

The following pages contain the schematic of the circuit. Pspice files and the output waveform as seen from the output of the transformer are included in the appendix.

5 volt Regulator

            Theory:

The 5-volt regulator circuit is of paramount importance to the frequency detection circuit.  The constant 5 volts it provides allow accurate measurement of voltages levels to within a few milli-volts regardless of different loading conditions and noise elsewhere in the circuit.  I had originally planned on using a 7805 3-terminal voltage regulator for this part of the circuit, but after discovering this circuit, I decided to use the opportunity to build my own simple voltage regulator.

Methodology:

            The reference voltage is set using a zener diode and a current biasing resistor.  The zener diode provides a constant 5 volts to the non-inverting terminal of the op-amp.  The emitter voltage is fed back to the op-amp into the inverting terminal and provides voltage regulation at the emitter of the pass transistor.  This regulation is accomplished by the feedback provided to the op-amp.  When the load is increased and the voltage at the emitter drops below the reference voltage set by the zener diode, the op-amp sees a decrease in voltage at its inverting terminal with respect to its non-inverting terminal.  This causes a rise in the output voltage of the op-amp, which turns the pass transistor on harder and allows more current to flow.  The output voltage of the op-amp continues to rise until the voltage at the emitter once again matches the voltage at the non-inverting terminal of the op-amp.  Likewise, when the voltage at the emitter is too high, the op-amp output voltage decreases, restricting the current flow through the transistor and lowering the emitter voltage. 

            Testing and evaluation:

            The circuit was bread boarded separately from the rest of the circuit and tested vigorously.  We monitored the output voltage of the regulator circuit as we rapidly varied the load that the regulator supplied.  We observed that the output was smooth and consistent as long as a filter capacitor was included on the output. Without the filter capacitor, the output had a tendency to jump rapidly above and below the regulation voltage when a load was quickly switched in or out of the circuit.  However, with the filter capacitor, the circuit operated reliably and consistently.

Schematics, Pspice and a sample output of the regulator for a 1K ohm load are provided in the appendix.  A filter capacitor is added to the emitter of the transistor to smooth out transients and sudden changes in the demands of the regulator.

Frequency Detection Circuit

Methodology:

I had first envisioned using a band pass filter with a high q around 60 hz fed into the input of a window comparator.  Through inspection though it was determined that there would be no way to identify if you were above or below 60 hz.  As the band pass filters magnitude is greatest around its center frequency and simply falls off to either side.  Thus if a threshold was set when the filter approached 60 hz from the low side, the comparator would trip and light an led or feed some logic. However if you continued past 60 hz toward the high frequency side the output would fall off again and the led would go out. You would have no idea if you were too high or to low.  Also, this circuit is sensitive to input voltage changes, causing an inaccurate reading if the input voltage fluctuates above or below its tuned voltage.  One problem could be solved by using another filter and comparator circuit to determine if you have gone above or below 60 hz but I wanted a simple one filter approach to the frequency indicator circuit. This led me to:

Design 2

With this set of design challenges to overcome I came up with this idea. Use a high pass filter and set the –3db point at 60 hz thus when the output of the filter is fed into a window comparator, if your frequency is too low then you will be below the threshold of both comparators and the circuit will output a logic “00”. As you increase the frequency the output of the filter will also increase causing the voltage to rise above the threshold of comparator 1 and outputting a logic “10”.  If you continue to increase the frequency the filter output would rise more thus exceeding the threshold voltage of comparator 1 and 2 and outputting a logic “11” indicating you have passed above 60 hz.  This circuit was constructed successfully however some problems were encountered.  The change in output voltage around 60 hz plus or minus 1% was approximately 75 mv, but this required the input voltage to remain constant.  If the input voltage were to shift at all the frequency measurement would be thrown askew.  Also, when a square wave input was applied the high pass filter blocked the lower fundamental frequency while passing the high frequency components.  This resulted in a spike and rendered the frequency detection circuit useless.  This behavior was completely undesirable because the main drive for our circuit would be a square wave and thus measuring the frequency of a square wave was most important. Which led me to:

Design 3:

Switch the filter around and make it a low pass filter! Then when a square wave input is encountered the high frequency components will be blocked but the lower fundamental frequency will be let through and will be attenuated by the filter as a function of its frequency. All that was required was a change in logic interpretation.  A logic “11” now means the frequency is to low, a logic “10” means we are within range and a logic “00” means we have passed above 60 hz. 

            This approach yielded promising results. a square wave input  now behaved in a much more predictable way thus its frequency could be measured.  However the frequency detection circuit was still extremely voltage dependent.  I attempted to rectify this by placing a zener diode before the low pass filter, thus I clipped the input signal to a known voltage as long as Vin was above the zener voltage.  This approach worked extremely well, and assured that as long as the voltage Vin stayed above the zener threshold, the circuit worked reliably, but it was observed that it was difficult to achieve an accuracy of better than +- 1 hz.  Which roughly translates into an accuracy of +- 1.6%.  Not quite good enough.  Which led me to:

Design 4:

I take a pretty radical leap from here.  I few things I have not previously discussed:

1.  Output flickered due to voltage only briefly passing above the threshold voltage.

This circuit works like so:  the comparator output goes high, charging the capacitor.  The capacitor holds the charge, while the comparator output goes low. (the diode is reversed biased so the cap is slowly discharged through the resistor which creates an RC circuit with a time constant many times longer than the period of the input spike.)  Thus ensuring that if the comparator goes high again within a 60 hz period (16ms) the capacitor will not be completely discharged and provide a logical 1 to the input of the buffer even while the comparator output is low, creating a flicker free output.  The circuit also responds quickly by turning off the output of the buffer within approx. 100 ms of receiving its last pulse.

2.      I decided that the input was still too voltage dependent and unpredictable with the zener low pass filter, so I opted to feed the input signal into another buffer, thus ensuring any input with a magnitude of 0 to greater than 3 volts would consistently provide me with a 0 to 5 volt square wave.  This 0 to 5 volt output was fed into a low pass filter and then into the window comparator circuit.  While observing the output on the scope, I noticed that the top and bottom of the wave was being attenuated, thus moving the dc component.  This was occurring because not only was the capacitor charging through the resistor, it was also discharging through the resistor, and since I was only measuring the positive going part of the wave, I felt accuracy could be increased by completely discharging the capacitor before charging it again, effectively changing the low pass filter approach into a time dependent charging curve.  This was accomplished my adding a diode to the RC network.

Notice the resistor marked “X” ohms.  I felt that by carefully selecting this value, the maximum range of voltages could be extracted from this time dependent charge circuit.  This was accomplished by first determining the times of most interest.  That would be the period of 59.5 hz and 60.5 hz giving me a period of 0.016807 sec. and 0.016529 sec. Respectively.  I then wrote this equation: 

  I then plotted Vdiff with respect to R and discovered that there was indeed an ideal resistor value.  The plot looked like this:

It was desirable to determine what value of R was at the peak, so, to determine that value, I took the derivative of the previous equation, set it equal to 0, and solved for R, yielding a value of 16.66 k Ohms.  An odd value, so I chose R = 15 K Ohms.  (Which was still very close to the peak of the curve.)  Thus, “X” = 15 K ohms. 

Testing and evaluation

The circuit was adjusted for the new setup, and tested.  The results were impressive.  The real world data is as follows:

Yielding an accuracy of +- 0.6 %.

            At this point, the circuit is input voltage stable, can be tuned to accept a square wave and has an accuracy of +- 0.6%.  On the following pages I provide a schematic of the circuit and show it operating at 60 hertz with the charge curve peaking between the two threshold voltages.  Also included is a Pspice file.

Decoding Logic

      Theory:

    The Logic Decoder is required to interpret the information coming out of the Frequency Detection circuit.  The design requirements state that only one LED may be on at any one time, and the logic decoding circuit takes care of that.  By analyzing the output states of the Frequency Detection circuit, it was determined that the yellow LED needed to be hooked up to an AND gate, the red LED needed to be hooked up to a NOR gate, and the green LED needed to be hooked up to an XOR gate, or equivalent since only 3 of the four possible states will ever be presented by the Frequency Detection circuit. 

Methodology:

            Diode-Transistor logic was used to create the three gates.  An AND gate was constructed using two diodes and a resistor.  When both inputs are high, both diodes are reverse biased and the resistor pulls the output high.  If either one of the inputs or both are pulled low, the diode will be forward biased and pull the output low.

            A NOR gate was created by using a two diode OR gate and then the output of the OR gate was fed into the input of an inverter (explained in the Power Driver Stage).

            The fake XOR gate was created by placing an inverter on the input of one of the inputs of an AND gate.  Thus the only time the gate output a logic ‘1’ was when one input was low, and the other was high, creating a fake XOR (since the other condition cannot happen). 

            Testing and Evaluation:

            The circuit was tested by applying all three possible logic conditions to the input of the circuit and observing its output on the three LED’s.  the circuit preformed exactly as it was designed to.

Following this page there is a schematic of the logic circuits and Pspice files are included in the appendix.

Conclusion

To reiterate the specifications set forth for this project:

Equipment:

Requirements Met / Not met and why:

1.      MET.  All components used in this project came either from our lab kit, or we obtained explicit approval to use in the project.

2.      MET.  Our circuit can be powered from 1 single supply of 12 volts.  Either a Tek PS280 or battery will suffice.

3.      NOT MET.  Our circuit will provide 120 VAC from 30 – 80 Hz, but below 30 Hz, our transformer begins to saturate and will damage components if allowed to operate below 30 Hz for any length of time.

4.      MET.  Our circuit will accurately indicate the frequency of operation to +- 1% of 60 Hz.  Also, only one LED is on at any time.  The Yellow LED is on for frequencies between 30 Hz and 59.5 Hz.  The Green LED is on between 59.5 Hz and 60.5 Hz.  And the Red LED is on for frequencies of 60.5 Hz to 80 Hz.

5.      MET.  The output voltage is measured at 123.8 Volts when connected to a motor.  The frequency is to within 1 Hz of the indicated frequency on our LED visual indicator.

6.      MET.  As specified above in 4.

Equipment used:

1.      Tektronix PS280 Power Supply

2.      Fluke 45 dual display multimeter

3.      HP 3312A function generator

4.      HP 54600A oscilloscope

5.      Design Kit

6.      Surplus equipment in lab

7.      radio shack components with instructor approval

Recommendations

            Given a second chance to do this circuit any way possible, I would prefer that the analog nature of the circuit be replaced with much simpler and more powerful digital devices.  With a 1 microcontroller, a simpler and cheaper device could be built with an accuracy greatly exceeding that of this analog circuit.  For example, with a microprocessor controlled VFPI the frequency could be display on an LCD screen to an accuracy better than .1 Hz.  Also, I would look for a different way to accomplish the inversion process.  A transformer capable of supplying the demands of a 10 horsepower motor would weigh in excess of 100 lbs.  this approach is simply not feasible for transportation where weight becomes an important design factor.  For relatively small loads, a transformer will suffice, but if more power is demanded of the circuit (power levels approaching 1000 watts) a transformer becomes expensive and heavy.  A different inversion process might also allow frequencies approaching 0 Hz.  With more time, and more money, I believe I could design a very stable and reliable device capable of powering a golf cart.

Results and Analysis

            The results of this circuit indicate that it is not suited for heavy loads, but rather it can handle small loads relatively easily.  However, I do not feel that I  have tested the true potential of this circuit.  It has been designed to operate at an output power of 60 watts, and not test I have conducted brought it near to that power output.  I do believe though that this circuit would not be able to light a 60 watt light bulb due to the non ideal transformer that we were using.  I feel that this transformer would saturate at loads approaching 30 watts, and further loading would simply result in wasted power.  I am however, very proud of actually designing and constructing the circuit and getting it to operate as required.