Goodnight Light Circuit Project Design and Solder - Jin

Through an introduction to electronics course, I was able to learn the basics of a night light circuit and went through the prototyping process using a breadboard and the soldering process that brought the circuit to life.

What is a Goodnight Light Circuit?

A Goodnight Light Circuit, also known as the night light circuit, senses the environment's light intensity and turns an LED on when it is relatively dark in the surroundings.

The photoresistor used in the circuit is responsible for detecting the light intensity. It reports different resistance readings, with high resistance as dark and lower resistance as bright.

To convert the photoresistor readings from resistance to a voltage value, we utilize the characteristic of voltage dividers in a circuit. When two resistors are connected in series and a wire is plugged in between the two resistors, making them parallel, the voltage going out from the wire would be the second resistor's resistance divided by the sum of the two resistances.

This is useful because by placing the photoresistor at the R2 position, we could derive a voltage output that can be compared.

To compare the voltage derived by our voltage divider for the photoresistor to a threshold that indicates bright and dark, we need to use an Op-Amp (operational amplifier). 

While an op-amp could have multiple usages such as applying different math equations to electronic signals or amplifying signals, we here uses it as a comparator. When the non-inverting input of the op-amp exceeds the inverting input, the output voltage will equal the positive voltage input of the op-amp.

Knowing this logic, we could set the non-inverting input as our threshold, while the inverting input can be the actual voltage that has been derived from the photoresistor's resistance value.

When it senses smaller resistance (it's bright enough), the resulting voltage is smaller than the threshold; the LED gets no voltage input (since V- is the ground), and vice versa.

The whole goodnight light circuit looks like the one above.

Breadboard Prototyping

Before designing the schematic and the actual PCB, prototyping is a good way for us to understand how the circuit should be wired to serve the function of the goodnight light.

Source: https://www.te.com/en/products/passive-active-electronic-components/resistors/intersection/resistor-color-codes.html

An interesting fact I have learned about resistors is that the resistance can be read from the stripes and colors on the resistor itself. The first few bands are the significant figures of the resistance, and the second-to-last band is the multiplier, while the last band is the tolerance.

Otherwise, done by measuring with a multimeter (which is usually faster), reading the resistance from the color may help you identify some really common resistances like 10 kΩ and can help you identify resistors that have the same values easily too.

My final circuit looks like the one above. After finishing the prototype, I began to wonder how it translates to an actual circuit and what each pin on the op-amps represents.

That led me to look up the datasheet for the op-amp that we used: LM358-N. I first tried to draw out the circuit and figure out which pin corresponds to which. To my surprise, my initial evaluation was right, and that gave me a great sense of accomplishment.

Then, I tried to draw out the circuit of the prototype that I built, and it connected to how circuits are read on the breadboard since it was in a completely different configuration. The calculated threshold and the dark and bright conditions for the photoresistor made sense, as the threshold is in between the bright and dark values.

To help myself understand better what the breadboard circuit looks like, I spent some time drawing out the board in Tinkercad based on my memory and refined it to fit the prototype.

Converting into Schematic and PCB Layout

The next step is to convert the breadboard prototype into a KiCad schematic layout that is later used to convert to the actual PCB layout.

A few differences were noted while converting. First of all, the battery header is required as part of the circuit because we have to tell the circuit where our power will be coming from. The power flags are the indications for where the battery wires will connect.

The next difference I have noticed is how the op-amp is divided into different units to display in the schematic instead of the normal op-amp display with 5 pins. This is because the LM358 op-amp we are using has 8 pins and has an A and B op-amp within, each providing 3 pins, while the V+ and V- pins are shared between the two (marked as Unit C).

Since we only used the A pins, only unit A and unit C of the op-amp are required. Unit C connects to V+, which in this case is the 9V battery, while V- connects to the ground. I have noted that a diode is required here because it helps prevent the circuit from crashing if the battery is accidentally plugged backwards. The forward-biased diode only allows current to flow when the anode is more positive than the cathode, which current only flows through if the anode is 9V in this case (connected correctly).

Additionally, a capacitor is needed between the 9V and the ground, especially closer to the battery source. This is a decoupling capacitor that helps cope with sudden surges in voltage demand when lighting up the LED. The charge inside the capacitor (charged to 9V at steady state) will discharge to meet the demands first to prevent damage to the battery.

I then convert the finished schematic into the PCB layout editor and lay it out after assigning footprints (how the actual component looks in PCB layout) to each component.

The steps are as follows:

1. Set the boundaries of the PCB to 1 inch x 1 inch
2. Arrange the components inside the schematic (put the LED further away from the photoresistor to avoid a "while true" loop, and put the decoupling capacitor closer to the battery)
3. Reference the schematic layout to know which components are connected to which
4. Connect the wires between components (leave ground and power source)
5. Fill the top layer as +9V, and the bottom layer as GND
6. Do the final wiring of the GND using vias to the second layer
7. Run DRC and fix any warnings and errors

Unfortunately, since there are PCB boards and components provided by the course, and there are usually minimum order requirements for manufacturing, my own layout wouldn't be used for final soldering.

Soldering and Testing

As instructed, I started with the surface-mounted components first, which are the resistors, the op-amp, the capacitor, and the diode.

I first started with the resistors. Since it's my first time soldering, I haven't quite grasped the sequence of applying solder to the components, and some of the "surface-mounted" resistors look like they're floating in the air.

After several attempts and trial and error (the solder is easy to fix by reheating it), I finished the resistors and moved on to the op-amp.

The op-amp has 8 very close pins, and the op-amp itself is also super close to other resistors. While soldering the op-amp, I alternated between sides and made sure that the pen was only touching one pin at a time. A resistor near an op-amp was once moved and turned because I accidentally touched a soldered resistor with the pen while soldering the op-amp, but it was an easy fix, as I reheated the resistor connection points and reapplied solder.

For the diode, instead of applying solder onto the component and the plate, I was taught to apply solder onto the plate first to provide a base for the diode to better connect to the board, as the pads for diodes were small.

Through-hole components were soldered using a different method. I had to heat up the holes first and push the solder into the hole to fill it up. It created a little volcano shape around the wires, indicating a successful solder with just the right amount of solder used.

Not surprisingly, the circuit did NOT work on the first try, as the LED wouldn't light up. With the help of the TAs, we investigated each connection using a multimeter to identify where the circuit might not have been completed, but we found none.

It came out to be the LED that was broken because by touching another LED to the board, it lighted up perfectly. I first attempted to remove the solder from the holes by heating it up and removing it a little by little, but it couldn't be fully cleared, and the LED still remained in place.

Collaborating with one of the TAs, we cut the LED down to only the wires and pulled the wires out while heating up the solder. Then we used a desoldering pump to suck out the solder from the holes, leaving us with a clean hole to place a new LED inside.

The final result was successful!!

The whole process, from understanding a circuit, building a prototype for the circuit, and finally soldering an actual circuit, allowed me to gain a lot of knowledge and experience. I really hope to carry the experience forward to making larger and larger projects in the future.

Engineering Group Project 1: Catapult

Since the release of our Project 1 instructions on August 28th, our four-person engineering group has held several meetings to discuss our design and testing, while keeping an electronic design notebook until the project demonstration day on September 16th.

The project spanned around 1 month, and our final project turned out to be something really valuable, so I decided to record this experience here.

Design Goals:

The design goals for this project are to shoot accurately while being able to shoot as far as possible. Two tests are conducted during the project demonstration: the accuracy test and the maximum distance test. 

During the accuracy test, the catapult needs to fire a projectile weighing 28.35 grams at a target 8ft away from the front-most part of the device while clearing a battlement with a height of 1.4 ft placed 1.5ft from the starting point.

During the maximum distance test, the catapult must launch the projectile as far as it can while still clearing the battlement placed at the same location as the first test.

The team's final score will be evaluated through a scoring equation:

Brainstorming and selecting the solution:

After our first meeting, we assigned the task of researching different catapult ideas to every group member, allowing all our ideas to contribute to the final idea, maximizing the advantage of group work.

The two most notable designs are shown above: a trebuchet with a winch and strings that pull down the catapult arm, and a regular catapult that launches by pulling a straw out.

Using an evidence-based decision matrix, we effectively selected the best solution that fits our needs. We weighed our customer needs that we identified for this project on a scale of 5 to 20, scaling from not-so-important (5) to very important (20). We then ranked each criterion of the two options on a scale from 1 to 10, with material usage, weight, and size being the smaller the higher the score, and structure simplicity, stability, and maximum distance being the greater the higher the score.

The total score of each option is evaluated by multiplying the weight by the scores and summing them up, resulting in our catapult being the better option.

However, we did modify some parts of the designs from the original brainstorming so that the best ideas are kept and combined into the best solution we think would be.

Solution Building Process:

Before constructing, we have 3D modeled the catapult in Onshape to get a hold of how the catapult would turn out to be like while confirming all the functionalities we have discussed are all included and feasible instead of having to build it out from scratch. 

Since material usage is also an important part of this project, we wanted to be as careful as possible before starting the actual construction.

Catapult Solution:

Our final solution is shown above. The process of translating a 3D model into a real-life object caused us to encounter some problems, as force and structural stability were not as clearly demonstrated in the model because every component just sticks together like super glue.

The most prominent problem was the base not sticking together with straws because there wasn't enough surface contact, and the straw cuts were not clean enough to be fully attached to the popsicle sticks.

The other problem encountered was the winch not being able to hold the increasing tension of the catapult while winding the strings down when the winch is made using a straw. Straw was selected initially because its round characteristic allows a smooth rotation. Despite trying to fill it with glue, it did not provide enough strength, so we had to switch to stacking popsicle sticks to make an axis.

Our Final Demonstration

This was one of our demonstrations on September 16th. The catapult shot very accurately at the targeted point, but it was unfortunately one of the attempts for the max distance test and not the accuracy test.

Executive Summary

Anchor & Accelerate Venture Pitch Competition: Finalists

My roommate Melina and I formed a team to compete in the Anchor & Accelerate Venture Pitch Competition, where we had to pitch for a company case under 3 minutes as if we were the company founders. We chose HeatHalo as our case, a company that sells wearable biomonitors and pairs them with a mobile app with an optional subscription to keep track of athletes' heat stress signals.

The competition has two rounds: the preliminary and the finals round. There are 30 teams in the preliminary round and down to 8 teams in the finals round, while the top three teams could earn prizes in cash. After our preliminary round, the judges gave us some positive feedback, and we thought we did great.

The finals schedule was supposed to be announced through emails, but we didn't receive any information. We even checked the competition website, but there wasn't any sign that we got into the finals. It was when Melina talked with her friend who also went to the competition that her friend showed her that we were on the list of finalists. We had to run to the competition as it had already started.

The run was terrible since I took my backpack that has almost everything, including my laptop, iPad, and books for next class. It resonated so much with our own pitch. "Encouraged to push through the pain", and "Doesn't feel the heat's impact until we feel sick", even though we didn't faint or anything; who knows if it had an effect on us.

It was a relief that we arrived on time while no one had started to present. Until we were called on stage, I was still unsure if we were supposed to be here since I still couldn't find any evidence of where we got in. We performed the pitch and took on the Q&A, and we were not on the list of the top 3.

The following are our slides and our pitch:

My name is Jin, and my name is Melina

Together, we are HeatHalo.

Heat kills more Americans than floods, tornadoes, and hurricanes combined. 

For athletes, heat stroke is a major factor in each of the top three causes of death in student athletes. Yet it’s entirely preventable. Athletes rarely rely on how they feel - encouraged to push through the pain -  and by the time they feel overheated, it’s already an emergency. This blind spot is where HeatHalo comes in, making the invisible visible with our seamless biosensors.

HeatHalo is a wearable biomonitor that continuously tracks the body’s heat-stress signals, which are early indicators that show up long before an athlete feels sick. Paired with a mobile app for players and an AI-driven dashboard for coaches, staff can monitor the entire team’s heat status in real time at a glance. By incorporating a scalable subscription-tiered model, organizations can add on sport-specific metrics based on their needs. With HeatHalo, all coaches can protect their athletes– before a heat emergency ever happens.

The timing has never been more urgent. 2024 recorded the hottest day in history. In 2023, daily ER visits for heat-related illness were at record highs. Now is the moment.

At HeatHalo, we serve every athlete. Sports in the US are growing rapidly. As of right now, we have over 8 million student athletes. That's over 420 thousand teams spanning U.S. high schools and colleges; that doesn't even take into account club teams, the military, and marching bands. 

We make money two ways: on an organizational scale, teams enter contracts with us and buy the wearable devices for their athletes. This is our primary revenue. We also incorporate, however, our tiered model as Jin described before. We take the baseline features and deepen the insights using a combination of sensors and AI-powered data on specific sports. For Heathalo, athletics is just the start — the same technology that can protect athletes can expand into workplace safety and heat-vulnerable patients, opening doors to new possibilities and shedding light on one of the most invisible killers.

Why us. Our CEO, Dr. Naomi Patel, is an ER physician who’s treated heat illness firsthand; that’s clinical credibility. Our CTO, Elena Ruiz, is a biomedical engineer specializing in wearable sensors. And our COO, Jennifer Crawford, has launched consumer health products; she takes them to market and helps us soar.

We’re raising $1.5 million in pre-seed funding to finalize the product, run pilots with our first athletic organizations, and prepare our first manufacturing run. In 18 months, that's 1,000 devices deployed, $300,000 in revenue, validated data showing fewer heat incidents, and our premium subscription live. 

Heat isn’t slowing down, and neither are we. Join us in making the invisible visible.

Overall, it was a great experience preparing. I loved the process of making the slides and practicing the speech to sound convincing and confident. The process of market and problem research to support the company's solution was the more difficult part, yet I still managed to find some sources that helped us. Melina was sick during fall break and was busy when she came back, so it was two days before the competition that we finalized our script and communicated our thoughts. We didn't have the time to practice the whole process of the pitch together before competition day, so we only practiced individually, but it was surprising that our pitch went really well.

If I could improve next time, I would like to start working on the pitch together earlier, so there wouldn't be that much stress before the competition.

For my personal skills, I feel like presentation has not been too much of a problem, and I could perform fairly well. However, my weak spot is definitely the Q&A session and talking with investors (judges) in a casual mindset, and speaking while I think.

Every time the judges ask a question, it doesn't always come clearly to mind what exactly they are asking for, and I need time to react and think before I respond. Maybe it is because I am not prepared enough to know what questions to expect from them; perhaps it would be better to prepare some responses beforehand. Compared to my robotics interviews and questions, I could handle them really well because I know what I am doing and am really confident in my work; maybe that's the difference because I don't have a comprehensive enough understanding of the company I am pitching for.

After the competition, the judges are really just normal people; my roommate and other contestants had really natural conversations with them, as if talking to a friend without a barrier. It was surprising to me because I had put them a level above me, and maybe that is why I wasn't able to speak as freely because I am afraid of making mistakes. I would like to start trying to talk to people as if they were my age and easy to approach; this may also help me navigate the future career fairs better.

Another thing I have learned is to shake the investor's hand before making the pitch, introducing myself to every single listener. Additionally, after hearing a question from the investor, it would be nice to reframe it like, "So the question is about if...?" That way, I would also have time to process the question while making sure I understood what they are asking. After responding to each question, it's also a good practice to ask if that answered their questions. "Thank you for the question" may also be a good response before starting to answer.

I am still not used to business-like environments like this pitch competition and career fairs. I really did learn a lot from this competition, and I am looking forward to putting myself in more uncomfortable situations to get used to these conversations. By doing so, I feel like it would open up many opportunities through conversations.

Overall, it was a great experience from preparation to competition. I don't know if I would want to do it again, but it would be nice to try.

Don't Worry Darling: Movie Review

Just a disclaimer, I am no professional writer, not to mention if it's gonna be a movie review. This article was written solely due to the fact that I seldom watch movies, but I am getting to like movies a little more than before, and because of that, I would like to document my thoughts on the movie so it's not only an experience but a memory.

Movie Title: Don't Worry Darling

Director: Olivia Wilde

Year: 2022

I watched this movie on a Saturday night with two of my roommates that loves horror movies. Despite not being brave and not liking to get scared, watching it with them reduced the tension of the movie, so I was excited to enjoy the movie night with them.

Initially, my roommates suggested watching Unfriended, which is, according to her, a movie about a Zoom call and supernatural events, but she thinks it would be too dark and scary for a novice horror movie watcher like me.

Plot Summary (HUGE SPOILER DISCLAIMER)

Starting with a simple plot summary, the story takes place in a living community called Victory Project, where multiple families live together in a neighborhood. All families consist of a wife and a husband, and some have children. The husbands go to work every morning at the same time and come back at night. While their jobs are confidential, they only describes it as the development of progressive materials. The wives are responsible for keeping the house clean, cooking, and simply enjoying their lives in the community; the only rule is to not go "outside" to the deserted area.

The main couple discussed here is Jack, the husband, and Alice, the wife. It all started with a normal life, where they enjoyed each day together inside the community, until Margaret, another woman from another family, once went out into the desert and came back saying that she saw something and that the world was lying to them, but of course, no one believed her and instead described her as psychotic.

Our main character, Alice, was once friends with Margaret but turned to ignore her when she called for help. One day, Alice went on a bus ride and witnessed a plane crash in the desert and walked towards it to help, and she touched the mirror on a building. Afterwards, she magically woke up in her bed and later on started to behave in the same way Margaret had, remembering something that she seemed to have forgotten.

After a forceful "reboot" by the community staff, Alice had her memory reset and behaved "normal" again. However, it was her husband Jack's usual humming of a song that ignited the lost memory that she once held.

She was a doctor who did 30-hour shifts of surgery every day, while his husband was unemployed, sitting in front of a computer all day and not being responsible for household chores. They used to worry about money and their circumstances. Apparently, his husband got information about this virtual reality world project, knocked her out, and joined the project just so they could live in a happily conditioned world without worrying about real-life difficulties. 

After she noticed the situation and she accidently killed her husband, she drove and ran towards the deserted area and exited the virtual world into reality through the mirrored building, returning to the real world.

Remarkable Scenes

Character Bunny

Bunny is another wife in the community, and they have two kids. She was friends with Alice since the start of the movie. She was the one who told the audience that Margaret was once also a part of their friend group before she had gone crazy. 

She strongly dislikes it when girls disobey the only rule of this community: not to go to the desert area. After Alice was rebooted and returned to normal, she was the one who first found out that Alice killed Jack. Surprisingly, she told her to run because Victory Project would go and kill her in real life after the husband died (the husband dies in the real world if they die in the virtual world). 

She admitted that she knew everything all along, but chose to stay in the world because that's the only way to still see her child.

Character Shelley

Shelley is the wife of Frank, the guy who created this virtual reality world and made it a product. After Alice started running, the world slowly started to collapse, while Shelley stabbed and killed Frank in one of the scenes.

This was surprising to me, as Shelley, who seemed to choose the type of life and govern the whole project, might have always been pretending and waiting for the time to end this project.

Joining the Victory Project

One key idea my roommate noticed, and I found interesting, was that the user of the Victory Project gets to select whether they have pre-existing relationships or not. Meaning, if they had no partner, the project may randomly abduct a girl to become their wives in the virtual world, which is kind of crazy.

Lesson Learned

After the movie, I despise the character Jack the most, as I feel like he is just a weak guy who could not possibly develop the skills and find a job to sustain the family, while he forces the Victory Project onto his wife, who was sufficiently independent and enjoys working for their better lives.

This whole worldview on the virtual reality idea kind of reminded me of Brave New World, but more unstable. They both contain a world where people live programmed lives while believing that what they are experiencing is happiness due to stability. The difference is that Victory Project failed due to maybe some constraints on modern technology, while also due to the fact that it's used on people who actually had a life and memories. Brave New World designed the world from when people were decanted, eliminating the fact that people would remember something that they shouldn't.

I feel bad for the wives who did not know anything prior to the project, since they would all be forcefully dragged into the project while having to abandon their previous jobs and freedom, even if their lives may have been difficult. If the project was done under mutual consent, it may be morally more acceptable, in some way?

It was quite interesting to slowly discover the secret behind the victory project. I was also quite shocked that no one prior to Alice had successfully escaped from the world, since it's a small community, and I thought it would be quite obvious that humans would search for truth when there are so many restrictions while having to live the same day every day.

自學第八年 回顧與自評 / Jin

回顧期間：2025 / 4 - 2026 / 3 填寫日期：2026 / 3 / 20


1. 這是我們自學的第 8 年，請介紹一下這一年你的課業學習。包括去年五月的 AP 結果，並以美國大學工程學系申請的角度，簡要總結一下你的申請特色，包括 GPA、SAT、TOEFL、AP、課外活動、社會回饋、特殊才藝等。

從上一次回顧自學到現在也已經過了一年，去年5月總共考了7 科AP，考的科目主要在比較難的自然科以及一些我感興趣的歷史和統計。因為去年沒有參加機器人比賽，一整年完全專注在課業的準備，很幸運的能夠在7個科目當中全部拿下5分滿分。這也向我自己證明，其實只要花時間，我是有實力可以讀得很好的。

比較特別的是，在考完我原本以為最難的 Physics C 之後，反而覺得題目的難度怎麼有些簡單，我當時還很怕當中會有一些陷阱沒有注意到。慶幸的是，最後結果出來都是好的，表示自己準備得非常充分。

從去年7月左右開始，準備美國大學相關申請的資料和統整國三到高三這4年來的成績。我的申請特色主要環繞在理工的背景，像是機器人比賽還有報考AP科目的自然科，另外還有一些單純興趣驅使的專長，像是日文和畫畫。

我的GPA Weighted 4.41，SAT 最高分是 1510 分，托福總分 106 分，AP總共考了 17 科，課外活動以機器人比賽為主，其他興趣幫忙更全面地展示我這個人。社會回饋的話我自己有經營一個YouTube頻道，用簡單的方式介紹一些比較複雜的 VEX 理論給初學者聽，這個部分也占了很多我在撰寫自我介紹時的篇幅。

我大學申請是以電機領域為主，第二志願都填機械系。我的主要賣點是4年來豐富的機器人比賽還有去到世界賽的成果，而和電機系比較相關的是，我參加過半導體的營隊以及在國三的時候製作的一臺乒乓球遊戲機。


2. 請介紹一下你所申請的美國大學列表，以及申請策略。

Early Action Schools:
MIT
Georgia Tech
University of Michigan
UIUC
Purdue
Virginia Tech
Arizona State University
The Ohio State University

Regular Decision Schools:
Stanford
Carnegie Mellon University
UT Austin
UCLA
UC Berkeley
University of Maryland, College Park
University of Washington, Seattle
UCSD

美國大學主要分成2個時段去申請。第一個是 Early Action，這是可以最早提交入學申請的時候。他的優點就是能夠向大學展示你的興趣和熱情，但缺點就是你會需要提早把很多事情完成。第二種是 Regular Decision，最一般最大眾的申請時段。根據不同學校，這2個時段的截止日期會不一樣，甚至有些學校只有一種申請時段。

提早申請有一些小小的限制，有些學校如果你有提早申請而且被錄取的話會強制你入學，而大部分的私立學校只會讓你申請一間私立大學的 Early Action。由於大部分公立學校可以一次投多家提早申請，而且不太會綁定入學，所以我的策略就是選擇一家最想去的私立大學，還有盡可能地塞越多公立學校越好。這樣可以分散準備申請的勞累程度。

這個申請策略事實證明對我比較有效，因為我在提早申請的學校池裡面被錄取的機率高非常多，並且錄取的學校理念和風氣比較符合我追求的。我覺得很慶幸自己有做這樣的規劃。

3. 這一年的 VEX V5 是第二年參賽，你進步了很多，也非常努力。請回顧一下今年的成績，並比較與去年的自己，進步了些什麼，跟我們分享。

這一年的機器人比賽，我從賽季開始之前，就研究如何讓整個比賽更順利，也讓隊員之間的爭吵可以越少越好。第一件我做的和之前不同的事情，是在挑選隊友的時候把工作分配考慮進去，甚至在確定新增隊員之前會和他們線上聊過天，讓我大概能了解他們對賽季的目標和期望。

在整個機器設計流程當中，我也努力地從每一個做得比較不好的地方學習，讓每一代機器都能成為下一代機器的養分。另外一個我自己覺得進步蠻多的，是我更懂得融會貫通，也看的懂更多機器結構。這讓我在最後一台世界賽機器的時候，可以完成一台基本上每件事都能做得很好的全能機。

今年成長蠻多的也有和隊友溝通的部分。賽季的初期我太常去猜測怎麼樣做會對隊員比較好，不過賽季後面我變得能勇敢去詢問他們的想法，並和他們討論出共同覺得不錯的方案，而不是要我自己扛著要決定所有事情的壓力。我覺得相較於第一次比賽的自己還有賽季初的自己，我的個性變得比較堅強一點。

今年因為整體規劃比較完善，包括工作日誌的排版和運作方式，都在賽季前就大概想好了，我們賽季當中一場比賽有拿到3個冠軍，這是我非常有成就感的一場比賽。在後面每一場比賽當中，我也能夠清楚的感受到自己在面試跟敘述團隊歷程方面相當完善。雖然評審可能因為不同考量，不一定會給我們我們心目中的獎項，但我覺得能夠全力把所有東西做完已經很棒了。相較於第一年比賽的我，今年在面試上花了更多時間，而且隊員都很樂意配合練習，也會將自己所要敘述的東西完整吸收成自己的想法。我們在面試過程中基本上不太會有空白的時間。


4. 過去一年，以大型語言模型 LLM 技術達成的 AI，改變了世界，越來越普及，也越來越聰明。請分享一下，在過去一年，你最常用那個 AI 工具，以及應用在哪些方面。你認為，AI 對你做為學生的學習需求上，是有用的嗎？

過去一年當中最常使用的AI工具是Chat GPT。最常使用的時間點是在撰寫大學申請作文的時候，常常需要一些回答題目的建議以及可能可以發揮的方向，同時也會詢問關於文法相關的問題。當時在比較Chat GPT跟Gemini兩者時，發現 Chat GPT的文筆比較好，不過互動久了之後，也發現一些重複性很高的文法和句型，很容易就看出AI的介入，所以後來我盡量就不會讓他介入太多。

在學習方面，因為今年沒有考AP考試，使用的機會就少了蠻多。目前大部分詢問如何申請大學或是準備大學入學相關的疑問都會偏向 Gemini，因為可以給予比較快和理性的回答。在自己的生活中，AI使用算相對比較少一點。


5. 這幾年，媽媽大概每年都會安排一些時間，全家到日本旅遊，請回顧這幾年的日本旅遊，分享一下從你自己的角度觀察，你自己有哪些改變？有哪些是你自己也很喜歡的進步？

在這幾年去日本的旅遊中，我發現自己進步最多的是溝通。以前的我可能會害怕店員和工作人員的回答我聽不懂或是和我認知的不一樣，於是就會害怕去詢問。不過可能整體生活的經驗變多了，我能大概預測對方回答的方向和會用到的詞彙，就算聽到不太熟悉的詞彙也能夠大概猜出他們在講什麼。感覺這部分的進步比較是生活經驗上的累積。

溝通上，另一方面的進步在於我不害怕去詢問我不懂的地方。和人聊天，或是詢問問題的時候，如果我聽不懂某些單字，我現在敢直接詢問他的意思，如果沒有聽清楚我也很敢直接請他再講一次。這個部分的進步在於個性上變的比較堅強一點，也不會怕自己表現出不懂的樣子，畢竟在正常的對話當中，就算用中文也可能聽不懂。

溝通以外，我在行程和時間安排上也有一些進步。我發現我變得更了解自己旅途中喜歡做什麼事情，和在看似一開始沒興趣的東西當中找到我有興趣的部分。日文的進步，讓我在去看歷史博物館的時候能讀懂牆上的解說，並且認真學習當中的內容，還蠻有趣的。

我也發現自己有時候喜歡購物行程，有時候想要去公園走走看看，也比較懂自己想吃什麼，變得更懂自己，也更敢表達自己想做的事情。

我最喜歡的進步在於變得厚臉皮一點點，讓自己不會那麼害怕失敗，也不會害怕講錯或聽不懂，變的比較敢問問題，個性也變得比較有韌性。希望這個改變可以繼續保持。


6. 這是最後一年的自學了，秋天你就要前往美國當個大學生，回顧這八年的自學歷程，如果你有機會跟小學的你自己說話，你會跟他說什麼呢？又，如果因為你的傑出表現，有即將上國中的小朋友來問你，是不是自己也可以自學，你會怎麼跟他說呢？

從國小到現在的所有經歷都造就了現在的我。如果能夠和國小的自己講一句話，我覺得會希望自己多去嘗試，並從每一次失敗和成功的經驗中學習。高中比較遺憾的地方，可能是在 11 年級的時候，沒有很好的平衡 V5 機器人比賽和6個AP考試，但我覺得就算提前和小學的自己講，當時的我應該也沒有辦法理解我這些體悟，主要是要有能夠從每個經驗中學習的能力才是最重要的。

我覺得這幾年下來我發現，自學最重要的點就是你要有一個目標去努力，這樣才能發揮自學最大的效益。自從我國中比完VEX IQ世界賽之後，我才真正的意識到自己要開始為大學的目標去努力，高中三年我也才真正喜歡上讀書。當我們有一個想要努力的目標時，就會很慶幸自己有辦法規劃讀書的進度，也才會覺得自學是對自己很有幫助的。我覺得自學要考慮的因素很多，而有些是沒辦法一開始就完全規劃好的，但花一點時間去嘗試也不是壞事。

2025-2026 VRC Push Back 高中組世界賽：比賽心得

每一年 VEX 機器人大賽都會在大約 4 到 5 月時舉辦世界賽，而今年我參加的是 VEX V5 高中組的比賽，日期是 4/21 到 4/24，為期四天，舉辦在聖路易斯的 America's Center 會議中心。每一個賽季的題目和比賽會隨著這世界賽的結束畫下一個完美的句點，而同時也會公布下一年的題目和所有大型比賽預計的時間和地點。

我們的團隊 66994B Unicorn Puncher，從去年四月左右就開始討論新賽季的分工，在去年的 5/25 則是我們第一次四個人實體的聚會。從那次聚會開始到這次世界賽，我們總共研發了 5 代機器，從一開始的籃子機到現在的 S 型機器，我們都經過很多場比賽的歷練並從中找到每台機器的弱點並改進。這次世界賽用的這台五代機可以說是基本上達到了所有機器該有的功能，做每個動作都很順暢。

為了這場世界賽，我們在機器研發和製作上面特別花了多一點的心思。首先，為了確保每一項功能都很好而且讓整台機器輕量化，用最簡單卻又穩固的連接方式把所有功能串接起來，我們先用了 Onshape 電腦輔助的 3D 繪圖去將我們的想法具象化出來，用電子的方式連接各個零件。這個過程大約花了 2 個禮拜的時間去把所有細節確認完，才能避免因為一些現實因素導致組裝時才發現結構脆弱或是孔洞對不上。

另外，為了讓機器看起來更簡潔乾淨，我們特別將一些鋁的零件送去電鍍，也將齒輪煮成黑色的，增加整體顏色一致性。原本的鋁零件都是灰色，而齒輪可能有綠色或紅色。最一開始在電腦上繪製機器的時候，就有將顏色考慮進去，事先模擬做出來的樣子。

出發世界賽之前，為了確保在會場有足夠的零件和工具維修機器，我們特別準備了兩個備用零件箱。為了讓我們四支隊伍都能夠有場地練習，我們準備了兩組場地來擺放在租借的會議室中。為了讓機器能夠完好無損的抵達美國也能平安抵達台灣，我特地設計了一個機器架，把機器鎖在這個架子上面，並把所有空隙塞滿泡泡紙。因為機器架會頂到大箱子的上下端和左右牆壁，只要機器鎖在上面，不管箱子怎麼晃，大架子都能夠保護機器，避免直接受到衝擊。

4/17 星期五下午兩點我們從台中搭乘遊覽車出發，大約下午 5 點多到桃園國際機場。我們將所有托運行李放上行李秤並完成報到，進去過安檢之前大家各自拍了許多張合照。飛行時長 14 個小時，從晚上 10 點多起飛到芝加哥是晚上 9 點多。

抵達芝加哥機場之後，首先將所有行李搬運到我們在機場的飯店後就去休息。隔天早上短暫的吃了早餐之後將所有行李移至遊覽車上，準備搭五個小時的車前往聖路易斯。途中，我們有停在超市採買後面幾天會需要的早餐或生活用品，也順便去附近的餐廳吃飯。

4/19-4/20 練習日

在世界賽開始前兩天，我們預留時間進行機器的校正和練習，也順便讓大家熟悉時差。這兩天下午我們也都有邀請許多隊伍來參加練習賽，用正式的比賽模式進行練習，讓各隊都能熟悉實際比賽的情形並依照狀況擬定正確的策略。

4/19 當天的練習賽邀請到的比較多是在台灣認識的隊伍，我們總共只有七隊，而且整體強度偏中等，大概是在台灣練習賽就能模擬出的強度。4/20 當天邀請到了很多美國名隊，像是 334C 和 334U，他們是在加州一直以來都非常有名的隊伍，整體強度難非常多，而且他們的等級和我們其實差了一段距離。當天幸運的是，我們和台灣另一支隊伍有打贏過他們 334 聯隊，經過那次之後我們整體信心就增加了不少。

4/21-4/24 比賽日

世界賽總共有四天，而我覺得這次的時間分配其實蠻平均而且很剛好。第一天的行程主要是報到和驗機，早上整個都是空的，提供給選手去測試機器、跑技能挑戰賽、也可以去其他隊伍的攤位走走看看。中午有一場一個小時的開幕式，當中最特別的就是國家遊行，我們台灣每一支隊伍會派一個代表上去。下午每一隊則會有一場練習賽以及一場資格賽。

第二天和第三天的行程全部都是資格賽，而若資格賽中間空閒時間較多則可能可以去練習場或是到處走走。攤位和比賽場地中間還有蠻遠的距離，然後去練習場基本上來回也要抓半小時，整體比賽下來基本上也都是一直走動，不太會有休息時間但也不會有需要在會場跑步趕場次的時候。第四天早上比完最後一場資格賽之後會開始聯隊選配，下午則馬上就會開始淘汰賽。

這次比賽中間我非常有時間去找其他在我們分區的隊伍聊天，也常常會去找一些之前比賽認識的隊伍交換紀念品，還會收集強隊的貼紙。前三天打下來時間上非常充裕，我們前兩天因為我弟弟發燒，稍微打得不如預期，但後面急起直追，最後排在資格賽排名第 15 名。

比賽期間的晚上，我們有時會和認識的隊伍簡單打個練習賽保持手感，也會檢查機器是否有問題。比較特別的是在第二天的時候我們加裝了一個側邊的裝置可以阻擋別隊的得分，這讓我們在第三天的時候打得輕鬆一些，不過若是早一點裝的話可能會發揮更大的效益。

最後我們和 18961B 聯隊組成第九種子，我們在 16 強賽當中獲勝，並出乎意料地打贏了八強賽，擊敗第一種子隊。可惜在四強賽當中因為隊友的機器和敵方機器糾纏，而沒有辦法打出應該要有的樣子，止步四強。

我們團隊在這次世界賽中也算是展露頭角，不僅在分區淘汰賽打到四強，也在分區中獲得發明獎 Create Award，是數一數二高的評審獎。在許多隊伍的口中也都能聽到我們隊伍的名字，也有非常多隊伍想要拿我們的貼紙和寶可夢卡牌。

這次整個賽季當中，我也負責經營團隊的 Instagram 帳號和 Youtube 頻道，記錄著我們從賽季開始到現在的會議還有比賽成果。從零開始，現在 Instagram 帳號累積到 460 多位追蹤者，其實是非常厲害也相當意料之外的，說明這個賽季當中有許多成績打響了我們的知名度。每一篇貼文我都有特別排版過照片也將色彩調整的比較鮮明，封面也都有統一製作邊框，讓別人能夠一看就知道這是 66994B 的貼文！我對於這個賽季經營的結果感到非常驕傲。

我覺得整體來說，這次的世界賽經驗非常值得，我們在台灣出動的隊伍當中算是綜合成績最好的，而且也透過世界賽認識了非常多隊伍，也和之前想見到的許多隊伍拍了合照。相較於之前的 V5 Over Under 世界賽我的成績也進步了非常多。很高興這次有機會來參加世界賽，我覺得有這場比賽當作這個賽季的結尾真的是再適合不過了。