The rapid growth of first person view (FPV) drone technology has revolutionized fields like cinematography, military reconnaissance, and competitive racing, sparking immense interest among enthusiasts worldwide. In the context of China FPV communities, however, the high cost of components often deters many from diving deeper into this hobby. A pivotal moment occurred in July 2020, when online platforms suspended sales of FPV drone parts for two months, highlighting the critical need for mastering core technologies locally. This realization prompted me to embark on developing my own flight control board for FPV drones, aiming to reduce reliance on expensive imports and enhance customization. Throughout this journey, I encountered numerous challenges in circuit design, testing, and optimization, which I document here from a first-person perspective, emphasizing the importance of innovation in the China FPV scene.
Designing the circuit schematic was the foundational step, as any minor error could lead to catastrophic failures like component burnout. I began by using common methods to check for open circuits, testing the wiring sequence of the flight control board systematically. However, this approach proved inadequate for comprehensive testing, and electrostatic discharge during the process damaged a commercial FPV drone flight control board. After some research, I discovered that by entering the “resource” command in the drone ground station CLI, I could access the pin definitions of the main control chip. This breakthrough allowed me to draft the first version of the circuit diagram, which I named JINHERC, and outsourced its fabrication to a professional PCB manufacturer. Unlike standard boards, this design featured an external power supply module attached beneath the flight control board, separating high-current power lines from sensitive control signals to minimize interference. The benefits of this approach can be summarized using the formula for signal-to-noise ratio (SNR), which is crucial for stable first person view experiences: $$SNR = \frac{P_{signal}}{P_{noise}}$$ where a higher SNR indicates better performance. By isolating power and signal paths, the design aimed to maximize SNR, ensuring smoother control in FPV drone operations.
To elaborate on the circuit design, I focused on key parameters such as voltage regulation and current handling. For instance, the power module needed to support various input voltages while maintaining stability. The relationship between input voltage ($V_{in}$), output voltage ($V_{out}$), and dropout voltage ($V_{drop}$) for a linear regulator can be expressed as: $$V_{out} = V_{in} – V_{drop}$$ where $V_{drop}$ depends on the load current. In this design, I prioritized components that could handle typical FPV drone loads, as shown in the table below summarizing the initial specifications:
| Component | Specification | Purpose |
|---|---|---|
| Main Control Chip | STM32 Series | Processing flight algorithms for first person view |
| Power Regulator | External LDO | Providing stable 3.3V supply |
| Gyroscope | MPU6000 | Motion sensing in FPV drone |
| USB Interface | Standard Type-B | Communication with ground station |
Upon receiving the PCB samples and components, I proceeded with soldering, starting with the main control section to conserve resources. After programming via ST-link, I connected the board to the ground station software, only to encounter two critical issues: the USB was not recognized, and the power indicator did not illuminate. This pointed to a oversight in the low-voltage USB power circuit, which I had omitted, leaving the chip unpowered. To diagnose this, I used a programmer to supply power separately, which lit the indicators but still failed to establish a connection. This setback led me to pause the project temporarily. Later, a discussion with online community members highlighted a potential hardware issue with the USB connection. Reflecting on a past experience with a similar problem, I realized that swapping the USB data lines might resolve it. After performing a fly-wire adjustment to interchange the signal lines, the computer successfully detected the board, and the ground station established a stable connection. This success marked a significant milestone in the development of my China FPV project, reinforcing the value of iterative testing in first person view drone systems.
The issues with the first version informed the design of the second iteration. I incorporated a high-performance LDO regulator to supply the main control chip, capable of delivering 500 mA at 3.3 V, with an input voltage as low as 4.5 V, making it ideal for USB-based power regulation. The output voltage stability is governed by the equation: $$V_{out} = V_{ref} \times (1 + \frac{R_1}{R_2})$$ where $V_{ref}$ is the reference voltage, and $R_1$ and $R_2$ are feedback resistors. Additionally, I corrected the USB signal lines to prevent recognition errors. The enhanced design also included components like an OSD character overlay chip and a gyroscope for better first person view functionality. For image filtering, I used tantalum capacitors, which improve video transmission quality in FPV drones by reducing noise. The capacitance ($C$) and its effect on filtering can be described by the impedance formula: $$Z = \frac{1}{j\omega C}$$ where $\omega$ is the angular frequency, and lower impedance at high frequencies results in cleaner signals. The table below compares the key improvements between the two versions:
| Aspect | First Version | Second Version |
|---|---|---|
| USB Power Circuit | Missing | Included with LDO regulator |
| Signal Line Configuration | Incorrect order | Corrected swap |
| Gyroscope Integration | Not tested | Fully functional with MPU6000 |
| Cost Efficiency | Low due to errors | Improved but still high |
Testing the second version involved multiple stages to ensure reliability for FPV drone applications. First, I evaluated the power system by configuring the ground station settings, switching the ESC protocol to dshot600 for efficient motor control. The dshot protocol’s digital signal timing can be modeled using the formula: $$T_{pulse} = \frac{1}{f_{dshot}}$$ where $f_{dshot}$ is the frequency, such as 600 kHz for dshot600. I conducted motor tests with propellers removed to avoid accidents, verifying that the thrust response aligned with expectations for first person view flights. Next, I assessed the communication system, confirming that data transmission between the flight control board and ground station was stable. Finally, I tested the image return and OSD character overlay, which are critical for real-time feedback in FPV drone operations. The OSD functionality relies on superimposing data onto the video signal, which can be expressed as: $$V_{out} = V_{video} + V_{OSD}$$ where $V_{video}$ is the original signal, and $V_{OSD}$ is the overlay data. Overall, the tests demonstrated satisfactory performance, with minimal latency and clear video output, essential for an immersive first person view experience in China FPV environments.
Outdoor flight tests were conducted to validate the flight control board under real-world conditions. I fine-tuned the PID parameters to suit this specific FPV drone, as PID control is fundamental to stable flight. The PID controller output $u(t)$ is given by: $$u(t) = K_p e(t) + K_i \int_0^t e(\tau) d\tau + K_d \frac{de(t)}{dt}$$ where $e(t)$ is the error signal, and $K_p$, $K_i$, and $K_d$ are the proportional, integral, and derivative gains, respectively. After adjustments, the drone exhibited smooth hovering and responsive maneuvers, with the first person view feed remaining stable even during aggressive flights. This outperformed many commercial boards, highlighting the potential for localized innovations in the China FPV market. The table below summarizes the PID parameters used and their effects:
| Parameter | Value | Effect on Flight |
|---|---|---|
| K_p (Proportional Gain) | 1.2 | Reduces steady-state error |
| K_i (Integral Gain) | 0.05 | Eliminates drift over time |
| K_d (Derivative Gain) | 0.8 | Dampens oscillations |
In summary, the development of this flight control board was largely successful, exceeding expectations in terms of performance for first person view applications. However, several shortcomings were identified. Firstly, the cost remained high due to the use of expensive components like the MPU6000 gyroscope, which is 15-20 times pricier than alternatives. To address this, I plan to explore substitutes such as MPU6050 or other ISP-compatible gyroscopes, which could be integrated via I2C interfaces. The cost comparison can be analyzed using the formula for total cost $C_{total}$: $$C_{total} = \sum_{i=1}^n C_i \times Q_i$$ where $C_i$ is the cost per component and $Q_i$ is the quantity. Secondly, the power system lacked support for higher voltages, such as 9 V, needed for digital HD systems like DJI FPV drones. Incorporating a wider input range regulator would enhance versatility. Lastly, the main control chip was not cost-effective; future iterations will trial the STM32F411CEU6 for better affordability. For advanced features, I aim to integrate sensors like barometers, magnetometers, and accelerometers, flashing INAV firmware to enable functions such as position hold, altitude stabilization, auto-return, and landing. These improvements would elevate the precision and accessibility of FPV drones in the China FPV community, fostering greater adoption and innovation.
Reflecting on this journey, I have gained a deeper appreciation for the intricacies of flight control systems in first person view drones. The process underscored the importance of resilience and community support in overcoming technical hurdles. As the China FPV landscape evolves, I hope that more enthusiasts will engage in similar projects, driving forward the localization of core technologies. By sharing these experiences, I aim to inspire others to explore the fascinating world of FPV drone customization, ultimately enriching the global first person view ecosystem with diverse, homegrown solutions.