
36
Data Device Corporation
www.ddc-web.com
BU-6174X/6184X/6186X
F-10/02-300
CLOCK IN
VALID
t7
t3
t8
t11
t13
t15
VALID
t10
t4
t9
t12
t19
;
;;
VALID
t16
t17
SELECT
(Note 2,7)
(Note 2)
(Note 3,4,7)
(Note 4,5)
STRBD
MEM/REG
RD/WR
IOEN
(Note 2,6)
(Note 6)
(Note 7,8,9)
READYD
A15-A0
D15-D0
;
;;
;;;
;;
t5
t1
t2
t6
t14
t18
FIGURE 13. CPU READING RAM / REGISTER (16-BIT BUFFERED, NONZERO WAIT)
NOTES:
1.
For the 16-bit buffered nonzero wait configuration, TRANSPARENT/BUFFERED must be connected to logic "0". ZERO_WAIT and DTREQ / 16/8
must be connected to logic "1". The inputs TRIGGER_SEL and MSB/LSB may be connected to either +5V or ground.
2.
SELECT and STRBD may be tied together. IOEN goes low on the first rising CLK edge when SELECT STRBD is sampled low (satisfying t1)
and the Enhanced Mini-ACE/-ACE's protocol/memory management logic is not accessing the internal RAM. When this occurs, IOEN goes low,
starting the transfer cycle. After IOEN goes low, SELECT may be released high.
3.
MEM/REG must be presented high for memory access, low for register access.
4.
MEM/REG and RD/WR are buffered transparently until the first falling edge of CLK after IOEN goes low. After this CLK edge, MEM/REG and
RD/WR become latched internally.
5.
The logic sense for RD/WR in the diagram assumes that POLARITY_SEL is connected to logic "1". If POLARITY_SEL is connected to logic "0",
RD/WR must be asserted low to read.
6.
The timing for IOEN, READYD and D15-D0 assumes a 50 pf load. For loading above 50 pf, the validity of IOEN, READYD, and D15-D0 is delayed
by an additional 0.14 ns/pf typ, 0.28 ns/pf max.
7.
The timing for A15-A0, MEM/REG and SELECT assumes that ADDR-LAT is connected to logic "1". Refer to Address Latch timing for additional
details.
8.
The address bus A15-A0 is internally buffered transparently until the first rising edge of CLK after IOEN goes low. After this CLK edge, A15-A0
become latched internally.
9.
to the system clock. When SELECT and STRBD do not meet the setup time of t1, but occur during the setup window of an internal flip-flop, an
additional clock cycle will be inserted between the falling clock edge that latches MEM/REG and RD/WR and the rising clock edge that latches
the Address (A15-A0). When this occurs, the delay from IOEN falling to READYD falling (t11) increases by one clock cycle and the address hold
time (t10) must be increased be one clock cycle.