Euresys Coaxlink CXP-12 Frame Grabbers — CoaXPress 2.0 Acquisition for High-Speed Vision

Specifications

Overview

CoaXPress is the interface machine-vision integrators choose when a GigE Vision camera cannot carry the data rate. Euresys' Coaxlink series is the reference implementation of that interface on PCIe, and the CXP-12 generation carries the standard to its current ceiling: 2.0 / 2.1 compliant links at 12.5 Gbps each, up to eight connections on one card depending on model.

The design decision that most simplifies a system is PoCXP. Power over CoaXPress delivers camera power down the same coax that carries data and trigger, so a multi-camera station does not need a separate power run per head, and camera synchronisation and triggering are handled over the same single cable. Connectors are 75-ohm micro-BNC (HD-BNC), supplied with both standard and low-profile brackets.

For high-throughput lines the Coaxlink Quad CXP-12 adds genuine on-card processing rather than functioning only as a transport. Each data path handles 2.5 GB/s of CoaXPress bandwidth and feeds a 250 MB FIFO buffer, and the pixel processor handles 8-, 10-, 12- and 14-bit data with selectable alignment into 16-bit containers — all at up to 250 Megapixels per second.

Where cable runs make copper impractical, the Coaxlink CXP-12 to QSFP+ Converter bridges CoaXPress-over-Fiber, accepting four device connections and carrying the same host-side interface. Across the family Euresys supplies up to 20 digital I/O lines, motion-encoder decoding on selected differential pairs, and the Memento Event Logging Tool for diagnosing acquisition problems in production.

QS Compute supplies Coaxlink frame grabbers, the CXP-12 to QSFP+ converter and matched camera cabling as a single machine-vision acquisition quote.

Key Benefits

12.5 Gbps per connection: the highest CoaXPress link speed, for high-resolution and high-frame-rate cameras. PoCXP camera power: one coax per camera carries data, trigger and power. Up to 8 connections per card: multi-head stations without multiple host cards. On-card image processing: 250 MPixels/s and 250 MB buffering per path offloads the host CPU.

Applications

High-speed machine vision inspection on production lines, semiconductor wafer and PCB inspection, flat-panel display and print inspection, high-resolution area-scan multi-camera stations, motion-synchronised acquisition with encoder feedback, line-scan and high-frame-rate imaging, and fiber-linked remote camera installations in large facilities.

Related Products