CY7C25632KV18
CY7C25652KV18
72-Mbit QDR® II+ SRAM Four-Word Burst Architecture
(2.5 Cycle Read Latency) with ODT
72-Mbit QDR® II+ SRAM Four-Word Burst Architecture (2.5 Cycle Read Latency) with ODT
■ JTAG 1149.1 compatible test access port
Features
■ Phase-locked loop (PLL) for accurate data placement
■ Separate independent read and write data ports
❐ Supports concurrent transactions
Configurations
■ 550 MHz clock for high bandwidth
With Read Cycle Latency of 2.5 cycles
CY7C25632KV18 – 4 M × 18
■ Four-word burst for reducing address bus frequency
■ Double data rate (DDR) interfaces on both read and write ports
(data transferred at 1100 MHz) at 550 MHz
CY7C25652KV18 – 2 M × 36
Functional Description
■ Available in 2.5 clock cycle latency
■ Two input clocks (K and K) for precise DDR timing
❐ SRAM uses rising edges only
The CY7C25632KV18 and CY7C25652KV18 are 1.8 V
Synchronous Pipelined SRAMs, equipped with QDR II+
architecture. Similar to QDR II architecture, QDR II+ architecture
consists of two separate ports: the read port and the write port to
access the memory array. The read port has dedicated data
outputs to support read operations and the write port has
dedicated data inputs to support write operations. QDR II+
architecture has separate data inputs and data outputs to
completely eliminate the need to “turn-around” the data bus that
exists with common I/O devices. Each port is accessed through
a common address bus. Addresses for read and write addresses
are latched on alternate rising edges of the input (K) clock.
Accesses to the QDR II+ read and write ports are completely
independent of one another. To maximize data throughput, both
read and write ports are equipped with DDR interfaces. Each
address location is associated with four 18-bit words
(CY7C25632KV18), or 36-bit words (CY7C25652KV18) that
burst sequentially into or out of the device. Because data is
transferred into and out of the device on every rising edge of both
input clocks (K and K), memory bandwidth is maximized while
simplifying system design by eliminating bus “turn-arounds”.
■ Echo clocks (CQ and CQ) simplify data capture in high-speed
systems
■ Data valid pin (QVLD) to indicate valid data on the output
■ On-die termination (ODT) feature
❐ Supported for D[x:0], BWS[x:0], and K/K inputs
■ Single multiplexed address input bus latches address inputs
for read and write ports
■ Separate port selects for depth expansion
■ Synchronous internally self-timed writes
■ QDR® II+ operates with 2.5 cycle read latency when DOFF is
asserted HIGH
■ OperatessimilartoQDRIdevicewith1cyclereadlatencywhen
DOFF is asserted LOW
■ Available in × 18, and × 36 configurations
These devices have an On-Die Termination feature supported
for D[x:0], BWS[x:0], and K/K inputs, which helps eliminate
external termination resistors, reduce cost, reduce board area,
and simplify board routing.
■ Full data coherency, providing most current data
[1]
■ Core VDD = 1.8 V ± 0.1 V; I/O VDDQ = 1.4 V to VDD
❐ Supports both 1.5 V and 1.8 V I/O supply
Depth expansion is accomplished with port selects, which
enables each port to operate independently.
■ HSTL inputs and variable drive HSTL output buffers
■ Available in 165-ball FBGA package (13 × 15 × 1.4 mm)
■ Offered in both Pb-free and non Pb-free packages
All synchronous inputs pass through input registers controlled by
the K or K input clocks. All data outputs pass through output
registers controlled by the K or K input clocks. Writes are
conducted with on-chip synchronous self-timed write circuitry.
Selection Guide
Description
Maximum Operating Frequency
Maximum Operating Current
550 MHz
550
500 MHz
500
450 MHz
450
400 MHz Unit
400
710
MHz
mA
× 18
× 36
920
850
780
1310
1210
1100
1000
Note
1. The Cypress QDR II+ devices surpass the QDR consortium specification and can support V
= 1.4 V to V
.
DD
DDQ
Cypress Semiconductor Corporation
Document Number: 001-66482 Rev. *C
•
198 Champion Court
•
San Jose, CA 95134-1709
•
408-943-2600
Revised September 21, 2012