Ghent University Academic Bibliography

Advanced

Analysis and design of highly efficient Class-E amplifiers for indoor ranging

Zhisheng Li UGent (2012)
abstract
The last few years, indoor positioning has gained more and more interest. Modern indoor positioning systems do not use Global Positioning System (GPS) satellites in any way because multipath reflections of the narrowband GPS signals limit the resolution to a few meters. However, for indoor ranging applications, submeter accuracy is required. The high resolution can be obtained by utilizing wideband signals. In the target ranging system, the wideband signal is generated by inducing a ranging pulse with a very short rise time at the beginning of each packet. In this system, similar to GPS, the ToA (Time of Arrival) technique is used to estimate the amount of time a signal takes to propagate from a transmitter to a receiver, and to calculate the distance between nodes. For a single-shot, one-way ranging technique, the ranging accuracy can be simply estimated by the following equation: d=c* tr, in which tr is the propagation time of the signal, c is the speed of light, and d is the estimated distance between the transmitter and receiver. Hence, d is highly dependent on the rise time of the ranging pulse. To achieve submeter ranging accuracy, the rise time should be less than 3.3 ns. To relieve the requirement on the rise time, a two-way ranging technique was applied. By utilizing this technique, the maximum rise time is doubled (6.5 ns) for submeter accuracy. However, generating the ranging pulse is a great challenge. A RF switch can be placed in front of the power amplifier (PA) to enable/disable the output RF signals fastly. However, it complicates the transmitter topology and requires extra power also. Hence we propose the innovative combination of the PA with an integrated broadband ranging switch. This dissertation focuses on the design of a Class-E PA with switch function for ranging applications. For the target ranging application, the envelope-rise time and the isolation between switched-on/ off states of the designed PA affects the ranging accuracy directly. Meanwhile, the performances of the PA should not be affected by including the ranging switch. And the PA should support both communication and ranging, which makes the PA design even more challenging. Even for a conventional wireless transceiver, a PA is one of the most important and challenging blocks. It consumes the most power in wireless systems, which limits the battery life time directly. Nowadays, most RF modules have mature CMOS solutions except the PA, mainly because of the intrinsic limitations of CMOS technologies, such as low transconductance, low breakdown voltage, low characteristic frequency, compared to the non-CMOS technologies like SiGe, GaAs and LDMOS. The integration of PAs in CMOS is challenging and has been a hot topic in recent years. In the target ranging application, power efficiency is a key specification for the PA. Since a constant envelope modulation scheme is applied in the ranging system, a Class-E PA is used because of its high efficiency. By utilizing the differential cascode topology, the voltage stress of the Class-E PA is relieved and no impedance transformation network is required, improving the efficiency. Power control is another important feature for modern wireless systems. The existing power control techniques of a Class-E PA cannot provide optimum efficiencies at low output power levels. In this thesis, by combining the dynamic supply voltage together with the dynamic cascode bias voltage, a wide power range and high drain efficiency can be obtained even at low output power levels. Because of the large output voltage swing, the typical diode ESD topology is not applicable to the Class-E PA anymore. In this work, by using series chains of diodes, the ESD circuit can sustain large voltage swing with little parasitic capacitance. The measurement results show that the proposed ESD circuit can handle ESD pulses higher than 2 kV. Even though the Class-E PA can achieve 100\% drain efficiency theoretically, the non-idealities (such as finite switch-on resistance, finite rise/fall time of the RF signals, finite Q factors of the used inductors) degrade the output power and efficiency significantly. In this dissertation, the analysis of the output power and losses of a Class-E PA with finite dc-feed inductance is presented, resulting in practical analytical expressions, which significantly simplify the design and optimization of a Class-E PA. Besides the typical performances of a PA, stability is another important aspect that should be cared about. The use of a differential topology in a Class-E PA arises differential-mode and common-mode stability problems. In this dissertation, the small-signal differential-mode and common-mode stability were analyzed. A new common-mode stabilization technique was proposed. By placing a resistor at the common-gate node of the cascode transistors, the Class-E PA was stabilized in common mode without degrading the PA performance. To prove the possibility of the co-design of a CMOS Class-E PA with the switch function, two PAs with switch functions were implemented in this thesis. In these two PAs, two switches are placed at the common-gate node of the cascode transistors to achieve fast switching function without degrading the PA performance. The first PA is designed in Silterra 0.18 um CMOS with the fast switch function only. The measurement results show that the proposed PA covers a power range from 2 dBm to 20 dBm with a power added efficiency (PAE) of 43.6% at the maximum 20 dBm output power. Meanwhile, the designed PA shows a rise time of merely 2.5 ns and an isolation of more than 70 dB, which enable submeter accuracy for the single-shot one-way ranging technique. The improved PA was designed in 0.13 um CMOS. This PA has a configurable switch, which can provide both fast switching for ranging applications and slow switching for communication. The measurements show that this PA covers a power range from 1.5 dBm to 20.5 dBm with a PAE as high as 52.5% at 20.5 dBm output power. The measured fast rise time of 4.5 ns and more than 70 dB isolation enable submeter ranging accuracy for two-way ToA ranging techniques. Both measurements of these PAs are state-of-the-art. In communication mode, the rise/fall time is reduced to 90 ns, which enhances the communication data rate. To estimate the capability of the PA for data communication, Return-to-Zero Binary Phase Shift Keying (RZ-BPSK) modulation is applied. The measurement results show that the maximum data rate for RZ-BPSK modulation is around 1 Mbps with the fast switch function only and can be up to more than 4 Mbps in communication mode, which proves the possibility of the co-design of a CMOS Class-E PA and a ranging switch for the target ranging application.
Please use this url to cite or link to this publication:
author
promoter
UGent and UGent
organization
alternative title
Analyse en ontwerp van hoogefficiënte klasse-E-versterkers voor indoorafstandsmetingen
year
type
dissertation (monograph)
subject
pages
137 pages
publisher
Ghent University, Department of Information technology
place of publication
Ghent, Belgium
defense location
Gent: Jozef-Plateauzaal (Jozef-Plateaustraat 22)
defense date
2012-09-28 15:00
ISBN
9789085785392
language
English
UGent publication?
yes
classification
D1
copyright statement
I have retained and own the full copyright for this publication
id
3007198
handle
http://hdl.handle.net/1854/LU-3007198
date created
2012-10-05 10:21:29
date last changed
2012-10-11 09:16:56
@phdthesis{3007198,
  abstract     = {The last few years, indoor positioning has gained more and more interest. Modern indoor positioning systems do not use Global Positioning System (GPS) satellites in any way because multipath reflections of the narrowband GPS signals limit the resolution to a few meters. However, for indoor ranging applications, submeter accuracy is required. The high resolution can be obtained by utilizing wideband signals. In the target ranging system, the wideband signal is generated by inducing a ranging pulse with a very short rise time at the beginning of each packet. In this system, similar to GPS, the ToA (Time of Arrival) technique is used to estimate the amount of time a signal takes to propagate from a transmitter to a receiver, and to calculate the distance between nodes. For a single-shot, one-way ranging technique, the ranging accuracy can be simply estimated by the following equation: d=c* tr, in which tr is the propagation time of the signal, c is the speed of light, and d is the estimated distance between the transmitter and receiver. Hence, d is highly dependent on the rise time of the ranging pulse. To achieve submeter ranging accuracy, the rise time should be less than 3.3 ns. To relieve the requirement on the rise time, a two-way ranging technique was applied. By utilizing this technique, the maximum rise time is doubled (6.5 ns) for submeter accuracy. However, generating the ranging pulse is a great challenge. A RF switch can be placed in front of the power amplifier (PA) to enable/disable the output RF signals fastly. However, it complicates the transmitter topology and requires extra power also. Hence we propose the innovative combination of the PA with an integrated broadband ranging switch. This dissertation focuses on the design of a Class-E PA with switch function for ranging applications. For the target ranging application, the envelope-rise time and the isolation between switched-on/ off  states of the designed PA affects the ranging accuracy directly. Meanwhile, the performances of the PA should not be affected by including the ranging switch. And the  PA should support both communication and ranging, which makes the PA design even more challenging. Even for a conventional wireless transceiver, a PA is one of the most important and challenging blocks. It consumes the most power in wireless systems, which limits the battery life time directly. Nowadays, most RF modules have mature CMOS solutions except the PA, mainly because of the intrinsic limitations of CMOS technologies, such as low transconductance, low breakdown voltage, low characteristic frequency, compared to the non-CMOS technologies like SiGe, GaAs and LDMOS. The integration of PAs in CMOS is challenging and has been a hot topic in recent years. In the target ranging application, power efficiency is a key specification for the PA. Since a constant envelope modulation scheme is applied in the ranging system, a Class-E PA is used because of its high efficiency. By utilizing the differential cascode topology, the voltage stress of the Class-E PA is relieved and no impedance transformation network is required, improving the efficiency. Power control is another important feature for modern wireless systems. The existing power control techniques of a Class-E PA cannot provide optimum efficiencies at low output power levels. In this thesis, by combining the dynamic supply voltage together with the dynamic cascode bias voltage, a wide power range and high drain efficiency can be obtained even at low output power levels. Because of the large output voltage swing, the typical diode ESD topology is not applicable to the Class-E PA anymore. In this work, by using series chains of diodes, the ESD circuit can sustain large voltage swing with little parasitic capacitance. The measurement results show that the proposed ESD circuit can handle ESD pulses higher than 2 kV. Even though the Class-E PA can achieve 100{\textbackslash}\% drain efficiency theoretically, the non-idealities (such as finite switch-on resistance, finite rise/fall time of the RF signals, finite Q factors of the used inductors) degrade the output power and efficiency significantly. In this dissertation, the analysis of the output power and losses of a Class-E PA with finite dc-feed inductance is presented, resulting in practical analytical expressions, which significantly simplify the design and optimization of a Class-E PA. Besides the typical performances of a PA, stability is another important aspect that should be cared about. The use of a differential topology in a Class-E PA arises differential-mode and common-mode stability problems. In this dissertation, the small-signal differential-mode and common-mode stability were analyzed. A new common-mode stabilization technique was proposed. By placing a resistor at the common-gate node of the cascode transistors, the Class-E PA was stabilized in common mode without degrading the PA performance. To prove the possibility of the co-design of a CMOS Class-E PA with the switch function, two PAs with switch functions were implemented in this thesis. In these two PAs, two switches are placed at the common-gate node of the cascode transistors to achieve fast switching function without degrading the PA performance. The first PA is designed in Silterra 0.18 um CMOS with the fast switch function only. The measurement results show that the proposed PA covers a power range from 2 dBm to 20 dBm with a power added efficiency (PAE) of 43.6\% at the maximum 20 dBm output power. Meanwhile, the designed PA shows a rise time of merely 2.5 ns and an isolation of more than 70 dB, which enable submeter accuracy for the single-shot one-way ranging technique. The improved PA was designed in 0.13 um CMOS. This PA has a configurable switch, which can provide both fast switching for ranging applications and slow switching for communication. The measurements show that this PA covers a power range from 1.5 dBm to 20.5 dBm with a PAE as high as 52.5\% at 20.5 dBm output power. The measured fast rise time of 4.5 ns and more than 70 dB isolation enable submeter ranging accuracy for two-way ToA ranging techniques. Both measurements of these PAs are state-of-the-art. In communication mode, the rise/fall time is reduced to 90 ns, which enhances the communication data rate. To estimate the capability of the PA for data communication, Return-to-Zero Binary Phase Shift Keying (RZ-BPSK) modulation is applied. The measurement results show that the maximum data rate for RZ-BPSK modulation is around 1 Mbps with the fast switch function only and can be up to more than 4 Mbps in communication mode, which proves the possibility of the co-design of a CMOS Class-E PA and a ranging switch for the target ranging application.},
  author       = {Li, Zhisheng},
  isbn         = {9789085785392},
  language     = {eng},
  pages        = {137},
  publisher    = {Ghent University, Department of Information technology},
  school       = {Ghent University},
  title        = {Analysis and design of highly efficient Class-E amplifiers for indoor ranging},
  year         = {2012},
}

Chicago
Li, Zhisheng. 2012. “Analysis and Design of Highly Efficient Class-E Amplifiers for Indoor Ranging”. Ghent, Belgium: Ghent University, Department of Information technology.
APA
Li, Zhisheng. (2012). Analysis and design of highly efficient Class-E amplifiers for indoor ranging. Ghent University, Department of Information technology, Ghent, Belgium.
Vancouver
1.
Li Z. Analysis and design of highly efficient Class-E amplifiers for indoor ranging. [Ghent, Belgium]: Ghent University, Department of Information technology; 2012.
MLA
Li, Zhisheng. “Analysis and Design of Highly Efficient Class-E Amplifiers for Indoor Ranging.” 2012 : n. pag. Print.