The Abaco Systems VMIPCI-5576-001 is a PCI-based fiber-optic reflective memory board that enables transparent, zero-overhead data sharing across up to 256 networked nodes in distributed computing systems. Data written to local SDRAM is automatically broadcast to corresponding memory addresses on all connected nodes over a daisy-chained fiber-optic ring, eliminating I/O bottlenecks in real-time control and data acquisition environments.

Technical Specifications

Memory & Data Transfer
• Configurable on-board SDRAM: 256 Kbyte, 512 Kbyte, or 1 Mbyte
• Memory access widths: D32:D16:D8
• Network data rate: 6.2 Mbyte/s (standard) or 3.2 Mbyte/s (redundant transfer mode)
• PCI target burst rate: 44 Mbyte/s
• PCI master DMA rate: 132 Mbyte/s
• Serial baud rate: 170 Mbaud

Network Architecture
• Maximum nodes: 256
• Maximum inter-node distance: 2,000 m (6,560 ft)
• Topology: Daisy-chained fiber-optic ring
• Fiber type: Multimode, 62.5 µm core
• Connector: ST-style bayonet
• Transmitter wavelength: 1,300 nm
• Attenuation range: 0.5 dB to 9 dB between nodes

Interrupt Architecture
• Three user-definable interrupts per node
• Any node can trigger interrupts to any other node or broadcast to all nodes
• Interrupts follow data delivery—acknowledgment only after reception confirmed
• Masked at power-up; enabled under software control

– Key Features

• Transmit and receive FIFOs buffer peak data rates, reducing CPU and bus overhead
• Error detection via TAXI chipset and parity encoding with redundant transmission support
• Configurable endian conversion for multi-architecture networks
• No processor overhead required for network operation

– Typical Applications

• Real-time distributed control systems
• Multi-node data acquisition networks
• Mission-critical applications requiring deterministic latency
• High-speed inter-system synchronization

– Compatibility & Integration

Networks with VMIC 5576 family boards (VMIVME-5576, VMIMBI-5576) and supports mixed-architecture systems combining PCI, VME, and Multibus platforms.