Parameter Extraction for a SPICE-like Delphi4LED Multi-Domain Chip-Level LED Model with an Improved Nelder–Mead Method
Round 1
Reviewer 1 Report
Comments and Suggestions for Authors1. The speed, accuracy and robustness of the method should be precise described in the conclusion and put appropriate data support to purpose the research objective.
2. There are a lot of “Error! Reference source not found.” in this paper. What this?
3. The introduction section should be simplified and the advantages and innovation of the method should be supplemented, which make this article more convincing
4. In the page 2, “The summary of our very first parameter extraction approach was presented in [21]”. The reference [21] should be briefly described
5. There are several typos and grammatical errors throughout the manuscript, the authors need to double-check again carefully.
Comments on the Quality of English Language1. There are several typos and grammatical errors throughout the manuscript, the authors need to double-check again carefully.
Author Response
Dear Reviewer,
Thank you for your review, especially given the issues with the document's formatting. I have now revised the manuscript to align with the journal's template.
Comment 1: The speed, accuracy and robustness of the method should be precise described in the conclusion and put appropriate data support to purpose the research objective.
Answer 1: We have enhanced the conclusion by including key results and runtime data to better support our findings and emphasize the method’s speed, accuracy, and robustness.
Comment 2: There are a lot of “Error! Reference source not found.” in this paper. What this?
Answer 2: We are not entirely sure what caused this issue, but it seems something went wrong during submission or when the editorial office pasted the content into the template. This was an oversight on my part, as I did not used the template at first.
Comment 3: The introduction section should be simplified and the advantages and innovation of the method should be supplemented, which make this article more convincing.
Answer 3: We was unable to significantly shorten the introduction, as I believe all sections are essential to properly position the chip-level multi-domain model within the context of LED modeling. It’s important that readers understand the relevance of such models and how they connect to existing models like the thermal and driver models. We have added a few lines to summarize these points, which I hope clarifies their importance. We also improved the readability of the introduction. (and the entire document)
Comment 4: In the page 2, “The summary of our very first parameter extraction approach was presented in [21]”. The reference [21] should be briefly described
Answer 4: I have replaced the reference with another source (the D4L report), and revised the sentence to:
"The first parameter extraction approach was presented in detail in [21] and is briefly described in this section."
Such changes have also been requested by the editor.
Comment 5: There are several typos and grammatical errors throughout the manuscript, the authors need to double-check again carefully.
Answer 5: Thank you for pointing this out. We will thoroughly review the manuscript to correct any typos or grammatical errors.
Reviewer 2 Report
Comments and Suggestions for AuthorsIn this work, the authors provided a novel parameter extraction procedure based on a modified Nelder-Mead optimization algorithm. This novel method is used to estimate the parameters of the Spice-like multi-domain model of LED chips developed by the previous Delphi4LED project. By using the modified algorithm, the parameters are estimated faster, compared to the previous algorithm. The modified parameter extraction procedure enhances the efficiency and simplifies the multi-domain LED characterization method originally proposed by the Delphi4LED project. I would recommend this manuscript for publication after minor revision. The following comments may help improve the manuscript.
1. Authors should demonstrate the parameter estimation error from the heat branch (a).
2. In Figure 3, authors cut off the maximum relative error at 20%, which makes the curve (N = 10) incomplete. I would suggest that the authors provide the complete curve of (N = 10).
3. Authors need to provide more details about the multiple advantages of the new parameter extraction procedure. In this paper, authors listed these advantages like faster operation, high level of robustness, and unsupervised operation.
Comments on the Quality of English LanguageMinor editing of English language required
Author Response
Dear Reviewer,
Thank you for your review, especially given the issues with the document's formatting. I have now revised the manuscript to align with the journal's template.
Comment 1: Authors should demonstrate the parameter estimation error from the heat branch (a).
Answer 2: The thermal model is not part of the chip-level multi-domain model. We apologize for the confusion. we have clarified this in the introduction:
"Using the extracted parameters and the thermal model described above (the process of thermal model generation is well-established and thus not further detailed here), a digital twin of the LED chip can be developed in a SPICE-like format. This digital twin can be seamlessly integrated with the driver's model at the electrical port and with the thermal model of the surrounding structure at the thermal port."
During the IVL measurements, the junction temperature is measured. Therefore, for model generation purposes, we do not need to know the junction-to-ambient thermal impedance.
During simulation, we can use R_th ≈ 0 (R_th = 0 would lead to errors) and set T_a = T_J, where T_a is the same as in the measurement. This allows us to measure model performance by comparing it with the measured values.
Comment 3: In Figure 3, authors cut off the maximum relative error at 20%, which makes the curve (N = 10) incomplete. I would suggest that the authors provide the complete curve of (N = 10).
Answer 3: Thank you for the suggestion. we have modified the figure to now display the full scale and the 0-10% range as well.
Comment 4: Authors need to provide more details about the multiple advantages of the new parameter extraction procedure. In this paper, authors listed these advantages like faster operation, high level of robustness, and unsupervised operation.
Answer 4:
Thank you for the comment. These are indeed the most important features, especially as they enable the investigation of LED aging, where hundreds of measurements must be processed at once.
However, we have added a few more advantages to the conclusion. One advantage is that the measurement process can be faster, as junction temperature control is no longer needed, and only thermal equilibrium needs to be established. Additionally, we can now include IVL measurement results that deviate from the junction temperature setpoint, often due to equipment limitations in achieving every measurement condition.
These further features however originated from the new extraction strategy that is made possible by the speed of the extractor.
Reviewer 3 Report
Comments and Suggestions for Authors
The authors estimated the parameters of the Spice-like multi-domain model of Light Emitting Diode (LED) chips developed and proposed by the past Del-7 phi4LED project. The modifications greatly improved the reliability of the optimization process, as the due to the applied heuristics the success rate 690 increased from 50% to 100%. This is an interesting paper, I recommend it is publishable after minor revision.
The following issues should be considered before publication.
1. The advantage of the proposed study should be added in the Abstract section.
2. There are many “Error! Reference source not found.”, the authors should check throughout the manuscript.
Such as Page 14, Line 585-586, please check and revise the following mistake.
“Error! Reference source 585 not found., Error! Reference source not found. and Error! Reference source not found”
3. The references should be revised according to the journal requirements.
Author Response
Dear Reviewer,
Thank you for your review, especially given the issues with the document's formatting. I have now revised the manuscript to align with the journal's template.
Comment 1: The advantage of the proposed study should be added in the Abstract section.
Answer 1: The abstract has been extended with:
"The speed and robustness of the new model eliminate the need for time-consuming junction temperature control during measurements by employing a novel extraction strategy that seeks the global minimum, rather than relying on the composition of marginal minima."
I believe this addition effectively highlights the remaining advantages of our study.
Comment 2: There are many “Error! Reference source not found.”, the authors should check throughout the manuscript.
Answer 2: We have thoroughly checked the manuscript. The original draft was not in the correct template, but we hope to have resolved this issue.
Comment 3: The references should be revised according to the journal requirements.
Answer 3: Thank you for the suggestion. We will review and revise all references to ensure they comply with the journal's requirements.
Reviewer 4 Report
Comments and Suggestions for AuthorsThe research discussed in the article is interesting. It concerns the modeling of LED characteristics in three areas: thermal, optical and electrical. The individual models are described in detail, including a mathematical description. The methods are also described in detail: the previous one and the newly proposed one for extracting parameters based on measurements.
1. Do the authors intend to make the program for calculating model parameters publicly available?
2. Do the authors intend to make the collected data publicly available?
3. If the article could be shortened and simplified, it would be more reader-friendly.
4. Abbreviations: EDA, IVL, OLED are not explained.
5. Errors occur in many places: Error! Reference source not found.
6. There are no graphs of sample measured characteristics with descriptions.
7. In my opinion, the word reliability is not used correctly in relation to algorithms.
However, I have the most doubts about Figures 7, 8 and 9. I understand the idea behind their creation, but they do not show much. The points overlap and that is all that can be concluded from the most important figures in the article. Similarly, the titles of these figures do not correspond to their content and the purpose of their presentation. In my opinion, this is a comparison of characteristics obtained from measurements with characteristics determined on the basis of modeling. In my opinion, the artificially added third dimension obscures these figures. For how many temperatures are the measurements presented? If for 4-5, specific values ​​can be placed instead of the third dimension.
The amount of work done on the article is significant. The article is worth publishing. However, I encourage the authors to make corrections and better present the final results.
Author Response
Dear Reviewer,
Thank you for your review, especially given the issues with the document's formatting. I have now revised the manuscript to align with the journal's template.
Comment 1: Do the authors intend to make the program for calculating model parameters publicly available?
Answer 1:
Yes, the following has been added to the conclusion:
"We have implemented an automated web-based model parameter extraction service available at https://extractor.ai-twilight.org/. This service has been publicly accessible and in use since March 2023 without any robustness issues."
Registration can be done with a false email address. However, as consortium members prefer to send files directly, the platform has not been updated for a while and currently only accepts HTML files, with luminous flux included, but not CSV or XLSX formats yet.
Comment 2: Do the authors intend to make the collected data publicly available?
Answer 2: Since the data was part of a round-robin measurement within the Delphi4LED project, we are unsure if it can be made public. Therefore, we are not publishing the data at this time.
Comment 3: If the article could be shortened and simplified, it would be more reader-friendly.
Answer 3: Thank you for the suggestion. We agree that the article should be simplified. We have revised the text to reduce complexity and eliminate repetitions wherever possible.
Comment 4: Abbreviations: EDA, IVL, OLED are not explained.
Answer 4: Thank you for pointing this out. We have now included explanations for these abbreviations.
Comment 5: Errors occur in many places: Error! Reference source not found.
Answer 5: These errors were due to not using the correct submission template. They should now be resolved.
Comment 6: There are no graphs of sample measured characteristics with descriptions.
Answer 6: We have now included these graphs. Please see Figure 2. Thank you for the suggestion!
Comment 7: In my opinion, the word reliability is not used correctly in relation to algorithms.
Answer 7: We have changed the word, we have used "reproducibility" and "robustness" instead.
Comment 8:However, I have the most doubts about Figures 7, 8 and 9. I understand the idea behind their creation, but they do not show much. The points overlap and that is all that can be concluded from the most important figures in the article. Similarly, the titles of these figures do not correspond to their content and the purpose of their presentation. In my opinion, this is a comparison of characteristics obtained from measurements with characteristics determined on the basis of modeling. In my opinion, the artificially added third dimension obscures these figures. For how many temperatures are the measurements presented? If for 4-5, specific values ​​can be placed instead of the third dimension.
Answer 8:
As you can see in the updated Figure 2, the junction temperatures do not always match the setpoint due to limitations in the measurement equipment. However, with the new strategy, these measurement points are included in the parameter extraction process. The model’s performance is also checked at these points, which is why the figures only showed plotted points.
In response to your comments, we have revised the figures and the data set. We collected values that match the target temperature and limited the plotted points to these, allowing us to draw isothermal lines, which enhance the visibility of the figures. We hope this revision makes the figures more informative. Additionally, we have extended the descriptions of the figures.
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsThe current state is basically fine and it can be accepted
Comments on the Quality of English LanguageThere are several typos and grammatical errors throughout the manuscript, the authors need to double-check again carefully.
