PIC
PIC microcontrollers from Microchip Technology are 8-, 16-, and 32-bit Harvard-architecture MCUs with on-board flash, SRAM, EEPROM, and a peripheral set (timers, ADC, DAC, comparators, UART, SPI, I²C, USB, CAN, Ethernet on top-end parts). PIC10/12/16/18 are the 8-bit lines; PIC24 / dsPIC are 16-bit; PIC32 is MIPS-based 32-bit.
PIC cores run a small RISC instruction set (35 instructions for early PIC16, 76 for PIC18) and use a single accumulator (W register). Peripherals are accessed through Special Function Registers (SFRs), and each chip has a configurable peripheral matrix — with newer parts the Configurable Logic Cell (CLC), Numerically Controlled Oscillator (NCO), and intelligent peripherals offload tasks the CPU would otherwise handle. Microchip's MPLAB X IDE plus XC8/XC16/XC32 compilers cover the entire family.
In plain terms
An 8-bit microcontroller catalog so deep that a 1995 design can still be ordered new in 2026 with the same pinout and tooling.
Why designers use it
- Long-life industrial, automotive, and appliance firmware where ABI / pinout stability matter.
- Battery-operated remote sensors and meters with intelligent peripherals running while the CPU sleeps.
- Designs that already standardised on Microchip tooling and won't migrate.
Best for
- Industrial firmware
- Remote sensors
- Long-life designs
Key specifications
- Cores: 8/16/32-bit Microchip / MIPS
- Clock: 1 – 200 MHz
- Flash: 0.5 KB – 2 MB
- SRAM: 16 B – 512 KB
- Voltage: 1.8 – 5.5 V
When not to use it
- On modern projects with C-based teams — the PIC's Harvard architecture and quirky banked memory cost iteration speed.
- Where ecosystem and library ubiquity matter — STM32 and ESP32 have orders of magnitude more open-source code.
- On any product that needs Wi-Fi or BLE — Microchip has it, but they're not the cheapest path.
Common mistakes
- Programming a 5 V PIC from a 3.3 V programmer and watching MCLR fail intermittently.
- Forgetting to set the CONFIG fuses correctly — the part runs but the watchdog or oscillator behaves randomly.
- Banking errors in assembly — writing to the wrong register because the bank select wasn't switched.
Where you will find it
- Carbon-monoxide detectors and smoke alarms in many homes use a PIC16 8-bit MCU because the chip has been in continuous production since the 1990s and the underlying alarm certifications are tied to the silicon.
- Automotive body-control modules in mid-tier vehicles (window lift, central locking) often use PIC18 or PIC24 chips for their AEC-Q100 qualification and 30-year availability commitment.
- Microchip's PIC32MZ runs Linux-friendly bootloaders for some niche embedded products that need the MIPS instruction set's bit-field manipulation features.
A short history
The PIC architecture was originally developed by General Instrument in 1975 as a Peripheral Interface Controller for the GI CP1600 microprocessor. General Instrument spun off its microelectronics division in 1985 as Microchip Technology, with the PIC line as the core product. PIC16C84 (1993) was the first PIC with on-board EEPROM that could be in-system reprogrammed, kicking off the modern PIC era. The PIC family expanded with 16-bit PIC24/dsPIC (2004) and 32-bit MIPS-based PIC32 (2007). Microchip's acquisition of Atmel in 2016 brought AVR alongside PIC — today both families are sold under the Microchip umbrella.
Good to know
- The PIC stands for 'Peripheral Interface Controller' — a name from 1975 when the chip was meant to control peripherals for GI's larger CP1600 CPU.
- Some PIC10F parts ship in a 6-pin SOT-23 — you can build a complete embedded design with one through-hole resistor and one PIC10.
- PIC's 'banked' memory model (paging registers between 8-bit address spaces) is a holdover from the GI CP1600 era and a constant headache for newcomers.