INVERTER

AC MOTOR CONTROLLER PAGE 1 APRIL 26, 2000

SINUSOIDAL PWM OPERATIONOF AN AC INDUCTION MOTOR CONTROLLER

CONTENTSAbstract

I. Introduction.

II. Design Overview.

A. AC-to-DC Converter.

B. PWM Generator.

C. Gate Driver.

D. DC-to-AC Inverter.

III. Theory of AC Motor Controller Operation.

A. Faraday’s Law.

B. Torque-speed characteristics.

C. Inductive reactance.

D. Duty Cycle.

E. Volts per Hertz Ratio.

F. Synchronous PWM.

G. Gate Driver.

H. Fundamental and Harmonics.

J. Switching Edges.

K. Overshoot.

L. Frequency-Speed Relationship.

IV. Efficiency and Reliability Considerations.

A. Efficiency Considerations.

1. Zero-voltage switching.

2. DC-to-DC conversion.

3. Deadtime distortion.

B. Reliability Considerations.

1.Snubber circuits.

2. Electrical isolation.

3. Overcurrent protection.

3

3

4

9

19


5. Source conversion.

6. Thermal protection.

7. Layout.

V. Results.

A. Converter Module.

B. PWM Module.

1. Variable Width PulseGeneration.

2. Switcher Pulse.

C. Inverter Module.

1. Analog Switchers.

2. Gate Drivers.

3. AC Synthesis.

VI. Conclusion.

References.

Appendix.

A. Specification.

B. Assembly Test Form

C. Performance Test Form

D. Component Data Sheets

25

35

36

37


SINUSOIDAL PWM OPERATION

OF AN AC INDUCTION MOTOR CONTROLLER

Curtis Nelson, Kelly McKeithan, Tom McDonough, Mark Gelazela

ABSTRACT.A single-phase ac induction motor controller is presented and the PWM (pulse-

width-modulated) frequency control part of the operation is verified experimentally. Theapplication for this ac motor controller is existing single-phase ac induction motors lessthan ½ hp. The ac motor controller is a VFD (variable frequency drive). Control is by avoltage source PWM inverter that uses IGBTs for power transistors. The IGBTs areconfigured in a full H-bridge with unipolar voltage switching. A variable frequency (45Hz to 75 Hz) output waveform is generated by the inverter to run a motor at variablespeeds that is directly proportional to this range of frequencies. The objective of thisproject is to construct an ac motor controller from scratch with as many base componentsas possible, and to demonstrate the sinusoidal PWM operation of the inverter part of thedesign that runs an ac induction motor at variable speeds.

I. INTRODUCTION. Numerous motor driven appliances operate in our homes and businesses today(refrigerators, air-conditioners, washers, dryers, basement water pumps etc.). Most ofthese appliances run on single-phase ac induction motors less than ½ horsepower. Andmost of those motors lack a proper motor controller in order to run the motor moreefficiently. Motors can run more efficiently by varying the speed of the motor to matchthe load. Motors are rated to operate best at full load. A motor controller that can vary thespeed of the motor automatically as the load changes will save energy. Fifty percent ofelectrical energy is consumed by motors. An estimated 10% of this is wasted at idle andan additional 5% to 10% is wasted when the motor operates at less than full load.

Therefore, the purpose of this paper is to describe an ac motor controller that canbe applied to existing single-phase ac induction motor appliances, and to demonstratesinusoidal PWM operation of the converter-inverter phase of the design. The prototype ofthis design experimentally verifies the key concept of pulse waveform generation toproduce the switching action of an inverter in order to form a synthesized sine wave thatruns an ac induction motor.

The motor controller of this design is a converter-inverter configuration. Theconverter supplies dc voltage to the inverter by rectifying the 120V, 60 Hz incoming acsignal from the wall outlet to a dc voltage. The ac voltage is also stepped down for thelow-voltage control circuits.

The PWM drives the gates of the power transistors in the inverter. The invertersynthesizes an ac sine wave from the dc voltage by the switching intervals of its power


transistors. The switching intervals are determined from the duration of the pulse fromthe PWM. The ac induction motor then synthesizes this chopped signal from the inverterinto an ac sine wave, because the motor’s inductance smoothes out the “notches” in thewaveform. The motor will then run at a speed proportional to the frequency of this signal.This design is intended to be an interface between the ac source and the motor in order toregulate the motor speed of an appliance to match the load efficiently. The loadrequirements are determined by feedback from sensors to the PWM waveformgenerators.

The prototype demonstrates waveform generation of the switching action of theinverter to run an ac induction motor. The small-signal waveforms from the PWM shapethe high voltage sinusoidal output waveform that runs the motor. The sine wavedetermines the frequency of the output waveform applied to the load. The triangle wavedetermines the switching frequency by the power transistors (IGBTs) of the inverter. Theresulting pulses from the comparison of the sine and triangle waves turns on the switches.The output waveform varies from 45 Hz to 75 Hz. The speed of the motor varies in directrelationship to the frequency.

II. DESIGN OVERVIEW.This design overview section will briefly review the relevant issues of motor

controller design. Fig.1 below is a block diagram of the ac motor controller. A sectioncovering theory of ac motor controller operation, then a section on efficiency andreliability considerations, and then the results of the experiment with the prototype willfollow this.

Fig.1. Block diagram of the ac motor controller showing the ac-to-dc conversion, thePWM pulse, and the dc-to-ac inverter.


The motor controller shown in Fig.1 operates by converting the ac input signalinto a dc signal through the converter. It then converts the dc signal back into acontrollable ac signal through the inverter in order to run the ac induction motor at avariable speed.

Part of the dc voltage from the converter operates wave generators in the PWMthat then send pulses in the form of square waves through the gate driver and into theinverter. There the pulses turn on and off transistor switches at a designed rate in order tosynthesize an ac sine wave at a variable frequency.

A. AC-TO-DC CONVERTER.The converter section of the motor controller converts ac to dc. Both high and low

voltage dc are required. The converter rectifies the incoming ac signal for the highvoltage requirement.

The converter also supplies the low voltage requirements. A transformer stepsdown the incoming ac signal and then this low ac voltage is rectified into a dc signal inorder to power the low voltage PWM generators and control circuits.

The purpose of the rectifier is to convert an ac signal into a dc voltage. However,unlike voltages cannot be directly connected or else KVL problems will result. If thediode bridge rectifier connects an ac voltage with a dc voltage, the input current can beextreme. The ratio between the peak input current and the average output current wouldbe high in this arrangement. The higher the capacitance, the briefer the total on-time, andthe higher the current spike will need to be to transfer the necessary energy. So asmoothing inductor acting as a current transfer source should be added to the output ofthe rectifier in order to improve its performance.

B. PWM GENERATOR.The purpose of the PWM component of the controller is to generate pulses that

trigger the transistor switches of the inverter. The pulse-width modulated signal is createdby comparing a fundamental sine wave from a sine-wave generator with a carrier trianglewave from a triangle wave generator as shown in Fig.2 below.

The variable width pulses from the PWM drives the gates of the switchingtransistors in the inverter and controls the duration and frequency that these switches turnon and off. The frequency of the fundamental sine wave of the PWM determines thefrequency of the output voltage of the inverter. The frequency of the carrier triangle waveof the PWM determines the frequency of the transistor switches and the resulting numberof square notches in the output waveform of the inverter.


Fig.2. PWM operation. V1 is compared to Vcarrier. For each time period, T, a squarepulse operates the switch of the inverter to output the fundamental waveform Vo1. [3,p.212].

A figure of the PWM waveforms together with the resulting pulse is shown belowin Fig.3.

Fig.3. PWM operation. The square pulse from the PWM is superimposed on the sine andtriangle waves as shown in this figure. The pulse is high during the interval when the sinewave is greater than the triangle wave. [2, p.223].

The square pulse waveform that is formed from the sine and triangle waves drivesthe gates of the transistor switches in the inverter and controls the duration and frequencythat these switches turn on and off. The dotted line sine wave in Fig.3 represents both thelow voltage PWM generated sine wave, and also the high voltage output waveform fromthe inverter that drives the motor.

C. GATE DRIVER.The gate driver receives the logic-level control signal generated by the PWM and

then conditions this signal to drive the gates of the power transistors of the inverter. Thegate drivers provide a floating ground for the high-side switching of the IGBTs in a fullH-bridge configuration. The gate voltage must be higher than the emitter voltage forhigh-side switching. The input components have to be level-shifted from common to theemitter voltage. This is accomplished by charging a bootstrap capacitor, which iscomposed of a capacitor and diode network, that gates the high-side IGBTs. The gate


driver also provides under-voltage protection to ensure the gate stays on for the durationof the turn-on period. The gate driver also provides fast switching edges, by minimizingturn-on and off times, in order to prevent the IGBTs from operating in a high dissipationmode and overheating.

D. DC-TO-AC INVERTER.The purpose of the inverter is to convert the dc signal into an ac signal with a

variable frequency. The output waveform from the inverter is a series of square wavesthat the motor ‘sees’ as a sine wave because the inductance of the motor smoothes outthis “chopped” waveform. The amplitude of the synthesized sine wave is determined bythe widths of these square waves. The relative widths of these square waves represent theapplied voltage. The wider the widths and the narrower the notches between the widths,the higher the amplitude of the synthesized sine wave because more voltage is beingapplied.

1. H-BRIDGE.The inverter consists of an H-bridge, which is a configuration of power

transistors. A full H-bridge for single-phase application using IGBTs (in the positionswhere switches are shown) for the power transistors is shown in Fig.4 below.

Fig.4. Single-phase full H-bridge inverter. The switches are IGBTs. The topology is forbipolar voltage switching. [3, p.212].

AT and BT are the IGBT switching transistors. AD and BD are free-wheelingdiodes. Free-wheeling diodes are used to clamp the motor's kickback voltage, as well asto steer the motor's current during normal PWM operation.

2. SWITCH.

The symbol for the transistor switch in the inverter that the PWM controls isshown in Fig.5 below:


Fig.5. Power transistor (IGBT) that does the switching in the inverter. The gate is turnedon by applying 15 Vdc to G. Current then conducts from C to E.

The high-voltage dc signal from the rectifier is applied to the collector (C). When thepulse from the PWM arrives at the gate (G), the switch is turned on and the transistorconducts for the duration of the pulse.


III. THEORY OF AC MOTOR CONTROLLER OPERATION.Faraday's Law sets the saturation flux for maximum energy conversion. The

torque-speed characteristics of a motor are affected by the applied voltage and frequency.Inductive reactance can increase current and heat up the motor. The duty cycle is ameasure of the ratio of the applied voltage and frequency to the motor. The volts-per-Hertz ratio is set by the sine and triangle waves of the PWM and controlled by a PLL.Synchronous PWM reduces harmonics. The gate driver conditions the pulse that drivesthe IGBT gates. Harmonics are a function of the frequency modulation index. Transistorswitching edges dissipate power. Overshoot occurs at switching. The synchronous speedof the motor is directly related to the applied frequency.

A. FARADAY'S LAW.

The ac induction motor requires a constant volts per Hertz ratio in the sinusoidalsignal that it receives in order to operate at saturation flux as reported in [1, 2, 3].Saturation flux represents the highest value for a machine to maximize the energyconversion process, so the motor can supply its rated torque.

The constant volts per Hertz ratio is explained by Faraday’s Law:

∫∫ •−=•sc

daBdtddSE

The line integral of the electric field intensity, E, around a closed contour is equal to thetime rate of change of the magnetic flux, B, linking that contour.

In magnetic structures with windings, like a motor, the E field in the wire isextremely small and can be neglected, so that the first term reduces to the negative of theinduced voltage, e, at the winding terminals. The flux in the second term is dominated bythe core flux φ. Since the winding (and hence the contour C) links the core flux N times,Faraday’s Law reduces to [1, p.10]:

tV

dtdNe

and ω

λφλ

φ

cose dtde

N

0

since

=

=

=

=


fV

t

tVso

πλ

ωω

λ

ωλ

2

sinV

:

cosdtd

0max

0

0

:isflux saturation theTherefore,

gintegratin

=

=

=

Where φ is the magnetic flux in Webers, and λ is the flux linkage in Weber-turns.

So in order to achieve maximum flux linkage, a constant volts per Hertz ratio must bemaintained.

B. TORQUE-SPEED CHARACTERISTICS.

The final speed of the motor is determined by the point in which the load torqueequals the generated torque of the motor as shown in Fig.6 below.

Fig.6. Torque-speed characteristics of a motor operating at saturation flux from stall to 0slip for operation with no change in voltage, frequency, or speed. [4. p.266].

From the figure it can be seen that the final speed of the motor occurs at a low slip ofaround 0.1.

1. EFFECTS OF CONSTANT VOLTAGE.

If the motor does not maintain a constant volts per Hertz ratio, the torque speedcurve will not maintain that straight line characteristic around 0 slip shown in Fig.6 that isrequired for saturation flux. The following figure (Fig.7) illustrates maintaining constantvoltage while varying the frequency.


Fig.7. Torque-speed characteristics for constant voltage and variable frequency. Themotor will increase speed but lose saturation flux at higher speeds. [4, p.270].

Notice from Fig.7 that the torque drops off at higher frequencies causing the motor tostall with an applied load.

2. EFFECTS OF CONSTANT FREQUENCY.

The next figure (Fig.8) illustrates the effects of constant frequency on the torque-speed curve.

Fig.8. This torque-speed curve illustrates the effects of varying the voltage whilemaintaining a constant frequency. The motor does not increase speed, but the torquedrops off as the voltage decreases. [4, p.270].

Notice from Fig.8 that the speed of the motor remains constant no matter the voltage; andthe motor loses its torque characteristics with decreasing voltage.

3. VARIABLE VOLTAGE AND FREQUENCY.

The figure below (Fig.9) illustrates the effects of varying the voltage andfrequency proportional to the duty cycle.


Fig.9. Torque-speed characteristics for a constant volts per Hertz ratio. The motormaintains saturation flux at all speeds. [4, p.271].

Notice from Fig.9 that the motor can now supply its rated torque and operate at saturationflux to provide the highest possible force over the full range of frequencies.

C. INDUCTIVE REACTANCE.

If the frequency is varied while the voltage remains constant, the motor willmaintain good torque-speed characteristics at low frequencies (as shown in the precedingFig.7). However excessive currents will be generated in the motor and heat it up. This isdue to the inductive reactance of the motor:

fL 2π=LX

The inductive reactance adds to the resistance and increases the current at lowfrequencies:

fL 2π+=

+=

RV

XRVI

L

So as the frequency decreases, the current increases. The figure below (Fig.10) shows thiseffect:

Fig.10. The Current, I, will increase at low frequencies due to the inductive reactance,XL, of the motor if a constant volts per Hertz ratio is not maintained. [4, p.126].


So, in order to prevent motor overheating due to current rise, the constant volts per Hertzratio must be maintained.

D. DUTY CYCLE.

The output ac sinusoidal wave that is generated by the switch action of thetransistor switches in the inverter matches the sine wave that is generated by the PWM.The duty cycle is a measure of the amplitude and frequency of this sine wave in relationto the amplitude and frequency of the triangle wave. The duty cycle changes based on thesinusoidal waveform as a reference.

The amplitude modulation index, am , is the ratio of the amplitude of thefundamental sine wave, 1V , to the amplitude of the carrier triangle wave, cV (from Fig.2).

am = cV

V1

The frequency modulation index, fm , is the ratio of the frequency of the carriertriangle wave, cf , to the frequency of the fundamental sine wave, 1f .

1ffm c

f =

The amplitude modulation index, am , is usually less than or equal to one, whichmeans that the amplitude of the carrier wave is greater than or equal to the amplitude ofthe sine wave as shown in Fig.11 below.

Fig.11. Ratio of the amplitude of the sine wave to the triangle wave and the resultingpulse intervals (square wave). [2, p.222].

The amplitude of the sine wave increases with am as shown in Fig.11. The amplitude ofthe triangle wave remains constant.

Overmodulation occurs for am greater than one and could result in the inverterwaveform degenerating into a square wave. Overmodulation is used for high powerlevels; in which case the magnitude of the dc input must be adjusted.


E. VOLTS PER HERTZ RATIO.The amplitude modulation index, am , should vary in proportion to the

fundamental frequency. For example (for a 75 Hz, 150 Vrms maximum operation), if thefundamental frequency reduces to 80% from 75 Hz to 60 Hz, then am should be .80 inorder to maintain a constant volts per hertz ratio (the output to the motor would then be60 Hz, 120 Vrms).

Choosing the frequency of the carrier triangle wave and the resulting switchfrequency is a trade-off between high frequency, for ease in filtering harmonics, and lowfrequency, for reduced switching losses. Frequencies in the audible range of about 6 kHzto about 15 kHz are usually avoided to reduce noise.

A constant volts per hertz ratio can be obtained using a phase-lock loop (PLL).The PLL drives the PWM. The PLL synchronizes the frequencies and regulates the dutycycle. A block diagram of the PLL is shown in Fig.12 below.

Fig.12. The PLL (Phase Lock Loop) inputs a frequency, ωin, and outputs a proportionalvoltage, Vo, and a synchronous frequency, ωout, to drive the waveform generators of thePWM.

The PLL operates by comparing an input frequency, ωin (as shown in Fig.12), tothe output frequency of the loop, ωout, through a multiplier, , by the phase detector,PD. The input signal isAsin(ωt+θi).The output signal isBcos(ωt+θo).The product isAsin(ωt+θi) Bcos(ωt+θo).The trigonometric identity for this expression isAB/2[sin(θi-θo) +sin(2ωt+θi+θo)].The second term of this expression is then filtered out by the lowpass filter and amplifiedto produce Vo proportional to ωin. The free running frequency, ωref, is initially set to amidrange frequency (60 Hz for a range of 45 Hz to 75 Hz). Then as Vo changes witheach loop, ωref changes proportionally. As the phase difference decreases (θi-θo), thephases lock and the outputs are proportional to the input. A multiplier, N, determines ωin,the input frequency of the carrier triangle wave, which is then divided out by the 1/Ndivider in the PLL.


F. SYNCHRONOUS PWM.The frequency modulation index, fm , should be chosen as an odd integer amount

(such as 21, 47, or 301) which results in odd symmetry as well as half-wave symmetry sothat the even harmonics disappear from the waveform. If the sine wave and the trianglewave vary together in proportion to the ratio fm , then the waveform is a synchronousPWM. Asynchronous PWM results in subharmonics that are undesirable. fm is oftenchosen to be greater than 21. For example, if cf = 2100 Hz and 1f = 60 Hz, then fm =35. A more realistic representation of the sine wave in relation to the generated pulse isshown in Fig.13 below.

Fig.13. Typical sine wave generation from the switch action of an inverter. There areusually a high number of switchings for every sine wave generated..[2, p.225].

This sine wave is the output waveform from the inverter that the motor ‘sees’ andmatches the sine wave generated from the PWM. The inductance of the motor smoothesout this waveform. The square waves are generated by the switch action of the inverter.They match the frequency of the triangle wave generated from the PWM. Notice fromFig. 13 that the sine wave is high when the widths of the square waves are wider at thetop, and low when wider at the bottom. This waveform has a fm of about 60 and a am ofabout 0.5.

G. GATE DRIVER.

The gate driver receives the logic-level control signal generated by the PWM andthen conditions this signal to drive the gates of the power transistors of the inverter. Thefunction of the gate driver for an IGBT or a MOSFET, with their voltage controlled gatecharacteristics, is capacitor charging and discharging. The transistor that is controlled bya gate driver is modeled with capacitors across each of its terminals. The gate driver mustprovide enough charge to account for the substantial current generated by the fast switchaction. And the gate driver must perform at the highest speeds possible.

The gate driver performs the following functions:1. It minimizes turn-on and off times.2. It provides adequate drive power to keep the power switch on.


3. It provides reverse bias in order to ensure the switching device remains in the off-state.4. It amplifies the control signal.5. It provides electrical isolation where required.6. It provides deadtime (blanking time).7. It provides protection from overcurrent (feedback would be required).8. It provides a large current for initiating turn-on, then large gate voltage at low current

levels for the duration of the turn-on period.The gate driver is the interface between the control circuit and the power switch.

The gate driver topology is determined by either unipolar or bipolar voltageswitching, electrical isolation, and either a parallel or series (cascode) connection.Unipolar voltage switching reduces harmonics. Bipolar voltage switching (BVS) is usedto speed up the switching. The power transistors will turn off as fast as they turn on withBVS. Bipolar voltage switching provides a reverse bias applied to the power switch thatwill ensure fast turn-off. Electrical isolation, when required, is specified as anoptocoupler, fiber optics, or a transformer. The drive circuit is usually shunt (parallel)connected so that the drive circuit conducts only a fraction of the current carried by thepower switch in the on state.

H. FUNDAMENTAL AND HARMONICS.The fundamental voltage and the harmonics for an example of

47m and 9. f ==am is calculated below for a normalized voltage and a 60 Hz supply.Since 9.0=am , the fundamental frequency is )60(9.01 Hzf = = 54 Hz. The harmonicsoccur at:

kHzfmkHzfmkHzfm

f

f

f

69.12)(5)(47)(54 )(5614.7)(3)(47)(54 )(3538.2)(1)(47)(54 )(

1

1

1

==

==

==

as shown in Fig.14 below.

54 Hz 2.5 kHz 7.6 kHz 12.7kHz

Fig.14. Fundamental and harmonic components for bipolar voltage switching of a full H-bridge.


J. SWITCHING EDGES.

Power dissipation in switching occurs at turn-on and turn-off as shown in figure15 below. There is also some power loss due to the small amount of on voltage while theswitch is on. Referring to the figure, prior to the on pulse to the switch, the voltage ishigh and the current is off. After the switch is turned on, there is a delay then the currentramps on as the voltage ramps off. This overlap of the voltage turning off and the currentturning on is a power loss, P = IV. This occurs at turn-off also when the current ramps offwhile the voltage ramps on.

Fig.15. The top graph represents the pulse from the PWM. The bottom graph is theresulting switch action of the inverter. The triangular areas where Vce and Ic ramp onand off (as well as the narrow area at the bottom between them, Von) represents powerloss, P=IV. [3, p.21].

The power dissipated due to the switching action is: )( )()(21

offConcSOdS ttfIVP +=

And the power due to the on voltage while the current is also on is: sonoonon ftIVP =

K. OVERSHOOT.

In addition to the losses due to the switching edges, there are losses due toovershoot. The ideal I-V trajectory (the current and voltage trajectory from turn-on toturn-off and back) is shown by the dashed line in Fig.16 below.

Fig.16. Preferred (dashed line) and actual trajectory of switching action and theresulting overshoot. When the switch is off, voltage is high but no current conducts. Whenthe switch is on, the voltage across the switch drops to near zero while the current ishigh.


However, during turn-on and turn-off there is an overshoot, shown by the arrowsin Fig.16, resulting in overcurrent and overvoltage. The value of this additional powerloss can be determined by subtracting the calculated values of S plus PonP from themeasured value of the actual switch loss.

L. FREQUENCY-SPEED RELATIONSHIP.

Variable speed operation of the motor is obtained by varying the frequency of thePWM pulses with sensory input to the sine wave generator in the PWM. Since the speedof an AC induction motor is directly proportional to this applied frequency,

Synchronous speed = 120 (frequency) / (number of poles),the motor speed can be varied by the frequency. The actual speed of the motor shaft isdetermined by the load which sets a slip in the motor.


IV. EFFICIENCY AND RELIABILITY CONSIDERATIONS.Since the ac motor controller operates with high voltages and currents, there are a

number of efficiency and reliability issues to consider.

A. EFFICIENCY CONSIDERATIONS.The ac signal can be rectified into dc voltage and then stepped up with a boost

converter to run the motor at voltages and speeds above the base inputs. Resonant filterscan be added to filter out unwanted harmonics. Since deadtime causes distortionproducing torque pulsations on the motor shaft, a correction term synchronized to thephase current signal with a current sensing device can be applied to the PWM waveformto counter-modulate the original PWM signal with the characteristics of this distortionand provide noise cancellation.

Choosing the frequency of the carrier triangle wave and the resulting switchfrequency is a trade-off between high frequency, for ease in filtering harmonics, and lowfrequency, for reducing the switching losses. Frequencies in the audible range of about 6kHz to about 15 kHz are usually avoided to reduce noise. The frequency modulationindex should be chosen as an odd integer. This results in odd symmetry as well as half-wave symmetry so that the even harmonics disappear from the waveform.

1. ZERO-VOLTAGE SWITCHING.A snubber circuit across the power transistor will improve the shaping of the

switching trajectory of the switch as it turns off. The turn-off snubber provides zerovoltage across the transistor while the current turns off. A configuration for a snubbercircuit around a power transistor is shown below in Fig.17.

Fig.17. The snubber circuit is composed of a diode, SD , a resistor, SR , and a capacitor,

SC . It protects the power transistor from heat stress. [3, p. 682]

The capacitor, SC of Fig.17, delays the voltage turning on long enough for thecurrent to turn off. The turn-off snubber capacitor value is:

d

foS V

tIC

2=


The capacitor energy, which is dissipated in the snubber resistor is:

2

2DS

RVCW =

Some of the advantages of zero-voltage switching using a turn-off snubber circuitare the following: The capacitor energy is dissipated in the resistor, which can be cooledeasier than the transistor. The transistor requires no additional energy due to the turn-off.And the transistor's peak current is not increased because of the turn-off snubber. Theturn-off snubber circuit also reduces dv/dt which minimizes EMI problems. A turn-onsnubber circuit is seldom used because the turn-on snubber inductance must carry theload current, which is expensive to implement.

2. DC-TO-DC CONVERSION.A technique that could be used to provide more efficient operation of the motor

controller would be to use the stored voltage from the snubber circuit described in thepreceding section to power the PWM generators. The snubber resistor, SR , shown inFig.17, could be replaced with a dc-to-dc converter. This dc voltage could then be usedby the PWM instead of being dissipated in the resistor.

3. DEADTIME DISTORTION.As reported in [4], deadtime distortion is a result of the deadtime added to the

PWM signal by the gate driver. This deadtime is necessary to prevent the top and bottomswitches of the inverter from shorting out, but it causes voltage and current waveformdistortion when the inverter is driving an inductive load like a motor as shown in Fig.18below. The net result is the production of torque pulsations on the motor shaft. Theproblem is more severe at low voltages due to SNR.

Fig.18. The effects of deadtime distortion on the voltage waveform of the inverter [4,p.308].

A bipolar square wave that is synchronized to the phase current signal can be usedto approximate this distortion of the phase voltage. Then the original PWM signal can becounter-modulated with the characteristics of this distortion in order to provide noisecancellation. This correction term is synchronized to the phase current signal with acurrent sensing device and applied to the PWM waveform.


B. RELIABILITY CONSIDERATIONS.There are also a number of product reliability factors to consider in order

to ensure end user satisfaction. Snubber circuits can be added to reduce stress on powersemiconductor devices by storing overvoltages from the switching action in the capacitorand then dissipating the power in the resistor instead of the switches. Transientovervoltage suppressors can be placed across the input of the controller and the inverterswitches for voltage spikes. Optocouplers can be included as part of the gate drivers forelectrical isolation between logic-level control signals and the power stage. Deadtimeprevents the switches from shorting out by spacing the turn-on time of oppositely biasedswitches, and can be built into the gate drivers with a few passive components. Protectionagainst overcurrent from stalled or shorted motor conditions can be provided with a shuntresistor in the ground path that feeds back to a comparator in the PWM in order to initiateshutdown of the gate signals. A transfer source in the form of a smoothing inductor canbe provided between the diode bridge rectifier and the rectifier capacitor for sourceconversion.

The junction temperature of the power transistor switches can be controlled with atemperature sensor to shut down the power stage at critical temperatures. The circuitlayout should be planned to minimize loop areas that can pick up noise, cancelelectromagnetic fields with parallel runs, avoid 90° angles that cause discontinuity andunwanted reflections, connect the control signal ground from the PWM to a single pointon the power stage of the inverter, and eliminate excessive conductor length to minimizeinductive voltage kickback. Insulation breakdown due to the switching action of thePWM operation may occur. Therefore, lower switching frequencies, a short cable lengthbetween the controller and the motor, or an inverter duty motor should be considered. Inaddition to this, transient overvoltage suppressors should be placed across the input andthe switches in order to ensure survival of the controller when voltage spikes occurduring prototyping.

1. SNUBBER CIRCUITS.Snubber circuits reduce stress on power semiconductor devices by storing

overvoltages from the switching action in the capacitor (shown as SC in Fig.17) and thendissipating the power in the resistor, SR . Resonant converters may also used to controlswitching stresses.

A snubber circuit should also be used for protection on a single-phase diodebridge rectifier as shown in Fig.19 below. For the case of continuous conduction, thefilter inductor (shown as smoothingL ) should be placed on the dc side.


Fig.19. An RC snubber circuit and a MOV used to protect a diode bridge rectifier. [3, p.677].

One RC snubber can be used to protect all the diodes. In addition to the RC snubber, aMOV (metal-oxice varistor) is used for transient overvoltage protection.

2. ELECTRICALLY ISOLATED DRIVE CIRCUITS.Electrical isolation between logic-level control signals and drive circuits can be

obtained from optocouplers, fiber optics, or transformers. Optocouplers are preferredover transformers because they are as effective but not as bulky. Optocouplers are usedwith electrical shields between the LED and the receiver transistor in order to avoidretriggering the power transistor at both turn-on and turn-off. The configuration shown inFig.20 below can be used for MOSFETs and IGBTs.

Fig.20. Drive circuit with optocoupler for an IBGT or a MOSFET. The optocouplerisolates the switch from the control circuitry. [3, p.707].

3.OVERCURRENT PROTECTION.Current limiting should be used to protect against stalled or shorted motor

conditions as reported in [3, 13]. There are three common overcurrent modes. One is ashort caused by the motor leads shorting out. Another is a ground fault from insulationbreakdown. The third is shoot through caused by false turn-on of an IGBT. Fuses do notact fast enough to protect power devices. Fast detection of overcurrents compared to alimit are needed in order to turn off the power transistor in the gate driver by means of aprotection network. The short-circuit current can be estimated from the transistor I-Vcharacteristics as shown in Fig.21 below.


Fig.21. I-V characteristics of a semiconductor device to determine overcurrentprotection. [3, p. 718].

A safe value to detect overcurrent would be about twice the continuous current rating ofthe transistor

A shunt resistor in the ground path can detect overcurrent conditions. From theInternational Rectifier product document, one of the traditional protection networks isdescribed: "One can detect the line-to-line short and shoot through currents by inserting aHall-effect sensor or a linear opto-isolator across the shunt resistor. The device should bein series with the negative dc bus line. For ground fault protection, an additional Hall-effect leakage current sensor could be placed either on the ac line input or on the dc bus.The protection circuit is then implemented by using fast comparators. The output of thesecomparators is 'Or'd' with the …PWM generator to initiate the shutdown of the gatesignals" [13]. The IR2137 integrated monolithic IC device is available to perform thissame function.

4. SOURCE CONVERSION.An ac motor controller performs source conversion. The input ac voltage is

converted to a dc voltage through the rectifier. However, this may cause KVL (Kirchoff’svoltage law) problems since unlike voltages appear to be connected. Therefore, a transfersource in the form of a smoothing inductor should be placed between the diode bridgerectifier and the rectifier capacitor as was shown in Fig.19. Since the inductor representsa current source, the incoming voltage converts to a current and then this current convertsto a voltage for the inverter.

Krein describes the process this way, “The source conversion concept is afundamental of power electronics. In a well designed power converter, both the input andoutput ought to have the characteristics of an ideal source. If the input is a voltage source,then the output should resemble a current source. If the input is a current source, then theoutput should have properties of a voltage source" [2, p.87].

5. THERMAL PROTECTION.One of the most critical reliability design criteria for a controller is to reduce the

junction temperature of the power transistor switches, as reported in [2, 4]. This will helplower the mechanical stress level and prolong the life of the transistor. For each 10°C risein the junction temperature, the long-term reliability of the transistor is reduced by 50%.

Thermal resistance, θR , is a measure of the temperature change of a semi-conductor material per the applied power level. The case-to-heatsink thermal resistance is


the thermal resistance of the interface material times the average material thicknessdivided by the area of the device mounted on the heatsink:

AtR CS ρθ =

The junction-to-case thermal resistance, JCRθ is equal to the change in temperaturedivided by the power dissipation:

PDtoTTR CJ

JC =θ

The power dissipation, PD, is equal to the current, CI , times the on voltage of the IGBT,.

ONCEV ,

ONCEC VIPD , =The current draw for calculating the thermal resistance occurs at locked rotor conditionand 100% duty cycle. This represents the worst case for IGBT power dissipation.

Silicone grease is used to ensure a contact surface for heat conduction but mayeventually dry out, so thermal conduction interface pads may be used instead. Atemperature sensor may be added to shut down the power stage at critical temperatures. Anatural air-convection heatsink is used to dissipate the unwanted heat from the powerstage.

6. LAYOUT.As reported in [3, 4], proper circuit layout is critical to the total design of a motor

controller. Valentine suggests the following rules for layout: Minimize loop areasbecause loops are antennas that can pick up noise and affect the power stage. Run thesignal and return close together in order to cancel the electromagnetic fields in the wires.Avoid 90° angles on wires that carry high-speed signals because this discontinuity willproduce unwanted reflections. And connect the control signal ground from the PWM to asingle point on the power stage of the inverter because transient voltage drops can besubstantial along power grounds due to the high values of di/dt that flows through a finiteinductance [4, p.233].

Inductive voltage kickback, PKV , is caused by conductors with excessive lengthand is calculated by the following equation,

dtdiLVPK =

where the change of current with time is determined by the controller’s switching speed.And the inductance of the conductors is calculated by

HdllL µ ) 75. 4log3026.2)(002(.

−××=

where l is the length of the conductor, and d is the diameter. Therefore, increasing thediameter of the conductor has a minimum effect on stray inductance compared todecreasing the length. In general, leads should be kept under an inch if possible.


V. RESULTS.This results section analyzes the three modules that comprise the ac motor

controller: the converter module, the PWM module, and the inverter module. Theconverter module consists of a step-down transformer and two diode bridge rectifiers.The PWM module consists of a triangle wave generator, a Wien bridge sine wavegenerator and a comparator. The inverter module consists of a full H-bridge with 4IGBTs, two analog switchers, and two gate drivers.

A. CONVERTER MODULE.The circuit diagram for the converter module is shown in Fig.22 below. The

incoming ac signal is sent to both the high and low-voltage sections of the converter. Thediode bridge rectifier of the high-voltage section uses diodes that are rated for 15 A, 250V operation. The diodes of the rectifier form a dc voltage across the resistor andcapacitor. The smoothing inductor, L1, is a transfer current source between the ac inputand the dc output to overcome source conversion. The PTCs are positive temperaturecoefficient thermistors. They dampen the inrush current at turn-on. When the current andresulting temperature rises, the resistance increases. When the current and resultingtemperature falls off with motor operation, the resistance falls off with it for normaloperation. The temperature rises within microseconds of the current rise. An 11kΩ powerresistor is required on the output of the high-voltage section due to the high currents.

Fig.22. This is the converter module of the ac motor controller. The high-voltage sectionsupplies the 170 V dc signal for the inverter. The low-voltage section supplies the +15Vdc and –15 Vdc signals to run the op-amps of the PWM module and the control circuitsof the inverter module.

S1

R4

C2

1.2mF

Dbreak

D9

D7L1

400uH

Dbreak

D8

0

PTC therm

TX1

D8

0

C4

1.2mF

C3

1.2mF

DbreakD7

R3

C1

1.2mF

V1

PTC therm

D10D9

R211k

C1

Dbreak

D10

HIGHVOLTAGESECTION

-170 Vdc

+

LOWVOLTAGESECTION

- 15 Vdc

+ 15 Vdc

CONVERTER MODULE


The low-voltage section of the converter supplies the small signal voltages for thePWM and the control circuits of the inverter. This section consists of a step-downtransformer, a diode bridge rectifier, a capacitor, and two resistors. The negative 15 Vdcis obtained by placing a ground between the output resistors as shown in the figure.

B. PWM MODULE.The PWM module is an analog, sinusoidal pulse generator. This module consists

of op-amps that generate the necessary waveforms and pulses to operate switchingtransistor gates. A schematic of the PWM module is shown in Fig. 23 below.

Fig.23. The PWM module of the ac motor controller showing the triangle wavegenerator, the variable frequency Wien bridge sine wave generator, the comparator, theswitcher pulse generator, and the PWM pulse inverter.

The triangle wave generator provides a high frequency triangle wave whoseamplitude is independent of the frequency. The generator consists of an integrator as aramp generator and a threshold detector with hysteresis as a reset circuit. The thresholddetector is implemented by using positive feedback around an op-amp. Frequency isdetermined from the capacitor and resistor circuit, R1 and C1 shown in Fig.23,

TRIANGLE WAVE

SINEWAVE

BUFFER

BUFFER

COMPARATOR

SWITCHER PULSE

10k

PULSE INVERTER

PWM PULSE

-

+

PWM MODULE

R1

C1

C2

C3

R2a

R3a

R2bR3b


f=1/(2πR1C1), as well as from the positive and negative saturation voltages of theamplifier. The frequency of the triangle wave is 2.1kHz.

The Wien bridge sine wave generator provides variable frequency pulses tooperate the motor at variable speeds. The frequency of the sine wave is determined fromthe resistor-capacitor networks shown in Fig.23. R2a and R3a are NTC (negativetemperature coefficient) thermistors for heat sensing. R2b and R3b are floatation switchresistors for water level sensing. These resistors vary with sensory input. The frequencyof the sine wave then varies with the change in resistance from the relationship,f=1/(2πRC). For heat sensing, the resistor-capacitor networks of R2a and C2, and R3aand C3 vary the frequency of the sine wave. For water level sensing, the resistor-capacitor networks of R2b and C2, and R3b and C3 vary the frequency of the sine wave.The frequency and speed of the motor varies with sensory input to these resistors.

The fundamental sine wave and the carrier triangle wave are both generated andcompared in the PWM section of the circuit. The two resistor-capacitor networks shownin lower left of the above figure determine the frequency of the fundamental sine wave;therefore, in order to meet the specification that the frequency of the inverter vary, one ofthe resistor-capacitor parameters in each of the networks has to vary. For this purpose, theresistors were chosen to be thermistors in combination with floatation switch resistors.The thermistors allow that the frequency of the wave applied to the motor will increase ordecrease with a temperature change. The floatation switch resistors allow the frequencyapplied to the motor to increase or decrease with an increase or decrease in surroundingwater height for a water pump application. These components allow for automaticsensory feedback to the PWM section in order to speed the motor up or down if, forinstance, the temperature in the room rises and the VFD is hooked to a fan or if the waterin the basement rises and the VFD is hooked to a water pump.

The buffer amplifiers are used to connect the source resistance to the loadresistance in order to prevent significant signal attenuation.

The comparator op-amp receives both the sine wave and the triangle wave signalsas inputs. The comparator then outputs a variable width, square wave pulse, generated bythe intersection of these two waveforms that is called a PWM pulse. This PWM pulseoperates the gates of selected switching transistors in the inverter to conduct current tothe motor in one direction. An inverting op-amp inverts this PWM pulse which thenoperates the diagonally opposite gates of the inverter to conduct current in the otherdirection.

Comparing the sine wave to ground generates a switcher pulse. The switcherpulse outputs a square wave pulse with a 50 % duty cycle. This pulse has the samefrequency as the sine wave and is used to direct the PWM pulse and the inverted PWMpulse through the motor.

1. VARIABLE WIDTH PULSE GENERATION.One of the functions of the PWM module is to generate a variable width pulse (called

a PWM pulse) that gates the switching transistors of the inverter in order to synthesize anac signal to run the motor. The mechanism for producing this pulse is shown in Fig.24below.


Fig.24. Variable width pulse generation with a sine wave and a triangle wave (PWMpulse).

A low frequency sine wave is compared to a high frequency triangle wave togenerate a variable width pulse. From Fig.24 it can be seen that the pulse is high (on)when the amplitude of the sine wave is greater than the amplitude of the triangle wave,and low (off) when the sine wave is less than the triangle wave. The width of the pulse isdetermined by the successive intersections of the waveforms as shown by the dashedlines in the figure. The pulses are wider when the sine wave is high (indicating a longerduration that the pulse is on), and narrower when the sine wave is low (indicating ashorter duration that the pulse is on).

The variable width pulses from the PWM control the duration and frequency thatthe switching transistors turn on and off. The frequency of the fundamental sine wave ofthe PWM determines the frequency of the output waveform to the motor. The frequencyof the carrier triangle wave of the PWM determines the frequency of the transistorswitches in the inverter and the resulting number of square notches in the outputwaveform.

2. SWITCHER PULSE.

In addition to the PWM pulse, the PWM module also generates a switcher pulseshown in Fig. 25 below. The switcher pulse operates at the same frequency as the outputwaveform to the motor, and is the mechanism that alternates the current to the motor.

PWM

PULSE

SINE WAVE TRIANGLE WAVE


Fig.25. The graphs on the left show the mechanism for generating the switcher pulse.When the sine wave is high, the switcher pulse is on (logic 1). When the sine wave is low,the switcher pulse is off (logic 0). The graphs on the right show the three pulsesgenerated by the PWM module to operate the inverter module. Only the wide pulses(representing maximum voltage output) are passed to the inverter, as indicated in thefigure.

The switcher pulse is constructed by comparing the sine wave generated by thePWM with ground; and thus the switcher pulse has the same frequency as this sine waveas well as the same frequency as the output waveform. The switcher pulse determineswhich gate driver is on. Notice from the graphs on the right of the figure above that thePWM pulses to the left of the red dashed line (over the period that the switcher pulse ison) are identical to the inverted PWM pulses to the right of the red dashed line (over theperiod that the switcher pulse is off). These three pulses are the pulses that form theoutput waveform.

C. INVERTER MODULE.

The function of the inverter module is to control the gating of the switchingtransistors in the proper sequence in order to synthesize an ac signal from a dc input. Theschematic of the inverter module is shown below in Fig.26.

The gate drivers condition the pulses from the PWM to provide a floating ground forhigh-side switching of the IGBTs in a full H-bridge configuration. Gate voltage must be10-15 volts higher than the emitter voltage to fully utilize the full switching capabilities.Being a high side switch, the gate voltage would have to be higher than the rail voltage,which is frequently the highest voltage available in the system. The gate voltage must becontrollable from the logic, which is normally referenced to ground. Thus, the controlsignals have to be level-shifted to the emitter of the high side power devices.

SWITCHER

PULSE

SINE WAVESWITCHER

PULSE

PWM

PULSE

INVERTED

PWM

PULSE

PWM OUTPUT PULSES

PWM

ON OFF

ON OFF

PASSED

PASSED


Fig.26. The inverter module of the ac motor controller consists of an H-bridge with fourIGBT power switching transistors, two analog switchers, and two gate drivers.

Under these constraints, several techniques are presently used to perform thesefunctions The design presented here utilizes a bootstrap circuit, which is the most simple andeconomical approach. The gate driver chosen was an International Rectifier control IC,IR21094, for high voltage, high speed power MOSFET and IGBT drivers with dependenthigh and low side referenced output channels. The logic inputs are fully compatible withstandard CMOS or LSTTL output. The output driver features a high pulse current bufferstage designed for minimum cross-conduction. The floating channel drives an N-channelIGBT in the high side configuration, which operates up to 600 volts.

The operation of the bootstrap circuit utilizes a capacitor and diode network to gatethe high IGBT. The Vbs voltage (the voltage difference between the Vb and Vs pins on thecontrol IC) provides the supply to the high side driver circuitry of the control IC that in mostcases is a high frequency square wave. The supply needs to be in the range of 10-20 volts toensure that the control IC can fully gate the IGBT. Under-voltage protection circuitry ensuresthat the IC does not drive the IGBT below a certain voltage. This prevents the IGBT fromoperating in a high dissipation mode.

When Vs is pulled down to ground (either through the low side field effect transistoror the load), the bootstrap capacitor (Cbs) charges through the bootstrap diode (Dds) from the15-Volt Vcc supply. Thus providing a supply to Vbs, as the figure above illustrates (Fig.26).

There are five factors which contribute to the supply requirements from the Vbscapacitor. These are:

1. Gate Charge required to enhance the FET

IN

LO

3

NC1 VSS

DT

V-

VSS

S2

100

14

IN

5

DT2200K

1uVSS1 S1

COM

NC3

VB

170 VDC

MOTOR

100

HO

IN VB

NC30

NC V+

NC2

Z1

V-

V+

D1

DT

Z2

S2

Z4

Z3

LO

10

V+

HO

100

D2

12

10

3 12

0

NC1

D2

5

NC2

SD

COM

VS1u

V+

SD

GND

100

IN

DT1200K

6

D1

GND

NC

9

H-BRIDGE

SWITCHERS

INVERTER MODULE GATE DRIVERS


2. Iqbs-queiscent current for the high side driver circuitry3. Currents within the level shifter of the control IC4. FET gate-emitter forward leakage current5. Bootstrap capacitor leakage current

Note: Factor 5 is only relevant if the bootstrap capacitor is an electrolytic capacitor, and canbe ignored if other types of capacitor are used. The following equation details the minimumcharge, which needs to be supplied by the bootstrap capacitor:

Qbs = 2*Qg + Iqbs(max)/f + Qls + Icbs(leak)/f (1)Where:Qg = gate charge of high side FETIcbs(leak) = Bootstrap capacitor leakage currentQls = level shift charge required per cycle = 5nC(500V/600V IC’s) or 20 nC (1200V IC’s)f = frequency of operation

The bootstrap capacitor must be able to supply this charge, and retain its full voltage,otherwise there will be a significant amount of ripple on the Vbs. voltage which could fallbelow the Vbsuv undervoltage lockout and cause the HO output to stop functioning. Thecharge in the capacitor must be a minimum of twice the above value. The minimumcapacitance value can be calculated from the equation below:

C > (2*(2*Qg + Iqbs(max)/f + Qls + Icbs(leak)/f)) (2)Vcc – Vf - VlsWhere:Vf = Forward voltage drop across the bootstrap diodeVls = Voltage drop across the low side FETNote: The capacitance above is the absolute minimum required, however, a general ruleshould apply by multiplying the above factor by 15 to ensure proper charging of thecapacitor.

The bootstrap diode (Dbs) needs to be able to block the full power rail voltage, whichis seen when the high side device is switched on. It must be a fast recovery device tominimize the amount of charge fed back from the bootstrap capacitor into the Vcc supply.Therefore:Diode characteristics:Vrrm = Power rail voltageMax. Irr =100nsIf =Qbs * f

Resistor-diode networks are on the input of each gate of the H-bridge. Thesenetworks serve two purposes, they prevent conduction overlap of switching intervals byadding dead-time, and reduce the peaking of the current spike during reverse recovery timeof the free-wheeling diodes (not shown) across the IGBTs. Conduction overlap occurs whenthe switches that conduct in one direction through the motor do not completely turn offbefore the switches that conduct in the opposite direction turn on. A shoot-through currentwould result and the switches would short out. Deadtime delays turn-on to prevent thiscondition. Deadtime is also inserted into the PWM signal by resistors DT1 and DT2 on theDT posts of the gate drivers.


The following analysis of ac synthesis divides the inverter into the left and right-sideswitchers and gate drivers. The left-side operates during the first half-cycle of the switcherpulse. The right-side operates during the second half-cycle of the switcher pulse.

1. ANALOG SWITCHERS.The analog switchers provide the mechanism for directing the current to the motor.

Diagrams of the analog switchers and their operation are shown below in figures 27 and 28.

a. FIRST HALF-CYCLE.The first half-cycle of the switcher pulse, when the switcher pulse is a logic one (on)

passes the PWM pulse, as shown below in Fig.27. Since the first half-cycle of the switcherpulse corresponds to the first half-cycle of the PWM sine wave, the PWM pulses are widest.Therefore, the gates of the IGBTs will turn on for a longer period of time to supply themaximum applied voltage to the motor.

Fig.27. During the first half-cycle of the switcher pulse, S1, of the left-side switcher, isswitched on when a logic 1 (on) pulse arrives at the IN post from the PWM module. Theright-side switcher does not conduct. Ground is provided to the non-conducting S2 of theleft-side switcher and S1 of the right-side switcher to reduce noise.

b. SECOND HALF-CYCLE.The second half-cycle of the switcher pulse, when the switcher pulse is a logic zero

(off) passes the inverted PWM pulse, as shown below in Fig.28. Since the second half-cycleof the switcher pulse corresponds to the second half-cycle of the PWM sine wave, theinverted PWM pulses are also at their widest over this period to maximize the appliedvoltage to the motor.

1S

2S2D

1D

INSWITCHER PULSE

INVERTED

PWM PULSE

1S

2S

INSWITCHER PULSE

PWM PULSE

2D

1D

ANALOG SWITCHERS

First half-cycle

LEFT-SIDE RIGHT-SIDE


Fig.28. During the second half-cycle of the switcher pulse, S2, of the right-side switcher, isswitched on when a logic 0 (off) pulse arrives at the IN post from the PWM module. The left-side switcher does not conduct.

2. GATE DRIVERS.The gate drivers provide the mechanism for turning on the switching transistors

(IGBTs) of the inverter. Diagrams of the gate drivers and their operation are shown below infigures 29 and 30.

a. FIRST HALF-CYCLE.The first half-cycle of the output waveform is synthesized by the H-bridge from the

left-side gate driver (as viewed from the previous Fig.26) shown in Fig.29 below.

Fig. 29. The left-side gate driver (in this analysis) forms the first half-cycle of the outputwaveform from the first half-cycle of the switcher pulse and the PWM pulse.

During the first half-cycle of the switcher pulse, the left-side gate driver receives thePWM pulse from the left-side switcher. The high output, Ho, conducts when the PWM pulse

1S

2S2D

1D

IN

INVERTED

PWM PULSE

SWITCHER PULSE

1S

2S 2D

1D

INSWITCHER PULSE

PWM PULSE

ANALOG SWITCHERS

Second half-cycle

LEFT-SIDE RIGHT-SIDE

First half ofSwitcher cycle

Vdc

MOTOR

GRD

PWM PULSEVb

Ho

Vs

Lo

V+

IN

DT

Vss

Comm

IR21094

+15 Vdc

Z1

Z4

H-Bridge

Left-Side Gate Driver OUTPUT WAVEFORM

First half-cycle


is on (logic 1). The low output, Lo, is open when Ho conducts. Lo defaults to normallyclosed when there is no pulse, or a logic 0 applied to the IN post.

Ho is attached to the high-side switching transistors (170 Vdc). Lo is attached to thelow-side switching transistors (ground). From the figure it can be seen that Z1 will be gatedon and off by the PWM pulse from Ho. Z4 will stay gated on over this half-cycle intervalbecause it is attached to the low output (Lo) of the right-side gate driver that is not receivinga pulse at this time. So the right-side Lo defaults to normally closed and the transistor, Z4, isgated on.

The result is that current will conduct from Z1 to Z4 through the motor. The outputwaveform shown in the figure above is identical to the PWM pulse but at a higher voltage.

b. SECOND HALF-CYCLE.The second half-cycle of the output waveform is synthesized by the H-bridge from

the right-side gate driver (as viewed from the previous Fig.26) shown in Fig.30 below.During the second half-cycle of the switcher pulse, the right-side gate driver receives

the inverted PWM pulse from the right-side switcher. The pattern of the inverted PWM pulseover this half-cycle interval is identical to the PWM pulse over the first half-cycle interval.

From Fig.30 it can be seen that Ho conducts pulses which gate Z3 on and off. Z2stays gated on over this half-cycle interval because the low output, Lo, in the left-side gatedriver is normally closed over this interval.

Fig.30. The right-side gate driver (in this analysis) forms the second half-cycle of the outputwaveform from the second half-cycle of the switcher pulse and the inverted PWM pulse.

The result is that current will conduct in the opposite direction through the motorfrom Z3 to Z2. The output waveform shown in the figure is identical to the inverted PWMpulse but at a higher voltage. Therefore, the conduction of current alternates directionthrough the motor every half cycle to synthesize an ac signal.

Second half ofSwitcher cycle

Vdc

MOTOR

GRD

Vb

Ho

Vs

Lo

IR21094

V+

IN

DT

Vss

Comm

+15 Vdc

PWM PULSE

Z3

Z2

H-Bridge

Right-Side Gate DriverOUTPUT WAVEFORM

Second half-cycle

INVERTEDPWM PULSE


3. AC SYNTHESIS.

The ac output waveform that is synthesized by the inverter module is shown below inFig.31.

Fig.31. Synthesized output waveform that runs the ac induction motor.

The full cycle of the output waveform to the motor looks like the figure above. Analternating current waveform is synthesized by the switching action of the IGBTs. Thisoutput waveform is a series of square waves that the motor ‘sees’ as a sine wave because theinductance of the motor smoothes out this “chopped” waveform. The amplitude of thesynthesized sine wave is determined by the widths of these square waves. The relative widthsof these square waves represent the applied voltage. Increasing the duty cycle results in ahigher synthesized sine wave amplitude because the average applied voltage is higher.

CONCLUSION.This design proves the principle of sinusoidal PWM operation of an IGBT based

inverter, as it applies to existing single-phase ac induction motor appliances. Construction ofthe prototype conforms to the design principles. The rectifier circuit develops a dc voltagewithin the range required by the inverter. The PWM circuit generates a variable width pulsethat turns on the switches in the inverter. The gate driver circuit conditions the PWM pulsewaveform for proper operation of the inverter gates. The inverter circuit utilizes theswitching action of the IGBTs to synthesize a sine wave. The resulting synthesized waveformruns the motor over the specified range of frequencies and speeds. This design demonstratesvariable speed control of single-phase ac induction motors in order to match the loadrequirements for efficient operation. The prototype controller operated only a small fractional

OUTPUT WAVEFORM


horsepower motor, but the design could be theoretically applied to larger motors with theaddition of the efficiency and reliability considerations to the circuit.

REFERENCES1. A.E Fitzgerald, Electric Machinery, 5th., McGraw-Hill, NY, 1990.

2. Philip T. Krein, Elements of Power Electronics, Oxford, NY, 1998.

3. Ned Mohan, Torre M. Undeland, William P. Robbins, Power Electronics, Converters,Applications, and Design, 2nd., Wiley, NY, 1995.

4. Richard Valentine, Motor Control Electronics Handbook, McGraw-Hill, NY, 1998.

5. Adel S. Sedra and Kenneth C. Smith, Microelectronic Circuits, Oxford University Press,NY, 1998.

6. Masayuki Morimoto, Katsumi Oshitani, Kiyotaka Sumito, Shinji Sato, Muneaki Ishida,and Shigeru Okuma, “New single-phase unity power factor PWM converter-invertersystem”, IEEE Power Electronics Specialists Conference, pp. 585-589, 1989.

7. Michael A. Boost, Phoivos D. Ziogas, “State-of-the-art carrier PWM techniques: a criticalevaluation”, IEEE Trans. Ind. Applicat., vol. 24, no. 2, pp. 271-280, March/April 1988.

8. J. F. Bangura, N. A. Demerdash, “Simulation of inverter-fed induction motor drives withpulse-width modulation by a time-stepping coupled finite element-flux linkage-based statespace model”, IEEE Transactions on Energy Conversion, Vol. 14, No. 3, September, 1999.

9. Omar Stihl, Boon-Tech Ool, “A single-phase controlled-current PWM rectifier”, IEEETransactions on Power Electronics, vol. 3, No. 4, October, 1988.

10. Ken Berringer, “AC motor drive using integrated power stage”, Motorola semiconductorapplication note AN1524/D, Motorola, Inc., 1996.

11. Norbert R. Malik, Electronic Circuits, Analysis, Simulation, and Design, Prentice Hall,Englewood Cliffs, NJ, 1995.

12. International Rectifier, "HV Floating MOS-Gate Driver Ics", Document INT978,International Rectifier.

13. International Rectifier, "Solving IGBT Protection in AC or BLDC Motor Drive",Document address: www.irf.com/product-info/motor/igbtprotect.pdf, International Rectifier.

http://www.irf.com/product-info/motor/igbtprotect.pdf


Documents

INVERTER