Harbin Institute of Technology: Their national competition is like a roller coaster

Everyone hopes that everything will go smoothly in the national competition (i.e. TI Cup 2019 National Undergraduate Electronic Design Competition), but the competition will always give you unexpected surprises. Maybe it is the anxiety when you can't answer the questions, maybe it is the confusion when you get the questions, maybe it is the compromise when there are disagreements, maybe it is the helplessness of staying up late and your brain is not working, and more often it is the frustration of encountering a difficult bug.

Yao Kai, Cao Libo, and Wen Zhaoliang, three friends from the School of Electrical Engineering and Automation of Harbin Institute of Technology (hereinafter referred to as "HIT"), met in an innovation base group dedicated to the competition when the school required the participants of the national competition to form a team on March 22, 2019. Originally, Yao Kai knew both of them and had some understanding of each other's temperament and technical strength, so they naturally came together.

From left: Wen Zhaoliang, Yao Kai, Cao Libo

Starting from June 20, the end of the semester, less than two months before the national competition, the team began relatively concentrated training. After several rounds of almost non-stop simulation competition questions, the three people became people who could rely on each other. Yao Kai was responsible for hardware. He could handle all the hardware problems in the team very well, and he was also a top-notch welder. Wen Zhaoliang was mainly responsible for theoretical analysis, simulation, and report writing. His strengths in simulation and theory allowed the team to make more theoretical attempts. Cao Libo was responsible for software, which was very reliable. Once the software was handed over to him, there was no need to worry about the follow-up. In theory, the strength of such a team without any shortcomings was quite outstanding among the participating teams.

However, despite this, they experienced a roller coaster-like experience during the 4 days and 3 nights of the national competition. At some points, they even felt that they would miss the national award. Looking back now, the excitement of the whole process is still vivid. Fortunately, they finally won the first prize of the national competition C, and this experience has become their deepest memory.

Should we choose C when we are undecided?

If you choose C, everyone who is used to answering multiple-choice questions will smile knowingly. However, the three friends did not choose C for this reason. "I am so upset." The moment they saw the question, their reaction was the same as that of many teams who prepared for the national competition based on power supply, because none of the 8 questions was a traditional power supply question.

He forced himself to cheer up and circled three questions, A, C, and D, from the existing questions. Question A is about the dynamic wireless charging system for electric cars, Question C is about the line load and fault detection device, and Question D is about a simple circuit characteristic tester. What they have in common is that they can all use relevant knowledge points about power supplies to a greater or lesser extent. Yao Kai recalled the considerations when everyone made the judgment at that time. He felt that Question D was relatively simple, there might be many candidates, and it was difficult to determine whether his team could stand out among many teams; the difficulty of Question A for them was that they did not have ready-made car models and charging hardware, and it was not easy to do a good job in power and efficiency. Question C is relatively new, and compared to other questions, there is no fixed classification direction. "This is equivalent to us being on the same starting line as all the teams doing Question C," and they guessed that there would not be too many teams choosing Question C.

Even though they recalled that their ideas were very clear, the three of them still made the final decision when the topic submission could not be delayed any longer. Because in this team that respects each other, any decision needs to be discussed and convinced by each other, and the choice is made only after repeated arguments. In fact, they later learned from their conversations with classmates that among the 62 participating teams from the Harbin Institute of Technology Innovation Base, only four teams chose topic C, which was similar to their previous judgment.

Tough Decisions

According to the analysis of Wang Zhijun, the responsible expert of the national competition expert group and professor of the School of Information Science and Technology of Peking University, in the competition question answering video, Question C focused on the students' analysis of the circuit characteristics of basic capacitors, inductors, and resistors. At the same time, combined with the artificial intelligence environment, it examined the students' ability to design agile systems to achieve real-time automatic monitoring.

In fact, Yao Kai's team tried two solutions throughout the game:

Solution 1: Apply a single frequency sine wave to the A end of the measurement loop, connect the B end to the ground through a sampling resistor, measure the voltage and current amplitude ratio and phase difference to determine the component type and calculate its value; determine the load type by measuring the load network impedance at a certain frequency and comparing it with the theoretical impedance of eight load networks composed of three components; calculate the fault distance by finding the relationship between the short-circuit loop resistance and the short-circuit fault point distance. This design method is simple in concept and has a simple circuit structure, but the actual load network judgment reliability is low, there is no obvious distinguishing feature, and it is easy to confuse. At the same time, it is difficult to grasp the accuracy of sampling due to the small magnitude of voltage and current, and the measurement error of the fault point distance is large.

Solution 2: Using the principle of digital bridge, the A and B points of the circuit to be tested are connected to the measuring end of the bridge. The phase angle is still used to determine the load type, and the volt-ampere method is used to measure the component value. When judging the network structure, a tree search is performed through the feature points. For nodes with fuzzy features, the binary method is used to find the resonance point, and the judgment is made through the resonance point features. This speeds up the detection speed and improves the detection accuracy. When locating the short-circuit fault point, a 10kHz sine wave is output, a push-pull circuit is used to output a large power, and a current closed loop is used to stabilize the load current to 200mA. The short-circuit impedance and the distance from the short-circuit fault point are used to locate the fault. The device is equipped with a second-order bandpass filter to filter out the environmental noise added in the question. This solution has a strong current carrying capacity and is highly adaptable to different load networks.

In fact, this process was not achieved overnight, and it also went through many twists and turns. On the first day, the three of them continued to do front-end power amplification based on Plan 1, and built a lot of experimental circuits using the power operational amplifiers at hand, but there was no substantial progress. But in the process of constantly flipping through books and looking up information, they accidentally gained a digital bridge circuit. Although they had never used a digital bridge before, they thought this structure was very clever in theory, so they decided to use a digital bridge circuit for the back-end processing, which is Plan 2. The difference between it and Plan 1 is basically the back-end circuit.

So, they built the digital bridge and adjusted the power amplifier circuit. It was 12:30 in the morning, but the circuit performance still did not meet expectations. "After choosing question C, there was only a rough solution, but in the actual process, I found that there were deeper problems that I had not thought of at the beginning," Wen Zhaoliang still has lingering fears when he thinks about it now. "At that time, my mentality was particularly bad. I felt that there was no hope of winning this year, and I began to question myself. At that time, both in terms of mentality and results, I was very frustrated." When encountering such a thing on the first day of the national competition, Yao Kai and his teammates can imagine how they felt. As the analysis deepened, more problems arose. After specific verification, it was found that the existing devices and circuit structure made the performance of the entire circuit unable to meet the requirements of the original solution, which was very difficult for the three brothers. How should fault judgment and detection be done? So the three of them gathered together to check information, discuss solutions, take out university textbooks, and recall traditional electrical knowledge, topological structure construction, etc. bit by bit. In the end, all bottlenecks pointed to the performance of the power amplifier.

Since there were no suitable ready-made devices on hand, staying up late was inevitable for the next two days, so on the first night, without a clear idea, the three of them went to bed with a "sense of guilt".

Breaking

During the four days and three nights of the national competition, it didn’t sound very reliable to go to the electronics market, but because the op amp models on hand could not meet the requirements, on the morning of the second day of the national competition, Yao Kai and Wen Zhaoliang went to the electronics market opposite the main gate of Harbin Institute of Technology to purchase, and then repeatedly verified the power, frequency, output waveform, etc. Unfortunately, the test results of the new power op amp were still not satisfactory, so another round of crazy flipping through books and looking up information suddenly found such a circuit: Finally, they decided to try to use the class A and B complementary power amplifier circuit built with op amps combined with transistors and diodes. But it just so happened that they didn’t have ready-made high-power transistors on hand, so Yao Kai and Wen Zhaoliang had to go to the electronics market for the second time. Thanks to Harbin Institute of Technology’s long tradition in the field of electronics, so that the electronics market was opened at the school gate, they didn’t waste too much time on purchasing.