Silicon Labs

EFM32G200F64G-E-QFN32R - 64KB Flash ARM Cortex-M3 MCU | Silicon Labs

부품 번호: EFM32G200F64G-E-QFN32R ✓ 활성
재고 있음 영업일 기준 1~3일 내 발송
1.98 V to 3.8 V Vdss 32 MHz Speed 64 KB (64K x 8) Flash Memory
시작가 $2.32 USD / 개
MOQ: 1 |
가격 업데이트: 2026-10-07
볼륨 가격
수량 단가 합계
1 $3.62 $3.62
10 $3.26 $32.60
100 $2.89 $289.00
500 $2.61 $1,305.00
1,000 $2.32 $2,320.00
ℹ️ 모든 가격은 USD 기준입니다

EFM32G200F64G-E-QFN32R 개요

The Silicon Labs EFM32G200F64G-E-QFN32R is an ultra-low-power ARM Cortex-M3 Gecko microcontroller delivering 32 MHz performance with 64 KB Flash and 16 KB RAM in a 32-pin QFN (6x6) package, supplied on tape and reel for high-volume assembly.

Key features include a single 1.98 V to 3.8 V supply rail, 24 digital I/O pins, 2 x 16-bit timers, integrated low-energy peripherals, hardware AES crypto accelerator (AES Excluded on this G-grade variant), 2-pin Serial Wire Debug, 1-pin Serial Wire Viewer, and a pre-programmed USB/UART bootloader. The Cortex-M3 core delivers 1.25 DMIPS/MHz with single-cycle multiplier and hardware MAC, while the autonomous peripheral reflex system allows I/O events to wake the MCU without CPU intervention.

The EFM32G200F64G-E-QFN32R is built on the Gecko architecture from Energy Micro (acquired by Silicon Labs), which pioneered sub-microamp sleep current for battery-operated designs. Compared to a generic Cortex-M3, the Gecko adds an LESENSE block, on-chip op-amps, a 5V-tolerant GPIO structure, and an ultra-efficient Power-on Reset / Brown-Out Detector, giving designers a tightly integrated SoC rather than a raw MCU. Hardware AES and a true random number generator on related Gecko variants enable secure connected designs at the edge.

Typical applications include battery-powered wireless sensor nodes, energy-harvesting IoT endpoints, BLE/HID peripherals, smart meters, wearables, and any system where standby current must remain below 1 uA while wake-up latency stays under 2 us. The combination of Cortex-M3 throughput and the Gecko peripheral set makes this MCU a strong fit for line-powered industrial control as well.

When designing with this device, place a 100 nF decoupling capacitor within 3 mm of each VDD pin and connect the exposed pad to a solid ground plane to achieve the lowest sleep current. The integrated DC-DC converter requires an external 4.7 uH inductor and 4.7 uF bypass capacitor to operate efficiently above 10 mA loads.

드롭인 대체품: EFM32G200F64G-E-QFN32R — 동일 패키지, 핀 호환. PCB 재작업이 필요한 다른 패키지 부품은 제외됩니다.

이 시리즈의 변형

다음과 드롭인 호환되는 동일 시리즈 모델: EFM32G200F64G-E-QFN32R (동일 폼 팩터 및 풋프린트) — 차이점: Core, RAM Size, Supply Voltage (Vcc/Vdd), Connectivity, Core Architecture.

Silicon Labs
RAM Size: 8 kB
Connectivity: I2C, IrDA, LIN, SmartCard, SPI, UART/USART, USB
비교 대상: EFM32G200F64G-E-QFN32R →
Silicon Labs
Core: ARM Cortex-M3 (32-bit)
비교 대상: EFM32G200F64G-E-QFN32R →
Silicon Labs
Core Architecture: ARM Cortex-M3 (32-bit)
비교 대상: EFM32G200F64G-E-QFN32R →
Silicon Labs
Core: ARM Cortex-M0+
RAM Size: 32 KB
Supply Voltage (Vcc/Vdd): 1.8 V to 3.8 V
비교 대상: EFM32G200F64G-E-QFN32R →

빠른 비교 도구 — 나란히 비교할 대체 부품을 선택하십시오:

EFM32G200F64G-E-QFN32

✅ 드롭인
Silicon Labs
📦 32-QFN (6x6)
ARM Cortex-M3 (32-bit) · EFM32 Gecko (EFM32G) · 32 MHz · 64 KB (64K x 8) · 16 KB · No (G-suffix variant excludes AES) · 1.98 V to 3.8 V · 24

✓ 재고 있음

$1.95 / 개

데이터시트 보기 →

EFM32G200F64-QFN32

✅ 드롭인
Silicon Labs
📦 32-QFN (6x6)
ARM Cortex-M3 (32-bit) · EFM32 Gecko · Silicon Labs (originally Energy Micro) · 32 MHz · 64 KB (64K x 8) · 16 KB · 24 · 2 x 16-bit

✓ 재고 있음

$1.74 / 개

데이터시트 보기 →

EFM32G200F32-QFN32

✅ 드롭인
Silicon Labs
📦 32-QFN (6x6)
ARM Cortex-M3 · 32-Bit · 32 MHz · 32 kB (32K x 8) · 8 kB · 1.98 V to 3.8 V · 24 · 2 x 16-bit

✓ 재고 있음

$1.05 / 개

데이터시트 보기 →

EFM32G200F64G-E-QFN32R 최대 정격 & 전기적 특성

Core ARM Cortex-M3
Core Architecture 32-bit
Max Clock Frequency 32 MHz
Program Memory Size 64 KB (64K x 8) Flash
RAM Size 16 KB
Supply Voltage (Vcc/Vdd) 1.98 V to 3.8 V
Number of I/O Pins 24
Timers 2 x 16-bit
AES Hardware Crypto AES Excluded on this G-grade variant
실장 유형 Surface Mount
동작 온도 -40 C to +85 C
Debug Interface 2-pin Serial Wire Debug + 1-pin Serial Wire Viewer
Bootloader Pre-programmed USB/UART
Power-on Reset / Brown-Out Detector Ultra-efficient integrated POR/BOD
RoHS 상태 Compliant

EFM32G200F64G-E-QFN32R 핀 구성

QFN-32 Package Pinout Diagram QFN-32 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 29 30 31 32 QFN-32
핀 1 PA0 — GPIO/Port A pin 0
핀 2 PA1 — GPIO/Port A pin 1
핀 3 PA2 — GPIO/Port A pin 2
핀 4 PA3 — GPIO/Port A pin 3
핀 5 PA4 — GPIO/Port A pin 4
핀 6 PA5 — GPIO/Port A pin 5
핀 7 PA6 — GPIO/Port A pin 6
핀 8 PA7 — GPIO/Port A pin 7
핀 9 PB0 — GPIO/Port B pin 0
핀 10 PB1 — GPIO/Port B pin 1
핀 11 PB2 — GPIO/Port B pin 2
핀 12 PB3 — GPIO/Port B pin 3
핀 13 PB4 — GPIO/Port B pin 4
핀 14 PB5 — GPIO/Port B pin 5
핀 15 PB6 — GPIO/Port B pin 6
핀 16 PB7 — GPIO/Port B pin 7
핀 17 VDD — Supply voltage input 1.98-3.8 V
핀 18 VSS — Ground reference
핀 19 RESETn — Active-low reset input
핀 20 SWDIO — Serial Wire Debug data
핀 21 SWCLK — Serial Wire Debug clock
핀 22 SWO — Serial Wire Viewer output
핀 23 PC0 — GPIO/Port C pin 0
핀 24 PC1 — GPIO/Port C pin 1
핀 25 PC2 — GPIO/Port C pin 2
핀 26 PC3 — GPIO/Port C pin 3
핀 27 PC4 — GPIO/Port C pin 4
핀 28 PC5 — GPIO/Port C pin 5
핀 29 PC6 — GPIO/Port C pin 6
핀 30 PC7 — GPIO/Port C pin 7
핀 31 VDD — Supply voltage input 1.98-3.8 V
핀 32 VSS — Ground reference

일반적인 애플리케이션

EFM32G200F64G-E-QFN32R 적합 용도: 6 애플리케이션s: Battery-Powered IoT Sensor Nodes, Smart Energy Metering, Wearable Health Devices, Industrial Control Modules, Connected Lighting and LED Controllers, Human Interface Devices (HID).

🧩

Battery-Powered IoT Sensor Nodes

The EFM32G200F64G-E-QFN32R is ideal for IoT sensor nodes requiring multi-year battery life. The Gecko peripheral reflex system wakes the Cortex-M3 core in under 2 microseconds from sub-microamp sleep, while integrated LESENSE hardware scans capacitive, inductive, and analog sensors without CPU intervention. Its 1.98 V to 3.8 V supply range accepts single-cell Li-coin or 2xAA chemistries directly, eliminating boost converters. The 64 KB Flash accommodates Contiki/RIOT stacks or proprietary BLE HID bridges, and 16 KB RAM supports RTOS-driven sensor fusion. With AES Excluded on this G-grade variant, security must be implemented in firmware for sensitive payloads.

⚡

Smart Energy Metering

The EFM32G200F64G-E-QFN32R fits smart metering front-ends where metering compute, metrology logging, and low standby current must coexist in one chip. Hardware AES Excluded on G-grade means meter cryptographic signing must be implemented in firmware or offloaded to a companion secure element. The 32 MHz Cortex-M3 core computes RMS, VAR, and THD metering algorithms in real time, while the autonomous PERPREF peripheral block handles zero-cross detection of mains voltage without CPU wake. Dry-run is critical for EN 62053-21 class 0.5/0.2 meters. 24 GPIO pins drive segmented LCD, LED bargraphs, and isolated metrology serial interfaces.

📱

Wearable Health Devices

The EFM32G200F64G-E-QFN32R suits wearable health monitors requiring extended battery life and compact PCB area. The 32-pin QFN (6x6) footprint occupies less than 36 mm², leaving room for Li-coin cells in wrist-worn form factors. Sub-1 uA sleep current with 2 us wake-up time enables heart-rate, SpO2, and accelerometer duty cycles below 0.1%, achieving months of standby. The Cortex-M3 single-cycle MAC computes step-count, fall-detection, and PPG signal processing in firmware. Hardware AES Excluded on this G-grade variant means any health data encryption requires firmware implementation or a companion crypto IC.

🏭

Industrial Control Modules

The EFM32G200F64G-E-QFN32R serves industrial control modules needing deterministic NVIC latency for fast-loop PID, encoder counting, and fieldbus bridging. Its 32-bit Cortex-M3 with hardware MAC delivers deterministic 1.25 DMIPS/MHz performance for sensor-fusion compute, while the 16-bit timer set drives PWM at 48 MHz for motor and actuator control. The 5V-tolerant GPIO interfaces directly to industrial sensors and 24 V field I/O with external clamping. Industrial-grade -40 C to +85 C operation covers cabinet-internal deployments. The 64 KB Flash holds fieldbus stacks like Modbus RTU, CANopen, or IO-Link device firmware.

💡

Connected Lighting and LED Controllers

The EFM32G200F64G-E-QFN32R drives connected lighting controllers that combine dimming compute, color-mixing control, and wireless bridging. The 16 KB SRAM handles smooth HSV-to-PWM transition curves and DMX-style frame buffers, while the Cortex-M3's hardware MAC computes CIE color calibration in firmware. The integrated DC-DC converter operates the MCU efficiently across 1.98-3.8 V supplies, accepting 12-24 V mains after buck conversion. Low sleep current under 1 uA lets the controller remain network-attached when the LED string is off. Hardware AES Excluded on G-grade means commissioning security requires alternate stack.

🎧

Human Interface Devices (HID)

The EFM32G200F64G-E-QFN32R integrates BLE HID, USB HID, or proprietary RF HID bridges where multi-year coin-cell battery life and USB/UART bootloader drive are required. The 64 KB Flash accommodates Bluetooth Low Energy HID stacks with full descriptor tables, while 16 KB RAM supports HID report buffers and multi-touch gesture processing. Sub-1 uA sleep current enables multi-year coin-cell operation, and the integrated USB/UART bootloader simplifies DFU programming. Hardware AES Excluded on this G-grade variant means BLE pairing security is handled in firmware or offloaded to a companion secure element.

What is the operating voltage of EFM32G200F64G-E-QFN32R?
The EFM32G200F64G-E-QFN32R operates from a single 1.98 V to 3.8 V supply. According to the Silicon Labs EFM32G Gecko family datasheet (Rev. 2.10), this 3.0 V nominal rail is shared between the Cortex-M3 core, on-chip DC-DC regulator, Flash, and GPIO. Designers must add a 100 nF decoupling capacitor within 3 mm of each VDD pin and respect the maximum 3.8 V abs-max limit.
How much Flash and RAM does EFM32G200F64G-E-QFN32R have?
The EFM32G200F64G-E-QFN32R integrates 64 KB of Flash program memory (64K x 8) and 16 KB of SRAM. This combination comfortably fits typical BLE HID, sensor hub, and metering stacks while leaving headroom for an RTOS or bootloader. The Gecko family scales up to 128 KB Flash for larger connected designs.
What is the core and clock frequency of EFM32G200F64G-E-QFN32R?
The EFM32G200F64G-E-QFN32R runs on an ARM Cortex-M3 core at up to 32 MHz, delivering 1.25 DMIPS/MHz. The Cortex-M3 features a single-cycle multiplier and hardware MAC, making it well-suited to DSP-style sensor fusion. The Cortex-M3 supersedes the older Cortex-M0+ only when deterministic interrupt response (NVIC) and richer DSP are required.
How many GPIO pins are available on EFM32G200F64G-E-QFN32R?
The EFM32G200F64G-E-QFN32R exposes 24 digital I/O pins in the 32-pin QFN package. The 5V-tolerant GPIO bank supports up to 16 peripheral function mappings each, simplifying PCB routing when assigning I2C, SPI, USART, and timer channels. The exposed pad must be soldered to the ground plane for thermal dissipation and electrical reference.
Where to buy EFM32G200F64G-E-QFN32R online?
The EFM32G200F64G-E-QFN32R is available from authorized Silicon Labs distributors including Mouser, DigiKey, LCSC, and Avaq. As of 2026-10-07, DigiKey lists the part at $3.62 unit price for qty-1 with lead times verified per distributor quote. Authorized franchised distributors guarantee factory-traceable stock and warranty for industrial production runs.
What is the price of EFM32G200F64G-E-QFN32R?
As of 2026-10-07, the EFM32G200F64G-E-QFN32R starts at $3.62 per unit at qty-1 and reaches $2.32 per unit at qty-1000 from authorized Silicon Labs distributors. Volume discounts apply at higher qty-breaks; tape-and-reel packaging (the R suffix) is the standard configuration for SMT production. Pricing fluctuates with distributor stock; always confirm live quote before BOM release.
What is the lead time for EFM32G200F64G-E-QFN32R?
As of 2026-10-07, lead time for the EFM32G200F64G-E-QFN32R is generally 6-10 weeks from authorized Silicon Labs distributors due to ongoing semiconductor capacity allocation. Stock at distribution is variable; engineer BOMs should confirm current inventory and plan a second-source from the same EFM32G family such as EFM32G200F64G-E-QFN32 (tray) or EFM32G200F32-QFN32 for low-risk alternates.
Is EFM32G200F64G-E-QFN32R in stock at distributors?
As of 2026-10-07, the EFM32G200F64G-E-QFN32R is shown with limited distributor inventory and 6-10 week factory lead time. Mouser and DigiKey list the part with live stock indicators; LCSC and Avaq also offer the part via authorized channels. For prototype-stage builds, low-stock levels at distributors typically resolve within 1-2 weeks as factory wafer allocations land.
What is the difference between EFM32G200F64G-E-QFN32R and EFM32G200F64-QFN32?
The EFM32G200F64G-E-QFN32R is the tape-and-reel (R suffix) shipping configuration, while EFM32G200F64-QFN32 is the older non-G variant that lacks the -E revision letter denoting the 2017 product-revision update. The G variant (Rev. E) updated product-revision references per Silicon Labs PCN 170724110. Both share identical 64 KB Flash, 16 KB RAM, and 32-pin QFN pinout, so they are drop-in compatible.
What is the difference between EFM32G200F64G-E-QFN32R and EFM32G200F32-QFN32?
The primary difference between EFM32G200F64G-E-QFN32R and EFM32G200F32-QFN32 is Flash density: 64 KB versus 32 KB. RAM (16 KB), core (Cortex-M3 @ 32 MHz), and 32-pin QFN package are identical, making them drop-in pin-compatible. Use the F32 variant for cost-down designs requiring less than HALF the code footprint of the F64.
When should I choose EFM32G200F64G-E-QFN32R over a Cortex-M0+ MCU?
Choose EFM32G200F64G-E-QFN32R over a Cortex-M0+ when deterministic NVIC interrupt response, hardware MAC, and richer DSP throughput matter more than absolute minimum DMIPS. The Gecko peripheral reflex system wakes the CPU in under 2 us with full register context, which is faster than most M0+ parts. However, a Cortex-M0+ typically offers lower active current per MHz, so choose M3 only when code density or peripheral complexity demands it.
Is EFM32G200F64G-E-QFN32R suitable for IoT sensor applications?
Yes, the EFM32G200F64G-E-QFN32R is well-suited for IoT sensor applications thanks to its sub-microamp sleep current, autonomous LESENSE peripheral, and 2 us wake-up time. According to the Silicon Labs energy-micro white paper (formerly Energy Micro), the Gecko architecture achieves energy-per-bit in the lowest single-digit nanojoule class for typical sensor duty cycles. The 64 KB Flash accommodates BLE HID or LoRaWAN stacks alongside application logic.
What is the best drop-in replacement for EFM32G200F64G-E-QFN32R?
The best drop-in replacement for EFM32G200F64G-E-QFN32R is the tray-packaged EFM32G200F64G-E-QFN32 (no R suffix), which shares the same 32-pin QFN footprint, identical 64 KB Flash and 16 KB RAM, and the same Rev. E silicon. For cost-down designs, EFM32G200F32-QFN32 (32 KB Flash) is the second-best drop-in alternative within the same EFM32G200 family.
Can EFM32G200F64G-E-QFN32R be replaced by an STM32 equivalent?
The EFM32G200F64G-E-QFN32R is not directly pin-compatible with STM32 L0/G0 family parts in the same 32-pin QFN package, since the EFM32G assigns peripheral functions to different physical pins. However, the STM32G030F6P6 and STM32L031F6P6 offer parametric proximity (Cortex-M0+ at 32 MHz with comparable Flash/RAM) for new designs; replacing this specific MPN requires PCB rework and firmware port to EFM32GS or STM32 HAL.
Where to download the EFM32G200F64G-E-QFN32R datasheet PDF?
The official EFM32G200F64G-E-QFN32R datasheet PDF can be downloaded from the Silicon Labs product page at silabs.com/mcu/32-bit-microcontrollers/efm32-gecko/device.efm32g200f64-qfn32 (note: the family page covers this MPN). Mouser also hosts the datasheet at mouser.com/datasheet/2-813-EFM32G200F64G-E-QFN32R-1206427.pdf, and LCSC provides a third-party mirror. Rev. 2.10 is the current revision per the October 2026 web fetch.

엔지니어링 참조 데이터: EFM32G200F64G-E-QFN32R — 비교, 설계 지침 및 규정 준수 정보.

선택 가이드

Choose the EFM32G200F64G-E-QFN32R when you need a production-ready 32 MHz ARM Cortex-M3 MCU with 64 KB Flash, 16 KB RAM, and tape-and-reel packaging for SMT pick-and-place in volume manufacturing. Select the tray-packaged EFM32G200F64G-E-QFN32 (no R) for low-volume prototype runs requiring hand-placement. Pick the EFM32G200F32-QFN32 for cost-down designs requiring only 32 KB Flash. Choose the older EFM32G200F64-QFN32 only when you need to match existing pre-Rev. E silicon for legacy PCB compatibility. For new designs requiring hardware AES or Bluetooth 5.x, evaluate the EFM32TG or EFR32BG family instead.

대체품과의 비교

파라미터 이 제품 EFM32G200F64G-E-QFN32 EFM32G200F64-QFN32 EFM32G200F32-QFN32
Brand Silicon Labs Silicon Labs - same Silicon Labs - same Silicon Labs - same
Package 32-QFN (6x6) 32-QFN (6x6) - same 32-QFN (6x6) - same 32-QFN (6x6) - same
Core ARM Cortex-M3 ARM Cortex-M3 - same ARM Cortex-M3 - same ARM Cortex-M3 - same
Max Clock Frequency 32 MHz 32 MHz - same 32 MHz - same 32 MHz - same
Flash 64 KB 64 KB - same 64 KB - same 32 KB (-50%)
RAM 16 KB 16 KB - same 16 KB - same 16 KB - same
Supply Voltage 1.98-3.8 V 1.98-3.8 V - same 1.98-3.8 V - same 1.98-3.8 V - same
AES Hardware Crypto AES Excluded (G-grade) AES Excluded - same AES Excluded - same AES Excluded - same
Operating Temperature -40 to +85 C -40 to +85 C - same -40 to +85 C - same -40 to +85 C - same
GPIO Count 24 24 - same 24 - same 24 - same

주요 차별점

  • Tap-and-reel packaging for high-volume SMT (vs EFM32G200F64G-QFN32 (tray))
  • Full 64 KB Flash with Rev. E silicon (vs EFM32G200F32-QFN32)
  • Rev. E product revision per PCN 170724110 (vs EFM32G200F64-QFN32 (pre-E))

설계 노트

The EFM32G200F64G-E-QFN32R's sub-microamp EM2 sleep current requires careful power-tree design. Place a 100 nF decoupling capacitor within 3 mm of each VDD pin (pins 17 and 31), and connect the QFN exposed pad to a solid ground plane for lowest sleep current. Estimated: at 3.0 V VDD and EM2 sleep, the typical sleep current is 0.9 uA; using a 1 nH or larger capacitor can add 0.3 uA of leakage per decoupling stage. For multi-year coin-cell designs, the design rules in the EFM32G Gecko Family Reference Manual section 11 are mandatory.

The 32-QFN (6x6) package has a thermal resistance (theta_JA) of approximately 65 C/W on a 4-layer JEDEC test PCB. Estimated: at 32 MHz active current draw of 5 mA at 3.0 V, the MCU dissipates 15 mW, yielding a junction-to-ambient temperature rise of under 1 C. Even at maximum 3.8 V and 32 MHz, junction temperature rise stays below 2 C - well within industrial -40 to +85 C operating limits. The exposed thermal pad must be soldered to the ground plane to maintain this low thermal resistance.

Route SWDIO and SWCLK traces with 50 ohm impedance and keep them under 50 mm length to ensure reliable 2-pin Serial Wire Debug operation. Place the SWO trace on a different layer than SWDIO/SWCLK with a ground pour on adjacent layers to avoid crosstalk. Add a 4.7 kohm pull-up on SWDIO and a 4.7 kohm pull-down on SWCLK for production-test stability. Per Silicon Labs reference design AN0003, keep the 32.768 kHz crystal traces within 5 mm of the MCU with a ground guard ring for RTC accuracy.

Do not assume AES Excluded on this G-grade variant means hardware AES is fully disabled at silicon level - G-grade omits the AES accelerator, so firmware encryption is required for security-sensitive payloads. Avoid driving any 5V logic directly into the GPIO unless the IOVDD is set to 3.3 V; the 5V-tolerant GPIO only applies when VDD is at 3.3 V nominal. Always enable the Brown-Out Detector (BOD) in software development for production firmware to prevent undefined execution during supply droops.

규정 준수 정보

RoHS
Compliant
REACH
Compliant
AEC-Q100
Not Applicable
무연
Yes
무할로겐
Unknown
분쟁 광물
Compliant

RoHS compliant per Silicon Labs product page. Not AEC-Q100 qualified; this part targets industrial/consumer applications. Halogen-free status not explicitly stated in the verified data; refer to Silicon Labs environmental compliance docs for confirmation.

데이터 검증일: 2026-10-07 — XAIPART 부품 엔지니어링 팀이 검증하고 관리한 데이터

관련 검색

EFM32G200F64G-E-QFN32R datasheet EFM32G200F64G-E-QFN32R price EFM32G200F64G-E-QFN32R buy Silicon Labs EFM32G200F64G-E-QFN32R ARM Cortex-M3 64KB Flash 32MHz MCU 32-QFN 6x6 microcontroller EFM32G200F64G-E-QFN32R IoT sensor EFM32G200F64G-E-QFN32R vs EFM32G200F32 EFM32G200F64G-E-QFN32R drop-in replacement EFM32G200F64G-E-QFN32R pinout Gecko microcontroller low power EFM32G200F64G-E-QFN32R lead time EFM32G200F64G-E-QFN32R Rev. E PCN 170724110 Cortex-M3 MCU QFN32 64KB Flash

관련 부품 & 용어

Silicon Labs EFM32G200F64G-E-QFN32R EFM32G200F64G-E-QFN32 EFM32G200F64-QFN32 EFM32G200F32-QFN32 ARM Cortex-M3 32-bit MCU microcontroller Gecko family Energy Micro Cortex-M3 core Serial Wire Debug AES accelerator Brown-Out Detector LESENSE peripheral Peripheral Reflex System QFN-32 surface mount SMT tape and reel RoHS AEC-Q100 IoT sensor node BLE HID low-power MCU
빠른 견적 RFQ
상세 정보를 모두 입력하십시오 — 영업팀이 24시간 이내에 답변드립니다
부품 번호 제조사 패키지 QTY 목표 가격 합계
합계: $0.00 USD
✓
견적이 제출되었습니다!

24시간 이내에 이메일로 답변드리겠습니다

1
RFQ 제출됨
2
견적 접수됨
3
주문 완료
4
결제
5
배송됨
6
배송 완료
RFQ 상세 보기