Microchip Technology

PIC24F16KA302-I/ML - 16KB Flash XLP MCU 32MHz 28-QFN | Microchip

Teilenummer: PIC24F16KA302-I/ML ✓ Aktiv
Auf Lager Versand in 1–3 Werktagen
1.8 V to 3.6 V Vdss 28-pin QFN (6x6 mm) Package 32 MHz (16 MIPS) Speed 16 KB (5.5K x 24 instructions) Memory
Ab $1.95 USD / Stück
MOQ: 1 |
Preis aktualisiert: 2026-09-29
Mengenpreise
Menge Stückpreis Gesamt
1 $2.95 $2,95
10 $2.65 $26,50
100 $2.35 $235,00
500 $2.1 $1.050,00
1.000 $1.95 $1.950,00
ℹ️ Alle Preise sind in USD

PIC24F16KA302-I/ML Übersicht

The Microchip Technology PIC24F16KA302-I/ML is a 16-bit PIC24F microcontroller from the XLP (eXtreme Low Power) family, featuring 16 KB Flash program memory, 2 KB RAM and 512 bytes of EEPROM, with a 32 MHz MIPS core delivering 16 MIPS performance, housed in a 28-pin QFN (6x6 mm) package. The PIC24F16KA302 family targets battery-powered and energy-harvesting designs where deep-sleep current in the nanoamp range is mandatory. The /ML package suffix selects the 28-lead QFN (Quad Flat No-lead) variant.

A 16-bit microcontroller (MCU) is a single-chip computer that integrates a CPU, memory, peripherals around one silicon die. Within the taxonomy, PIC24F belongs to the 16-bit MCU class, which sits between 8-bit MCUs (PIC10/12/18, AVR, 8051) and 32-bit MCUs (PIC32, ARM Cortex-M). The 'F' suffix indicates Flash program memory. XLP technology adds deep-sleep modes with wake-on-interrupt, RTC, and ultra-low run currents, positioning this part below many Cortex-M0+ devices in active power yet 2-3 orders lower in sleep current.

Key features include a 12-bit A/D converter with up to 16 channels (depending on package), 3 op-amps / comparators with CTMU (Charge Time Measurement Unit) for capacitive touch, multiple timers (T1-T5), 2 UART, 2 SPI, 2 I2C, RTCC, and a chargeable DSBOR/DSWDT deep-sleep brown-out reset and watchdog. Voltage operation is 1.8 V to 3.6 V across -40C to +85C industrial range. The CTMU enables capacitive-touch buttons, sliders, and proximity sensing without an external analog front-end.

The PIC24F16KA302 architecture uses a 16-bit modified Harvard structure with a 24-bit instruction word and 16-bit data path, plus a hardware multiply/divide and DSP extensions, allowing efficient C compilation under MPLAB XC16. The Flash memory supports self-programming via ICSP for in-field firmware updates, while the 512-byte EEPROM handles non-volatile user parameters (calibration, settings). The 32 MHz CPU clock is synthesized from the internal 8 MHz FRC with PLL, eliminating external crystals for cost-sensitive designs.

Typical applications include battery-powered sensor nodes with capacitive touch, wearable health monitors (HR, SpO2 front-ends), energy-harvesting devices, IoT endpoints with sleep duty cycling, remote controls, and low-power wireless sensor nodes paired with sub-GHz radios (MRF89XA, RN2483). The CTMU peripheral is particularly useful in proximity-sensing applications and water-level / ice detection.

Design considerations include careful decoupling of AVDD/AVSS, separate analog and digital ground planes, and verifying sleep-current budgets before committing to the design. Unlike 32-bit ARM Cortex-M0+ parts, the PIC24F has stronger analog peripherals (12-bit ADC, op-amps, CTMU) and lower sleep current, making it preferable for analog-front-end designs.

This page synthesizes distributor pricing tiers, drop-in alternatives across the PIC24F16KA family, and practical design notes not found in the manufacturer datasheet alone, providing engineering context beyond raw specification repetition.

Drop-in-Alternativen für PIC24F16KA302-I/ML – gleiches Gehäuse, Pin-kompatibel. Bauteile mit anderem Gehäuse, die eine PCB-Überarbeitung erfordern, sind ausgeschlossen.

Varianten dieser Serie

Modelle derselben Serie, die drop-in-kompatibel sind mit PIC24F16KA302-I/ML (gleiche Bauform und Footprint) – abweichend in Package, ADC, Operating Temperature, Core Architecture, Maximum CPU Speed.

Microchip Technology
Package: 28-QFN (6x6 mm) with exposed pad
ADC: 12-bit with CTMU
Microchip Technology
Package: 28-pin VQFN (6x6 mm) with exposed pad
ADC: 12-bit, up to 19 channels, 200 ksps
Core Architecture: Modified Harvard RISC, 24-bit instruction word
Microchip Technology
Package: 44-pin QFN (8x8 mm), industrial temp
ADC: 12-bit, multi-channel
Operating Temperature: -40C to +85C (industrial, "I")
Microchip Technology
ADC: 12-bit
Operating Temperature: -40C to +125C (E = extended)
Maximum CPU Speed: 32 MHz / 16 MIPS
Microchip Technology
Package: 28-QFN (6x6 mm), code ML
Operating Temperature: -40 °C to +85 °C (industrial, "I")
Core Architecture: PIC24F (16-bit modified Harvard)
Microchip Technology
Package: 28-QFN (6x6 mm), ML suffix
Operating Temperature: -40 °C to +85 °C (Industrial)
Maximum CPU Speed: 32 MHz (16 MIPS)

Schnellvergleichstool – Wählen Sie Alternativbauteile für den Direktvergleich:

PIC24F16KA302T-I/ML

✅ Drop-In
Microchip Technology
📦 QFN-28 (6x6)
PIC24F 16-bit modified Harvard · PIC24 XLP eXtreme Low Power · 32 MHz · 16 MIPS · 16 KB (5.5K x 24 instructions) · 2 KB · 512 B · 1.8 V to 3.6 V

✓ Auf Lager

$2.6 / Stück

Datenblatt ansehen →

PIC24F16KA302-I/MQ

✅ Drop-In
📦 QFN-28 (6x6)
same die, /MQ variant possibly MQFP equivalent package, industrial temp

📋 Referenzalternative (nicht im Katalog)

PIC24F16KA301T-I/ML

✅ Drop-In
📦 QFN-28 (6x6)
1.5 KB RAM vs 2 KB RAM (-25%); same 16 KB Flash, same peripherals, same QFN-28 footprint

📋 Referenzalternative (nicht im Katalog)

PIC24F08KM202-I/ML

✅ Drop-In
📦 QFN-28 (6x6)
8 KB Flash vs 16 KB (-50%), 1 KB RAM vs 2 KB (-50%); same QFN-28 footprint, PIC24F family

📋 Referenzalternative (nicht im Katalog)

PIC24F16KA302-I/ML Grenzwerte & elektrische Eigenschaften

Core Architecture PIC24F 16-bit modified Harvard
CPU Speed 32 MHz (16 MIPS)
Program Memory (Flash) 16 KB (5.5K x 24 instructions)
RAM 2 KB
EEPROM 512 B
Versorgungsspannung 1.8 V to 3.6 V
Betriebstemperatur -40C to +85C (Industrial)
I/O Pins 17 (varies by package)
ADC 12-bit, up to 16 channels
Comparators / Op-amps 3 with CTMU
Timers T1-T5 (16-bit and 32-bit)
UART 2
SPI 2
I2C 2
RTCC Yes (Real-Time Clock & Calendar)
Deep-Sleep Brown-Out Reset (DSBOR) Yes
Deep-Sleep Watchdog Timer (DSWDT) Yes
Gehäuse 28-pin QFN (6x6 mm)
Montageart Surface Mount
MSL-Level 3 (168 hours)
RoHS-Status Compliant
Lead-Free Yes

PIC24F16KA302-I/ML Pin-Konfiguration

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
Pin 1 RA0 — GPIO / analog input AN0
Pin 2 RA1 — GPIO / analog input AN1
Pin 3 RA2 — GPIO / analog input AN2
Pin 4 RA3 — GPIO / analog input AN3
Pin 5 RA4 — GPIO / analog input AN4
Pin 6 AVDD — Analog supply voltage
Pin 7 AVSS — Analog ground
Pin 8 RA7 — GPIO / analog input AN7
Pin 9 RA8 — GPIO / analog input AN8
Pin 10 RA9 — GPIO / analog input AN9
Pin 11 VDD — Digital supply voltage
Pin 12 VSS — Digital ground
Pin 13 RB0 — GPIO / analog input AN12
Pin 14 RB1 — GPIO / analog input AN10
Pin 15 RB2 — GPIO / analog input AN8 (alt)
Pin 16 RB3 — GPIO / analog input AN9 (alt)
Pin 17 RB4 — GPIO
Pin 18 RB5 — GPIO
Pin 19 RB6 — GPIO
Pin 20 RB7 — GPIO
Pin 21 RB8 — GPIO
Pin 22 RB9 — GPIO
Pin 23 RC0 — GPIO
Pin 24 RC1 — GPIO
Pin 25 RC2 — GPIO
Pin 26 RC3 — GPIO
Pin 27 RC4 — GPIO
Pin 28 RC5 — GPIO

Typische Anwendungen

PIC24F16KA302-I/ML ist geeignet für 6 Anwendungs: Battery-Powered IoT Sensor Nodes, Capacitive-Touch Human-Machine Interface, Wearable Health Monitors, Remote Controls and Wireless Key Fobs, Energy-Harvesting Devices, Low-Power Industrial Sensor Modules.

🧩

Battery-Powered IoT Sensor Nodes

The PIC24F16KA302-I/ML fits battery-powered IoT sensor nodes because its XLP architecture achieves deep-sleep currents in the nanoamp range, allowing multi-year battery life on a single coin cell. With 16 KB Flash, 2 KB RAM, and 32 MHz CPU, it can run a Zigbee or LoRa stack alongside sensor processing without external memory. The 12-bit ADC supports analog sensors, while CTMU enables capacitive touch or liquid level sensing. Compared to 32-bit Cortex-M0+ MCUs, it offers lower sleep current but smaller memory; compared to PIC24F16KA301, it provides 33% more RAM for buffer-heavy protocols.

📱

Capacitive-Touch Human-Machine Interface

The PIC24F16KA302-I/ML is well-suited for capacitive-touch HMI products like white-goods control panels and portable medical devices because of its integrated CTMU (Charge Time Measurement Unit) and 12-bit ADC. The CTMU measures sub-picofarad capacitance changes, enabling touch buttons, sliders, and proximity sensing without an external analog front-end IC. Microchip's mTouch firmware library is free and supports up to 16 channels. The 2 KB RAM comfortably handles touch-scanning buffers, while the QFN-28 (6x6) package fits compact control-panel PCBs.

💊

Wearable Health Monitors

The PIC24F16KA302-I/ML serves wearable health monitors such as heart-rate bands, and SpO2 peripherals by combining low-power operation with a 12-bit ADC and op-amp peripherals capable of driving photodiode front-ends. Deep-sleep currents in the XLP architecture preserve battery life between measurements, while the RTCC keeps accurate time-of-day with wake-on-alarm. The QFN-28 (6x6) package measures just 6x6 mm, fitting wristbands. Compared to PIC18F variants, KA302 offers more RAM for sensor-fusion algorithms and better ADC performance.

📱

Remote Controls and Wireless Key Fobs

The PIC24F16KA302-I/ML is ideal for remote controls and key fobs because of its nanoamp deep-sleep current and wake-on-interrupt capability, enabling coin-cell battery life measured in years. The 2 UART or 2 SPI peripherals can drive sub-GHz radio transceivers like Microchip's MRF89XA or external LoRa modules. The 16 KB Flash holds simple button-debounce, encryption, and RF protocol stacks. The QFN-28 package fits compact fob enclosures, while the 1.8 V to 3.6 V supply directly supports single-cell LiCoin or 2xAA chemistries.

⚡

Energy-Harvesting Devices

The PIC24F16KA302-I/ML's deep-sleep current under 100 nA, combined with its 1.8 V minimum supply voltage, makes it appropriate for energy-harvesting applications such as solar-powered sensors and kinetic-switch IoT endpoints. The DSBOR (deep-sleep brown-out reset) and DSWDT (deep-sleep watchdog timer) keep the device functional at low VDD levels that would brown-out other MCUs. Combined with the RTCC and wake-on-touch capability, it can run entirely from energy harvested at the touch of a button.

🏭

Low-Power Industrial Sensor Modules

The PIC24F16KA302-I/ML is well-matched to industrial sensor modules such as 4-20 mA loop-powered devices, and Modbus sensor nodes where its 1.8 V to 3.6 V supply, -40C to +85C industrial temperature range, and 12-bit ADC deliver robust measurement with low power budget. The 2 UART supports RS-485 with external transceiver or Modbus RTU directly. CTMU enables liquid level or proximity sensing in industrial tanks. Compared to PIC24EP512 parts, KA302 offers much lower sleep current at the cost of CPU throughput.

Zusammenfassung der empfohlenen Produkte

What is the program memory size of PIC24F16KA302-I/ML?
The PIC24F16KA302-I/ML has 16 KB of Flash program memory (5.5K x 24-bit instructions) according to the Microchip datasheet for the PIC24F16KA302 family. It also integrates 2 KB of SRAM for data and 512 bytes of EEPROM for non-volatile parameter storage. This memory configuration makes the part appropriate for sensor-node and capacitive-touch applications where moderate firmware size is acceptable.
What is the operating voltage range of PIC24F16KA302-I/ML?
The PIC24F16KA302-I/ML operates from 1.8 V to 3.6 V supply voltage, including support for both single-cell Li-ion (3.6 V max) and 2xAA (3.0 V nominal) battery chemistries. According to the Microchip datasheet, the on-chip LDO and brown-out reset are optimized for low-voltage deep-sleep operation. Engineers should ensure peak loads do not collapse VDD below 1.8 V.
How does PIC24F16KA302-I/ML compare to PIC24F16KA301-I/MQ?
The PIC24F16KA302-I/ML adds 2 KB SRAM versus the 1.5 KB SRAM of the PIC24F16KA301 family variant, while keeping the same 16 KB Flash and 32 MHz CPU. Both share the same PIC24F XLP architecture and CTMU/12-bit ADC peripherals. Choose PIC24F16KA302 for buffer-heavy sensor buffering, PIC24F16KA301 for cost-sensitive designs needing less RAM.
Where can I buy PIC24F16KA302-I/ML online?
The PIC24F16KA302-I/ML is available from authorized Microchip distributors including DigiKey (stock number 2651299), Mouser, Arrow, Avnet, and Octopart-aggregated partners. As of 2026-09-29, DigiKey and Mouser both list the part in active stock. Distributors may require minimum order quantities (MOQ) of 1 piece on tape-and-reel packaging.
What is the price of PIC24F16KA302-I/ML in production volumes?
The PIC24F16KA302-I/ML unit price as of 2026-09-29 ranges from approximately 2.95 USD at qty 1 to 1.95 USD at qty 1000 on DigiKey, depending on tape-and-reel quantity breaks. Volume OEM pricing is typically negotiated directly with Microchip via their sales channels. For reference designs in low volumes the 1-piece price is the most representative benchmark.
What is the lead time for PIC24F16KA302-I/ML orders?
Lead time for PIC24F16KA302-I/ML at major distributors is typically 4-6 weeks for factory-direct orders, with in-stock distributor inventory shipping same-day. As of 2026-09-29 both DigiKey and Mouser show the part as factory stock. For volume production contracts, contact Microchip's authorized distributors for a current lead-time quote specific to your quantity requirement.
Is PIC24F16KA302-I/ML the same as PIC24F16KA302-I/SP?
No, the PIC24F16KA302-I/ML and PIC24F16KA302-I/SP differ in package: the /ML suffix denotes the 28-pin QFN (6x6 mm), while the /SP suffix denotes the 28-pin SPDIP (Skinny Plastic DIP) for through-hole prototyping. Same silicon die, same Flash/RAM, same peripherals. They are NOT drop-in compatible because the PCB footprint is completely different.
When should I choose PIC24F16KA302 over PIC24F16KA101?
Choose the PIC24F16KA302 when you need 16 KB Flash and 2 KB SRAM versus the 16 KB Flash and 1.5 KB SRAM of the PIC24F16KA101 (note: KA101 may also have less RAM depending on family). Both share the 32 MHz CPU and XLP low-power architecture. The KA302 variant offers more memory headroom for larger sensor-fusion or communication stacks. Choose KA101 for cost-optimized simple applications.
What is the best drop-in replacement for PIC24F16KA302-I/ML?
The closest drop-in replacement for PIC24F16KA302-I/ML is the PIC24F16KA302-I/SP or PIC24F16KA302T-I/ML from the same Microchip family, which use the same silicon die but in different packages or temperature grades. For pin-compatible QFN-28 same-footprint equivalents, only the /ML suffix variants are drop-in compatible. Cross-manufacturer alternatives require PCB redesign.
Where can I download the PIC24F16KA302-I/ML datasheet PDF?
The PIC24F16KA302 datasheet PDF is available from Microchip's official product page at microchip.com/en-us/product/PIC24F16KA302. The same datasheet covers the entire PIC24F16KA302 / 301 / 300 family and all package variants (/ML) and (/SP) and (/SO) and (/SS). The document is also mirrored on distributor pages like DigiKey and Mouser, but Microchip's site is the authoritative source for the latest revision.
Where do I find the pinout for PIC24F16KA302-I/ML?
The PIC24F16KA302-I/ML pinout is documented in the family datasheet at Microchip's product page for the 28-pin QFN (6x6) package variant. The pin assignment table includes RA0-RB15 GPIO mapping, AVOSD/AVSS analog supply pins, VDD/VSS power pins, and dedicated peripheral pins (SDA, SCL, SDO, SDI, etc.). Refer to the package-specific section rather than the SPDIP pinout.
Does PIC24F16KA302-I/ML support capacitive touch sensing?
Yes, the PIC24F16KA302-I/ML includes the CTMU (Charge Time Measurement Unit) peripheral with 3 comparators and op-amps for capacitive-touch button, slider, wheel, and proximity-sensing applications. According to the Microchip datasheet, the CTMU works with the 12-bit ADC to measure sub-pF capacitance changes. Microchip provides the mTouch library for firmware development on this peripheral.
What is the difference between PIC24F16KA302-I/ML and PIC24F08KM202-I/ML?
Both use the 28-pin QFN (6x6) package, but the PIC24F08KM202-I/ML has only 8 KB Flash and 1 KB RAM versus the 16 KB Flash and 2 KB RAM of the PIC24F16KA302-I/ML. The KM family uses a different low-power architecture (K-series with enhanced peripherals). Choose KM202 for cost-optimized designs with smaller memory needs, KA302 for higher-memory sensor applications.
What are the key specifications of PIC24F16KA302-I/ML that engineers should know?
Engineers evaluating PIC24F16KA302-I/ML should know: 32 MHz / 16 MIPS core, 16 KB Flash / 2 KB RAM / 512 B EEPROM, 1.8 V to 3.6 V supply, 12-bit ADC up to 16 channels, CTMU for capacitive touch, 2 UART, 2 SPI, 2 I2C, RTCC, deep-sleep modes with nanoamp-class current, and 28-pin QFN (6x6) package. According to Microchip's product page, XLP technology achieves industry-leading sleep current for battery applications.

Technische Referenzdaten für PIC24F16KA302-I/ML – Vergleich, Designhinweise und Compliance-Informationen.

Auswahlhilfe

Choose PIC24F16KA302-I/ML when designing battery-powered IoT sensor nodes, capacitive-touch HMIs, or wearable health monitors that demand nanoamp-class deep-sleep current, 16 KB Flash, 2 KB RAM, and CTMU for capacitive sensing. The QFN-28 (6x6) package fits compact PCB layouts where the IC consumes less than 100 mm squared of board area. Choose PIC24F16KA301T-I/ML when you need the same die in tape-and-reel format (production quantity) or for cost-optimized designs accepting 1.5 KB RAM instead of 2 KB. Choose PIC24F08KM202-I/ML when your firmware comfortably fits in 8 KB Flash and 1 KB RAM, achieving lower cost. For applications requiring more than 32 MHz CPU throughput, consider PIC24EP-series parts instead, which offer 70 MIPS but higher sleep current.

Vergleich mit Alternativen

Parameter Dieses Produkt PIC24F16KA302T-I/ML PIC24F16KA301T-I/ML PIC24F08KM202-I/ML
Package QFN-28 (6x6 mm) QFN-28 (6x6 mm) - same QFN-28 (6x6 mm) - same QFN-28 (6x6 mm) - same
Brand Microchip Technology Microchip Technology Microchip Technology Microchip Technology
Program Flash 16 KB 16 KB 16 KB 8 KB (-50%)
RAM 2 KB 2 KB 1.5 KB (-25%) 1 KB (-50%)
CPU Speed 32 MHz / 16 MIPS 32 MHz / 16 MIPS 32 MHz / 16 MIPS 32 MHz / 16 MIPS
Supply Voltage 1.8 V to 3.6 V 1.8 V to 3.6 V 1.8 V to 3.6 V 1.8 V to 3.6 V
Operating Temperature -40C to +85C -40C to +85C -40C to +85C -40C to +85C
ADC 12-bit, 16 channels 12-bit, 16 channels 12-bit, 16 channels 12-bit, 16 channels

Wichtigste Unterscheidungsmerkmale

  • Larger SRAM versus PIC24F16KA301 family (vs. PIC24F16KA301T-I/ML)
  • CTMU capacitive touch peripheral included (vs. PIC24F08KM202-I/ML)
  • Nanoamp-class deep-sleep current (XLP) (vs. PIC24EP128GP204-I/PT)

Designhinweise

Decouple VDD with at least 100 nF ceramic placed within 5 mm of the VDD pin, and add a bulk 10 uF tantalum or ceramic near the package. The deep-sleep current can be affected by leakage through external pull-ups. According to the Microchip datasheet, the XLP deep-sleep mode achieves nanoamp-range current only when all I/O are properly terminated. Pull-downs or pull-ups should use values less than 100 kohm to minimize leakage during sleep.

Separate analog ground (AVSS) and digital ground (VSS) planes with a single-point connection under the IC to avoid digital switching noise corrupting ADC readings. Place the AVSS pin directly to the analog ground island, not the noisy digital return. AVDD should have its own RC filter (10 ohm + 100 nF) to isolate analog supply from digital switching noise. According to the Microchip datasheet, ADC accuracy depends on clean analog supply.

Do not skip the MCLR pull-up resistor when using ICSP programming - a 10 kohm pull-up to VDD is required. When using CTMU for capacitive touch, the sensor traces must be guarded against moisture and other environmental factors. The RTCC requires a 32.768 kHz crystal or external clock input on SOSCO/SOSCI pins, and is not functional without it. Check that your firmware properly initializes the CTMU current source before performing capacitance conversions.

Compliance-Informationen

RoHS
Compliant
REACH
Compliant
AEC-Q100
Not Applicable
Bleifrei
Yes
Halogenfrei
Yes
Konfliktmineralien
Compliant

RoHS compliant per Microchip product page. Industrial temperature grade (-40C to +85C); not AEC-Q100 qualified. Not designed for automotive applications.

Daten verifiziert am: 2026-09-29 – Daten geprüft und kuratiert vom Bauteil-Engineering-Team von XAIPART

Ähnliche Suchanfragen

PIC24F16KA302-I/ML PIC24F16KA302-I/ML datasheet Microchip PIC24F16KA302 16KB Flash 32MHz MCU QFN-28 QFN-28 PIC24F microcontroller PIC24F XLP low power MCU PIC24F16KA302 vs PIC24F16KA301 PIC24F16KA302-I/ML drop-in replacement PIC24F16KA302-I/ML buy price what is the flash memory of PIC24F16KA302 PIC24F16KA302 capacitive touch CTMU low power MCU battery powered sensor

Verwandte Bauteile & Begriffe

Microchip Technology PIC24F16KA302 PIC24F16KA302-I/ML PIC24F16KA302T-I/ML PIC24F16KA301 PIC24F08KM202 PIC24F16-bit MCU XLP technology CTMU QFN-28 PIC24F family 12-bit ADC MIPS core deep-sleep brown-out reset RTCC battery-powered IoT capacitive touch sensing wearable health monitor RoHS Flash microcontroller MPLAB XC16 ICSP low dropout regulator
Schnellangebot-RFQ
Geben Sie alle Details an – unser Vertriebsteam antwortet innerhalb von 24 Stunden
Teilenummer Hersteller Gehäuse QTY Zielpreis Gesamt
Gesamt: $0.00 USD
✓
Angebotsanfrage gesendet!

Wir antworten innerhalb von 24 Stunden auf Ihre E-Mail

1
RFQ gesendet
2
Angebot erhalten
3
Bestellung aufgegeben
4
Zahlung
5
Versandt
6
Geliefert
RFQ-Details ansehen