MG74P

Oki's 0.25m Application-Specific Integrated Circuit (ASIC) products are available in both Sea Of Gates (SOG) and Customer Structured Array (CSA) architectures. Both the SOG-based MG115P series and the CSA-based MG75P series use a five-layer metal process on 0.25m drawn (0.18m L-effective) CMOS technology. The SOG MG113P/114P series uses the same SOG base-array architecture as the MG115P series, but offers four and three metal layers, respectively. The MG73P/74P CSA series uses three and four metal layers, respectively. The semiconductor process is adapted from Oki's production-proven 64- Mbit DRAM manufacturing process.

The 0.25m family provides significant performance, density, and power improvement over previous 0.30 and 0.35m technologies. An innovative 4-transistor cell structure, licensed from In-Chip Systems, Inc., provides 30 to 50% less power and 30 to 50% more usable gates than traditional cell designs. The Oki 0.25m family operates using 2.5-V V _DD core with optimized 3-V I/O buffers. The 3-, 4-, and 5-layer metal MG113P/114P/115P SOG series contains 4 array bases, offering up to 588 I/O pads and over 2.4M raw gates. The 3-, 4-, and 5-layer metal MG73P/74P/75P CSA series contains 21 array bases, offering up to 868 I/O pads and over 5.4M raw gates. These SOG and CSA array sizes are designed to fit the most popular quad flat pack (QFP), low profile QFPs (LQFPs), thin QFPs (TQFPs), and plastic ball grid array (PBGA) packages.

The MG113P/114P/115P series SOG architecture allows rapid prototyping turnaround times (TATs),additionally offering the most cost-effective solution for pad-limited circuits (particularly the 3-layer metal MG113P series). The 3-layer-metal MG73P, 4-layer-metal MG74P and 5-layer-metal MG75P CSA  series contains 21 array bases, offering a wider span of gate and I/O counts than the SOG series. Oki uses the Artisan Components memory compiler which provides high performance, embedded synchronous single- and dual-port RAM macrocells for CSA designs. As such, the MG73P/74P/75P series is suited to memory-intensive ASICs and high-volume designs where fine tuning of package size produces significant cost or real-estate savings.