Effective Digital Predistortion (DPD) on a Broadband Millimeter-Wave GaN Power Amplifier Using LTE 64-QAM Waveforms

Round 1

Reviewer 1 Report

This paper studied the digital distortion of broadband millimeter-wave power amplifier. I have the following comments:

- The most interesting results of this paper are the measurement results after using GMP DPD. However, the authors should still highlight the novel contribution in algorithm design. GMP DPD has been studied in the literature.

- Besides the meaurement results, the authors should include more analysis and try to give more insights.

- When further extending the bandwidth, memory effect becomes more significant. Does the proposed method applicable in wideband signaling? Please refer to the folloing papers and the references therein, Model identification for digital predistortion of power amplifier with signed regressor algorithm, mwcl. Digital predistortion of wideband power amplifier with single undersampling adc, mwcl. 

Author Response

Dear Reviewer,

Thank you for your valuable comments of our initial submitted manuscript “Effective Digital Predistortion (DPD) on a Broadband Millimeter-Wave GaN Power Amplifier Using LTE 64-QAM Waveforms” to the Electronics Journal under the special issue “Microwave/ Millimeter-Wave Power Amplifier”. We appreciate the time and efforts you have dedicated to providing your valuable feedbacks on our manuscript. We have incorporated changes to address most of the suggestions provided by the reviewers. We have highlighted the changes within the manuscript.

Here are the point-by-point responses to the reviewers’ comments and concerns.

Reviewer #1:

This paper studied the digital distortion of broadband millimeter-wave power amplifier. I have the following comments:

• Comment 1: The most interesting results of this paper are the measurement results after using GMP DPD. However, the authors should still highlight the novel contribution in algorithm design. GMP DPD has been studied in the literature.

• Response: We thank the reviewer’s insightful comments. One of the things we have done differently in the GMP DPD algorithm is that we have used a relatively high lag ( and lead (  memory depth of ‘5” for this paper. We have mentioned this in section 2. A lead and lag memory depth of “2” is commonly used for GMP DPD experiments [7-8]. This may have made our GMP DPD yielding better performance than many GMP DPD reported in literature, but it took less than 30 seconds to perform the GMP DPD. The GMP DPD in this paper can be easily applied in the NI testbench in a very user-friendly way. We have also included this point in the revised manuscript for clarifications: “A lead and lag memory depth of “2” is commonly used for the GMP DPD experiments in literature [7-8]. In our paper, however, a lead and lag memory depth of “5” is used, which may have made our GMP DPD yielding better performance than those in literature while tackling wideband signals. It took less than 30 seconds to perform the GMP DPD with the higher depths in our work. The GMP DPD used in this paper was also applied in a very user-friendly NI testbench”.

• Comment 2: Besides the measurement results, the authors should include more analysis and try to give more insights.

• Response: Yes, we appreciate the reviewer’s insightful comments. Besides extensive measurement data, one insightful analysis we have provided on the effectiveness of the GMP DPD is dependent on the wider modulated signal BW from carrier aggregation (Figs. 9,11 and 12), and this highlights the challenges one will face in linearizing mm-Wave broadband PAs. The other insight is on the linearity improvement is mostly independent of the operation frequency of the broadband mm-Wave PAs tested here, which we have added this sentence in the revised manuscript :“The linearity improvement is mostly independent of the operation frequency of the broadband mm-Wave PAs tested in this work, which is encouraging as this demonstrates the effective of the GMP DPD model used”.

Additionally, on the analysis of the specific AM/AM and AM/PM improvement vs. the modulated signal BW, we have found that the distortion increases with the signal BW, as the higher BW signals are achieved by carrier aggregation (CA). With wider BW signals, the AM/ Gain linearity improves after the GMP DPD. We also see that the ACLR and EVM after GMP DPD reduced. A considerable linearity improvement is seen at higher output power levels as the mm-Wave GaN PAs are required to have larger P_{OUT, Linear} for 5G small cell applications and we thus added this sentence in the revised manuscript: “Wideband modulated 5G signals are achieved using carrier aggregation. The distortion worsens as the signal BW increases, with higher AM/AM and the rms phase error of AM/PM rises as the signal BW increases, as well as worsened ACLR and EVM values. However, considerable linearity improvements are  seen after the GMP DPD, especially relevant and critical to extend mm-Wave GaN PAs for mm-Wave applications that require larger P_{OUT, Linear} such as 5G small cells”.

On the analysis of the specific concerns of the DPD speed using the GMP DPD vs. the modulated signal BW or the operation signal BW, we have found that the GMP DPD all finished very fast in less than 30 seconds and thus as discussed above in response to Comment #1, we have added this sentence in the revised manuscript “It took less than 30 seconds to perform the GMP DPD with the higher depths in our work.”.

• Comment 3: When further extending the bandwidth, memory effect becomes more significant. Does the proposed method applicable in wideband signalling? Please refer to the following papers and the references therein, Model identification for digital predistortion of power amplifier with signed regressor algorithm, mwcl. Digital predistortion of wideband power amplifier with single undersampling adc, mwcl. 

• Response: We thank the reviewer for this insightful comment. Yes, the proposed method is applicable in wideband signals. We have tested wideband signals such as 80 MHz/ 100 MHz for this experiment. We are also planning to test further higher wideband signals such as 400 MHz in future using the PAs designed in our lab and the initial data looks promising. We have also checked in the suggested insightful MWCL papers and added them as Refs. 23 and 24.

Author Response File: Author Response.pdf

Reviewer 2 Report

In this study, the authors showed the DPD algorithm to compensate distortion of GaN PA. Using DPD, the test results showed improved performance. I think it can be published in this journal.

However, I think it might be better that some explanation is added for the clarity of the study.

1. Why should DPD be used for GaN PA? I think there are several ways to compensate for the distortion of GaN PA. The advantages of using DPD should be added compared to other compensation algorithms.

Author Response

Dear Reviewer,

Thank you for your valuable comments of our initial submitted manuscript “Effective Digital Predistortion (DPD) on a Broadband Millimeter-Wave GaN Power Amplifier Using LTE 64-QAM Waveforms” to the Electronics Journal under the special issue “Microwave/ Millimeter-Wave Power Amplifier”. We appreciate the time and efforts you have dedicated to providing your valuable feedbacks on our manuscript. We have incorporated changes to address most of the suggestions provided by the reviewers. We have highlighted the changes within the manuscript.

Here are the point-by-point responses to the reviewers’ comments and concerns.

Reviewer #2:

In this study, the authors showed the DPD algorithm to compensate distortion of GaN PA. Using DPD, the test results showed improved performance. I think it can be published in this journal. However, I think it might be better that some explanation is added for the clarity of the study.

• Comment 1: Why should DPD be used for GaN PA? I think there are several ways to compensate for the distortion of GaN PA. The advantages of using DPD should be added compared to other compensation algorithms.

• Response: We appreciate the reviewer for these insightful and positive comments. Yes, we agree with the reviewer that when applying DPD on the GaN PA, even though it comes with improved linearity vs. PAE design trade-off significantly, does come with a price as the baseband DPD processing also adds on the power and may slow down the operation. So far, we have only been using the NI testbench to characterize the PAs and we have not worked on the option of using a FPGA to implement the DPD algorithms so we have not evaluated the extra power and other issues that may be associated with adding DPD into mm-Wave PA as a commercial product, nor have we investigated analog predistortion. Therefore, we have added the following sentences in revised manuscript in the end to clarify: “We would like to mention that applying GMP DPD on the mm-Wave GaN PAs, even though as shown in this work providing clear improvement in linearity vs. PAE design trade-offs significantly, would also come with a price as the DPD done at baseband also adds on the power consumption and complicated the baseband signal processing. So far, we have only been using the NI testbench to characterize the PAs with DPD algorithms as a feasibility study, but we have not worked on evaluating the extra power practically needed by adding DPD using a FPGA or into an IC to implement the algorithms, nor have we investigated analog predistortion options, but we would like to explore these other linearization techniques further in the future when funding allows”.

Author Response File: Author Response.pdf

Reviewer 3 Report

This paper presents a linearization method on a mmW broadband MMIC GaN PA at 5G FR2 band using DPD. Measurements show that significant improvement of linearity has been achieved by using GMP DPD. The paper can be accepted for publication after addressing the following questions and comments.

1. The authors should double check the English grammar before submission. Errors have been found across the paper. For instance, in Line 77, the "of" before "for all" should be deleted. In Line 381, there should be a space between "of" and "average". Besides, in Line 385, the "-" between "state-of-the-art" and "mm-Wave" should be deleted. In addition, in Line 100, a "are" is missed between "changes" and "detected". There are other errors, and the authors should all correct them.

2. Although the authors present the algorithm of their DPD technique in Section 2, it would be much improved if the authors can show how they used this algorithm to extract parameters using the actual PA measurement data in Section 3.

See comments above.

Author Response

Dear Reviewer,

Thank you for your valuable comments of our initial submitted manuscript “Effective Digital Predistortion (DPD) on a Broadband Millimeter-Wave GaN Power Amplifier Using LTE 64-QAM Waveforms” to the Electronics Journal under the special issue “Microwave/ Millimeter-Wave Power Amplifier”. We appreciate the time and efforts you have dedicated to providing your valuable feedbacks on our manuscript. We have incorporated changes to address most of the suggestions provided by the reviewers. We have highlighted the changes within the manuscript.

Here are the point-by-point responses to the reviewers’ comments and concerns.

Reviewer #3:

This paper presents a linearization method on a mmW broadband MMIC GaN PA at 5G FR2 band using DPD. Measurements show that significant improvement of linearity has been achieved by using GMP DPD. The paper can be accepted for publication after addressing the following questions and comments.

• Comment 1: The authors should double check the English grammar before submission. Errors have been found across the paper. For instance, in Line 77, the "of" before "for all" should be deleted. In Line 381, there should be a space between "of" and "average". Besides, in Line 385, the "-" between "state-of-the-art" and "mm-Wave" should be deleted. In addition, in Line 100, a "are" is missed between "changes" and "detected". There are other errors, and the authors should all correct them.

• Response: We appreciate the reviewer for these insightful comments. We have applied the changes that the reviewer carefully mentioned. We also checked and corrected this entire manuscript for other grammar errors too.

• Comment 2: Although the authors present the algorithm of their DPD technique in Section 2, it would be much improved if the authors can show how they used this algorithm to extract parameters using the actual PA measurement data in Section 3

• Response: We thank the reviewer for this comment. To clarify, we have indeed used the DPD algorithms discussed in Section 2 to have extracted parameters using the actual PA measurement data in Section 3. Also, the process of extracting parameters happens under the hood of the NI software in LabView, so we are sorry it is almost impossible to show the extraction process with these parameters. We hope to have clearly explained on how the extraction happens in section 2. However, to clarify further, as explained to the Comments from Reviewer #1, a lead and lag memory depth of “2” is commonly used for GMP DPD experiments, but we used a level of “5” [7-8]. This may have made our GMP DPD yielding better performance than many GMP DPD reported in literature, and it took less than 30 seconds to perform the GMP DPD. We have thus also included this point in the revised manuscript for clarifications: “A lead and lag memory depth of “2” is commonly used for the GMP DPD experiments in literature [7-8]. In our paper, however, a lead and lag memory depth of “5” is used, which may have made our GMP DPD yielding better performance than those in literature while tackling wideband signals. It took less than 30 seconds to perform the GMP DPD with the higher depths in our work. The GMP DPD used in this paper was also applied in a very user-friendly NI testbench”.

Author Response File: Author Response.pdf

Reviewer 4 Report

The paper presents a good quality of simulation and practical work supported by mathematical modelling. The research work presented is quite well written. Despite some minor sentence issues, no major revision is needed. I recommend its publication.

I recommend to proof read it once before submitting again. Thanks

Author Response

Dear Reviewer,

Thank you for your valuable comments of our initial submitted manuscript “Effective Digital Predistortion (DPD) on a Broadband Millimeter-Wave GaN Power Amplifier Using LTE 64-QAM Waveforms” to the Electronics Journal under the special issue “Microwave/ Millimeter-Wave Power Amplifier”. We appreciate the time and efforts you have dedicated to providing your valuable feedbacks on our manuscript. We have incorporated changes to address most of the suggestions provided by the reviewers. We have highlighted the changes within the manuscript.

Here are the point-by-point responses to the reviewers’ comments and concerns.

Reviewer #4:

• Comment 1: The paper presents a good quality of simulation and practical work supported by mathematical modelling. The research work presented is quite well written. Despite some minor sentence issues, no major revision is needed. I recommend its publication.
• Response: Thanks for the recommendation. We will proofread the paper once again! Thanks.

Author Response File: Author Response.pdf

Reviewer 5 Report

1: The author should avoid using star marks to denote multiplication in all equations and inline mathematical equations.

2: The authors can include a list of abbreviations to facilitate better reading.

3: Sometimes "labeled" and sometimes "labelled" are used. Please use consistent English.

4: The abstract should be written better, avoiding many mathematical symbols. 

5: The authors can highlight the contributions in the introduction by a better approach.

6: The introduction should be improved to introduce the research gap and the solution proposed in this study.

7: I need help accessing the given website link to access the resource, as shown in the ref. 2.

Author Response

Dear Reviewer,

Thank you for your valuable comments of our initial submitted manuscript “Effective Digital Predistortion (DPD) on a Broadband Millimeter-Wave GaN Power Amplifier Using LTE 64-QAM Waveforms” to the Electronics Journal under the special issue “Microwave/ Millimeter-Wave Power Amplifier”. We appreciate the time and efforts you have dedicated to providing your valuable feedbacks on our manuscript. We have incorporated changes to address most of the suggestions provided by the reviewers. We have highlighted the changes within the manuscript.

Here are the point-by-point responses to the reviewers’ comments and concerns.

Reviewer #5:

• Comment 1:The author should avoid using star marks to denote multiplication in all equations and inline mathematical equations.

• Response: We thank the reviewer for this comment. We have changed according to the reviewer’s suggestion.
• Comment 2:The authors can include a list of abbreviations to facilitate better reading.
• Response: We thank the reviewer for this comment. We have changed according to the reviewer’s suggestion and included the abbreviations in the revised manuscript: “ 4G (4^th Generation); 5G (5^th Generation); Adjacent Channel Leakage Power Ratio (ACLR); Carrier Aggregation (CA); Digital Pre-distortion (DPD); Error vector Magnitude (EVM); Gallium Nitride (GaN); Generalized Memory Polynomial (GMP); Long-Term Evolution (LTE); Millimeter-Wave (mm-Wave); Power Amplifier (PA); Quadrature Amplitude Modulation (QAM)”.
• Comment 3:Sometimes "labeled" and sometimes "labelled" are used. Please use consistent English.
• Response: Thanks for pointing out. We have made consistent throughout the paper. Thanks!!
• Comment 4: The abstract should be written better, avoiding many mathematical symbols. 
• Response: We thank the reviewer for this comment. We have changed according to the reviewer’s suggestion. The abstract is now revised as “We demonstrate in this work effective linearization on a millimeter-wave (mm-Wave) broadband monolithic Gallium Nitride (GaN) power amplifier (PA) using Digital Predistortion (DPD). The PA used is a 2-stage common-source (CS)/ 2-stack PA that operates in the mm-Wave 5G FR2 band, and it is linearized with the Generalized Memory Polynomial (GMP) DPD and tested using a 4G (4^th Generation) Long-Term Evolution (LTE) 64-QAM (Quadrature Amplitude Modulation) modulated signals with a PAPR (peak-to-average-power ratio) of 8 dB. Measurement results after implementing GMP DPD indicate considerable broadband improvement on the Adjacent Channel Leakage Power Ratio (ACLR) of 16.9 dB/ 17.3 dB/ 16.5 dB/ 15.1 dB at 24 GHz/ 28 GHz/ 37 GHz/ 39 GHz, respectively, with a common averageP_OUT of 15 dBm using a 100 MHz LTE 64-QAM input signal. At a fixed frequency of 28 GHz, the GaN PA after GMP DPD achieved signal bandwidth-dependent ACLR improvement and root-mean-square (rms) EVM (Error Vector Magnitude) reduction using 20 MHz/ 40 MHz/ 80 MHz/ 100 MHz LTE 64-QAM waveforms with a common average P_OUT of 15 dBm. The GaN PA thus achieved very good linearization results compared to other state-of-art mm-Wave PA DPD studies in literature, suggesting GMP DPD should be rather effective for linearizing mm-Wave 5G broadband GaN PAs to improve P_{OUT, Linear}”.
• Comment 5: The authors can highlight the contributions in the introduction by a better approach.
• Response: We thank the reviewer for this comment. We have made changes according to the reviewer’s suggestion. The Introduction section is now revised as: “The 5^th Generation (5G) mobile network currently offers both the Sub-6 GHz Frequency Range 1 (FR1) Band and the millimeter-wave (mm-Wave) Frequency Range 2 (FR2) Band of 24.25 – 52.6 GHz to achieve up to 10 Gb/s download speeds for Enhanced Mobile Broadband (eMBB) and other applications [1]. For mm-Wave power amplifiers (PA) designed at the higher 5G FR2 band, the trade-off in linearity vs. Power-Added-Efficiency (PAE) can become considerably more complicated, making it more difficult and increasingly expensive to construct mm-Wave 5G small cells and base stations (i.e., macrocells, microcells and picocells) [2-3]. For a mm-Wave RF transmitter (TX), the performance of the PA can often dominate the overall TX performance in heat dissipation, P_OUT, linearity, efficiency, reliability, etc. For example, modulated input signals with high a peak-to-average-power-ratio (PAPR) can degrade a PA’s average PAE and linear output power P_{OUT, Linear} and cause TX overheating [4]. Therefore, highly efficient PA design with high P_{OUT, Linear} is crucial for 5G applications. Digital Predistortion (DPD) can play a pivotal role here as it can linearize the mm-Wave PAs and significantly improve their P_{OUT, Linear} and design margins. The effectiveness of various DPD techniques can be evaluated based on their improvement to the Adjacent Channel Leakage Power Ratio (ACLR) in the PA’s output spectrum, and the modulation quality determined from constellation diagrams (i.e., error vector magnitude (EVM)) [2]. Many DPD techniques have been reported in recent years, such as the popular Look-Up-Table (LUT)-based DPD [5], Memory Polynomial DPD [6], Generalized Memory Polynomial DPD [7] and High I/Q Imbalance DPD [8]. For example, the DPD model reported in [8] relies on constructing the multi-toned signal consisting of the I (In-phase) signal whose frequency components do not correlate with the Q (Quadrature) signal components. Using interleaved multi-toned signals allows faster detection of the predistortion coefficients, and the coefficients of the I/Q model are estimated and compared until the error is zero [8]. Another possible benefit of using multi-tone signals is that the output signals can have a high Spurious Free Dynamic Range (SFDR).

This work, however, uses a relatively simpler single-tone GMP DPD model implemented in a user-friendly National Instruments (NI) mm-Wave PA testbench and not needing an external FPGA (Field Programmable Gate Arrays) [5]. This work applies a GMP DPD algorithm to linearize a 2-stage Common Source (CS)/ 2-stack Gallium Nitride (GaN) monolithic mm-Wave PA designed in our lab [9]. This GMP DPD achieves impressive performance on the broadband mm-Wave PA, compared vs. other similar mm-Wave PA studies in the literature [6-8], as discussed in section II. The DPD algorithm and the experimental mm-Wave hardware DPD testbench setup are also discussed in section II. Section III presents the measurement results of the PA tested using Long-Term Evolution (LTE) Quadrature Amplitude Modulation (64-QAM) modulated signals of different signal bandwidth (BW) at various output power levels and frequencies. We conclude by comparing our work with several state-of-the-art mm-Wave DPD PAs in literature”.

• Comment 6:The introduction should be improved to introduce the research gap and the solution proposed in this study.
• Response: We thank the reviewer for this comment. We have made changes according to the reviewer’s suggestion. The Introduction is now revised as “The 5^th Generation (5G) mobile network currently offers both the Sub-6 GHz Frequency Range 1 (FR1) Band and the millimeter-wave (mm-Wave) Frequency Range 2 (FR2) Band of 24.25 – 52.6 GHz to achieve up to 10 Gb/s download speeds for Enhanced Mobile Broadband (eMBB) and other applications [1]. For mm-Wave power amplifiers (PA) designed at the higher 5G FR2 band, the trade-off in linearity vs. Power-Added-Efficiency (PAE) can become considerably more complicated, making it more difficult and increasingly expensive to construct mm-Wave 5G small cells and base stations (i.e., macrocells, microcells and picocells) [2-3]. For a mm-Wave RF transmitter (TX), the performance of the PA can often dominate the overall TX performance in heat dissipation, P_OUT, linearity, efficiency, reliability, etc. For example, modulated input signals with high a peak-to-average-power-ratio (PAPR) can degrade a PA’s average PAE and linear output power P_{OUT, Linear} and cause TX overheating [4]. Therefore, highly efficient PA design with high P_{OUT, Linear} is crucial for 5G applications. Digital Predistortion (DPD) can play a pivotal role here as it can linearize the mm-Wave PAs and significantly improve their P_{OUT, Linear} and design margins. The effectiveness of various DPD techniques can be evaluated based on their improvement to the Adjacent Channel Leakage Power Ratio (ACLR) in the PA’s output spectrum, and the modulation quality determined from constellation diagrams (i.e., error vector magnitude (EVM)) [2]. Many DPD techniques have been reported in recent years, such as the popular Look-Up-Table (LUT)-based DPD [5], Memory Polynomial DPD [6], Generalized Memory Polynomial DPD [7] and High I/Q Imbalance DPD [8]. For example, the DPD model reported in [8] relies on constructing the multi-toned signal consisting of the I (In-phase) signal whose frequency components do not correlate with the Q (Quadrature) signal components. Using interleaved multi-toned signals allows faster detection of the predistortion coefficients, and the coefficients of the I/Q model are estimated and compared until the error is zero [8]. Another possible benefit of using multi-tone signals is that the output signals can have a high Spurious Free Dynamic Range (SFDR).

This work, however, uses a relatively simpler single-tone GMP DPD model implemented in a user-friendly National Instruments (NI) mm-Wave PA testbench and not needing an external FPGA (Field Programmable Gate Arrays) [5]. This work applies a GMP DPD algorithm to linearize a 2-stage Common Source (CS)/ 2-stack Gallium Nitride (GaN) monolithic mm-Wave PA designed in our lab [9]. This GMP DPD achieves impressive performance on the broadband mm-Wave PA, compared vs. other similar mm-Wave PA studies in the literature [6-8], as discussed in section II. The DPD algorithm and the experimental mm-Wave hardware DPD testbench setup are also discussed in section II. Section III presents the measurement results of the PA tested using Long-Term Evolution (LTE) Quadrature Amplitude Modulation (64-QAM) modulated signals of different signal bandwidth (BW) at various output power levels and frequencies. We conclude by comparing our work with several state-of-the-art mm-Wave DPD PAs in literature”.

• Comment 7: I need help accessing the given website link to access the resource, as shown in the ref. 2.
• Response: We thank the reviewer for this comment. We will include the file (ref #2) as an attachment along with this response letter. Thanks!!

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have addressed my concerns. I have no further comments.