nrrL3325.00
08/06/2001
IBM Corporation. All rights reserved.
Use is further subject to the provisions at the end of this document.
Page 1 of 24
IBM0436A11NLAA
IBM0418A11NLAA
Preliminary
32Kx36 & 64Kx18 SRAM
Features
32Kx36 or 64Kx18 organizations
0.25 Micron CMOS technology
Synchronous Pipeline Mode of Operation with
Self-Timed Late Write
Differential PECL Clocks or 2.5V LVTTL swing
with one clock tied to V
DDQ
/2
+3.3V Power Supply, Ground, 2.5V V
DDQ
2.5V LVTTL Input and Output levels
Registered Addresses, Write Enables, Synchro-
nous Select, and Data Ins
Registered Outputs
30 Ohm Drivers
Common I/O
Asynchronous Output Enable
Synchronous Power Down Input
Boundary Scan using limited set of JTAG
1149.1 functions
Byte Write Capability and Global Write Enable
7 x 17 Bump Ball Grid Array Package with
SRAM JEDEC Standard Pinout and Boundary
SCAN Order
Description
IBM0436A11NLAA and IBM0418A11NLAA are 1Mb
Synchronous Pipeline Mode, high-performance
CMOS Static Random Access Memories (SRAM).
These SRAMs are versatile, have a wide input/out-
put (I/O) interface, and can achieve cycle times as
short as 3.0ns. Differential K clocks are used to ini-
tiate the read/write operation; all internal operations
are self-timed. At the rising edge of the K clock, all
address, write-enable, sync select, and data input
signals are registered internally. Data outputs are
updated from output registers off the next rising
edge of the K clock. An internal write buffer allows
write data to follow one cycle after addresses and
controls. The device is operated with a single +3.3V
power supply and is compatible with 2.5V LVTTL I/O
interfaces.
.
IBM0436A11NLAA
IBM0418A11NLAA
32Kx36 & 64Kx18 SRAM
Preliminary
IBM Corporation. All rights reserved.
Use is further subject to the provisions at the end of this document.
Page 2 of 24
nrrL3325.00
08/06/2001
x36 BGA Pinout
(Top View)
1
2
3
4
5
6
7
A
V
DDQ
SA
SA
NC
SA
SA
V
DDQ
B
NC
NC
NC
NC
NC
NC
NC
C
NC
SA
SA
V
DD
SA
SA
NC
D
DQ19
DQ18
V
SS
NC
V
SS
DQ9
DQ10
E
DQ22
DQ20
V
SS
SS
V
SS
DQ11
DQb13
F
V
DDQ
DQ21
V
SS
G
V
SS
DQ12
V
DDQ
G
DQ24
DQ23
SBWc
NC
SBWb
DQ14
DQb15
H
DQ25
DQ26
V
SS
NC
V
SS
DQ17
DQb16
J
V
DDQ
V
DD
NC
V
DD
NC
V
DD
V
DDQ
K
DQ34
DQ35
V
SS
K
V
SS
DQ8
DQ7
L
DQ33
DQ32
SBWd
K
SBWa
DQ5
DQ6
M
V
DDQ
DQ30
V
SS
SW
V
SS
DQ3
V
DDQ
N
DQ31
DQ29
V
SS
SA
V
SS
DQ2
DQ4
P
DQ28
DQ27
V
SS
SA
V
SS
DQ0
DQ1
R
NC
SA
M1*
V
DD
M2*
SA
NC
T
NC
NC
SA
SA
SA
NC
ZZ
U
V
DDQ
TMS
TDI
TCK
TDO
NC
V
DDQ
* M1 and M2 are clock mode pins. For this application, M1 and M2 need to connect to V
SS
and V
DD
,
respectively.
x18 BGA Pinout
(Top View)
1
2
3
4
5
6
7
A
V
DDQ
SA
SA
NC
SA
SA
V
DDQ
B
NC
NC
NC
NC
NC
NC
NC
C
NC
SA
SA
V
DD
SA
SA
NC
D
DQ14
NC
V
SS
NC
V
SS
DQ0
NC
E
NC
DQ15
V
SS
SS
V
SS
NC
DQ1
F
V
DDQ
NC
V
SS
G
V
SS
DQ2
V
DDQ
G
NC
DQ16
SBWb
NC
NC
NC
DQ3
H
DQ17
NC
V
SS
NC
V
SS
DQ4
NC
J
V
DDQ
V
DD
NC
V
DD
NC
V
DD
V
DDQ
K
NC
DQ13
V
SS
K
V
SS
NC
DQ8
L
DQ12
NC
NC
K
SBWa
DQ7
NC
M
V
DDQ
DQ10
V
SS
SW
V
SS
NC
V
DDQ
N
DQ11
NC
V
SS
SA
V
SS
DQ6
NC
P
NC
DQ9
V
SS
SA
V
SS
NC
DQ5
R
NC
SA
M1
V
DD
M2
SA
NC
T
NC
SA
SA
NC
SA
SA
ZZ
U
V
DDQ
TMS
TDI
TCK
TDO
NC
V
DDQ
* M1 and M2 are clock mode pins. For this application, M1 and M2 need to connect to V
SS
and V
DD
, respectively.
IBM0436A11NLAA
IBM0418A11NLAA
Preliminary
32Kx36 & 64Kx18 SRAM
nrrL3325.00
08/06/2001
IBM Corporation. All rights reserved.
Use is further subject to the provisions at the end of this document.
Page 3 of 24
Pin Description
SA
Address Input
16 Addresses for 512Kx18
15 Addresses for 256Kx36
TDO
IEEE 1149.1 Test Output (LVTTL level)
DQ0-DQ35
Data I/O
DQ0-DQ17 for 512K x 18
DQ0-DQ35 for 256K x 36
G
Asynchronous Output Enable
K, K
Differential Input Register Clocks
SS
Synchronous Select
SW
Write Enable, Global
M1, M2
Clock Mode Inputs. Selects Single or Dual Clock
Operation.
SBWa
Write Enable, Byte a (DQ0-DQ8)
V
DD
Power Supply (+3.3V)
SBWb
Write Enable, Byte b (DQ9-DQ17)
V
SS
Ground
SBWc
Write Enable, Byte c (DQ18-DQ26)
V
DDQ
Output Power Supply
SBWd
Write Enable, Byte d (DQ27-DQ35)
ZZ
Synchronous Sleep Mode
TMS, TDI, TCK
IEEE
1149.1 Test Inputs (LVTTL levels)
NC
No Connect
Ordering Information
Part Number
Organization
Speed
Leads
IBM0418A11NLAA - 3
64Kx18
1.8ns Access / 3.0ns Cycle
7 x 17 BGA
IBM0418A11NLAA - 3F
64Kx18
2.0ns Access / 3.3ns Cycle
7 x 17 BGA
IBM0418A11NLAA - 3N
64Kx18
2.0ns Access / 3.7ns Cycle
7 x 17 BGA
IBM0418A11NLAA - 4
64Kx18
2.25ns Access / 4.0ns Cycle
7 x 17 BGA
IBM0418A11NLAA - 5
64Kx18
2.5ns Access / 5.0ns Cycle
7 x 17 BGA
IBM0436A11NLAA - 3
32Kx36
1.8ns Access / 3.0ns Cycle
7 x 17 BGA
IBM0436A11NLAA - 3F
32Kx36
2.0ns Access / 3.3ns Cycle
7 x 17 BGA
IBM0436A11NLAA - 3N
32Kx36
2.0ns Access / 3.7ns Cycle
7 x 17 BGA
IBM0436A11NLAA - 4
32Kx36
2.25ns Access / 4.0ns Cycle
7 x 17 BGA
IBM0436A11NLAA - 5
32Kx36
2.5ns Access / 5.0ns Cycle
7 x 17 BGA
IBM0436A11NLAA
IBM0418A11NLAA
32Kx36 & 64Kx18 SRAM
Preliminary
IBM Corporation. All rights reserved.
Use is further subject to the provisions at the end of this document.
Page 4 of 24
nrrL3325.00
08/06/2001
Block Diagram
SBW
Ro
w De
co
d
e
Col Decode
Read/Wr Amp
DOC_Array0
SA0-SA18
K
ZZ
G
SW
SS
DQ0-DQ35
REG
REG
SB
W
2:
1
M
U
X
DOC_MUX0
WRI
T
E1
A
DD REG
WRI
T
E
0
ADD REG
RE
AD
AD
D REG
REA
D
WRI
T
E
MA
TC
H
MA
TC
H
1
LATCH
LATCH0
WR
_
B
U
F
1
WR
_
B
U
F
0
2:1 MUX
DOC_MUX1
2:1 MUX
DOC_MUX2
SB
W0
SW0
SW1
REG
REG
DOC_
DOUT0
REG
REG
SS1
SS0
IBM0436A11NLAA
IBM0418A11NLAA
Preliminary
32Kx36 & 64Kx18 SRAM
nrrL3325.00
08/06/2001
IBM Corporation. All rights reserved.
Use is further subject to the provisions at the end of this document.
Page 5 of 24
SRAM Features
Late Write
The Late Write function allows for write data to be registered one cycle after addresses and controls. This fea-
ture eliminates one bus-turnaround cycle, necessary when going from a read to a write operation. Late Write
is accomplished by buffering write addresses and data so that the write operation occurs during the next write
cycle. When a read cycle occurs after a write cycle, the address and write data information are stored tempo-
rarily in holding registers. During the first write cycle preceded by a read cycle, the SRAM array is updated
with address and data from the holding registers. Read cycle addresses are monitored to determine if read
data is to be supplied from the SRAM array or the write buffer. The bypassing of the SRAM array occurs on a
byte-by-byte basis. When only one byte is written during a write cycle, read data from the last written address
has new byte data from the write buffer and remaining bytes from the SRAM array.
Mode Control
Mode control pins M1 and M2 are used to select four different JEDEC-standard read protocols. This SRAM
supports single clock, pipeline operation (M1 = V
SS
, M2 = V
DD
). This datasheet describes single clock pipe-
line functionality only. Mode control inputs must be set at power up and must not change during SRAM oper-
ation. This SRAM is tested only in the pipeline mode.
Sleep Mode
The sleep mode is enabled by switching the synchronous signal ZZ High. When the SRAM is in the sleep
mode, the outputs go to a High-Z state and the SRAM draws standby current. SRAM data is preserved and a
recovery time (t
ZZR
) is required before the SRAM resumes normal operation.
Power-Up Requirements
To ensure the optimum internally regulated supply voltage, the SRAM requires 4
s of power-up time after
V
DD
reaches its operating range.
Power-Up and Power-Down Sequencing
The power supplies must be powered up in the following order: V
DD
, V
DDQ
and Inputs. The power-down
sequence must be in the reverse order. V
DDQ
may not exceed V
DD
by more than 0.6V.
PDF 
ZIP