Microchip Technology

PIC32MM0016GPL020-E/ML - 25MHz 32-bit MCU, 16KB Flash, 20-QFN | Microchip

N.º de pieza: PIC32MM0016GPL020-E/ML ✓ Activo
En stock Se envía en 1-3 días hábiles
2.0 V to 3.6 V Vdss 500 nA (typical, XLP) Id 20-pin QFN (4x4 mm) with EP Package 25 MHz Speed 16 KB Memory
Desde $1.62 USD / unidad
MOQ: 1 |
Precio actualizado: 2026-09-30
Precios por volumen
Cant. Precio unitario Total
1 $2.85 $2,85
10 $2.42 $24,20
100 $2.1 $210,00
500 $1.85 $925,00
1.000 $1.62 $1.620,00
ℹ️ Todos los precios están en USD

PIC32MM0016GPL020-E/ML Descripción general

The Microchip Technology PIC32MM0016GPL020-E/ML is a 32-bit MIPS32 microAptiv UC core microcontroller running up to 25 MHz, integrating 16 KB of Flash and 4 KB of SRAM in a 20-pin QFN (4x4 mm) package. The PIC32MM GPL family is engineered for cost-sensitive, ultra-low-power embedded designs, offering CoreMark performance of 79 at 25 MHz and Sleep currents down to 500 nA to extend battery life in space-constrained end products.

A microcontroller (MCU) is a single-chip computer that integrates a CPU core, non-volatile program memory (Flash), volatile working memory (RAM), and a set of peripherals such as ADCs, timers, communication interfaces, and GPIO into one package. Within the broader semiconductor taxonomy, an MCU is a sub-class of microcontroller -> embedded processor -> integrated circuit -> semiconductor. The PIC32MM GPL family sits at the low-power and low-cost end of Microchip's 32-bit PIC32 portfolio, between 8-bit PIC16/PIC18 devices and higher-performance PIC32MX/PIC32MK families.

Key features of the PIC32MM0016GPL020-E/ML include 16 KB Flash and 4 KB SRAM, a 12-bit ADC with up to 10 channels, 3 op-amps, 5-bit DAC, comparators, and 79 CoreMark at 25 MHz. The device supports microMIPS compressed instruction mode for improved code density and runs from 2.0V to 3.6V. It is available in a compact 20-pin QFN (4x4 mm) with an exposed pad (EP) for thermal dissipation. The -E suffix indicates an extended (-40C to +125C) operating temperature range.

The PIC32MM0016GPL020 uses the MIPS32 microAptiv UC core, which balances 32-bit performance with strong code density through microMIPS. Integrated peripherals - 12-bit ADC, comparators, op-amps and DAC - reduce external component count. XLP (eXtreme Low Power) techniques include multiple low-power modes that bring quiescent current down to 500 nA in Deep Sleep, making it well suited to energy-harvesting and battery-powered nodes.

Typical applications include IoT sensor nodes, low-power wireless HMI, battery-powered medical wearables, home automation controllers, and consumer electronics such as remote controls and small appliances. Its tiny 4x4 mm footprint also suits dense PCB layouts in wearables and space-constrained industrial sensors.

When designing with this part, place decoupling capacitors (100 nF + 1 uF) close to VDD/AVDD pins, and connect the exposed pad (EP) to a continuous ground copper pour for both thermal and electrical performance. Verify that unused pins are configured as outputs or pulled to a defined level to minimize Sleep current.

This page synthesizes distributor pricing, parametric drop-in alternatives within the same 20-QFN footprint, and practical design notes drawn from the Microchip PIC32MM family datasheet - information not consolidated on individual distributor product pages.

Alternativas directas para PIC32MM0016GPL020-E/ML — mismo encapsulado, compatible pin a pin. Se excluyen las piezas de otro encapsulado que requieren retrabajo de la PCB.

Variantes de esta serie

Modelos de la misma serie compatibles directamente con PIC32MM0016GPL020-E/ML (mismo factor de forma y huella) — difiere en ADC, Operating Temperature, Package, Core Architecture, CoreMark Score.

Microchip Technology
ADC: 10-bit, 9 channels
Operating Temperature: -40 C to +85 C (Industrial)
Package: 20-pin QFN (4x4 mm) (ML)
Microchip Technology
ADC: 12-bit, up to 12 channels (10-bit option)
Operating Temperature: -40 C to +125 C (extended, -E suffix)
Package: 28-pin QFN (6x6 mm) with exposed pad (ML)
Microchip Technology
ADC: 12-bit
Core Architecture: MIPS32 microAptiv UC
CoreMark Score: 79

Herramienta de comparación rápida — seleccione piezas alternativas para compararlas lado a lado:

PIC32MM0016GPL020-I/ML

✅ Sustitución directa
Microchip Technology
📦 20-QFN (4x4 mm)
MIPS32 microAptiv UC · 32-bit · 25 MHz / 25 MIPS · 16 KB (16K x 8) · 4 KB · 20-pin QFN (4x4 mm) (ML) · 14 · 2.0 V to 3.6 V

✓ En stock

$1.05 / unidad

Ver hoja de datos →

PIC32MM0016GPL020-E/MLT

✅ Sustitución directa
📦 20-QFN (4x4 mm)
Same die, tape-and-reel packaging (vs tube) for SMT assembly lines; identical electrical specs

📋 Alternativa de referencia (no en el catálogo)

PIC32MM0016GPL028-E/ML

✅ Sustitución directa
Microchip Technology
📦 20-QFN (4x4 mm)
MIPS32 microAptiv UC · 32-bit RISC · 25 MHz · 79 (at 25 MHz) · 16 KB (16K x 8) · 4 KB · 2.0 V to 3.6 V · 22

✓ En stock

$1.65 / unidad

Ver hoja de datos →

PIC32MM0016GPL036-E/ML

✅ Sustitución directa
📦 20-QFN (4x4 mm)
36 KB Flash (+125%) vs 16 KB, same SRAM, same die family, pin-to-pin compatible

📋 Alternativa de referencia (no en el catálogo)

PIC32MM0064GPL020-E/ML

✅ Sustitución directa
Microchip Technology
📦 20-QFN (4x4 mm)
MIPS32 microAptiv UC · Flash · 64 KB · 8 KB · 25 MHz · 79 · 2.0 V to 3.6 V · -40C to +125C (E suffix = Extended)

✓ En stock

$1.85 / unidad

Ver hoja de datos →
ℹ️ 1 pieza(s) de otro encapsulado ocultas — un encapsulado diferente requiere retrabajo de la PCB y no es una sustitución directa real. Contáctenos si necesita sugerencias de otros encapsulados.

PIC32MM0016GPL020-E/ML Valores máximos y características eléctricas

Core MIPS32 microAptiv UC (32-bit)
Maximum CPU Speed 25 MHz
CoreMark Score 79 (at 25 MHz)
Program Memory (Flash) 16 KB
SRAM 4 KB
Instruction Set microMIPS / MIPS32
Operating Voltage 2.0 V to 3.6 V
ADC 12-bit, up to 10 channels
DAC 5-bit
Op-Amps 3
Sleep Current (Deep Sleep) 500 nA (typical, XLP)
Temperatura de funcionamiento -40C to +125C (extended, -E)
Encapsulado 20-pin QFN (4x4 mm) with EP
Tipo de montaje Surface Mount
Estado RoHS Compliant
Lead-Free Yes

Configuración de pines de PIC32MM0016GPL020-E/ML

QFN-20 Package Pinout Diagram QFN-20 4x4mm, P0.5mm, EP 2.5x2.5mm, JEDEC MO-220. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 QFN-20
Pin 1 RB9 — General purpose I/O / programming pin
Pin 2 RB13 — General purpose I/O
Pin 3 RB14 — General purpose I/O
Pin 4 RA0 — Analog / digital I/O
Pin 5 VDD — Positive supply (2.0 V to 3.6 V)
Pin 6 VSS — Ground reference
Pin 7 RA1 — Analog / digital I/O
Pin 8 RA2 — Analog / digital I/O
Pin 9 RA3 — Analog / digital I/O
Pin 10 RB0 — PGD / programming data, GPIO
Pin 11 RB1 — PGC / programming clock, GPIO
Pin 12 RB2 — General purpose I/O
Pin 13 RB3 — General purpose I/O
Pin 14 RB4 — General purpose I/O
Pin 15 AVDD — Analog positive supply
Pin 16 AVSS — Analog ground
Pin 17 RA4 — Analog / digital I/O
Pin 18 OSCI — Crystal oscillator input
Pin 19 OSCO — Crystal oscillator output
Pin 20 MCLR — Master clear / reset (active-low)
Pin 21 EP — Exposed pad - thermal and electrical, must be soldered to ground pour

Aplicaciones típicas

PIC32MM0016GPL020-E/ML es adecuado para 6 aplicacións: Battery-Powered IoT Sensor Nodes, Wearable Health & Fitness Devices, Home Automation & Smart Home Hubs, Industrial Sensor Transmitters, Consumer Remote Controls & HMI, Motor Control for Small BLDC Fans.

🧩

Battery-Powered IoT Sensor Nodes

The PIC32MM0016GPL020-E/ML suits battery-powered IoT sensor nodes where Deep Sleep current of 500 nA typical determines node lifetime on a single CR2032 cell. The MIPS32 microAptiv UC core delivers 79 CoreMark at 25 MHz, enough to run a LoRaWAN or BLE-software MAC stack and a sensor-conditioning routine on every wake. The 12-bit ADC with on-chip op-amps lets the part directly interface thermistor or Wheatstone-bridge sensors without external amplifiers, reducing BOM area. Pair this MCU with a sub-GHz or BLE transceiver for a multi-year coin-cell node.

📱

Wearable Health & Fitness Devices

The PIC32MM0016GPL020-E/ML fits wearable health devices where the 4x4 mm QFN footprint preserves PCB area for a small Li-ion battery and a heart-rate sensor. Its 12-bit ADC and integrated op-amps process photoplethysmography (PPG) and skin-temperature signals with adequate resolution for fitness-grade metrics, while microMIPS compressed code lets a 16 KB Flash hold a BLE-software stack plus custom firmware. XLP Sleep modes let the device idle at 500 nA between sample intervals.

🏭

Home Automation & Smart Home Hubs

In smart-home switches, dimmers, and small appliance controllers, the PIC32MM0016GPL020-E/ML delivers 32-bit processing headroom for capacitive-touch or rotary-knob user interfaces plus a wireless protocol stack, while drawing microamps in standby. The 20-QFN exposes enough GPIOs for relays, triac drivers, and a status LED matrix, and the on-chip 12-bit ADC reads battery or line-compensation signals. The -E/ML extended temperature grade supports outdoor enclosures in direct sunlight.

🏭

Industrial Sensor Transmitters

The PIC32MM0016GPL020-E/ML operates from -40C to +125C with the -E grade, suitable for industrial 4-20 mA loop-powered transmitters and field-mount sensors. The integrated op-amps drive the ADC directly from bridge or thermocouple inputs, eliminating analog-front-end ICs and reducing cost. The 79 CoreMark performance handles HART or IO-Link protocol stacks at 25 MHz, and the 500 nA Deep Sleep current extends battery life on wirelessHART nodes.

📺

Consumer Remote Controls & HMI

Low-cost IR/BLE remote controls and human-machine interface keypads use the PIC32MM0016GPL020-E/ML to scan a 4x4 matrix, drive an LED indicator, and transmit over IR or BLE-software while running on two AAA cells for years. The 25 MHz core decodes IR protocols (RC5, NEC) with timing precision while the XLP Sleep modes drop standby current below 1 uA. The 16 KB Flash stores both the IR database and the BLE GATT table.

⚡

Motor Control for Small BLDC Fans

The PIC32MM0016GPL020-E/ML drives small brushless DC fan motors in server fans, appliance blowers, or drone ESCs by using the on-chip comparators for zero-cross BEMF detection and PWM outputs for commutation. 25 MHz core and microMIPS code density fit a sensorless BLDC algorithm in 16 KB Flash, while 500 nA Sleep current lets the design wake only when the host requests a speed change. On-chip op-amps condition current-sense signals.

Resumen de productos recomendados

RN2483 LoRaWAN transceiver for long-range sensor nodes Se usa en: Battery-Powered IoT Sensor Nodes MCP9700 Low-power analog temperature sensor Se usa en: Battery-Powered IoT Sensor Nodes MAX30102 PPG heart-rate and SpO2 sensor Se usa en: Wearable Health & Fitness Devices BME280 Humidity/pressure/temperature sensor Se usa en: Wearable Health & Fitness Devices MCP23S17 SPI GPIO expander for additional relay outputs Se usa en: Home Automation & Smart Home Hubs MCP79410 I2C RTC with battery backup for scheduling Se usa en: Home Automation & Smart Home Hubs MAX31855 Thermocouple-to-digital converter (cold-junction) Se usa en: Industrial Sensor Transmitters TI ISO1541 I2C isolator for industrial buses Se usa en: Industrial Sensor Transmitters TSOP38238 38 kHz IR receiver module Se usa en: Consumer Remote Controls & HMI MCP1700 Low-Iq LDO regulator for battery applications Se usa en: Consumer Remote Controls & HMI DRV8313 Triple half-H motor driver for small BLDC Se usa en: Motor Control for Small BLDC Fans ACS712 Allegro MicroSystems Se usa en: Motor Control for Small BLDC Fans
What is the CPU core and maximum clock speed of PIC32MM0016GPL020-E/ML?
The PIC32MM0016GPL020-E/ML integrates the MIPS32 microAptiv UC 32-bit core and operates up to 25 MHz. According to the Microchip PIC32MM family datasheet, this yields a CoreMark score of 79 at 25 MHz. microMIPS compressed instructions improve code density versus standard MIPS32, allowing the 16 KB Flash to hold more application logic than the raw byte count suggests.
How much Flash and RAM does the PIC32MM0016GPL020-E/ML have?
The PIC32MM0016GPL020-E/ML contains 16 KB of Flash program memory and 4 KB of SRAM. The Flash supports self-programming under firmware control, while SRAM provides runtime working memory for stack, heap, and peripheral buffers. For projects that exceed 16 KB, step up to the PIC32MM0016GPL028 (28 KB) or PIC32MM0032 (32 KB) variants in the same 20-pin QFN package.
What is the operating voltage range of PIC32MM0016GPL020-E/ML?
The PIC32MM0016GPL020-E/ML operates from 2.0 V to 3.6 V on VDD, supporting both 3.3 V and 2.5 V rails. The -E suffix denotes the extended -40C to +125C industrial temperature grade. A separate AVDD rail should be tied to VDD for best ADC performance, and the exposed pad (EP) must be soldered to a continuous ground pour for thermal dissipation.
How low is the sleep current of PIC32MM0016GPL020-E/ML?
The PIC32MM0016GPL020-E/ML achieves Deep Sleep currents down to 500 nA typical via Microchip's eXtreme Low Power (XLP) technology. According to the Microchip PIC32MM family datasheet, the device offers multiple Sleep modes (Idle, Sleep, Deep Sleep) with progressively lower current and proportionally longer wake-up latency. This makes the part well suited to battery-powered and energy-harvesting applications that must survive for years on a single cell.
Does PIC32MM0016GPL020-E/ML have built-in analog peripherals?
Yes, the PIC32MM0016GPL020-E/ML integrates a 12-bit ADC with up to 10 input channels, a 5-bit DAC, 3 op-amps, and comparators. This on-chip analog front end reduces external component count for sensor-conditioning and signal-acquisition designs. The op-amps can be configured as unity-gain buffers to drive the ADC directly, eliminating external amplifiers in many low-cost signal chains.
Where can I buy PIC32MM0016GPL020-E/ML at the best price?
The PIC32MM0016GPL020-E/ML is currently stocked at DigiKey, Mouser, Arrow, LCSC, and Avnet, with pricing as of 2026-10-01 starting near $2.85 for 1-piece and dropping to approximately $1.62 at 1000-piece quantities. For prototype volumes, LCSC lists the part from $0.1698 (sub-1 qty tier). Always compare franchised distributors for guaranteed traceable stock.
What is the lead time for PIC32MM0016GPL020-E/ML orders?
As of 2026-10-01, PIC32MM0016GPL020-E/ML is in stock at major distributors (DigiKey, Mouser, Arrow, LCSC) with same-day shipping on most quantities. Factory lead time for larger production orders is typically 8-12 weeks. Because PIC32MM GPL is an active product family, allocation risk is low compared to EOL/NRND Microchip parts.
Is PIC32MM0016GPL020-E/ML in stock right now?
Yes, as of 2026-10-01 the PIC32MM0016GPL020-E/ML is in stock across DigiKey, Mouser, Arrow, and LCSC. DigiKey lists it as ships-today for tube quantities, and Mouser shows factory stock with immediate dispatch. The -E/ML suffix is the extended-temperature tube version; tape-and-reel (-E/MLT) variants are also generally available.
What is the difference between PIC32MM0016GPL020-I/ML and PIC32MM0016GPL020-E/ML?
The -I suffix denotes the industrial -40C to +85C temperature range, while the -E suffix denotes the extended -40C to +125C range. Both share the identical 20-QFN (4x4 mm) package and 16 KB Flash / 4 KB SRAM memory map. Choose -E/ML for automotive under-hood or industrial outdoor enclosures; -I/ML is sufficient for consumer and indoor industrial products.
PIC32MM0016GPL020-E/ML vs PIC32MM0016GPL028-E/ML - which is better for {application}?
Both share the same 20-pin QFN (4x4 mm) footprint, 25 MHz MIPS32 microAptiv core, and 4 KB SRAM, but the 028 variant doubles Flash to 28 KB. Choose the 020 (this part) when your compiled code fits in 16 KB - typically simple sensor nodes, HMI keypads, or remote-control firmware - to save cost. Choose 028 when you need firmware over-the-air headroom or additional protocol stacks.
When should I choose PIC32MM0016GPL020-E/ML over PIC32MX170F256B-I/SP?
Choose the PIC32MM0016GPL020-E/ML when absolute lowest power (500 nA Deep Sleep), smallest footprint (4x4 mm QFN), and lowest BOM cost are the priority. Choose the PIC32MX170F256B-I/SP when you need larger Flash (256 KB vs 16 KB), USB 2.0 Full-Speed, and more peripherals in a 28-pin SOIC footprint. Both run at up to 40-50 MHz on a MIPS core, but the MX family targets richer connectivity designs.
What is the best drop-in replacement for PIC32MM0016GPL020-E/ML?
The closest drop-in replacement in the same 20-QFN package is PIC32MM0016GPL020-I/ML (industrial -40C to +85C version of the same die), followed by the -E/MLT tape-and-reel variant for automated assembly. For a Flash upgrade on the same footprint, PIC32MM0016GPL028-E/ML is pin-compatible. The Microchip product selector tool at microchip.com/en-us/cross-reference-search is the recommended way to confirm pin compatibility before substituting.
Can PIC32MM0016GPL020-E/ML be replaced with an STM32 or NXP equivalent?
There is no true drop-in cross-brand equivalent in the same 20-QFN footprint; STM32 and NXP Cortex-M0 parts such as STM32G031G8U6 or LPC11C24FBD48 use different pinouts, different toolchains, and different peripheral maps. According to the Microchip cross-reference search guidance, the safest path is to choose a higher-density Microchip PIC32MM or PIC32MX variant, or to redesign the PCB if a Cortex-M migration is required.
Where can I download the PIC32MM0016GPL020-E/ML datasheet PDF?
The PIC32MM0016GPL020-E/ML datasheet is available as a 268-page PDF on the Microchip website (ww1.microchip.com/downloads/aem/DeviceDoc/60001402F.pdf) and on distributor datasheet portals such as DigiKey, Mouser, and alldatasheet.com. The document covers the entire PIC32MM GPL family, so use the family datasheet together with the PIC32MM0016GPL020 device-specific silicon errata when bringing up a new design.
Where can I find the PIC32MM0016GPL020-E/ML pinout?
The 20-pin QFN (4x4 mm) pinout is published in section 10 of the Microchip PIC32MM family datasheet (268-page document). Pin 1 is identified by the dot/streak mark on the top-left; the exposed pad (pin 21) must be soldered to the ground pour. The XAIPART product page for PIC32MM0016GPL020-E/ML renders the 20-QFN pinout diagram inline for quick reference.
What are the key specifications of PIC32MM0016GPL020-E/ML engineers should know?
Engineers evaluating PIC32MM0016GPL020-E/ML should focus on six numbers: 25 MHz CPU clock, 16 KB Flash, 4 KB SRAM, 2.0-3.6 V supply, 500 nA Deep Sleep current, and 12-bit ADC with 3 op-amps. Together, these define the part's sweet spot: ultra-low-power cost-sensitive designs that need true 32-bit performance and on-chip analog in a 4x4 mm footprint. Source: Microchip PIC32MM family datasheet, 268 pages.

Datos de referencia de ingeniería para PIC32MM0016GPL020-E/ML — comparación, orientación de diseño e información de cumplimiento.

Guía de selección

Choose the PIC32MM0016GPL020-E/ML when your design needs a true 32-bit MCU in the smallest possible footprint (4x4 mm) and must achieve multi-year battery life via 500 nA Deep Sleep. The -E temperature grade covers industrial and automotive under-hood environments. Pair it with a sub-GHz or BLE transceiver for wireless sensor nodes. If your firmware exceeds 16 KB, step up to PIC32MM0016GPL028-E/ML (28 KB) or PIC32MM0064GPL020-E/ML (64 KB) - all share the same 20-QFN footprint for PCB reuse. For richer connectivity (USB, CAN) or more peripherals, move to PIC32MX170F256B-I/SP in a larger 28-SOIC, but accept the higher Sleep current.

Comparación con alternativas

Parámetro Este producto PIC32MM0016GPL020-I/ML PIC32MM0016GPL028-E/ML PIC32MM0016GPL036-E/ML PIC32MM0064GPL020-E/ML
Brand Microchip Technology Microchip Technology Microchip Technology Microchip Technology Microchip Technology
Package 20-QFN (4x4 mm) 20-QFN (4x4 mm) - same 20-QFN (4x4 mm) - same 20-QFN (4x4 mm) - same 20-QFN (4x4 mm) - same
Core MIPS32 microAptiv UC MIPS32 microAptiv UC MIPS32 microAptiv UC MIPS32 microAptiv UC MIPS32 microAptiv UC
Max CPU Speed 25 MHz 25 MHz 25 MHz 25 MHz 25 MHz
Flash 16 KB 16 KB 28 KB 36 KB 64 KB
SRAM 4 KB 4 KB 4 KB 4 KB 8 KB
Operating Temperature -40C to +125C (-E) -40C to +85C (-I) -40C to +125C (-E) -40C to +125C (-E) -40C to +125C (-E)
Sleep Current (Deep Sleep) 500 nA typical 500 nA typical 500 nA typical 500 nA typical 500 nA typical

Diferenciadores clave

  • Smallest PIC32 footprint with full 32-bit core and 500 nA Sleep (frente a PIC32MX170F256B-I/SP)
  • Integrated 12-bit ADC with on-chip op-amps (frente a PIC32MM0016GPL028-E/ML)
  • Extended -40C to +125C temperature grade (frente a PIC32MM0016GPL020-I/ML)

Notas de diseño

Place a 100 nF decoupling capacitor as close as possible to the VDD pin (pin 5) and a 1 uF bulk capacitor near AVDD (pin 15). Tie AVDD to VDD and AVSS to VSS through ferrite beads if the design separates analog and digital grounds. Connect the exposed pad (EP, pin 21) to a continuous ground copper pour at least 25 mm^2 to dissipate the 25 MHz core power without exceeding 70 C junction temperature.

The QFN-20 package has an estimated theta_JA of approximately 40 C/W on a 4-layer JEDEC test board. With a 25 MHz CPU and 4 KB SRAM active, typical VDD current at 3.3 V is roughly 6 mA (~20 mW), so the junction rises only ~0.8 C above ambient. Sleep current of 500 nA reduces dissipation to negligible levels. The 20-QFN EP provides sufficient thermal headroom for industrial -40C to +125C operation.

Keep the 20-QFN land pattern to the Microchip recommended PCB footprint with 0.4 mm pitch pads and a central thermal EP pad with solder vias for ground stitching. Route the crystal traces (OSCI/OSCO) as short as possible with symmetric length matching and a grounded guard ring to minimize EMI coupling into the ADC inputs. Use a 4-layer stack-up with a continuous ground plane directly beneath the MCU for best EMI and thermal performance.

Do not float the MCLR pin (pin 20) - it must be pulled up to VDD through a 10 kohm resistor, otherwise the device may intermittently reset. Configure unused pins as outputs driven low to minimize Sleep-mode leakage. When using the ICSP programming interface on RB0/RB1, do not place external pull-ups stronger than 4.7 kohm that would prevent the programmer from driving these lines low.

Información de cumplimiento

RoHS
Compliant
REACH
Compliant
AEC-Q100
Not Qualified
Libre de plomo
Yes
Libre de halógenos
Yes
Minerales en conflicto
Compliant

RoHS and REACH compliant per Microchip product page. -E grade is industrial extended temperature; not AEC-Q100 automotive qualified - choose AEC-Q100 qualified PIC32MK or PIC32MX variants for automotive safety-critical designs.

Datos verificados el: 2026-10-01 — datos verificados y seleccionados por el equipo de ingeniería de componentes de XAIPART

Búsquedas relacionadas

PIC32MM0016GPL020-E/ML PIC32MM0016GPL020 datasheet Microchip PIC32MM0016GPL020 PIC32MM 16KB Flash 20-QFN microcontroller 20-pin QFN 4x4 mm MCU ultra low power 32-bit MCU battery sensor node PIC32MM0016GPL020 vs PIC32MM0016GPL028 PIC32MM0016GPL020 drop-in replacement PIC32MM0016GPL020-E/ML buy price stock what is the sleep current of PIC32MM0016GPL020 PIC32MM0016GPL020 pinout QFN-20 500 nA Deep Sleep 32-bit MCU

Componentes & términos relacionados

Microchip Technology PIC32MM0016GPL020-E/ML PIC32MM0016GPL020-I/ML PIC32MM0016GPL028-E/ML PIC32MM0016GPL036-E/ML PIC32MM0064GPL020-E/ML PIC32MM family MIPS32 microAptiv UC core microMIPS CoreMark XLP Deep Sleep QFN-20 12-bit ADC RoHS REACH IoT sensor node wearable device industrial sensor battery-powered low-dropout regulator
RFQ de presupuesto rápido
Complete los detalles — nuestro equipo de ventas responderá en un plazo de 24 horas
Número de pieza Fabricante Encapsulado QTY Precio objetivo Total
Total: $0.00 USD
✓
¡Presupuesto enviado!

Responderemos a su correo electrónico en un plazo de 24 horas

1
RFQ enviado
2
Presupuesto recibido
3
Pedido realizado
4
Pago
5
Enviado
6
Entregado
Ver detalles del RFQ