ST AN2794 APPLICATION NOTE

AN2794

Application note

1 kW dual stage DC-AC converter based on the STP160N75F3

Introduction

This application note provides design guidelines and performance characterization of the STEVAL-ISV001V1 demonstration board.

This board implements a 1 kW dual stage DC-AC converter, suitable for use in batterypowered uninterruptible power supplies (UPS) or photovoltaic (PV) standalone systems.

The converter is fed by a low DC input voltage varying from 20 V to 28 V, and is capable of supplying up to 1 kW of output power on a single-phase AC load. These features are possible thanks to a dual stage conversion topology that includes an efficient step-up pushpull DC-DC converter, which produces a regulated high-voltage DC bus and a sinusoidal H- Bridge PWM inverter to generate a 50 Hz, 230 Vrms output sine wave. Other key features of the system proposed are high power density, high switching frequency and efficiency greater than 90% over a wide output load range

Figure 1. 1 kW DC-AC converter prototype

January 2012

Doc ID 14827 Rev 2

1/39

www.st.com

Contents	AN2794

Contents

1	System description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .		. 5
2	Design considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .		. 8
	2.1	Layout considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .	20
3	Schematic description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .		23
4	Experimental results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .		26
5	Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .		30
6	Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .		30
Appendix A		Component list. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .	31
Appendix B Product technical specification . . . . . . . . . . . . . . . . . . . . . . . . . . . .			35
7	Revision history . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .		38

2/39	Doc ID 14827 Rev 2

AN2794	List of tables

List of tables

Table 1. System specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Table 2. Push-pull converter specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Table 3. HF transformer design parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Table 4. Output inductor design parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Table 5. Power MOSFET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Table 6. Diode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Table 7. Bill of material (BOM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Table 8. Document revision history . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

Doc ID 14827 Rev 2

3/39

List of figures	AN2794

List of figures

Figure 1. 1 kW DC-AC converter prototype . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Figure 2. Block diagram of an offline UPS system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Figure 3. Possible use of a DC-AC converter in standalone PV conversion . . . . . . . . . . . . . . . . . . . . 5 Figure 4. Block diagram of the proposed conversion scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Figure 5. Push-pull converter typical waveforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Figure 6. Distribution of converter losses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Figure 7. Distribution of losses with 3 STP160N75F3s paralleled . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Figure 8. Component placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Figure 9. Top layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Figure 10. Bottom layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Figure 11. Schematic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Figure 12. Characteristic waveforms (measured at 24 V input voltage and 280 W resistive load) . . . 26 Figure 13. Characteristic waveforms (measured at 28 V input voltage and 1000 W resistive load) . . 26 Figure 14. MOSFET voltage (ch4) and current (ch3) without RC snubber . . . . . . . . . . . . . . . . . . . . . 27 Figure 15. MOSFET voltage (ch4) and current (ch3) with RC snubber . . . . . . . . . . . . . . . . . . . . . . . . 27 Figure 16. Rectifier diode current (ch3) and voltage (ch4) without RDC snubber . . . . . . . . . . . . . . . . 27 Figure 17. Rectifier diode current (ch3) and voltage (ch4) with RDC snubber. . . . . . . . . . . . . . . . . . . 27 Figure 18. Ch1, ch3 MOSFETs drain current, ch2, ch4 MOSFET drain-source voltage . . . . . . . . . . . 28 Figure 19. Startup, ch2, ch3 inverter voltage and current, ch4 DC bus voltage . . . . . . . . . . . . . . . . . 28 Figure 20. DC-DC converter efficiency with 20 V input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Figure 21. DC-DC converter efficiency with 22 V input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Figure 22. DC-DC converter efficiency with 24 V input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Figure 23. DC-DC converter efficiency with 26 V input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Figure 24. DC-DC converter efficiency with 28 V input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Figure 25. Converter efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Figure 26. Technical specification for 1.5 mH 2.5 A inductor L4 (produced by MAGNETICA) . . . . . . 35 Figure 27. Technical specification for 1 kW, 100 kHz switch mode power transformer TX1

(produced by MAGNETICA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Figure 28. Dimensional drawing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

4/39	Doc ID 14827 Rev 2

AN2794	System description

1 System description

In a UPS system, as shown in Figure 2, a DC-AC converter is always used to convert the DC power from the batteries to AC power used to supply the load. The basic scheme also includes a battery pack, a battery charger which converts AC power from the grid into DC power, and a transfer switch to supply the load from the mains or from the energy storage elements if a line voltage drop or failure occurs.

Figure 2. Block diagram of an offline UPS system

AC/DC	DC/AC	SWITCH
	Battery

Another application where a DC-AC converter is always required is shown in the block diagram of Figure 3. In this case, the converter is part of a conversion scheme commonly used in standalone photovoltaic systems. An additional DC-DC converter operates as a battery charger while performing a maximum power point tracking algorithm (MPPT), which is necessary to maximize the energy yield from the PV array. The battery pack is always present to store energy when solar radiation is available and release it at night or during hours of low insolation.

Figure 3. Possible use of a DC-AC converter in standalone PV conversion

DC/DC

Battery

Charger

+

MPPT

DC/AC

LC Filter

Batteries

Load

A possible implementation of an isolated DC-AC converter, which can be successfully used in both the above mentioned applications, is given in the block diagram of Figure 4. It consists of three main sections:

1.The DC-DC converter

2.The DC-AC converter

3.The power supply section

Doc ID 14827 Rev 2

5/39

System description	AN2794

Figure 4. Block diagram of the proposed conversion scheme

	3TEP UPUSTAGES 0USH 0ULL			)NVERTERE3TAGE3 ( "RIDGE
				34'7 .# 7$$
	340 . &		,
	48		$
		$		:	:
			#	,		,	6OUT
6IN		$	$
?				:	:
	-	-					?
	3'	344( 2		34 &LITE
		.
	,$				6	0OWER
						3UPPLY
				,	6 3ECTION
		34. .& ,
						!-V

The DC-DC section is a critical part of the converter design. In fact, the need for high overall efficiency (close to 90% or higher) together with the specifications for continuous power rating, low input voltage range leading to high input current, and the need for high switching frequency to minimize weight and size of passive components, makes it a quite challenging design.

Due to the constraints given by the specifications given in Table 1, few topology solutions are suitable to meet the efficiency target. Actually, since the input voltage of the DC-AC converter must be at least equal to 350 V, it is not feasible to use non-isolated DC-DC converters. Moreover, the output power rating prevents the use of single switch topologies such as the flyback and the forward. Among the remaining isolated topologies, the half bridge and full bridge are more suitable for high DC input voltage applications and also characterized by the added complexity of gate drive circuitry of the high side switches.

Table 1.	System specifications
	Specification		Value
Nominal input voltage			24 V
	Output voltage		230 Vrms, 50 Hz
	Output power		1kW
	Efficiency		90%
	Switching frequency		100 kHz (DC-DC); 16 kHz (DC-AC)

Due to such considerations, the push-pull represents the most suitable choice. This topology features two transistors on the primary side and a center tapped high frequency transformer, as shown in the step-up section in Figure 4. It is quite efficient at low input voltage making it widely used in battery powered UPS applications. Both power devices are ground referenced with consequent simple gate drive circuits. They are alternatively turned

6/39	Doc ID 14827 Rev 2

AN2794	System description

on and off in order to transfer power to each primary of the center tapped transformer. Contemporary conduction of both devices must be avoided by limiting the duty cycle value of the constant frequency PWM modulator to less than 0.5. The PWM modulator should also prevent unequal ON times for the driving signals since this would result in transformer saturation caused by the "Flux Walking" phenomenon.

The basic operation is similar to a forward converter. In fact, when a primary switch is active, the current flows through the rectifier diodes, charging the output inductor, while when both the switches are off, the output inductor discharges. It is important to point out that the operating frequency of the output inductor is twice the switching frequency.

A transformer reset circuit is not needed thanks to the bipolar flux operation, which also means better transformer core utilization with respect to single-ended topologies.

The main disadvantage of the push-pull converter is the breakdown voltage of primary power devices which has to be higher than twice the input voltage. In fact, when voltage is applied to one of the two transformer primary windings by the conduction of a transistor, the reflected voltage across the other primary winding puts the drain of the off state transistor at twice the input voltage with respect to ground. This is the reason why push-pull converters are not suitable for high input voltage applications.

For the above mentioned reasons, the voltage fed push-pull converter, shown in Figure 4, is chosen to boost the input voltage from 24 V to a regulated 350 V, suitable for optimal inverter operation. The high voltage conversion ratio can be achieved by proper transformer turns ratio design, taking into account that the input to output voltage transfer function is given by:

Equation 1

Vout = 2 N2 DVin

N1

The duty cycle is set by a voltage mode PWM regulator (SG3525) to keep a constant output DC bus voltage. This voltage is then converted into AC using a standard H-bridge converter implemented with four ultrafast switching IGBTs in PowerMESH™ technology, switching at

16 kHz. The switching strategy, based on PWM sinusoidal modulation, is implemented on an 8-bit ST7lite39 microcontroller unit. This allows the use of a simple LC circuit to obtain a high quality sine wave in terms of harmonic content.

The power supply section consists of a buck-boost converter to produce a regulated 15 V from a minimum input voltage of 4 V. The circuit can be simply implemented by means of a L5973 device, characterized by an internal P-channel DMOS transistor and few external components. In this way, it is possible to supply all the driving circuits and the PWM modulator. A standard linear regulator, L7805, provides 5 V supply to the microcontroller unit.

Doc ID 14827 Rev 2

7/39

Design considerations	AN2794

2 Design considerations

The basic operation of a voltage fed push-pull converter is shown in Figure 5, where theoretical converter waveforms are highlighted. In practice, significant overvoltages across devices M1, M2 and across the four rectifier diodes are observed in most cases due to the leakage inductance of the high frequency transformer. As a consequence, the breakdown voltage of primary devices must be greater than twice the input voltage, and the use of snubbing and/or clamping circuits is often helpful.

Special attention has to be paid to transformer design, due to the difficulties in minimizing the leakage inductance and implementing low-voltage high-current terminations. Moreover, imbalance in the two primary inductance values must be avoided both by symmetrical windings and proper printed circuit board (PCB) layout. While transformer construction techniques guarantee good symmetry and low leakage inductance values, asymmetrical layout due to inappropriate component placement can be the source of different PCB trace inductances. Whatever the cause of a difference in peak current through the switching elements, transformer saturation in voltage mode push-pull converters can occur in a few switching cycles with catastrophic consequences.

Figure 5. Push-pull converter typical waveforms

8/39	Doc ID 14827 Rev 2

AN2794	Design considerations

Starting from the specifications in Table 2, a step-by-step design procedure and some design hints to obtain a symmetrical layout are given below.

Table 2.	Push-pull converter specifications
	Specification	Symbol	Value
Nominal input voltage		Vin	24 V
Maximum input voltage		Vinmax	28 V
Minimum input voltage		Vinmin	20 V
Nominal output power		Pout	1000 W
Nominal output voltage		Vout	350 V
Target efficiency		η	> 90%
Switching frequency		f	100 kHz

A switching frequency of f = 100 kHz was chosen to minimize passive components size and weight, then the following step-by-step calculation was done:

●Switching period:

Equation 2

T =	1	=		1	=10 μs
			105
	f

●Maximum duty cycle

The theoretical maximum on time for each phase of the push-pull converter is:

Equation 3

t*on = 0.5T = 5 μs

Since deadtime has to be provided in order to avoid simultaneous device conduction, it is better to choose the maximum duty cycle of each phase as:

Equation 4

Dmax = 0.9 t*on = 0.45 T

This means a total deadtime of 1μs at maximum duty cycle, occurring for minimum input voltage operation.

●Input power

Assuming 90% efficiency the input power is:

Equation 5

Pin = Pout =1111W

0.9

Doc ID 14827 Rev 2

9/39

Design considerations				AN2794
● Maximum average input current:
Equation 6
I =	Pin	=	1111	= 55.55 A
in	Vinmin	20

●Maximum equivalent flat topped input current:

Equation 7

I	=	Iin	=	55.55	= 61.72 A
pft		2Dmax	0.9

●Maximum input RMS current:

Equation 8

IinRMS =Ipft 2Dmax = 58.55A

●Maximum MOSFET RMS current:

Equation 9

IMosRMS = Ipft Dmax = 41.4A

●Minimum MOSFET breakdown voltage:

Equation 10

VBrk Mos = 1.3 • 2 • VinMax = 72 .8 V

● Transformer turns ratio:

Equation 11

N =	N2		=		Vout		=19
	N1				2Vinmin Dmax
● Minimum duty cycle value:
Equation 12
Dmin =					Vout	= 0.32
				2NVinmax

● Duty cycle at nominal input voltage:

Equation 13

Dmin = Vout = 0.38

2NVin

●Maximum average output current:

Equation 14

Iout = Pout = 2.86 A

Vout

10/39

Doc ID 14827 Rev 2

AN2794	Design considerations

●Secondary maximum RMS current

Assuming that the secondary top flat current value is equal to the average output value the rms secondary current is:

Equation 15

IsecRMS = Iout Dmax =1.91 A

●Rectifier diode voltage:

Equation 16

Vdiode = NVinMax = 532 V

●Output filter inductor value:

Equation 17

		≥(	N		ton
	Lmin		2	Vin - Vout )	Max
			N1		I

Assuming a ripple current value I= 15% Iout = 0.43A, the minimum value for the output filter inductance is:

Equation 18

Lmin =1.109 mH

With this value of inductance continuous current mode (CCM) operation is guaranteed for a minimum output current of:

Equation 19

IoutMin = 2I = 0.215A

which means a minimum load of 75 W is required for CCM operation. The chosen value for this design is L=1.5 mH.

●Output filter capacitor value:

Equation 20

C =	1		IL	T
8				s
			V0

Considering a maximum output ripple value equal to:

Equation 21

V0 = 0.1%Vout = 0.35 V

Doc ID 14827 Rev 2

11/39

Design considerations	AN2794

the minimum value of capacitance is:

Equation 22

Cmin =1.53 μF

and the equivalent series resistance (ESR) has to be lower than:

Equation 23

ESR =	V0	= 0.81Ω
max	IL

●Input capacitor:

Equation 24

	Cin	=ICrms	TonMax
			Vin

where Icrms is the RMS capacitor current value given by:

Equation 25

ICrms = IIn2 Rms - Iin2 =19A

and

Equation 26

Vin = 0.1%VinMax = 0.028V

then

Equation 27

	C	in	=I	Crms	TonMax	= 3053 μF
					Vin

12/39

Doc ID 14827 Rev 2