Translated from Huawei Xincheng Community:
The power module is the "heart" of the equipment, undertaking the important role of electric energy conversion. About eight years ago, China's power module industry was still in the stage of technological catch-up. Now, this field has achieved a transformation from "introduction and absorption" to "self-reliance and strength," and Lang Fengqun, a digital energy module welding technology expert at Huawei, is also one of the core promoters and important contributors to this technological advancement.
With deep experience in the development of power module technology, he broke through the ice and embarked on the technical breakthrough journey of digital energy power modules from "zero to one": building an industrial-grade module packaging welding technology platform to lay the process foundation for mass delivery; building a dual-side cooling IGBT (Insulated Gate Bipolar Transistor) module welding process platform for automotive MCUs (Motor Control Units), breaking through the high reliability threshold of automotive standards; building a single-side cooling silicon carbide module manufacturing process platform, paving the way for the large-scale application of third-generation semiconductors, and ensuring the mass production of millions of modules.
How to meet the dual requirements of mass production and extreme reliability? How to apply the craftsmanship spirit of "dedication, focus, and pursuit of excellence" to work and contribute to Huawei's self-developed modules being at the forefront? With these questions, we interviewed Dr. Lang Fengqun to decode the secrets behind the transition of power modules from "following" to "leading."
1. Welding Engineers Are the "Masons" of the Product Tower
Q: What契机 led you to choose Huawei and the field of power module welding technology?
Lang Fengqun: About 20 years ago, I was engaged in the development of power modules abroad and deeply realized that welding is one of the most critical core processes in power modules. As the core carrier of digital energy conversion, power modules need to operate stably under high voltage, high current, and high-frequency scenarios, and the welding process directly determines their conductivity, heat dissipation, and stress resistance.
The complexity of welding goes beyond imagination. The high-voltage terminal welding area must withstand vibration stress, thermal stress caused by temperature changes, and the combined effect of high-temperature oxidation. Any failure in any link may lead to module-level faults. From the perspective of material metallurgy, the thickness of intermetallic compounds formed between solder and base material must be precisely controlled below several tens of micrometers to reduce stress. A deviation of more than 10 micrometers in the thickness of the solder sheet and the component to be welded could potentially cause batch virtual welding... These factors make welding a systemic challenge that integrates materials, thermal mechanics, and mechanical engineering.
Welding is important and has many topics and challenges. Based on my professional knowledge in materials and previous research in welding, I chose this field and like it. In my view, welding engineers are like the "masons" of the product tower, seemingly basic but determining the stability of the entire tower.
My connection with Huawei dates back to my doctoral studies abroad, where I often heard about Huawei's technological breakthroughs. Huawei's focus on technological development and continuous progress deeply impressed me. During international academic conferences, I also met friends working at Huawei, who impressed me with their professionalism. Through further communication, I genuinely felt the company's dedication to technological research and development, which highly aligned with my own research初心. Moreover, Huawei's technological layout always aimed at the industry's cutting edge, which further solidified my decision to join Huawei.
At that time, the domestic power module industry was still in its early stages, but the new energy industry was experiencing explosive growth. As a core component in fields such as photovoltaic inverters, electric drives, and energy storage, power modules saw a surge in demand, presenting a rare opportunity. So, when I received an invitation, I immediately decided to join Huawei, hoping to contribute to China's power module transitioning from "following" to "leading."
2. To Do It, Do It Better, More Reliable, and Higher Quality
Q: How did the "self-reliance battle" of energy modules begin?
Lang Fengqun: In 2016, with the diversification of photovoltaic inverter types and gradual performance improvements, the power modules available in the industry could no longer fully adapt to various derivative products of photovoltaic inverters. To meet industry needs and ensure the robust development of power supply products, Huawei Digital Energy (then the Network Energy product line) decided to establish a module R&D department, focusing on root technologies of chips and modules. After years of hard work, we developed from a few people and a few guns to a department of hundreds of people, successfully achieving the shipment of millions of power modules.
At that time, the industry generally used simple structures such as silicone gel encapsulation modules. However, in high humidity environments, the modules were prone to moisture absorption, affecting device reliability and causing product short circuits and damage, especially posing a risk of failure for photovoltaic equipment exposed to extreme conditions such as high altitude, high temperature, and high humidity.
After in-depth discussions, the team decided to take a differentiated approach, abandoning the conventional solution and adopting a "plastic encapsulant + high-strength structure." This "cement and rebar" approach offers anti-moisture, anti-humidity, anti-high-temperature properties, as well as resistance to mechanical impact, high reliability, and good heat dissipation. Our goal was clear: to do better, more reliable, and higher-quality designs, building product core competitiveness through technological leadership.
Image source: "Huawei People"
Q: The path of new product development is inevitably full of difficulties. How did the first generation of products come into being?
Lang Fengqun: At that time, large-power plastic-encapsulated modules for photovoltaic inverters were a first in the industry, with no references. The core challenges were two-fold: one was the development of low void rate high-reliability internal welding technology for modules, including the development of high-reliability chip welding technology that can withstand multiple reflow processes; the other was the development of large-area plastic encapsulation technology.
Low void high-reliability welding is like laying a road for chips, requiring both strong adhesion and fast heat dissipation. When the power chip works, it instantly becomes a "small furnace," and the heat rapidly increases, relying entirely on the welding layer to quickly dissipate heat. However, if there are bubbles or voids in the welding layer, it's like having obstacles on the road, causing heat to get stuck halfway, eventually frying the chip! To eliminate these obstacles, we thoroughly studied the welding materials, conducted repeated experiments, and finally selected a high-reliability solder specifically for vacuum reflow welding, which not only adheres strongly but also doesn't "fail" due to secondary high temperatures. At the same time, we also cracked the "culprits" behind void formation, optimized the welding process, and successfully overcame this technical challenge, allowing heat dissipation to be as unimpeded as possible on the chip!
The plastic encapsulation of large-power modules is like putting a "armor" on precision electronic components, using special materials to encase the modules tightly, like pouring cement, to protect against moisture, dust, and shocks, ensuring the devices remain stable even in harsh environments. But this "armor" is difficult to create; if the encapsulant is too hard, it cracks like a brittle cracker, and if it's too soft, it can't withstand high temperatures. If there's a mistake during encapsulation, the inside might peel off like a layered pastry. Worse yet, the ceramic substrate as the "skeleton" might develop cracks due to encapsulation stress.
We repeatedly experimented with the encapsulating materials like preparing a secret formula, and optimized the process parameters and module structure like "micro-carving" to finally create a "protective cover" that is both sturdy and snug.
During the development of the first generation of modules, problems arose one after another. There was once a technical breakthrough at the EMS factory. At that time, we were focused on solving the issue of chip welding voids, but upon reaching the production line, we discovered another tricky problem: workers frequently performed rework welding on the terminals of the plastic-encapsulated modules.
This raised my alertness. After investigation, we found that the terminals were misaligned during welding, causing them to hang in the air and resulting in poor welding. It was like "pushing down one gourd and another rising up." At that time, the issue of chip welding voids had not been completely resolved, and the problem of terminal welding defects emerged. Facing dual challenges, we had to tackle two technical breakthroughs simultaneously.
To overcome the challenges as soon as possible, we formed a joint task force with the EMS factory, analyzing, experimenting, and optimizing day and night. After nearly two months of effort, we finally identified the root cause: the mismatch between welding process parameters and positioning accuracy, and successfully found a solution.
This experience made me deeply realize that true technological breakthroughs often come from the front lines. Experts must go to the site, observe carefully, and accurately identify problems. Only through open collaboration and teamwork with partners can we efficiently overcome difficulties.
After 1.5 years of collaborative efforts, we successfully developed the industry's first large-power plastic-encapsulated module and continuously improved it to build a platform capability for industrial plastic-encapsulated modules. Based on this platform, we developed dozens of industrial modules and successfully achieved mass production.
3. Technical Breakthroughs Are Like "Building Fortresses," Proceeding Step by Step and Taking Multiple Paths
Q: From following to leading, what efforts have we made?
Lang Fengqun: The modules in photovoltaic inverters are like the "heart," and their quality must be perfect, and they must be supplied adequately and safely.
As the power handled by industrial-grade modules increases, the modules generate more heat. Traditional module cooling methods rely on thermal paste between the module and the heat sink for heat dissipation. However, as the product's service life increases, the thermal performance of the thermal paste degrades, leading to reduced heat dissipation and potential module failure due to overheating. Therefore, the cooling performance of the module is particularly critical.
We analyzed the product pain points repeatedly and proposed a new solution: welding the large-power plastic-encapsulated module to the heat sink. Because the thermal conductivity of the solder is at least one order of magnitude higher than that of thermal paste, it's like walking barefoot on iron in winter compared to wearing shoes—much cooler and faster heat transfer! Additionally, the module and heat sink become a single unit through welding, significantly extending their lifespan.
The core technology of this solution is the large-area welding of the module and the heat sink. The main challenges are two: one is the high reliability of large-area welding, and the other is the low void rate of large-area welding. Our goal is to reduce the void rate through the coordinated optimization of materials, processes, and structures, while withstanding long-term cyclic thermal stress, enabling photovoltaic modules to operate safely outdoors for years.
First, we needed to solve the high reliability problem faced by modules during long-term operation. Due to a maximum temperature difference of up to 190 degrees Celsius between the module chip and the outdoor environment, the large-area welding of the module and the heat sink generates significant stress. If the welding layer cannot withstand this stress, it will crack, leading to module and inverter failure.
At that time, there were no mature experiences in the industry to refer to, so to ensure the success of the project, we selected some alternative solders and designed multiple technical defenses, starting from materials, surface treatment, and structural adjustments. However, the initial experimental results were still unsatisfactory, with a high void rate in the welding, and the project faced serious challenges.
To address the reliability technical issues, we also prepared a solution for interface stress relief. We learned that a certain surface treatment method from a domestic university might be worth trying, so we quickly went to collaborate with the university and ultimately successfully established the interface stress relief technology. We then repeatedly adjusted the solder composition, going through multiple formulation iterations until we found an alloy that met the reliability requirements.
After the first success, we continued material optimization and process experiments on the trial production line, eventually forming a complete solution and applying for multiple core patents. Technical breakthroughs are like "building fortresses and fighting stubborn battles," requiring step-by-step progress and thorough implementation, taking multiple paths simultaneously.
The second challenge was to develop the "low void rate large-area welding" process for the module substrate and heat sink. Photovoltaic inverter modules have high power density, and each module must withstand currents as high as six to seven hundred amperes, with chip temperatures reaching over 150 degrees Celsius during operation. If there are voids in the weld points, they can be punctured, causing heat to accumulate and the chip to "overheat" and fail, or even cause a short circuit by puncturing the chip. Therefore, the void rate in welding is extremely high, and even a void rate of less than 3% in the core welding area of the module and heat sink would significantly reduce the heat dissipation performance.
There are many factors that cause welding voids, making it a systematic engineering problem. I led the team to conduct technical breakthroughs around the clock, using the "people, machines, materials, methods, and environment" analysis method to optimize solder composition, adjust various process parameters in the vacuum reflow process, improve fixtures, optimize equipment performance, and establish appropriate "thermal field control" for the product...
After repeated experiments, ultrasonic testing still showed that the samples had large welding voids, and the team encountered repeated setbacks. Finally, many people began to doubt whether this problem could be solved. As an expert, I should give confidence to my colleagues at critical moments. I told them, "When you feel there's no way out, that's when there is a way. Those untraveled roads might be the path to success."
We worked on the breakthrough for two months, conducting over a hundred experiments, and based on previous experiments, proposed a new thermal field solution, ultimately solving the edge void problem—significantly reducing the void rate and achieving industry-leading levels. This breakthrough filled us with immense excitement and joy, and after months of relentless effort, the results finally came to fruition! These breakthroughs laid a solid foundation for subsequent development and continuous competitiveness. Today, this root technology has been maturely applied to the mass production of multiple modules and is now stably produced on multiple production lines.
Working on the production line. Image source: "Huawei People"
4. Quality Is Not "Important," It Is "The Only One"
Q: During the process of power modules moving from "leading" to "mass production," have you experienced any memorable stories?
Lang Fengqun: I remember one incident clearly. After solving the problem, the yield fluctuated, and we were just one step away from mass production. However, we couldn't find the reason for the unstable yield, and we checked every direction, but still found nothing.
To find the root cause of the problem, I went to the production line for experiments. During the experiment, despite wearing plastic gloves, I still felt that the solder thickness was wrong: "Two different thicknesses of solder sheets may have been mixed!" This sounds incredible, and those around me were also doubtful. How could differences of dozens of micrometers be detected by touch?
I took out the two different thickness solder sheets and asked two groups of people to measure them using different methods. The result surprised everyone: "They were really mixed!" They were curious how I discovered it. I joked, "After working in this field for so many years, it's like a traditional Chinese doctor diagnosing a patient, and you can tell the problem just by touching." In reality, the secret is just two words: getting close to the front line and getting your hands dirty. Problems and answers on the production line are hidden in details, and we need to explore and solve problems through practice and analysis.
We immediately stopped the line for inspection and found that the main issue was human error causing the mixing of solder sheets. The incoming solder sheets were not marked by batch, and the warehouse was not stored in separate areas.
We sealed the mixed batch and re-verified with new materials. When the yield stabilized at the level required for mass production, everyone was very happy and relieved. In the future, to prevent such issues, we must implement physical error-proofing, process standardization, and personnel empowerment. Details determine success or failure, and even a few micrometers of error can escape the system, but not the touch of an experienced craftsman. However, truly reliable mass production cannot rely solely on a "human defense," but must build a firewall through "error-proof design + process rules."
Q: 2019 was a special year. Faced with a severe external environment, how did the team respond?
Lang Fengqun: As the external environment became more severe, some advanced production factors became unavailable, and some modules needed to be quickly taken over. The module technology platform we built from scratch and the deep-rooted module fundamental technology we had cultivated for years now came into play. Moreover, as we continued to develop, we achieved industry-leading performance in heat, electricity, and reliability, which can be considered as "turning the backup into the main," and our past hardships were all worth it.
Previously, our modules were in the development stage. The shipment volume might have been a few thousand modules, but now the power module has vertically increased, and the shipment volume has reached the level of millions. Not only does the market performance exceed that of competitors, but the scale has also reached the forefront of the industry. Some say we were lucky, but I believe it is due to the company's forward-looking vision and early planning, which is the result of the brothers' hard work, turning the "back road" into the "front road."
Q: What do you think are the most important traits of a technical expert?
Lang Fengqun: Experts should be down-to-earth. I am accustomed to solving various problems on the production line. For example, checking the production date of solder paste on the production line or checking the standardization of operational procedures. These "butchering" experiences actually make me understand better how to use a "scalpel" to accurately tackle technical challenges.
Normally, when I have free time, I like to read papers and attend academic conferences, but the core remains rooted in the front line. After all, the problems on the production line won't be written in literature. Only by staying on-site can we grasp the details and determine the direction, giving the team a "peace of mind."
Image source: "Huawei People"
Q: How do you understand the weight of quality in technological research and development?
Lang Fengqun: "Quality is the lifeline" is not just a slogan; it's a truth accumulated through countless lessons. Photovoltaic inverters are installed on mountain tops thousands of meters high in the Qinghai-Tibet Plateau and in the high-temperature deserts of Xinjiang. Once a problem occurs, the maintenance cost is huge. In these extreme scenarios, quality is not "important," it is "the only one," and quality is the precise stacking of every detail.
The most direct comparison is the fire test of energy storage equipment. Once the equipment catches fire, the consequences are severe, but our equipment can achieve "fire without spreading"—even if a single equipment box burns for seven hours, the surrounding equipment can still operate normally.
Quality is a fragile item in technology; one mistake can destroy ten years of reputation. Winning in technology may lose in details. Therefore, when designing products, we must consider all kinds of problems that may arise during mass production in advance and prepare accordingly, ensuring end-to-end supervision of the production process. Details decide success or failure, and we need to deeply explore the technical details of each process.
5. Looking Ahead: Innovation Brings New Hope
Q: How do you view the evolution of power module welding technology in digital energy?
Lang Fengqun: We have always tried to expand welding technology from a systems perspective. In terms of material compatibility, power modules are evolving towards high density and high temperature, but packaging technology has not kept up with this demand. Together with relevant departments in the company, we continue to invest in new materials and new processes, developing new technologies such as high thermal conductivity and high reliability welding, sintering technology, and diffusion welding technology. In addition, in terms of AI applications, to better control welding voids, we have tried to introduce AI automatic identification technology to improve efficiency and prevent missed inspections.
Q: What advice do you have for young engineers?
Lang Fengqun: Young colleagues are all excellent and can endure hardship. Their resilience and work ability are no less than ours. The times are really moving fast, but some things haven't changed, such as attention to detail, respect for quality, and persistence in R&D. No matter which generation of engineers, once they enter the workforce, they must engrave rigor into their bones.
The most important points I want to share are a few aspects that should complement each other: first, regard R&D work as a pleasure and enjoy it; second, details determine success or failure; third, get close to the front line and tackle tough challenges. If you have the opportunity, actively participate in exhibitions, academic conferences, and seminars, which can not only broaden your horizons and absorb cosmic energy but also help solve problems encountered in product development or production processes.
This article is an exclusive contribution from Observer, and unauthorized reproduction is prohibited.
Original: https://www.toutiao.com/article/7521942941816914474/
Statement: The article represents the personal views of the author. Please express your opinion by clicking the [Up/Down] buttons below.