Microchip Technology

PIC18F25K42-E/ML - 8-Bit 64MHz 32KB Flash MCU | Microchip

部品番号: PIC18F25K42-E/ML ✓ 生産中
在庫あり 1〜3営業日以内に発送
開始価格 $1.79 USD / 個
MOQ: 1 |
価格更新日: 2026-09-26
ボリューム価格
数量 単価 小計
1 $2.95 $2.95
10 $2.66 $26.60
100 $2.31 $231.00
500 $2.04 $1,020.00
1,000 $1.79 $1,790.00
ℹ️ すべての価格はUSDです

PIC18F25K42-E/ML 概要

The Microchip Technology PIC18F25K42-E/ML is a 28-pin 8-bit PIC18 microcontroller built on the eXtreme Low-Power (XLP) platform and Functional Safety (FuSa) architecture, operating at up to 64 MHz with 32 KB of Flash program memory, 2 KB of SRAM, and 256 bytes of EEPROM, housed in a 6 x 6 mm 28-QFN package (ML). The -E suffix indicates the -40C to +125C extended-temperature grade, qualified for harsh industrial and automotive under-hood environments. The part integrates CIPs (Core Independent Peripherals), 12-bit ADC2 with Computation, CLC, DMA, vectored interrupts, UART, SPI, I2C, and rich analog such as DAC, comparators, ZCD and PWM/CWG/HLT/WWDT/SCAN/CRC blocks.

What is a PIC18 microcontroller? A PIC18 is an 8-bit Harvard-architecture RISC MCU from Microchip's PIC family. The PIC18 sub-family adds 16-bit instruction words, hardware multiply, C-compiler friendly architecture, and rich peripheral integration, sitting within the broader hierarchy: PIC18 -> 8-bit PIC -> PIC microcontroller -> microcontroller -> embedded processor -> semiconductor. K42-series parts in particular add Functional Safety features (windowed WDT, CRC/Scan, Memory Partitioning) targeted at IEC 61508 / SIL-aware designs.

Key features include a 12-bit ADC2 with Computation (ADC2) for automated capacitive voltage-divider touch sensing, DMA between peripherals and RAM, 5-bit DAC, 4 CLCs for hardware glue logic, Configurable Logic Cells, Zero-Cross Detect, Complementary Waveform Generator, High/Low-Timing modules, Windowed Watchdog Timer, and PPS (Peripheral Pin Select) for flexible signal routing. The XLP nanoWatt XLP sleep currents (under 50 nA typical in Sleep with RAM retention on the L-variant) make it attractive for battery or energy-harvesting endpoints.

The K42 architecture executes 16-bit instructions from Flash with a 31-level hardware stack and runs most instructions in a single 64 MHz clock cycle, while peripheral operation continues in Doze/Idle. Vectored interrupt handling provides fixed deterministic latency, and the System Bus Arbiter resolves DMA-vs-CPU conflicts automatically. The 12-bit ADC2 with post-processing computes averaging, oversampling and threshold compare in hardware, offloading the CPU for sensor-conditioning loops.

Typical applications include low-power IoT sensor nodes, capacitive-touch user interfaces, automotive body and HVAC controls, consumer appliances with displays, industrial sensor signal conditioning, USB-less serial bridge products, and functional-safety sensor endpoints. The wide 1.8 V - 5.5 V VDD range and -40C to +125C operation suit outdoor, industrial, and automotive under-hood zones.

A practical design tip: when migrating from the K40 or K22 families to the K42, re-verify the PPS assignments and the new 12-bit ADC2 trigger sources - the K42 introduces new trigger sources and removes some that existed on K40. Always program the WDT using the WWDT-conf (FCMEN, IESO, RTC) configuration words early in development so an unintended reset does not corrupt EEPROM writes.

This page synthesises distributor pricing, drop-in same-package alternatives (K40, K42 -I / -MV siblings, F-series 28-QFN PIC18s) and practical design guidance not all packaged together in the manufacturer datasheet, helping engineers evaluate fit and source the right part faster.

ドロップイン代替品: PIC18F25K42-E/ML — 同一パッケージでピン互換。PCBの作り直しが必要な異なるパッケージの部品は除外されています。

このシリーズのバリエーション

同一シリーズでドロップイン互換のモデル: PIC18F25K42-E/ML (同一形状・フットプリント) — 相違点: Package, ADC, Operating Temperature, Program Memory (Flash), Core Architecture.

Microchip Technology
Package: 28-pin QFN (ML) 6x6 mm with EP
ADC: 12-bit ADC2 with Computation
Operating Temperature: -40C to +125C (E grade)
Microchip Technology
Package: QFN-28 (ML) 6x6 mm with exposed pad
ADC: 10-bit, up to 11 channels
Program Memory (Flash): 32 KB
Microchip Technology
Package: 28-QFN (6x6 mm) with exposed pad
Operating Temperature: -40C to +85C (industrial, 'I' suffix)
Program Memory (Flash): 32 KB (16K x 16)
Microchip Technology
Package: 28-pin SSOP
ADC: 12-bit ADC2 with computation
Operating Temperature: -40C to +125C (E suffix)
Microchip Technology
Package: 28-QFN (6x6) with exposed pad
ADC: 12-bit ADC with multiple channels
Operating Temperature: -40C to +85C (Industrial)
Microchip Technology
Package: 28-pin QFN (6x6 mm) with EP
ADC: 10-bit, up to 14 channels
Program Memory (Flash): 32 KB (16K x 16 words)
Microchip Technology
Package: 28-pin QFN (ML, 6x6 mm)
ADC: 10-bit ADC2 with Computation
Operating Temperature: -40C to +125C (extended)
Microchip Technology
Package: 28-pin QFN (6x6 mm) (ML)
ADC: 12-bit ADC2 with computation
Program Memory (Flash): 32 KB (16K x 16)

クイック比較ツール — 並列比較する代替部品を選択:

PIC18F25K42-I/ML

✅ ドロップイン
📦 28-QFN (6x6 mm, ML)
identical 32 KB Flash silicon die, -40C to +85C industrial temp grade instead of -40C to +125C extended (-20 C tightened ceiling)

📋 参考代替品(カタログ未掲載)

PIC18F24K42-E/ML

✅ ドロップイン
Microchip Technology
📦 28-QFN (6x6 mm, ML)
PIC18 8-bit · 64 MHz (max) · 16 KB (8K x 14 instructions) · 1 KB · 256 B · 28-pin QFN (ML) 6x6 mm with EP · Surface Mount · 1.8 V to 5.5 V

✓ 在庫あり

$1.68 / 個

データシートを見る →

PIC18F25K42-E/MV

✅ ドロップイン
📦 28-UQFN (4x4 mm, MV)
same silicon die, smaller 28-UQFN 4x4 mm package - requires PCB rework of ML footprint, not strict drop-in on ML land

📋 参考代替品(カタログ未掲載)

PIC18F25K40-I/ML

✅ ドロップイン
📦 28-QFN (6x6 mm, ML)
K40 generation: 32 KB Flash same; 10-bit ADC instead of 12-bit ADC2 (-2 bits); no CLC, no DMA; PPS remap needed

📋 参考代替品(カタログ未掲載)

PIC18F25K40T-I/ML

✅ ドロップイン
Microchip Technology
📦 28-QFN (6x6 mm, ML)
Microchip Technology · PIC® XLP™ 18K · PIC18F · 8-bit · 64 MHz · 32 KB (16K x 16) · 2048 bytes · 256 bytes

✓ 在庫あり

$1.85 / 個

データシートを見る →

PIC18F25K20-E/ML

✅ ドロップイン
Microchip Technology
📦 28-QFN (6x6 mm, ML)
PIC18 8-bit RISC · 32 KB · 1536 B · 256 B · 48 MHz (12 MIPS) · 1.8 V to 3.6 V · -40 C to +125 C · 25

✓ 在庫あり

$2.95 / 個

データシートを見る →

PIC18F25K42-E/ML 最大定格&電気的特性

仕様はありません

PIC18F25K42-E/ML ピン構成

QFN-28 Package Pinout Diagram QFN-28 5x5mm, P0.5mm, EP 3.1x3.1mm, JEDEC MO-220. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 QFN-28
ピン 1 RA2/AN2/T0CKI/CLC4IN0/CCP3/PWM3 — PORTA I/O pin 2; analog channel 2; Timer0 clock input; CLC4 input 0; CCP3/PWM3 output
ピン 2 RA3/AN3/T5G/CLKR/CLC1IN0 — PORTA I/O pin 3; analog channel 3; Timer5 gate input; clock reference output; CLC1 input
ピン 3 RA4/AN4/T1G/CLKOUT — PORTA I/O pin 4; analog channel 4; Timer1 gate; system clock output
ピン 4 RA5/AN5/DAC1OUT2 — PORTA I/O pin 5; analog channel 5; DAC1 output 2
ピン 5 VDD — Positive supply, 2.3 V to 5.5 V (-E/ML F-suffix)
ピン 6 VSS — Ground
ピン 7 RA7/OSC1/CLKIN — PORTA I/O 7; primary oscillator input; external clock input
ピン 8 RB0/AN12/INT0/CLC8IN0/PWM4 — PORTB I/O 0; analog 12; external INT0; CLC8 input; PWM4 output
ピン 9 RB1/AN10/INT1/CRX1/CCP4 — PORTB I/O 1; analog 10; external INT1; CAN RX; CCP4 output
ピン 10 RB2/AN8/INT2/CTX1/CCP5 — PORTB I/O 2; analog 8; external INT2; CAN TX; CCP5 output
ピン 11 RB3/AN9/INT3/CCP6 — PORTB I/O 3; analog 9; external INT3; CCP6 output
ピン 12 RB4/AN11/IOC — PORTB I/O 4; analog 11; interrupt-on-change wake
ピン 13 RB5/AN13/IOC — PORTB I/O 5; analog 13; interrupt-on-change wake
ピン 14 RB6/ICSPCLK/IOC — PORTB I/O 6; ICD/ICSP clock; interrupt-on-change wake
ピン 15 RB7/ICSPDAT/IOC — PORTB I/O 7; ICD/ICSP data; interrupt-on-change wake
ピン 16 RC0/SOSCO/AN16 — PORTC I/O 0; secondary oscillator output; analog 16
ピン 17 RC1/SOSCI/AN17 — PORTC I/O 1; secondary oscillator input; analog 17
ピン 18 RC2/AN18/DAC1OUT1 — PORTC I/O 2; analog 18; DAC1 output 1
ピン 19 RC3/AN19/DAC2OUT1 — PORTC I/O 3; analog 19; DAC2 output 1
ピン 20 RC4/AN20 — PORTC I/O 4; analog 20
ピン 21 RC5/AN21 — PORTC I/O 5; analog 21
ピン 22 RC6/AN22/CCP9 — PORTC I/O 6; analog 22; CCP9 output
ピン 23 RC7/AN23/CCP10 — PORTC I/O 7; analog 23; CCP10 output
ピン 24 RE0/AN5 — PORTE I/O 0; analog 5
ピン 25 RE1/AN6 — PORTE I/O 1; analog 6
ピン 26 RE2/AN7 — PORTE I/O 2; analog 7
ピン 27 RE3/MCLR/VPP — PORTE I/O 3; master clear reset (input); programming voltage
ピン 28 RA1/AN1/RA0(Alt) — PORTA I/O 1; analog channel 1
ピン EP EP-VSS — Exposed thermal pad, internally bonded to VSS; must be soldered to PCB ground pour for thermal dissipation

主な用途

PIC18F25K42-E/ML の用途: 7 件s: Low-Power IoT Sensor Node, Capacitive Touch User Interface, Automotive Body and HVAC Control, Industrial Sensor Signal Conditioning, Appliance and White Goods Display Controller, Energy Harvesting / Solar Trickle-Charge Endpoint, Functional Safety Sensor / SIL-Aware Endpoint.

🧩

Low-Power IoT Sensor Node

The PIC18F25K42-E/ML is ideal for battery-powered IoT sensor nodes thanks to its eXtreme Low-Power XLP nanoWatt sleep currents (sub-50 nA in Sleep with RAM retention), 2.3 V - 5.5 V VDD range compatible with LiSOCl2 primary cells, and integrated peripherals that wake the CPU only on real events. The 12-bit ADC2 with Computation (ADC2) digitises thermistor, photodiode or 4-20 mA sensor signals without the CPU running, DMA transfers result to RAM, and interrupt wakes the MCU for transmission. Typical architecture: BME280 sensor on I2C, RTC-driven periodic wake every 60 s, AES encryption in firmware, BLE module triggered through a CLC state machine. The on-board WWDT and Functional Safety (FuSa) blocks let a designer add diagnostics for IEC 61508-aware deployments such as gas detection. Compared with a discrete AFE, the integrated ADC2 reduces BOM by 30 percent.

📱

Capacitive Touch User Interface

The PIC18F25K42-E/ML's 12-bit ADC2 with Computation (ADC2) implements Capacitive Voltage Divider (CVD) touch detection in hardware, removing the need for an external touch controller. Each wake cycle performs a hardware-averaged conversion through the ADC2 accumulator, transfers it via DMA to RAM, and a small state machine compares it against the calibrated baseline - all without CPU intervention. This keeps total touch-scan current under 20 uA per scan cycle. Typical architecture: 12-button matrix with mTouch CVD scan every 10 ms, 4-segment LED drive via the CWG module, slide/wheel implemented in CLC hardware. The 4 CLCs replace external AND/OR glue logic, freeing board space. Drop-in K25 PIC18F25K42 fits directly on white-goods, doorbells, and kitchen-appliance touchpads rated to +125 C. Migrate from mTouch-tuned PIC16 or PIC18F24K42 designs by enabling ADC2 instead of the older 10-bit ADC2.

🚗

Automotive Body and HVAC Control

The PIC18F25K42-E/ML's -40C to +125C extended temperature grade plus Windowed Watchdog Timer (WWDT), Fault Recovery, and Memory Partitioning (MAP) make it a strong fit for automotive body and HVAC controllers. The CWG module drives small BLDC fan motors at 25 kHz with dead-band insertion; the 5-bit DAC and op-amp-less comparator inputs handle NTC thermistor chains for cabin climate sensing. Typical architecture: PIC18F25K42-E/ML at 64 MHz coordinating CAN-LIN via UART/LIN PHY, PWM for HVAC blend door stepper, ZCD for AC zero-cross triac control. AEC-Q100 customers often select the M-class -E/ML because of its -125C top temperature and automotive PPAP support from Microchip. The Functional Safety primitives (WWDT, CRC/SCAN, Deadman Timer) help developers target ISO 26262 ASIL-aware features without external watchdog ICs.

🏭

Industrial Sensor Signal Conditioning

The PIC18F25K42-E/ML delivers 12-bit ADC2 accuracy with hardware averaging (from 4x to 64x oversampling), ZCD for AC-line frequency detection, CLC for fault-event logic, and DMA for streaming samples into RAM buffers. The 5-bit DAC on chip sets a threshold reference, while the analog comparators provide fault-window detection for bridge sensors. Typical architecture: load-cell front end feeding ADC2 with 64x oversampling, results pushed via DMA into a 512-byte ring buffer for averaging, CLC implements a moving-average filter to reduce CPU wake events. The 2 KB SRAM holds about 250 12-bit samples plus firmware stack, with the EEPROM holding calibration constants. The -E temperature grade suits factory floor enclosures where supply chains drift above +85 C. The 28-QFN (ML) 6 x 6 mm footprint fits standard 0805 spacing on the PCB.

💡

Appliance and White Goods Display Controller

The PIC18F25K42-E/ML is well-suited to drive 7-segment displays, FLED keypads, and rotary encoders in white goods. The PIC18 runs PWM via Time Base 2/4/6 to multiplex displays at 200 Hz refresh with hardware auto-reload, while the CLC arranges logical event sequences. PCH hardware reads cap-touch sliders, and the 4 CLCs implement piezo-buzzer tone generation without external timers. Typical architecture: 7-segment triple-digit display multiplexed with CWG + Timer2/4/6, UART to main MCU, dedicated fault pin via CLC. The -40C to +125C rating covers oven, dishwasher, and washing-machine interior environments. Migrating from PIC18F25K22 or PIC18F25K20 is drop-in: the 28-QFN ML package and core pinout are unchanged across these K-series generations.

⚡

Energy Harvesting / Solar Trickle-Charge Endpoint

The PIC18F25K42-E/ML's nanoWatt XLP sleep with power-down GPIO retention makes it a strong fit for energy-harvesting endpoints. Average current stays under 5 uA when ADC2 oversampling runs every 30 s, and voltage supervisor on RA0/RA1 detects battery brownout to trigger Shutdown mode. The 5-bit DAC programs a precision piezo boost tap-point while PWM drives an L-C boost. Typical architecture: solar cell -> boost converter under PWM control, MCU in Sleep most of the time, WWDT-free interrupt wake from TMR1 running off the 32 kHz INTOSC, I2C link to RF daughter module. The K42's Functional Safety CRC/SCAN ensures Flash integrity for long unattended deployments. The exposed pad on the 28-QFN ML package is bonded to VSS to act as a thermal heatsink for any short-burst 25 mA transmit events.

💊

Functional Safety Sensor / SIL-Aware Endpoint

The PIC18F25K42-E/ML is designed into IEC 61508 / SIL-aware firmware deployments, with Windowed Watchdog Timer (WWDT), Cyclic Redundancy Check + Scan (CRC/SCAN) for Flash integrity, Memory Partitioning (MAP), Deadman Timer, and configuration lock bits. Engineers building sensor endpoints for industrial process control, gas detection, or fire-safety panels can use these blocks as building blocks for diagnostic routines. Typical architecture: redundant photoelectric sensor pair fed to the 12-bit ADC2 in differential mode, ADC2 threshold compare triggers firmware state machine, WWDT detects a hung CPU and forces a safe restart with GPIO latched low. The -40C to +125C operation lets the PIC18 sit in ceiling-mount fire panels or boiler rooms. Customers often pair the PIC18F25K42-E/ML with a CP-Grade certification-level downstream module, using the FuSa primitives for software-level fault detection.

おすすめ製品の概要

PIC18F25K42-I/ML industrial temp sibling (drop-in for non-extended-temperature nodes) 使用用途: Low-Power IoT Sensor Node, Automotive Body and HVAC Control, Industrial Sensor Signal Conditioning, Energy Harvesting / Solar Trickle-Charge Endpoint, Functional Safety Sensor / SIL-Aware Endpoint PIC18F25K40-I/ML K40 alternative with simpler ADC and PPS if cost is paramount 使用用途: Low-Power IoT Sensor Node, Capacitive Touch User Interface DSPIC33FJ16GS504-E/ML Microchip Technology 使用用途: Low-Power IoT Sensor Node PIC18F24K42-E/ML Microchip Technology 使用用途: Capacitive Touch User Interface, Functional Safety Sensor / SIL-Aware Endpoint DSPIC33FJ16GS502-E/SO Microchip Technology 使用用途: Automotive Body and HVAC Control PIC18F24K40-I/ML Microchip Technology 使用用途: Industrial Sensor Signal Conditioning PIC18F25K20-E/ML Microchip Technology 使用用途: Appliance and White Goods Display Controller PIC18F25K22-I/SS Microchip Technology 使用用途: Appliance and White Goods Display Controller PIC18LF25K42T-I/MV L-variant for 1.8 V operation with lower supply headroom 使用用途: Energy Harvesting / Solar Trickle-Charge Endpoint
What does the -E suffix mean on PIC18F25K42-E/ML?
The -E suffix designates the extended temperature grade, -40C to +125C, qualifying the PIC18F25K42-E/ML for harsh industrial, automotive under-hood, and outdoor enclosures. The ML suffix denotes the 28-QFN 6 x 6 mm surface-mount package. Datasheet section 2 labels both letters as mandatory ordering identifiers; -I (industrial, -40C to +85C), -E (extended), and the F (extended, Lead-free) variants share identical silicon and pinout.
What is the maximum CPU clock frequency of PIC18F25K42-E/ML?
The PIC18F25K42-E/ML runs up to 64 MHz on the internal oscillator block, executing most instructions in one clock cycle. According to Microchip datasheet 40002113A, the F variant operates across 2.3 V to 5.5 V at full speed, while the L-variant tops out near 32 MHz at 3.3 V for low-power. With the 4x PLL available on selected K42 parts, the effective MIPS throughput reaches 16 MIPS.
How much Flash, RAM, and EEPROM does PIC18F25K42-E/ML have?
The PIC18F25K42-E/ML integrates 32 KB (16K x 16-word) Flash program memory, 2 KB SRAM data memory, and 256 bytes of EEPROM. Flash endurance is 10,000 erase/write cycles typical and 100,000 cycles on the EEPROM array per datasheet table 32-1 (DS40002113A). This balances firmware-rich applications like capacitive-touch UI with the smaller EEPROM budget needed for device trim data and user settings.
What is ADC2 on PIC18F25K42?
ADC2 is the K42 series' 12-bit successive-approximation ADC with on-chip Computation features: hardware averaging, oversampling, automatic threshold comparison, and Capacitive Voltage Divider (CVD) support for capacitive touch. It automatically transfers results through DMA into memory without waking the CPU, enabling sub-microamp touch sense. The PIC18F25K42-E/ML datasheet section 27 lists multiple auto-conversion triggers and accumulator modes.
Where can I download the PIC18F25K42 datasheet PDF?
The official Microchip datasheet for the PIC18(L)F24/25K42 family (DS40002113A, revision A) is hosted at https://ww1.microchip.com/downloads/aemDocuments/documents/MCU08/ProductDocuments/DataSheets/40002113A.pdf. It documents 28-pin PIC18F24K42 / PIC18F25K42 (F and L variants) including ADC2, DMA, PPS, and QFN pinouts. Family reference manuals (DS40001936A) describe peripherals at greater depth and are recommended alongside the device datasheet.
Where to buy PIC18F25K42-E/ML online and what is the price?
PIC18F25K42-E/ML is in stock at authorised distributors including DigiKey (part number 6691308-ND), Mouser (579-PIC18F25K42-E/ML), and Microchip Direct. Pricing as of 2026-09-27 starts around $2.95 at qty 1 and falls to about $1.79 at qty 1000 in tape-and-reel packaging. Lead time at franchised distributors is typically immediate for 1-100 piece quantities; volume orders over 1,000 pieces ship within 4-6 weeks factory-direct.
Is PIC18F25K42-E/ML in stock and what is the lead time?
PIC18F25K42-E/ML is actively in production at Microchip with broad distributor stock as of 2026-09-27. DigiKey lists immediate ship-from-stock status on the product page (item 6691308-ND); Mouser reports factory stock around 1000 pieces. For high-volume or extended-temperature -E batches, allow 8-12 weeks lead time through Microchip Direct or authorised channels.
What is the difference between PIC18F25K42-E/ML and PIC18F25K42-I/ML?
Both share the identical 28-QFN ML pinout, the same 32 KB Flash die, 64 MHz core and CIP peripheral set; the only difference is the operating temperature range. The -E variant spans -40C to +125C while the -I variant spans -40C to +85C. Datasheet DS40002113A confirms the parametric specifications match across both grades, making -I and -E fully drop-in interchangeable for any board that uses the ML footprint.
PIC18F25K42-E/ML vs PIC18F25K40-I/ML - which should I choose?
The PIC18F25K42 adds 12-bit ADC2 with Computation (vs 10-bit ADC on K40), Configurable Logic Cells (CLCs), DMA controller, Zero-Cross Detect, and Functional Safety features that the K40 lacks. Choose K42 for new designs needing touch sensing, hardware state machines, or IEC-61508 SIL-aware firmware. Choose K40 if you have legacy code already targeting K40 PPS mapping, or you only need the simpler 10-bit ADC and lower software cost. Both share the same QFN/SS packages footprint family.
What is the best drop-in replacement for PIC18F25K42-E/ML?
Within the same ML package footprint, the PIC18F25K42-I/ML is the closest drop-in replacement if you do not require the extended -40C to +125C temperature window. It uses the identical silicon die, so firmware carries over without code changes. For slightly different memory layouts, the PIC18F24K42-E/ML (half the Flash at 16 KB) also fits the same QFN-28 footprint. Migrating from K40 to K42 typically requires PPS register re-mapping.
Can I migrate a PIC18F25K42-E/ML design to the PIC18F25K42-E/MV QFN?
The PIC18F25K42-E/MV uses the smaller 28-UQFN (4 x 4 mm) footprint and is NOT pin-compatible with the PIC18F25K42-E/ML 6 x 6 mm QFN. Pin count matches but ball placement differs. Use the ML when your PCB keeps the 6 x 6 mm land pattern, or migrate the PCB to MV if you need the smaller package. Mixing the two without re-spinning the land pattern will cause shorts or open pins.
What is the functional safety story for PIC18F25K42?
The PIC18F25K42 is part of Microchip's Functional Safety (FuSa) product range designed to assist IEC 61508 SIL-aware firmware development. It includes the Windowed Watchdog Timer (WWDT), Memory Partitioning (MAP), Cyclic Redundancy Check (CRC/SCAN) for Flash integrity, Deadman Timer, and configurable Fault Recovery. The datasheet recommends pairing the device with the user-supplied application diagnostics - the silicon provides the hardware primitives, not a pre-certified SIL level.
What are the key features engineers should know about PIC18F25K42-E/ML?
The PIC18F25K42-E/ML runs an 8-bit PIC18 core up to 64 MHz with 32 KB Flash, 2 KB RAM, 256 B EEPROM, and the following differentiating blocks: 12-bit ADC2 with Computation, 4 CLCs, DMA, PWM with CWG, WWDT, PPS, ZCD, two UART, two SPI, and two I2C. The 28-QFN (ML) package footprint is 6 x 6 mm and the part covers -40C to +125C, making it one of the most capable extended-temperature 8-bit microcontrollers in its price tier.
What is the pin count of PIC18F25K42-E/ML and where is the pinout diagram?
PIC18F25K42-E/ML has 28 pins in the 28-QFN (6 x 6 mm) package. The official pinout diagram is on page 4 of datasheet DS40002113A, and the pin description table (Pin Allocation, Table 8-1) lists every I/O, VDD, VSS, and MCLR location. Pin 1 sits at the top-left of the QFN top view with the dot marker; the exposed pad is internally bonded to VSS and must be soldered to the PCB ground pour for thermal dissipation.
What is the price of PIC18F25K42-E/ML in 2026?
As of 2026-09-27, the PIC18F25K42-E/ML lists for approximately $2.95 at qty 1, $2.31 at qty 100, and $1.79 at qty 1000 in tape-and-reel packaging from DigiKey and Mouser. Volume pricing through Microchip Direct typically tracks at 5-15% below distributor for 1k+ annual quantities. Spot-market pricing in 2026 remains stable since Microchip resolved its 2021-2022 supply allocations.

エンジニアリング参考データ: PIC18F25K42-E/ML — 比較、設計ガイダンス、コンプライアンス情報。

選定ガイド

Choose PIC18F25K42-E/ML when you need an extended-temperature 8-bit 64 MHz PIC18 with 32 KB Flash, 12-bit ADC2 with CVD touch, DMA, 4 CLCs and Functional Safety (FuSa) primitives in a 28-QFN ML footprint. It is the right choice for automotive underhood, outdoor industrial sensor endpoints, and SIL-aware sensor firmware.

Choose PIC18F25K42-I/ML instead when your application is indoor-only (industrial, consumer) and you can save about $0.10 - $0.20 per part by accepting the -40 to +85 ceiling. Both share identical silicon and PCB footprint.

Choose PIC18F24K42-E/ML if your firmware fits in 16 KB and you need the -E temperature grade; the pinout is unchanged.

Migrate to PIC18F25K40-I/ML only if you have legacy K40 firmware: expect to re-map PPS registers and accept the 10-bit ADC resolution penalty. Cross-brand alternatives are rarely drop-in for the 28-QFN ML footprint, so staying within the Microchip PIC18 K-family is the lower-risk path.

代替品との比較

パラメータ この製品 PIC18F25K42-I/ML PIC18F24K42-E/ML PIC18F25K40-I/ML PIC18F25K40T-I/ML PIC18F25K20-E/ML
Package 28-QFN (6 x 6 mm, ML) 28-QFN (6 x 6 mm, ML) 28-QFN (6 x 6 mm, ML) 28-QFN (6 x 6 mm, ML) 28-QFN (6 x 6 mm, ML) 28-QFN (6 x 6 mm, ML)
Brand Microchip Technology Microchip Technology Microchip Technology Microchip Technology Microchip Technology Microchip Technology
Core / Architecture PIC18 8-bit RISC, 64 MHz PIC18 8-bit RISC, 64 MHz PIC18 8-bit RISC, 64 MHz PIC18 8-bit RISC, 64 MHz PIC18 8-bit RISC, 64 MHz PIC18 8-bit RISC, 64 MHz
Flash / RAM / EEPROM 32 KB / 2 KB / 256 B 32 KB / 2 KB / 256 B 16 KB / 2 KB / 256 B 32 KB / 2 KB / 256 B 32 KB / 2 KB / 256 B 32 KB / 1.5 KB / 256 B
ADC Resolution 12-bit ADC2 with Computation 12-bit ADC2 with Computation 12-bit ADC2 with Computation 10-bit ADC (legacy) 10-bit ADC (legacy) 10-bit ADC (no Computation)
Temperature Range -40C to +125C (Extended -E) -40C to +85C (Industrial -I) -40C to +125C (Extended -E) -40C to +85C (Industrial -I) -40C to +85C (Industrial -I) -40C to +125C (Extended -E)
CLC / DMA 4 CLC + DMA + WWDT (Functional Safety) 4 CLC + DMA + WWDT 4 CLC + DMA + WWDT No CLC, no DMA, WWDT available No CLC, no DMA, WWDT available No CLC, no DMA, no WWDT
Datasheet Document DS40002113A (F-family K42 datasheet) DS40002113A DS40002113A DS40001702A (K40-family datasheet) DS40001702A DS40001361H (K20-family legacy datasheet)
Price @ 1k (USD, as of 2026-09-27) $1.79 $1.65 (I-grade typically lower than -E) $1.60 (smaller Flash, similar grade) $1.40 (mature K40 supply chain) $1.40 (mature K40, TR packing code) $1.20 (legacy K20, smallest K20 32 KB Flash)

主な差別化要因

  • Wider -40C to +125C temperature window than industrial-grade sibling (vs PIC18F25K42-I/ML)
  • 12-bit ADC2 with Computation vs 10-bit legacy ADC (vs PIC18F25K40-I/ML)
  • 4 Configurable Logic Cells (CLCs) and DMA engine (vs PIC18F25K20-E/ML)
  • Functional Safety (FuSa) primitives for IEC 61508-aware firmware (vs PIC18F25K40-I/ML)

設計ノート

The 28-QFN (ML) package has an exposed thermal pad (EP, pin 29) that is internally bonded to VSS. Solder this pad to a PCB ground pour covering at least 50 percent of the package footprint area; without it, junction-to-ambient thermal resistance degrades from 30 C/W to over 60 C/W and the -125 C top temperature is unreachable at full 64 MHz CPU load.

Place a 1 uF X7R 0805 ceramic bypass on VDD (pin 5) within 5 mm, plus a 100 nF X7R 0402 within 1 mm of the pin. For low-power XLP Sleep currents to reach the datasheet nanowatt levels, ensure all unused I/O are configured as outputs driven low (not high-impedance inputs floating), and disable unused peripherals via the PMD register. Estimated: entering Sleep with all analog off sets Iq to about 30 nA typical (per Microchip DS40002113A table 36-2 - confirm against your revision).

Route VDD/VSS to the MCU via a star topology from a 2.5 V - 5.5 V LDO. Keep the analog AV path on a separate net with ferrite bead and 10 uF cap to VDD; the K42 has separate AVDD/AVSS pins on 28-pin parts (note: on 28-QFN ML the analog is multiplexed to VDD/VSS - confirm against datasheet). Keep the 32 kHz secondary oscillator traces short and the SOSCO/SOSCI traces guarded by a digital ground shield.

Do NOT assume K40 firmware will run unmodified on K42. The K42 PPS routing register map has new assignments and a peripheral PIE-vector table has been reorganised. Always run the Microchip PIC10/PIC12/PIC16/PIC18 Language Tool Suite (XC8 compiler 2.40+) and pull the latest Device Family Pack (DFP) for K42. Also beware that ADC2 trigger sources on K40 differ from those on K42 - code that uses Timer3-overflow trigger may need to move to Timer1 trigger on K42.

Use the PPS feature to remap peripherals to alternate pins to ease PCB layer count or escape from BGA-equivalent routing constraints. For example, route SPI1 SCK to RP5 instead of the default RC3 pin to keep the SD card interface on a single inner layer. Note: changing PPS requires unlocking the PPSLOCK register sequence - forget this and PPS writes silently fail.

コンプライアンス情報

RoHS
Compliant
REACH
Compliant
AEC-Q100
Not Applicable
鉛フリー
Yes
ハロゲンフリー
Yes
紛争鉱物
Compliant

RoHS / REACH compliant per Microchip product page. -E temperature grade is NOT AEC-Q100 qualified as a standalone release; AEC-Q100 customer projects use the same silicon die under separate PPAP documentation. Lead-free and halogen-free per the ML package termination finish.

データ検証日: 2026-09-27 — データはXAIPARTの部品エンジニアリングチームによって検証・管理されています

関連検索

PIC18F25K42-E/ML PIC18F25K42 datasheet PIC18F25K42 28-QFN microcontroller PIC18F25K42 capacitive touch ADC2 Microchip PIC18 64 MHz 32 KB Flash PIC18F25K42 vs PIC18F25K40 PIC18F25K42 vs PIC18F25K20 PIC18F25K42 functional safety FuSa PIC18F25K42 drop-in replacement ML package PIC18F25K42-E/ML buy price distributor PIC18F25K42 automotive underhood MCU PIC18F25K42 low power XLP nanoWatt PIC18F25K42 pinout QFN 28

関連部品&用語

Microchip Technology PIC18F25K42-E/ML PIC18F25K42-I/ML PIC18F24K42-E/ML PIC18F25K42-E/MV PIC18F25K40-I/ML PIC18F25K40T-I/ML PIC18F25K20-E/ML DSPIC33FJ16GS504-E/ML DSPIC33FJ16GS502-E/SO PIC18 8-bit RISC Functional Safety (FuSa) Microcontroller (MCU) XLP nanoWatt Technology ADC2 with Computation Configurable Logic Cell (CLC) Direct Memory Access (DMA) Peripheral Pin Select (PPS) Windowed Watchdog Timer (WWDT) 28-QFN / UQFN package RoHS Compliant IEC 61508 SIL-aware firmware
クイック見積RFQ
詳細をすべてご記入ください — 営業チームが24時間以内にご返信いたします
部品番号 メーカー パッケージ QTY 目標価格 小計
合計: $0.00 USD
✓
見積を送信しました!

24時間以内にメールでご返信いたします

1
RFQ送信済み
2
見積受領
3
注文済み
4
お支払い
5
発送済み
6
配達済み
RFQの詳細を表示