Harmonic Distortion Peculiarities of High-Frequency SiGe HBT Power Cells for Radar Front End and Wireless Communication

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

electronics-3662096

 

Review Report

 

The Manuscript (MS) entitled “Harmonic Distortion Peculiarities of High Frequency SiGe HBT Power Cells for Radar Front End and Wireless Communication” by P. Sakalas, and A. Mukherjee presents an experimental study of harmonic distortion and power cells based on SiGe HBTs for RF applications. A frequency-dependent increase of the 3rd harmonic of Pout at 10 GHz was measured for the two different technology PwCs considered. An analysis was performed in terms of several parameters (avalanche multiplication, non-linear diffusion capacitances, …). The origin of the observed lowering of the 3rd harmonic of the output power found in the study is attributed to the resonant behaviour of the substrate capacitance. Additionally, an important result reported in the MS is that both SiGes PwC studied exhibited good linearity under matched conditions in both scalar and vectorial load-pull tests.

The paper is well structured and well written and results presented are timely, relevant and sound in a very competitive field of research.

Despite minor details such as typos (for example, “Rescent” in line 23 on page 1) and the missing definition of PwC in the main body of the text (it is only defined in the Abstract of the paper), I recommend the paper for publication in its present form.

Author Response

Reply to the reviewer 1

Thanks a lot for revising the manuscript and for the positive evaluation. We corrected all your remarks and even more: we corrected and improved the whole manuscript style for easy reading.

Reviewer 2 Report

Comments and Suggestions for Authors

The paper presents SiGe power cell characterization.

Questions/suggestions:
1) If understood correctly, IHP’s SG13G2 process was used for measurement and characterization, but the thing that is unclear is whether the structure in Figure 1 is a standard cell for the latter process or is this the modified configuration. The same goes with Figure 2, as two configurations are given. This has to be clarified;
2) Related to question 1 - if the presented power cell is not standard, then a comparison to the standard cell performance would be good to include. Obviously, manufacturing or measuring currently might not be possible, but at least mentioning the problems with standard cell components that are being addressed and solved should be mentioned.
3) Continuing with the first two questions - the end of the introduction section should clearly define what is addressed and solved in this paper. If it is a proposed cell within the process, then specific solutions or differeces from available standard cells should be given. If the latter measurements are done with a standard cell, but they are not present in process documentation, this should be highlighted. Because currently it seems that the paper lacks purpose of the large work that has been done.
4) Question about the setup - were the setup connections (cables) de-embedded and were additional component ( circulator, bias-T) losses excluded from the final results presented in plots? This should be mentioned in the description of the setup, as the measurements are not trivial.
5) The conclusions do not resemble the purpose/aim of the work or problem that was solved.
6) The plot size could be bigger, as the figures are currently small and there is a lot going on there;
7) The real test setup picture could be given along with Figure 3 block diagram if possible.

I'd suggest accepting the paper after clarifying the above, but the level of the overall work is high.

Author Response

Reply to the comment 1)

• SG13G2 process does not contain optimized power cells or standard high frequency power cells. We have designed them and optimized tof power application at high frequencies. There are power devices with less doped collector and with extended BVCEO but low fT, fmax. Normally pdk (e.g IHP, Infineon) provides so called “power transistor” with higher BVCEO but much lower fT and fmax. So you can’t use them in higher bands (V or W).
• Thanks for your comments! Yes, these are not IHP standard transistor available in their SG13G2 pdk. They are stripe-transistors, additionally put on the wafer from our side.

IHP pdk transistors are normally CEB in configuration (like a “H” shaped transistor). Following is the basic structure of a IHP pdk transistor

 

 

 

 

 

 

 

The emitter length can be increased by the factor “NX” which is nothing but a multiplication factor, increasing the emitter length just by adding the above basic cell in X-direction. Below is the IHP pdk transistor with NX=5

 

 

 

 

 

 

 

 

In this we have used stripe transistors CBEBEBC etc (where C, B and E contacts are parallel to each other) designed by our group.

 

 2) Related to question 1 - if the presented power cell is not standard, then a comparison to the standard cell performance would be good to include. Obviously, manufacturing or measuring currently might not be possible, but at least mentioning the problems with standard cell components that are being addressed and solved should be mentioned.

Reply to comment 2)

• Very good point. Unfortunately, the standard Power cells are not existing. So, we designed an optimized. Anyway, I did comparison to PDK device, which is single SiGe HBT but not for power applications, Fig. 17.

3) Continuing with the first two questions - the end of the introduction section should clearly define what is addressed and solved in this paper. If it is a proposed cell within the process, then specific solutions or differeces from available standard cells should be given. If the latter measurements are done with a standard cell, but they are not present in process documentation, this should be highlighted. Because currently it seems that the paper lacks purpose of the large work that has been done.

Reply to comment 3)

• Thanks for your comments!

Since normal transistors layout are not optimized for handling high currents required for power amplifiers, in this paper we have tried to summarize the different optimization procedure for the power-cell design for microwave frequencies and finally their measured characteristics are given. One main objectives of power cell design are to make normal transistor work at its best by improving required layout which is unavailable in the pdk.

•  
• This paper analyzes power behavior of the high frequency power cell, optimized to drive high currents for high frequency application in V and W frequency bands. The cells were designed by Anindya Mukherjee. As mentioned before, usually power transistors (with higher BVCEO but lower fT and fmax) is provided by pdk but you can’t use them at higher frequencies (V, W band). So, we have taken high speed transistors (with lower BVCEO but higher fT and fmax) for our power cell design where we optimize the layout to reduce parasitic capacitances etc. and to have reasonable fT and fmax (to be able to use in V and W bands).
• In this work authors wanted to investigate and model with compact model power cells. It appeared that the third harmonic demonstrates strange behavior, which was not captured by the model. Such behavior was also found in another technology power cells, designed by us. We have measured all available configuration power cells and did find the same behavior. This behavior was investigated and explained by a numerous measurements and simulations.
• So the work task was to design and characterize high speed power cells and to realize within SiGe technology and model them. The observed harmonic power drop out was found and explained. This third harmonic could be used for Third harmonic peaking amplifiers, following those publications:

1. Arar, “Designing a Third-Harmonic Peaking Class F Amplifier for Maximum Efficiency”,
2. Sepahvand, A. Sheikhi, “Analysis and design of the novel five harmonic peaking Class-EF power amplifier”, Scientific reports, open in www.nature.com/scientificreports.
3. Y. Mortazavi and K. -J. Koh, "A 38 GHz inverse class-F power amplifier with 38.5% peak PAE, 16.5 dB gain, and 50 mW Psat in 0.13-μm SiGe BiCMOS," 2015 IEEE Radio Frequency Integrated Circuits Symposium (RFIC), Phoenix, AZ, USA, 2015, pp. 211-214.

 

4) Question about the setup - were the setup connections (cables) de-embedded and were additional component ( circulator, bias-T) losses excluded from the final results presented in plots? This should be mentioned in the description of the setup, as the measurements are not trivial.

• All losses of the cabling and probes, BT etc. were de-embedded during the power calibration.

5) The conclusions do not resemble the purpose/aim of the work or problem that was solved.

• We improved the conclusions to be better cited.

6) The plot size could be bigger, as the figures are currently small and there is a lot going on there;

• The plots were increased to be more readable. Thank you for the comment.

7) The real test setup picture could be given along with Figure 3 block diagram if possible.

• This is a standard setup built on a probe station. Due to space limitation we did not show the photo. One setup was with PNA-X NVNA. Another was with Focus impedance tuners.

I'd suggest accepting the paper after clarifying the above, but the level of the overall work is high.

We appreciate the reviewer 2 for very scientific and deep comments. This forced us to improve the whole manuscript. We also made style corrections through all the paper to make it easier readable.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

In this paper by P. Sakalas et al. authors have contributed to the study of HBT power cells for Radar front end and wireless communications. The results are interesting, and serve as key factor for wireless transmission. Here are my comments:

1. In Fig. 1, I suggest authors to present work principle of SiGe HBT, such as small signal model and PN junction for carrier transmit.
2. It is commercially available with the low-frequency power amplifier, I suggest authors to include discussions why the spectral usage with studied in this work. In my view, the particular challenge is located at THz frequency, some references can be added to improve discussions with  Research. 2023;6:0274. DOI:10.34133/research.0274, Research. 2024;7:0514.DOI:10.34133/research.0514, Nat. Commun. 16, 4736 (2025).
3.  Why the second harmonic disppear? By the way how to improve the cut-off frequency?

 

Author Response

Rev 3

In this paper by P. Sakalas et al. authors have contributed to the study of HBT power cells for Radar front end and wireless communications. The results are interesting, and serve as key factor for wireless transmission. Here are my comments:

1. In Fig. 1, I suggest authors to present work principle of SiGe HBT, such as small signal model and PN junction for carrier transmit.

• Re to comment 1). We have used a compact model HiCuM L2, which equivalent circuit can be found elsewhere. Large signal model easily reduces to a small signal. Nevertheless the small signal circuit is not capable top account large signal related effects. Small signal model could be used for noise explanation and s-parameter behavior. In our case we needed large signal model.

1. It is commercially available with the low-frequency power amplifier, I suggest authors to include discussions why the spectral usage with studied in this work. In my view, the particular challenge is located at THz frequency, some references can be added to improve discussions with  Research. 2023;6:0274. DOI:10.34133/research.0274, Research. 2024;7:0514.DOI:10.34133/research.0514, Nat. Commun. 16, 4736 (2025).

• Re to remark No 2). This is a good question. Such analysis could be used for generating further work. This manuscript is already overloaded with results and discussions.

1.  Why the second harmonic disappear? By the way how to improve the cut-off frequency?

Re to remark 3) The second harmonic was measured and modeled. It is not so interesting like 3^rd harmonic, which is difficult to get rid of. The second harmonic can be easily canceled.

Attached is a figure where the second and 4^th harmonics are seen.

Thank s for the reviewers. Your work enabled significant improvement of submitted manuscript. We hope that it will be cited.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you for the updates, all suggestions were considered.