Build an Open-Source 3D-Printed Robot with ESP32 and FreeRTOS

GPIO (General Purpose Input/Output) serves as the nervous system connecting microcontrollers to the external world. The ESP32 GPIO system consists of multiple hardware layers: physical pins, silicon buffers with input/output drivers, GPIO controller blocks containing registers, memory mappers, and software APIs. Critical pin restrictions include: pins 6-11 reserved for flash memory (never use for I/O), pins 34-39 input-only for analog/digital inputs, and pins 1-4 for UART serial console. Proper configuration follows a four-step workflow: reset the pin, set direction, configure pull-up/pull-down resistors, and read/write pin state. For LEDs, connect through 220Ω resistors to limit current to ~12mA. For buttons, enable internal pull-up resistors. Implement 100ms debouncing to handle mechanical switch bounce. Interrupts enable immediate response to hardware events without constant polling, using gpio_install_isr_handler() to register handlers.

This section covers the ESP32 PWM servo control and GPIO pin configuration. The PWM unit is initialized with 50Hz frequency and 0 duty cycle for continuous rotation servo. For input pins (push button), the pin is configured as input with pull-up resistance enabled and pull-down disabled. For output pins (LED), the pin is configured as output with both pull resistances disabled and initial state set to low. The servo dispensing function turns on the LED, rotates the servo for 2 seconds at specified PWM rate, reverses for 0.5 seconds to prevent jamming, then dispenses for an additional 2 seconds before stopping.

The ESP32 S3 has 45 GPIO pins, but only 21 are completely safe for general use (pins 1-18, 21, 47-48); additional pins can be used under specific conditions (30 total with strapping pins usable as outputs after boot, 35 if external RAM is absent), while pins 00, 26-32, and 43-44 are strictly prohibited as they serve boot mode configuration, external flash memory, and programming functions respectively.

Programming the ESP32 requires using GPIO (General Purpose Input/Output) pin numbers rather than counting pins from the board. For example, the 6th physical pin corresponds to GPIO 22. Reserved pins marked with red arrows (for OLED display and LoRa antenna) should not be used for other purposes. A typical program consists of setup() for configuration (setting pins as INPUT/OUTPUT) and loop() for continuous operation. The example demonstrates reading a push button input and controlling relay outputs with a 1-second delay, illustrating fundamental concepts of digital input reading and output control in embedded programming.

The ESP32 microcontroller features 40 GPIO pins that can be configured as digital inputs, outputs, or high-impedance open-drain outputs using the SDK-IDF native library; pins 34-39 can only be configured as inputs, and some pins are multiplexed with internal peripherals requiring explicit configuration before use. The SDK-IDF provides functions like gpio_pad_select_gpio() to configure pins as digital I/O, gpio_set_direction() to set input/output mode, and gpio_set_level() to control output states, enabling developers to create applications such as LED blinking through task-based programming with FreeRTOS.

Bare metal systems require manual management of everything with a single main loop (super loop) calling functions sequentially combined with interrupt handling, simulating multitasking via flags checked in the loop—limited to single-threaded execution, difficult to scale, and less predictable. RTOS systems provide true priority-based preemptive multitasking with the kernel managing scheduling, allowing multiple independent tasks with their own stacks, contexts, and priorities to run concurrently; this enables better scalability, modularity, and deterministic timing for complex systems like IoT devices, medical equipment, and automotive systems.

Choosing between bare metal and Real-Time Operating System (RTOS) is a critical decision in embedded development. Bare metal means your code is the only thing running—every line of C or assembly executes directly. RTOS provides a kernel, scheduler, heap, synchronization primitives (locks, events), and concurrency features. RTOS enables task-based concurrency where the scheduler decides which task runs next, creating the illusion of parallelism through rapid context switching. The choice depends on complexity: simple applications suit bare metal while complex ones benefit from RTOS features. However, RTOS adds significant complexity and context management overhead.

Bare metal programming uses a main function with hardware initializations followed by an infinite loop executing functions sequentially (polling method), causing delays as one function waits for another. ISRs are present but the main disadvantage is execution delays. In contrast, RTOS provides a structured approach with hardware interrupts, tasks, and a kernel that manages task execution based on priority. Tasks are the basic entities in RTOS, implementing functions and utilizing features like binary semaphores or mutexes managed by the kernel. Based on the scheduling algorithm, tasks are switched to execute one after another according to priority levels, enabling concurrent execution without polling delays.

A Real-Time Operating System (RTOS) is a lightweight operating system designed for microcontrollers that enables concurrent execution of multiple tasks while guaranteeing strict timing deadlines, unlike general-purpose operating systems which prioritize human responsiveness and may have non-deterministic scheduling; RTOS is particularly useful when developers need to run several independent threads simultaneously on a processor, such as handling user input, network communication, sensor reading, and display updates concurrently on devices like the ESP32, though simpler projects with minimal concurrency needs may be better suited for bare-metal super loop architectures.

This segment covers the fundamental principles of multitasking operating systems. Key concepts include: (1) Multitasking is multi-programming with time sharing, where CPU time is allocated to multiple processes in rapid succession; (2) The system is also known as time-sharing or interactive operating system; (3) Fair share scheduling ensures every process receives equal CPU execution time; (4) The trade-off is increased average response time compared to multi-programming systems; (5) Real concurrency cannot be achieved on a single CPU, as only one process executes at a time.

Precision is not an inherent machine property but a system function encompassing the entire manufacturing ecosystem—including machines, tools, workpieces, operators, and software. Tolerances and clearances enable assembly, movement, and quality differentiation between products. In 3D printing, precision depends on printer parameters, material properties, model design, slicing software, and operator skill. Calibration transforms budget printers into production-capable tools capable of producing interchangeable parts.

Assembly (어셈블리) is the process of joining multiple 3D printed components together. In 3D printing, assembly involves taking two modeled parts and putting them together to form a functional unit. Three main assembly types exist: (1) Rectangular slot with square pin for linear reciprocating motion, (2) Round slot with round shaft for rotational motion, and (3) Round hole with round shaft for rotational motion. Each type applies to specific drawing numbers in the public drawing set. Proper assembly requires specifying tolerances with plus/minus values and providing 1mm clearance to ensure smooth movement without interference.

Tolerance is the small gap between two 3D printed parts that allows them to fit together properly; without it, parts require excessive force to assemble. There are two main approaches to adding tolerances: (1) in the design software (Fusion 360) using sketch offset or the Press Pull tool, which is the professional method, and (2) in the slicer (Bambu Studio) using contour compensation, which is useful for quick prototyping adjustments. The correct tolerance depends on your machine and application—typically 0.05mm for tight fits, 0.1mm for standard fits, and 0.15-0.2mm for loose fits or rotating parts. To determine the optimal tolerance for your specific printer, print test pieces with different hole sizes and test which tolerance works best for your intended application.

Effective 3D printing design minimizes support material by orienting parts flat on the print bed and designing features that don't require overhangs. Mechanical assembly relies on understanding interference fits (tight connections) versus clearance fits (loose connections), with differences as small as 0.1mm in diameter. Designers must account for real-world variations like material shrinkage and machine variance that cause CAD dimensions to differ from physical reality. Using left-handed and right-handed threading ensures correct assembly without additional hardware. Registration pins help align parts during assembly by providing precise reference points. Designers should test fits empirically, adjusting dimensions based on actual print results rather than relying solely on CAD specifications.

When designing mechanical parts for 3D printing, you must add tolerances (typically 0.1-0.25mm depending on printer quality) to holes and pegs to account for printing imperfections, as 3D printers cannot achieve perfect precision; additionally, using angled faces (around 42 degrees) instead of vertical walls helps prevent support material issues and ensures better-fitting parts.