Design of GaN-Based MHz Totem-Pole PFC Rectifier - IEEE Xplore

15 downloads 0 Views 3MB Size Report
Abstract—Totem-pole bridgeless power factor correction. (PFC) rectifier is recently recognized as a promising front end candidate for applications like server and ...
Design of GaN-based MHz Totem-pole PFC Rectifier Zhengyang Liu, Fred C. Lee, Qiang Li, Yuchen Yang Center for Power Electronics Systems The Bradley Department of Electrical and Computer Engineering Virginia Polytechnic Institute and State University Blacksburg, VA 24061 USA Abstract—Totem-pole bridgeless power factor correction (PFC) rectifier is recently recognized as a promising front end candidate for applications like server and telecommunication power supply. In this paper the advantage, enabled by emerging high-voltage gallium-nitride (GaN) devices, of totem-pole PFC rectifier comparing to traditional PFC rectifier is discussed in the beginning. Critical mode operation is used by the totem-pole PFC rectifier in order to achieve both high frequency and high efficiency. Then several high frequency issues and detailed design considerations are introduced including zero-voltageswitching (ZVS) extension for entire line-cycle ZVS operation; variable on-time strategy for zero-crossing distortion suppression; and interleaving control for ripple current cancellation. The volume reduction of differential-mode (DM) electro-magnetic interference (EMI) filter significantly benefited from MHz high frequency operation and multi-phase interleaving is also presented. Finally, a 1.2kW GaN-based MHz totem-pole PFC rectifier is demonstrated with 99% peak efficiency and 200W/in3 power density. Keywords—GaN; Totem-pole PFC; MHz; CRM;

I. INTRODUCTION With the advent of 600V gallium-nitride (GaN) power semiconductor devices, the totem-pole bridgeless power factor correction (PFC) rectifier, which was a nearly abandoned topology, is suddenly become a popular solution for applications like front-end converter for server and telecommunication power supply. This is mostly attributed to the significant performance improvement of the GaN highelectron-mobility transistor (HEMT) compared to silicon (Si) metal–oxide–semiconductor field-effect transistor (MOSFET), particularly better figure-of-merit and significantly smaller body diode reverse recovery effect. GaN-based hard-switching totem-pole PFC rectifier is demonstrated several times in literature [1-3]. As the reverse recovery charge of the GaN HEMT is much smaller than the Si MOSFET, hard-switching operation in totem-pole bridge configuration turned to be practical. By limiting switching

frequency around or below 100kHz, the efficiency could be above 98% for a 1kW level single-phase PFC rectifier. Even the simple topology and high efficiency are attractive, the system level benefit is limited because the switching frequency is still similar to Si-based PFC rectifier. Based on previous study, soft switching truly benefits the cascode GaN HEMT [4-7]. As the cascode GaN HEMT has high turn-on loss and extremely small turn-off loss due to the current-source turn-off mechanism [5], critical mode (CRM) operation is very suitable. A GaN-based critical mode (CRM) boost PFC rectifier is also demonstrated which shows the high-frequency capability of the GaN HEMT and significant system benefits as the volume of the boost inductor and the DM filter is dramatically reduced [8,9]. With a similar system-level vision, the cascode GaN HEMT is applied in the totem-pole PFC rectifier while pushing frequency to above 1MHz. Several important highfrequency issues, which used to be less significant at lowfrequency, are emphasized and the corresponding solutions are proposed and experimentally verified. To address these issues, the advantages of the totem-pole PFC rectifier is summarized at first, while the difference between hard-switching and soft-switching, and between Si MOSFET and GaN HEMT are also illustrated in section II. After that, detailed design considerations are presented in section III including ZVS extension in order to solve switching loss caused by non-ZVS valley switching; variable on-time control to improve the power factor, particularly the zero-crossing distortion caused by traditional constant ontime control; and interleaving control for input current ripple cancellation. The volume of DM filter is reduced significantly by pushing frequency to several MHz and further multi-phase interleaving. Section IV is hardware prototype and preliminary experimental results. In the last is summary and conclusion.

This work was supported in part by the Power Management Consortium (PMC) in CPES, Virginia Tech aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa aaaaaaaaaaaaaaa

978-1-4673-7151-3/15/$31.00 ©2015 IEEE

682

II. TOPOLOGY COMPARISON BETWEEN TOTEM-POLE BRIDGELESS PFC RECTIFIER AND DUAL-BOO OST BRIDGELESS PFC RECTIFIER Bridgeless PFC rectifier has clear advanttages [10, 11] by eliminating the diode rectifier bridge so thaat the conduction loss of power semiconductor devices is reduced. Among boost-type bridgeless PFC rectifier, dual--boost bridgeless PFC rectifier is popular in industry produccts because it has less conduction loss compared to boost PFC C rectifier; lower common-mode (CM) noise compared to conventional bridgeless boost PFC rectifier; and only low w-side gate driver compared to others requiring high-side floaating gate driver. However, from the view of topology, totem-pole PFC rectifier is even simpler than dual-boost bridgeless PFC rectifier. As a side-by-side comparison liisted in Table I, totem-pole PFC rectifier eliminates the uusage of siliconcarbide (SiC) schottky diode and requires onnly four switches and one inductor. Therefore it is most sim mplified topology among boost-type bridgeless PFC rectifiers. Even the topology of the totem-pole PFC C rectifier is very simple, it is seldom used due to significannt drawbacks that cannot be overcome by Si MOSFET. With continues-current mode (CCM) hard-switching operation, it can hardly work because there is tremendous turn-on loss andd parasitic ringing due to the reverse recovery effect of the aanti-parallel body diode. CRM totem-pole PFC has been repoorted in literature [12-14], although the previous issue is alleviated by ZCS turn-off of the body-diode and ZVS turn-oon of the control switch, the increased current ripple still leads to higher conduction loss and higher turn-off loss. So Si-based CRM mode totem-pole PFC rectifier is usuallyy limited to low frequency and low power level.

Table I. Component count of dual-boost bridgeless PFC rectifier and totem-pole bridgeless PFC C rectifier Topology Dual-boost Totem-pole 2 (S1/S2) Active Switch 2 (S1/S2) SiC Schottky Diode 2 (D1/D2) 0 Rectifier Diode 2 (DN1/DN2) 2 (DN1/DN2) Total Switch Count 6 4 Inductor Count 2 1

III. DETAILED DESIGN CONSIDER RATIONS OF GAN-BASED MHZ TOTEM-POLE PFC C RECTIFIER A. Valley switching and ZVS extensiion The CRM PFC rectifier utilizess the resonance between inductor and device junction capaccitors to achieve ZVS or valley-switching. For boost-type CRM C PFC rectifier, ZVS can only be achieved when the inp put voltage is lower than one half of the output voltage, assu uming negligible damping effect which is often true with good g design and limited resonant cycles. So when the input voltage v is higher than one half of the output voltage, the drain n source voltage can only resonate to a valley point which iss equals to (2Vin-Vo). So (0.5CV2) loss occurs at next turn-on n instant. The non-ZVS energy of each vallley switching is calculated at each operating point of a halff line cycle according to oss in a half line cycle is equation (1). Then the non-ZVS lo also derived as the product of non-Z ZVS energy and switching frequency according equation (2). The T last step is to do linecycle averaging so that the line-cyclle averaged non-ZVS loss at different input voltage is known according a to equation (3).

The high-voltage GaN HEMT is ablee to extend the application of the totem-pole PFC rectifierr due to its good characteristics. The significantly reduced rreverser-recovery charge of the cascode GaN HEMT makess CCM operation practical at certain frequency range (e.g. 500kHz or 100kHz). Furthermore, as previously discovered that the turn-off loss of the cascode GaN HEMT is extremely ssmall, with CRM operation the switching frequency is able to be pushed to above 1MHz while achieving good efficienncy. A 98% peak efficiency is reported in a MHz CRM booost PFC rectifier which is a clear example of how CRM soft sswitching benefits the cascode GaN HEMT. So the followinng part focus on detailed design aspects of the MHz totem-poole PFC rectifier.

S2

D1

D2 +

L1

VIN

C

L2 DN2

DN1

S1

R

VO -

VIN

(c) Fig. 2. Calculation of non-ZVS loss (a) non-ZVS loss; (b) frequency; and (c) line-cycle averaged no on-ZVS loss

DN2

L

C

R

(b)

(a)

+ VO -

S1

DN1

S2

Eoss (Vin, t) = ͲǤͷைௌௌ ሺʹξʹ௜௡ •‹ɘ– െ ܸ௢ ሻଶ (1) Poss (Vin, t) = ͲǤͷைௌௌ ሺʹξʹ௜௡ •‹ɘ– െ ܸ௢ ሻଶ ݂௦ ሺ௜௡ ǡ ‫ݐ‬ሻ (2) ೟శ೅

(a) (b) Fig. 1. Comparison between (a) dual-boost bridgeless PFC rectifier and (b) totem-pole bridgeless PFC rectiffier

Poss_ave (Vin) =

683

ሾ‫׬‬೟

଴Ǥହେೀೄೄ ሺଶξଶ୚೔೙ ୱ୧୬ன ன୲ି௏೚ ሻమ ௙ೞ ሺ୚೔೙ ǡ௧ሻሿ ்

(3)

As this loss is directly related to the swittching frequency, so when the frequency is pushed to multi-MH Hz level, the nonZVS loss is significant and dominant in thhe total converter loss as shown in Figure 3.

Figure 3. Line-cycle averaged Non-ZVS loss vvs. input voltage

In order to solve this issue, ZVS extensioon strategy [12-14] is used. The concept is to modify the operatiion from CRM to quasi-square-wave (QSW) mode. So insteaad of turning off the synchronous rectifier (SR) right befoore the inductor current zero crossing, a delay is purposely addded so that there is enough initial energy stored in the inductoor to help achieve ZVS after SR turn off. The control of the ZVS extension is crritical because if there is too much SR extra on time, then therre would be more circulating energy; on the other hand if theere is not enough SR extra on time, then ZVS cannot be achievved. The accurate calculation is enhanced by a trajectory annalysis (Figure 4) which clearly illustrates the resonant statuus for CRM and QSW.

Figure 5. Minimum negative current and SR R extra on-time in half line cycle to achieve ZVS exttension

According to this ZVS extension n control, Figure 6 shows the simulated half line-cycle inductor current without and with ZVS extension, while the experimental waveforms (Figure 7) with entire line-cycle ZVS validates the ZVS extension strategy. The saved switching loss is significant because the total efficiency is increaased by 0.3% to 1% from full load to half load which is shown n in Chapter IV.

(a) (b) Figure 6. Half line cycle inductor currrent simulation waveform (a) without ZVS extension and (b) with w ZVS extension

Figure 4. Trajectory of resonance for CRM and Q QSW operations

According to the trajectory, the minnimum required negative current to achieve ZVS is calculated as equation (4). Then the required extra SR conduction time is further calculated with equation (5) in order to achieeve such negative current. The calculation results are also drawn in Figure 5. Within the two dash-line is the non-ZVS zonne which requires the ZVS extension control. ݅௠௜௡ ሺ‫ݐ‬ሻ ൌ

ඥሾଶ௏೔೙ ሺ௧ሻି௏೚ ሿ௏೚ ට௅Τଶ஼೚ೞೞሺ೟ೝሻ

‫ݐ‬ௌோ̴௘௫௧௥௔ ሺ‫ݐ‬ሻ ൌ





௅ඥሾଶ௏೔೙ ሺ௧ሻି௏೚ ሿ௏೚ ሾ௏೚ ି௏೔೙ ሺ௧ሻሿට௅Τଶ஼೚ೞೞሺ೟ೝሻ

(4) (5)

Figure 7. Experimental verification of ZVS extension strategy

684

B. Zero-crossing distortion and variable on-ttime control The second high frequency issue is relaated to the power quality and harmonics emission. Ideally, the CRM mode PFC offers unity power factor with voltage moode (constant ontime) control. Since the on time is constannt, the envelop of the inductor peak current follows the shaape of the input voltage. Then if ignoring the negative ccurrent, the peak current of the inductor is always twice of thee average current, which means the input current always following the input voltage. However, when the frequency is inncreased to MHz range, the negative current during resonaant period is not negligible thus there is notable difference between the shape of peak inductor current and average induuctor current. In addition, there is also a non-energy transfer time around line voltage zero crossing in which the average innductor current is zero. Both of them lead to increased harm monics and poor power factor as shown in Figure 8 and Figuree 9.

Variable on-time control is intrroduced in [15]. Similar concept is used in this paper but with improved and more accurate implementation by digital control in order to solve this issue. The concept is illustrated in Figure 10. By increase the on time near zero crossing, thee input current is able to achieve good power factor agaain. Figure 11 is the experimental verification. The calculation of variable on n time involves massive mathematical works. Different implementations are possible with the trade-off between accuraacy and MCU resources occupation. A real-time calculation n method is practical but requires high-end MCU which increases the total cost of the LUT) is an alternative system. Instead, look-up-table (L solution which preloads several tablles for different input and output conditions.

(a) (a)

(b) (b) Figure 8. Frequency impact on power factor and haarmonics (a) 100kHz constant on-time CRM PFC and (b) 1MHz constant oon-time CRM PFC

Figure 10. Concept diagram of CRM PFC C rectifier with (a) constant ontime control and (b) variable on-time control

(a)

(b) Figure 9. Three operation modes in half line cyclee with voltage mode constant on time operation.

c on-time control and (b) Figure 11. Experimental verification (a) constant variable on-time control

685

An analytical model is built to accurattely calculate the required on time for each operating point in a half line cycle. Figure 12 shows the two operation modes annd corresponding trajectories in a switch cycle, in which Zn is the characteristic impedance in the resonance as shown in eqquation (6). Since the output junction capacitor (COSS) has non-linear characteristics and is a function of voltage, time equivalent output junction capacitor (COSS(tr)) is ussed to calculate resonant time and impedance. The control sswitch on state is defined as the starting point state I. After state I, all the remaining status is determined in mode 1 critical mode operation and mode 2 quasi square wave moode operation. As the trajectory is unique with given cirrcuit design and input/output parameters like Vin and Vo, then the instantaneous current in every switch cycle ccan be derived as a function of on time. So the average currrent can also be derived in a further step. After that, the avverage current is equaled to a sinusoidal reference so thhat the on time distribution in a half line cycle can be ccalculated as the required variable on time table to achieve unnity power factor. ܼ݊ ൌ ට

௅ ଶ஼௢௦௦ሺ௧௥ሻ



(6)

deal with this issue, 2-phase interleeaving structure is used to effectively reduce the DM noise by ripple cancellation effect. r The impact of high frequency on the DM filter is shown in Figure 13 and Figure 14. By pushing frequency 10 times higher, the volume of the DM filteer is reduced at least 50% and a simple one-stage filter is goo od enough to suppress the noise below standard. By good intterleaving, the volume is reduced by another 50%. So in totall, the DM filter is just one quarter size compared to 100kHz DM D filter. More analysis regarding EMI filter design for thiis MHz totel-pole PFC is included in paper [9]. c to achieve good Interleaving control is very critical interleaving and maintain small enough e phase error. The waveform in Figure 15 shows good d interleaving is achieved. So even the current ripple in each phase p is always more than two times higher than the phase in nput current, but the total input current ripple is significantly y reduced by interleaving. The interleaving control is usually not n an issue for frequency below 100kHz but become a ch hallenge for multi-MHz variable frequency CRM PFC. To address a this issue, another paper is in progress.

(a)

Figure 13. Frequency impact on DM M filter corner frequency

Figure 14. DM filter for 100kHz and 1M MHz totem-pole PFC rectifier

(b) RM and (b) QSW Figure 12. Two operations and trajectory (a) CR

C. Dual-phase interleaving and ripple canceellation Another drawback of the CRM PFC recctifier is the high current ripple, which leads to not only higheer conduction loss but also high DM noise compared to CCM PFC rectifier. To

686

Figure 15. Ripple cancellation effect with two phase interleaving

D. Digital control implementation

S12

As no commercial CRM PFC controllerr supports multiMHz operation, it is quite a challenge to impplement the MHz CRM totem-pole PFC and all functions m mentioned above. MCU based digital control is used for quickk implementation. The control diagram is shown in Figure 16. More analysis on this topic will be presented in future publicattion.

S22

SN2

L1

+

L2

C

R

VO -

VIN S11

S21

SN1

(b) Figure 17. MHz totem-pole PFC (a) prototype and (b) topology

Figure 16. MCU configuration for CRM MHz totem-pole PFC

d single phase efficiency Figure 18. Measured and predicted

IV. PROTOTYPE AND EXPERIMENTA AL RESULTS The system diagram and prototype is shoown in Figure 17. The power rating is 1.2kW and the power density achieved by this prototype is around 200W/in3. Tessted efficiency is close to 99% peak with ZVS extension strateegy. The inductor design is another key factor to achieve bothh high efficiency and high power density, which has been elaaborated in paper [16]. The discrete cascode GaN HEMT T has capacitor mismatch issue [7] which causes extra losss. With proposed solution, a full-bridge GaN module [17, 18]] is built and will be applied in the future. So that the prediccted efficiency is shown as the green curve (Figure 18) withh peak efficiency higher than 99%.

SIONS V. CONCLUS A 1.2kW 1MHz GaN-based duaal-phase interleaved CRM totem-pole PFC rectifier is dem monstrated in this paper. Several high frequency issues are id dentified and tackled with proposed solutions. With ZVS soft switching, s high efficiency is able to be achieved even at MHz frequency range. Attributing to the high frequency, the volume of passive components like inductor and DM filter is shrunk d of the system is dramatically so that the power density increased significantly. GMENT ACKNOWLEDG

This work was conducted with the use of GaN device samples donated in kind by Traansphorm of the CPES Industry Consortium Program. This work was conducted witth the use of SIMPLIS donated in kind by Simplis Tecchnologies of the CPES Industry Consortium Program.

CES REFERENC [1]

[2]

(a)

[3]

687

Liang Zhou, Yi-Feng Wu, Umesh Mishra, M “True Bridgeless Totempole PFC based on GaN HEMTs”, PCIM Europe 2013, 14-16 May, 2013, pp.1017-1022. Liang Zhou, and YiFeng Wu, “99% Effficiency True-Bridgeless Totem-Pole PFC Based on GaN HEMT Ts”, http://www.transphormusa.com/sites/d default/files/transphorm/news/To tem-pole%20paper_0.pdf GaN Systems, application note, “Tottem pole bridgeless PFC using GaN transistors”,

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

http://www.gansystems.com/_uploads/whitepapers/425686_App%20 Note%20%20Totem%20pole%20bridgeless%20PFC%20using%20GaN%20tra nsistors.pdf \ Huang, X.; Liu, Z.; Li, Q.; Lee, F.C., "Evaluation and Application of 600 V GaN HEMT in Cascode Structure," Power Electronics, IEEE Transactions on , vol.29, no.5, pp.2453,2461, May 2014 Huang, X.; Li, Q.; Liu, Z.; Lee, F.C., "Analytical Loss Model of High Voltage GaN HEMT in Cascode Configuration," Power Electronics, IEEE Transactions on , vol.29, no.5, pp.2208,2219, May 2014 Zhengyang Liu; Xiucheng Huang; Lee, F.C.; Qiang Li, "Package Parasitic Inductance Extraction and Simulation Model Development for the High-Voltage Cascode GaN HEMT," Power Electronics, IEEE Transactions on , vol.29, no.4, pp.1977,1985, April 2014 Huang, X.; Du, W.; Lee, F.C.; Li, Q.; Liu, Z., "Avoiding Si MOSFET Avalanche and Achieving Zero-Voltage-Switching for Cascode GaN Devices," Power Electronics, IEEE Transactions on , vol.PP, no.99, pp.1,1 Zhengyang Liu; Xiucheng Huang; Mingkai Mu; Yuchen Yang; Lee, F.C.; Qiang Li, "Design and evaluation of GaN-based dual-phase interleaved MHz critical mode PFC converter," Energy Conversion Congress and Exposition (ECCE), 2014 IEEE , vol., no., pp.611,616, 14-18 Sept. 2014 Yuchen Yang; Zhengyang Liu; Lee, F.C.; Qiang Li, "Analysis and filter design of differential mode EMI noise for GaN-based interleaved MHz critical mode PFC converter," Energy Conversion Congress and Exposition (ECCE), 2014 IEEE , vol., no., pp.4784,4789, 14-18 Sept. 2014 Jovanovic, M.M.; Yungtaek Jang, "State-of-the-art, single-phase, active power-factor-correction techniques for high-power applications - an overview," Industrial Electronics, IEEE Transactions on , vol.52, no.3, pp.701,708, June 2005 Huber, L.; Yungtaek Jang; Jovanovic, M.M., "Performance Evaluation of Bridgeless PFC Boost Rectifiers," Power Electronics, IEEE Transactions on , vol.23, no.3, pp.1381,1390, May 2008 Bin Su; Junming Zhang; Zhengyu Lu, "Totem-Pole Boost Bridgeless PFC Rectifier With Simple Zero-Current Detection and Full-Range ZVS Operating at the Boundary of DCM/CCM," Power Electronics, IEEE Transactions on , vol.26, no.2, pp.427,435, Feb. 2011 Bin Su; Zhengyu Lu, "An Interleaved Totem-Pole Boost Bridgeless Rectifier With Reduced Reverse-Recovery Problems For Power Factor Correction," Power Electronics, IEEE Transactions on , vol.25, no.6, pp.1406,1415, June 2010 Marxgut, C.; Krismer, F.; Bortis, D.; Kolar, J.W., "Ultraflat Interleaved Triangular Current Mode (TCM) Single-Phase PFC Rectifier," Power Electronics, IEEE Transactions on , vol.29, no.2, pp.873,882, Feb. 2014 Kim, J.W.; Choi, S. M.; Kim, K. T., "Variable On-time Control of the Critical Conduction Mode Boost Power Factor Correction Converter to Improve Zero-crossing Distortion," Power Electronics and Drives Systems, 2005. PEDS 2005. International Conference on , vol.2, no., pp.1542,1546, 28-01 Nov. 2005 Mingkai Mu; Lee, F.C., "Comparison and optimization of high frequency inductors for critical model GaN converter operating at 1MHz," Electronics and Application Conference and Exposition (PEAC), 2014 International , vol., no., pp.1363,1368, 5-8 Nov. 2014 Shuojie She; Wenli Zhang; Xiucheng Huang; Weijing Du; Zhengyang Liu; Lee, F.C.; Qiang Li, "Thermal analysis and improvement of cascode GaN HEMT in stack-die structure," Energy Conversion Congress and Exposition (ECCE), 2014 IEEE , vol., no., pp.5709,5715, 14-18 Sept. 2014 Zhang, W.; Huang, X.; Liu, Z.; Lee, F.C.; She, S.; Du, W.; Li, Q., "A New Package of High-Voltage Cascode Gallium Nitride Device for Megahertz Operation," Power Electronics, IEEE Transactions on , vol.PP, no.99, pp.1,1

688