Features
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Fast Access Time: 30/45 ns
Wide Temperature Range: -55°C to +125°C
Separate Upper Byte and Lower Byte Control for Multiplexed Bus Compatibility
Expandable Data Bus to 32 bits or More Using Master/Slave Chip Select When Using
More Than One Device
On-chip Arbitration Logic
Versatile Pin Select for Master or Slave:
– M/S = H for Busy Output Flag On Master
– M/S = L for Busy Input Flag On Slave
INT Flag for Port to Port Communication
Full Hardware Support of Semaphore Signaling Between Ports
Fully Asynchronous Operation From Either Port
Battery Back-up Operation: 2V Data Retention
TTL Compatible
Single 5V + 10% Power Supply
No Single Event Latch-up below a LET Threshold of 80 MeV/mg/cm
2
Tested up to a Total Dose of 30 krads (Si) according to MIL STD 883 Method 1019
Quality grades: QML Q and V with SMD 5962-91617 and ESCC with 9301/050
Rad. Tolerant
High Speed
8 Kb x 16
Dual Port RAM
M67025E
Introduction
The M67025E is a very low power CMOS dual port static RAM organized as 8192 bit
×
16. The product is designed to be used as a stand-alone 16-bit dual port RAM or as
a combination MASTER/SLAVE dual port for 32-bit or more width systems. The Atmel
MASTER/SLAVE dual port approach in memory system applications results in full
speed, error free operation without the need of an additional discrete logic.
Master and slave devices provide two independent ports with separate control,
address and I/O pins that permit independent, asynchronous access for reads and
writes to any location in the memory. An automatic power down feature controlled by
CS permits the on-chip circuitry of each port in order to enter a very low stand by
power mode.
Using an array of eight transistors (8T) memory cell, the M67025E combines an
extremely low standby supply current (typ = 1.0 µA) with a fast access time at 30 ns
over the full temperature range. All versions offer battery backup data retention capa-
bility with a typical power consumption at less than 5 µW.
For military/space applications that demand superior levels of performance and reli-
ability the M67025E is processed according to the methods of the latest revision of the
MIL PRF 38635 (Q and V) and/or ESCC 9000.
4146N–AERO–04/07
Block Diagram
Notes:
1. (MASTER): BUSY is output. (SLAVE): BUSY is input.
2. LB = Lower Byte
UB = Upper Byte
2
M67025E
4146N–AERO–04/07
M67025E
Pin Configuration
Pin Description
Left Port
CS
L
R/W
L
OE
L
A
0L – 12L
I/O
0L – 15L
SEM
L
UB
L
LB
L
INT
L
BUSY
L
Right Port
CS
R
R/W
R
OE
R
A
0R – 12R
I/O
0R – 15R
SEM
R
UB
R
LB
R
INT
R
BUSY
R
M/S
Vcc
GND
Names
Chip select
Read/Write Enable
Output Enable
Address
Data Input/Output
Semaphore Enable
Upper Byte Select
Lower Byte Select
Interrupt Flag
Busy Flag
Master or Slave Select
Power
Ground
3
4146N–AERO–04/07
Functional
Description
The M67025E has two ports with separate control, address and I/O pins that permit independent
read/write access to any memory location. These devices have an automatic power-down fea-
ture controlled by CS. CS controls on-chip power-down circuitry which causes the port
concerned to go into stand-by mode when not selected (CS high). When a port is selected
access to the full memory array is permitted. Each port has its own Output Enable control (OE).
In read mode, the port’s OE turns the Output drivers on when set LOW. Non-conflicting
READ/WRITE conditions are illustrated in Table 1.
The interrupt flag (INT) allows communication between ports or systems. If the user chooses to
use the interrupt function, a memory location (mail box or message center) is assigned to each
port. The left port interrupt flag (INT
L
) is set when the right port writes to memory location 1FFE
(HEX). The left port clears the interrupt by reading address location 1FFE. Similarly, the right
port interrupt flag (INT
R
) is set when the left port writes to memory location 1FFF (HEX), and the
right port must read memory location 1FFF in order to clear the interrupt flag (INT
R
). The 16-bit
message at 1FFE or 1FFF is user-defined. If the interrupt function is not used, address locations
1FFE and 1FFF are not reserved for mail boxes but become part of the RAM. See Table 3 for
the interrupt function.
Arbitration Logic
The arbitration logic will resolve an address match or a chip select match down to a minimum of
5 ns determine which port has access. In all cases, an active BUSY flag will be set for the inhib-
ited port.
The BUSY flags are required when both ports attempt to access the same location simulta-
neously. Should this conflict arise, on-chip arbitration logic will determine which port has access
and set the BUSY flag for the inhibited port. BUSY is set at speeds that allow the processor to
hold the operation with its associated address and data. It should be noted that the operation is
invalid for the port for which BUSY is set LOW. The inhibited port will be given access when
BUSY goes inactive.
A conflict will occur when both left and right ports are active and the two addresses coincide.
The on-chip arbitration determines access in these circumstances. Two modes of arbitration are
provided: (1) if the addresses match and are valid before CS on-chip control logic arbitrates
between CS
L
and CS
R
for access; or (2) if the CS are low before an address match, on-chip con-
trol logic arbitrates between the left and right addresses for access (refer to Table 4). The
inhibited port’s BUSY flag is set and will reset when the port granted access completes its oper-
ation in both arbitration modes.
Data Bus Width
Expansion
Expanding the data bus width to 32 or more bits in a dual-port RAM system means that several
chips may be active simultaneously. If every chips has a hardware arbitrator, and the addresses
for each arrive at the same time one chip may activate in L BUSY signal while another activates
its R BUSY signal. Both sides are now busy and the CPUs will wait indefinitely for their port to
become free.
To overcome this “Busy Lock-Out’ problem, Atmel has developed a MASTER/SLAVE system
which uses a single hardware arbitrator located on the MASTER. The SLAVE has BUSY inputs
which allow direct interface to the MASTER with no external components, giving a speed advantage
over other systems.
When dual-port RAMs are expanded in width, the SLAVE RAMs must be prevented from writing
until after the BUSY input has settled. Otherwise, the SLAVE chip may begin a write cycle during a con-
flict situation. Conversely, the write pulse must extend a hold time beyond BUSY to ensure that a write
cycle occurs once the conflict is resolved. This timing is inherent in all dual-port memory systems where
more than one chip is active at the same time.
4
M67025E
4146N–AERO–04/07
M67025E
The write pulse to the SLAVE must be inhibited by the MASTER’s maximum arbitration time. If a
conflict then occurs, the write to the SLAVE will be inhibited because of the MASTER’s BUSY
signal.
Semaphore Logic
The M67025E is an extremely fast dual-port 4 Kb
×
16 CMOS static RAM with an additional loca-
tions dedicated to binary semaphore flags. These flags allow either of the processors on the left or right
side of the dual-port RAM to claim priority over the other for functions defined by the system software.
For example, the semaphore flag can be used by one processor to inhibit the other from accessing a
portion of the dual-port RAM or any other shared resource.
The dual-port RAM has a fast access time, and the two ports are completely independent of
each another. This means that the activity on the left port cannot slow the access time of the
right port. The ports are identical in function to standard CMOS static RAMs and can be read
from, or written to, at the same time with the only possible conflict arising from simultaneous writ-
ing to, or a simultaneous READ/WRITE operation on, a non-semaphore location. Semaphores
are protected against such ambiguous situations and may be used by the system program to
prevent conflicts in the non-semaphore segment of the dual-port RAM. The devices have an
automatic power-down feature controlled by CS, the dual-port RAM select and SEM, the semaphore
enable. The CS and SEM pins control on-chip-power-down circuitry that permits the port concerned to
go into stand-by mode when not selected. This conditions is shown in Table 1 where CS and SEM are
both high.
Systems best able to exploit the M67025E are based around multiple processors or controllers
and are typically very high-speed, software controlled or software-intensive systems. These sys-
tems can benefit from the performance enhancement offered by the M67025 hardware
semaphores, which provide a lock-out mechanism without the need for complex programming.
Software handshaking between processors offers the maximum level of system flexibility by per-
mitting shared resources to be allocated in varying configurations. The M67025E does not use
its semaphore flags to control any resources through hardware, thus allowing the system
designer total flexibility in system architecture.
An advantage of using semaphores rather than the more usual methods of hardware arbitration
is that neither processor ever incurs wait states. This can prove to be a considerable advantage
in very high speed systems.
How The
Semaphore Flags
Work
The semaphore logic is a set of eight latches independent of the dual-port RAM. These latches
can be used to pass a flag or token, from one port to the other to indicate that a shared resource
is in use. The semaphore provides the hardware context for the “Token Passing Allocation’
method of use assignment. This method uses the state of a semaphore latch as a token indicat-
ing that a shared resource is in use. If the left processor needs to use a resource, it requests the
token by setting the latch. The processor then verifies that the latch has been set by reading it. If
the latch has been set the processor assumes control over the shared resource. If the latch has
not been set, the left processor has established that the right processor had set the latch first,
has the token and is using the shared resource. The left processor may then either repeatedly
query the status of the semaphore, or abandon its request for the token and perform another
operation whilst occasionally attempting to gain control of the token through a set and test oper-
ation. Once the right side has relinquished the token the left side will be able to take control of
the shared resource.
The semaphore flags are active low. A token is requested by writing a zero to a semaphore
latch, and is relinquished again when the same side writes a one to the latch.
The eight semaphore flags are located in a separate memory space from the dual-port RAM in
the M67025E. The address space is accessed by placing a low input on the SEM pin (which acts
as a chip select for the semaphore flags) and using the other control pins (address, OE and R/W) as nor-
5
4146N–AERO–04/07
京公网安备 11010802033920号
EEWORLD
