Copyright 2008 by Dr.K.K.

Thyagharajan 1
3 3. . I In nt te er rf fa ac ci in ng g M Me em mo or ry y w wi it th h 8 80 08 85 5
A microprocessor will have its own address space for accessing physical memory.
The memory locations that are directly addressed by the microprocessor is called physical
memory space. For example a microprocessor like 8085 has 16 address lines, and it can
access a physical memory space of 64K starting from 0000H to FFFFH as shown in
Table 3.1. The process of interfacing memories to microprocessor and allocating address
to each memory location is called memory mapping. The complete address space may be
considered as a single memory block. But practically, when ICs are used as memory
devices, instead of a single IC (Integrated Chip), few devices that fit into the address
space will be used. This is due to the fact a microprocessor based system requires at least
one ROM/EPROM and a RAM. For example instead of using a memory device of size
64KB (Kilo Bytes), we can use 8 memory devices with a capacity of 8KB each. This will
reduce the chip replacement cost while servicing the microprocessor-based system. When
memory blocks are used, to access all the locations in a block n address lines are needed
so that

2
n
= 8 K
= 2
3
K
= 2
3
x 2
10

= 2
13

Therefore n = 13

3 3 . . 1 1 A Ad dd d r r e e s s s s D De e c c o o d di i n ng g a a n nd d M Me e m mo o r r y y M Ma a p pp pi i n n g g

Memory address decoding is nothing but to assign an address for each location in
the memory chip. The data stored in the memory is accessed by specifying its address.
Memory address can be decoded in two ways
i) Absolute or Fully decoding and
ii) Linear Select or Partial decoding

There are many advantages in absolute address decoding. They are
i) Each memory location has only one address, there is no duplication in the
address
ii) Memory can be placed contiguously in the address space of the
microprocessor
iii) Future expansion can be made easily without disturbing the existing
circuitry

There are few disadvantages in this method
i) Extra decoders are necessary
ii) Some delay will be produced by these extra decoders.
Copyright 2008 by Dr.K.K.Thyagharajan 2
The main advantage of linear select decoding is its simplified decoding circuit.
This reduces the hardware design cost. But there are many disadvantages in this
decoding.
i) Multiple addresses are provided for the same location
ii) Complete memory space of the microprocessor is not efficiently used
iii) Adding or interfacing ICs with already existing circuitry is difficult.


3 3. .1 1. .1 1 A Ab bs so ol lu ut te e A Ad dd dr re es ss s D De ec co od di in ng g

The 8085 microprocessor has 16 address lines. Therefore it can access 2
16

locations in the physical memory. If all these lines are connected to a single memory
device, it will decode these 16 address lines internally and produces 2
16
different
addresses from 0000H to FFFFH so that each location in the memory will have a unique
address as shown in table 3.1 and figure 3.1. This is called absolute address decoding.
Figure 3.1 is called the memory map of 8085 address space and it shows how each
location in the physical memory gets an address from the address space of the
microprocessor.


A
15
A
14
A
13
A
12
A
11
A
10
A
9
A
8
A
7
A
6
A
5
A
4
A
3
A
2
A
1


A
0
Hex Address
0 0 0 0

0 0 0 0

0 0 0 0

- - -
- - -


1 1 1 1

1 1 1 1
0 0 0 0

0 0 0 0

0 0 0 0

- - -
- - -


1 1 1 1

1 1 1 1
0 0 0 0

0 0 0 0

0 0 0 0

- - -
- - -


1 1 1 1

1 1 1 1
0 0 0 0

0 0 0 1

0 0 1 0

- - -
- - -


1 1 1 0

1 1 1 0
0000H

0001H

0002H

- - -
- - -


FFFEH

FFFFH


Table 3.1: Memory Address

Copyright 2008 by Dr.K.K.Thyagharajan 3
0000 H
8000 H
0001 H
0002 H
7FFF H
8001 H
FFFD H
FFFE H
FFFF H


Figure 3.1: 64KB Memory Map
(Single Block)

Instead of connecting all the 16 address lines to a single 64 KB memory device,
we can also use memory devices of lower capacity. In such cases the physical memory
space of the microprocessor is divided into smaller memory spaces or memory blocks.
Each block is selected by a block select address signal and the memory locations with in
a block are accessed by the processors address lines. This is called address portioning.
For example if the memory device (memory block) has only 13 address lines, we can use
the lower 13 address lines (A
0
A
12
) to access the locations with in a memory block and
the remaining 3 lines (A
13
, A
14
and A
15
) to access 8 such blocks. These three lines are
called the block select address signals and this is called address partitioning.
Depending on the number of memory blocks, the number of block select address lines
will change. The starting address and ending address of each block can be found as
shown in Table 3.2. Since each block is an Integrated Chip (IC), and each chip has a
Chip Select signal, the block select address lines must be used to select the ICs. We have
to produce eight chip select signals from the three address lines. Therefore we have to
decode these lines using a 3 to 8 decoder or 1 of 8 decoder. In this case the lower thirteen
address lines of the processor are connected to the 13 address lines of the memory chip
and hence they are internally decoded in the memory. The higher 3 address lines (A
13
,
A
14
and A
15
) are externally decoded by a 3 to 8 decoder. Since three lines can provide a
maximum of 8 addresses, in this case they are said to be fully decoded and hence each
location in each block has a specific unique address as shown in table 3.2. This is called
as absolute address decoding. The diagrammatic representation of each block with its
address range shown in figure 3.2 is called the memory map of the microprocessor
system.

Copyright 2008 by Dr.K.K.Thyagharajan 4
Block
No
A
15
A
14
A
13
A
12
A
11
A
10
A
9
A
8
A
7
A
6
A
5
A
4
A
3
A
2
A
1
A
0

Hex
Address
range
1 0 0 0 0

0 0 0 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0000H to

1FFFH
2 0 0 1 0

0 0 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
2000H to

3FFFH
3 0 1 0 0

0 1 0 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
4000H to

5FFFH
4 0 1 1 0

0 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
6000H to

7FFFH
5 1 0 0 0

1 0 0 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
8000H to

9FFFH
6 1 0 1 0

1 0 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
A000H to

BFFFH
7 1 1 0 0

1 1 0 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
C000H to

DFFFH
8 1 1 1 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
E000H to

FFFFH

Table 3.2: Address Space for Memory Blocks

Copyright 2008 by Dr.K.K.Thyagharajan 5
Memory
Block1(8KB)
Memory
Block2(8KB)
Memory
Block3(8KB)
Memory
Block4(8KB)
Memory
Block5(8KB)
Memory
Block6(8KB)
Memory
Block7(8KB)
Memory
Block8(8KB)
0000
1FFF
2000
5FFF
4000
3FFF
6000
7FFF
8000
9FFF
A000
BFFF
C000
DFFF
E000
FFFF
EPROM
RAM 1
RAM 2
RAM 3
RAM 4
RAM 5
RAM 6
RAM 7


Figure 3.2: 64KB Memory Map
(Eight Blocks)

We use 74LS138 address decoder to generate the chip select signals for each
memory block. In this decoder when the address lines A
13
, A
14
and A
15
are 000, the
output line Y
0
will be activated as shown in figure 3.3. This in turn selects the first
memory block. Similarly when these lines are 001 (C=0, B=0 and A=1) Y
1
will be
activated and the second memory block will be selected. When the 8085 microprocessor
is reset the contents of the program counter will be 0000H. Since this content is placed on
the address bus while accessing the memory, the microprocessor will start executing the
program from the address 0000H. A monitor program must be placed starting from this
address, so that it reads the inputs and takes action based on that program. The input may
even be a function key that requires a specific action to be carried over. Since this
program is fixed and developed by the manufactures of the product it is called as
firmware and it is stored in a ROM or in an EPROM. Therefore, in any microprocessor
based system an EPROM should be placed at the reset address (0000H). In addition to
EPROM, any microprocessor based system would require at least few RAM locations to
store temporary data. Specifically there should be a RAM for stack operations. So,
though we are using eight memory chips of 8KB size each, at least one of the eight ICs
must be an EPROM and it should be placed at the starting address of 0000H. The
memory map for this case can be given as shown in figure 3.2. The EPROM chip must be
selected using the chip select signal CS1 and the RAM1 chip must be selected using CS2.
The memory interface diagram may be given as shown in Figure 3.4.




Copyright 2008 by Dr.K.K.Thyagharajan 6























Figure 3.3: Memory Block Decoder

In this type of memory interfacing, all the address lines (A
0
to A
15
) have been
used. Each location in the memory will have a single address. This type of address
decoding is called as absolute or fully decoded addressing.

Most of the microprocessor based systems do not use the complete 64 KB
memory space. Even one EPROM and a RAM will be sufficient. For example in the
memory map shown in figure 3.2, if only the EPROM and RAM 2 are used in the
practical system, the memory map of such a system can be given as shown in figure 3.5.
Still each location has single address. Therefore it is also called absolute address
decoding. The main advantage of this type of decoding is, you can add memory devices
for this system without disturbing the already connected devices. The memory interface
diagram for this case is given in figure 3.6.


A13
A14
A15
IO/M
CS1
CS2
CS3
CS4
CS5
CS6
CS7
CS8
+5V
Y0
Y1
Y2
Y3
Y4
Y5
Y6
Y7
A
B
C
G1
G2
G3
74LS138

Block
Select
Address
Signals
Block
Select
Signals
}
Copyright 2008 by Dr.K.K.Thyagharajan 7



Copyright 2008 by Dr.K.K.Thyagharajan 8
Memory
Block1(8KB)
Memory
Block3(8KB)
0000
1FFF
2000
5FFF
4000
3FFF
6000
7FFF
8000
9FFF
A000
BFFF
C000
DFFF
E000
FFFF
EPROM
RAM 2
Unused memory space


Figure 3.5: Memory Map with EPROM and RAM 2

A13
A14
A15
IO/M
CS1
CS2
CS3
CS4
CS5
CS6
CS7
CS8
+5V
Y0
Y1
Y2
Y3
Y4
Y5
Y6
Y7
A
B
C
G1
G2
G3
74LS138
CS CS
EPROM
RAM 2
A0 - A12 Address Bus


Figure 3.6: Interfacing EPROM and RAM

Instead of using 3 to 8 decoder, NAND/AND/OR gates can also be used to
generate chip select signals as shown in figure 3.7. In this case, CS1 will be activated to
select the EPROM, when all the three inputs to the OR gate (OR1) are zero. Similarly the
chip select signal CS3 will be activated to select RAM 2 only when A
13
= 0, A
14
= 1 and
A
15
= 0. This is also an absolute address decoding method. But, here we need more
Copyright 2008 by Dr.K.K.Thyagharajan 9
decoders (gates) when we want to add more memory chips. The main advantage of the
absolute address decoding is the contiguous placement of the memory locations. So no
memory space of the microprocessor is wasted without being used.

CS CS
EPROM
RAM 2
A0 - A12 Address Bus
A13
A14
A15
CS1
OR 1
A13
A14
A15
CS3
OR 2
8K 8K


Figure 3.7: Address Decoding using OR Gates


3 3. .1 1. .2 2 L Li in ne ea ar r S Se el le ec ct t A Ad dd dr re es ss s D De ec co od di in ng g

In the circuit given in figure 3.8, the address line A
15
of the microprocessor has
not been used for decoding the chip select signals. The outputs Y
0
and Y
2
of the 74LS138
decoder will be activated as given in table 3.3. Since the address line A
15
is not connected
to the decoder and the pin C of the decoder has been grounded, irrespective of the signal
A
15
(i.e.A
15
is 0 or 1), the pin C is 0. So, you will get Y
0
selected twice i.e. when A
15
= 0
and when A
15
= 1. Similarly Y
2
will also be selected twice. Therefore the address map for
the EPROM and RAM is found as shown in Table 3.4.

A
14
A
13

A
15
C

B A
Y
0
Y
2

0 0 0 0 0 1
0 0 1 0 1 0
1 0 0 0 0 1
1 0 1 0 1 0

Table 3.3: Linear Select Address Decoding Chip Select Signals

Copyright 2008 by Dr.K.K.Thyagharajan 10
A13
A14
IO/M
CS1
CS2
CS3
CS4
CS5
CS6
CS7
CS8
+5V
Y0
Y1
Y2
Y3
Y4
Y5
Y6
Y7
A
B
C
G1
G2
G3
74LS138
CS CS
EPROM
RAM 2
A0 - A12 Address Bus


Figure 3.8: Linear Select Address Decoding - Circuit Diagram

Block
No
A
15
A
14
A
13
A
12
A
11
A
10
A
9
A
8
A
7
A
6
A
5
A
4
A
3
A
2
A
1
A
0

Hex
Address
range
1
0 0 0 0

0 0 0 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0000H
to
1FFFH
3
0 1 0 0

0 1 0 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
4000H
to
5FFFH
5
1 0 0 0

1 0 0 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
8000H
to
9FFFH
7
1 1 0 0

1 1 0 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
C000H
to
DFFFH

Table 3.4: Linear Select Decoding Memory Address Ranges

The same EPROM is selected both for the address range 0000H to 1FFFH and
8000H to 9FFFH. This is due to the signals A
13
and A
14
which have same values in both
the cases and hence activates the same chip select signal CS1. Similarly the RAM is
selected both for the address range 4000H to 5FFFH and C000H to DFFFH. The memory
map for this case is as shown in figure 3.9.

Copyright 2008 by Dr.K.K.Thyagharajan 11
Memory
Block1(8KB)
Memory
Block3(8KB)
0000
1FFF
2000
5FFF
4000
3FFF
6000
7FFF
8000
9FFF
A000
BFFF
C000
DFFF
E000
FFFF
EPROM 1
RAM 2
Unused memory
space
Memory
Block5(8KB)
EPROM 1
Memory
Block7(8KB)
RAM 2


Figure 3.9: Linear Select Address Decoding Memory Map

CS CS
EPROM RAM 2
A0 - A12 Address Bus
A13
A14
CS1
OR 1
A13
A14
CS3
OR 2
8K 8K


Figure 3.10: Linear Select Address Decoding using OR gates

In this type of address decoding, if we want to add a RAM chip at the address
space 8000H to 9FFFH, which was unnecessarily occupied by the first EPROM, we have
to change the decoding circuit. We may not simply add additional circuitry without
disturbing the already existing circuit. In this method each location in the physical
memory is accessed by more than one address (here two addresses) i.e. address space is
wasted. This type of address decoding is called as Linear Select or partially decoded
addressing. This happens when we have few address lines without being used for
decoding. The same circuit can also be implemented using OR gates as shown in figure
3.10. In this circuit the OR gates require only two inputs. This is possible, since we have
left A
15
address line unconnected. Therefore, linear select addressing reduces the
hardware requirements.


Copyright 2008 by Dr.K.K.Thyagharajan 12
3 3 . . 2 2 B Bu u s s C Co o n nt t e e n n t t i i o o n n

Consider the timing diagram shown in figure 3.11 that reads a byte (opcode) from
the memory. At the end of the ALE signal, all the 16 bits of the address will be available
on the bus. The decoder (74LS138) produces the chip select signals from these address
lines after a small decoding delay. The chip select signal will be generated with in 30 ns
after the trailing edge of ALE signal. When the chip select signal enables the memory
chip, since the address has been already placed on the address pins of the memory, the
memory will place the data on the data lines after a small access delay called the memory
access time. The output buffer of the memory will be enabled within 20 ns after getting
CS signal. The memory access time is the time delay between the address placed on the
address bus by the address decoder and the data placed by the memory on the data bus. If
the data are placed before the end of the T
1
state, both the lower order address and the
data from the memory will be placed on the AD
0
AD
7
lines i.e. both the microprocessor
and the memory try to access the same lines as shown in figure 3.11 by hashed lines. If
the microprocessor tries to place 1 (+5V) and the memory tries to place 0 (0 V or
Ground), then the microprocessor pin will be shorted through the memory. This may
even damage the processor. This is called bus contention. So, the bus contention will
occur when two ICs try to place data on the same bus at the same instant of time. This
effect will be very adverse in the case of high speed memories interfaced with slow
processors. Since the RD or WR signal is issued by the processor only after the first T-
state, if these signals are used along with the output of the address decoders to select the
chip, the content of the memory will not be placed before the start of the T
2
state.

Copyright 2008 by Dr.K.K.Thyagharajan 13
T
1
CLK
41H
00H
Low-Order address
4FH Opcode
Unspecified High-Order Memorry Address
A
15
A
8
AD
7
AD
0
ALE
T
2
T
3
T
4
RD
CS
Delay in the decoder
4FH Opcode
Data from
the Memory
Access Time
Bus Contention
Bus Contention
CS
CS
Avoiding Bus
Contention
Data from memory
B
D
E
H
A
C
G
F

Figure 3.11: Bus Contention Reading Opcode from the Memory


Copyright 2008 by Dr.K.K.Thyagharajan 14
A13
A14
A15
IO/M
CS1
CS2
CS3
CS4
CS5
CS6
CS7
CS8
+5V
Y0
Y1
Y2
Y3
Y4
Y5
Y6
Y7
A
B
C
G1
G2
G3
74LS138
CS CS
EPROM
RAM 2
A0 - A12 Address Bus
RD
WR


Figure 3.12: Avoiding Bus Contention

By using a standard 3 to 8 decoder as shown in figure 3.12, the CS signal can be
delayed until the end of T1 state. The address decoder is enabled by the G
3
signal and
hence the CS signal is generated only when RD or WR is low. In figure 3.11, point A
activates point B and point B in turn activates point C to avoid bus contention.

If you are using logic gates to decode the addresses, bus contention may be
avoided by combining the output of the final stage of the decoder with RD or WR signal
to produce an active low chip select signal.

NB: Instead of connecting the RD or WR signals to the input or output of the
address decoder, the best practice to avoid bus contention is to connect the RD signal to
the OE pin of EPROM/RAM and WR signal to R/W pin of RAM.

Example 3.1:

Interface a 4K EPROM, one 4K RAM and one 8K RAM to a microprocessor with
the following Memory Map.





Copyright 2008 by Dr.K.K.Thyagharajan 15




Figure 3.13: Memory Map


A memory chip select decoder is used to provide chip select signal for each
memory device (IC). This will decide the address range that is allotted for each memory
IC. 74LS138 is a 3 to 8 decoder and it can be used for this purpose. In this example the
minimum memory block size is 4K. To access 1K locations 10 address lines must be used
(2
10
=1K = 1024 locations). So to access 4K locations (4 X 1K = 2
2
X 2
10
= 2
12
) 12
address lines (A
0
A
11
) must be used. Since 8085 has 16 address lines the decoding can
be indicated as shown below.




Figure 3.14: Variable Address Lines


While accessing 4 K locations the lower 12 bits (A
0
A
11
) can have either 0 or 1.
If all the 12 bits are 0 then that will be the starting address of that memory block if all 12
bits are 1 that will be the end address of the block. The remaining 4 address lines (A
12
-
A
15
) are the block select address signals, which decide the memory block number. For
example, if A
15
A
12
is 0000 then memory block that can be accessed by the remaining
address lines is block 0. If it is 0001 the memory block that can be accessed is 1. Note
that the memory blocks 5 and 6 are combined because a 8-K RAM is to be placed there.
The following table shows the starting and ending address of each memory block.

Copyright 2008 by Dr.K.K.Thyagharajan 16
Block
No
A
15
A
14
A
13
A
12
A
11
A
10
A
9
A
8
A
7
A
6
A
5
A
4
A
3
A
2
A
1
A
0

Hex
Address
Range
0 0 0 0 0

0 0 0 0
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0000H
to
0FFFH
1 0 0 0 1

0 0 0 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
1000H
to
1FFFH
2 0 0 1 0

0 0 1 0
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
2000H
to
2FFFH
3 0 0 1 1

0 0 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
3000H
to
3FFFH
4 0 1 0 0

0 1 0 0
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
4000H
to
4FFFH
5 0 1 0 1

0 1 0 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
5000H
to
5FFFH
:
:
:
15 1 1 1 1

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
0 0 0 0

1 1 1 1
F000H
to
FFFFH

Table 3.5: Memory Map

The 74LS138 decoder has three input pins marked as CBA when all these three
pins are 0 then the output pin O
0
will be activated i.e. O
0
becomes zero. When CBA is
001 the O
1
will be activated and all other output pins will be at high state. i.e. any one of
the output pins O
0
to O
7
is selected based on the input A, B and C and this will happen
only if the control pins G
1
, G2 and G3 are properly activated. Table 3.5 shows that
address line A
15
is always zero for the address range 0000H to 6FFFH. So this can be
connected to G3 pin of the 74LS138. The output pins of 74LS138 must be selected only
for memory operations. During memory operations IO/M = 0. So IO/ M is connected to
G2 pin of the 74LS138 and hence for any IO operation the chip 74LS138 will not
activate any of the output pins. The pin G1 is connected to +5V through a resistor to
enable 74LS138.
The O
0
pin of 74LS138 is connected to the chip select (CS) signal of the 4K
EPROM. This allows selecting this EPROM chip when the microprocessor sends
address in the range 0000H to 0FFFH. Similarly O
1
pin of 74LS138 is connected to the
CS of 4K RAM. Since no memory chip is connected in the address range 2000H to
3FFFH, the output pin Q
2
is left free. The 8K RAM requires address from 4000H to
Copyright 2008 by Dr.K.K.Thyagharajan 17
5FFFH and these addresses are covered by Q
4
and Q
5
pins. So these two pins are
connected to a two input AND gate. The output of the AND gate is connected to the chip
select pin of the RAM and this CS will be activated (active low) if either Q
4
or Q
5
is low.
i.e. the output of the AND gate will be 0 for the addresses 4000H to 5FFFH and hence the
8K RAM chip is selected for this address range.

C B A Output pin Activated
0 0 0 O
0

0 0 1 O
1

0 1 0 O
2

0 1 1 O
3

1 0 0 O
4

1 0 1 O
5

1 1 0 O
6

1 1 1 O
7


Table 3.1: Function Table of 74LS138

The address lines A
0
to A
11
of 8085 MPU must be connected to the corresponding
address pins of the memory chip and the data lines D
0
to D
7
of MPU must be connected
to the corresponding data pins of all the memory chips. Since 8K RAM has 13 address
lines in addition to the twelve address lines (A
0
to A
11
) A
12
of MPU must also be
connected to the A
12
pin of the 8K RAM. You note that bus contention is not addressed in
this circuit.

A13
A14
A12
IO/M
+5V
Y0
Y1
Y2
Y3
Y4
Y5
Y6
Y7
A
B
C
G1
G2
G3
74LS138
CS CS
EPROM RAM
A15
0000 - 0FFF
1000 - 1FFF
2000 -2FFF
3000 - 3FFF
4000 - 4FFF
5000 - 5FFF
6000 - 6FFF
7000 - 7FFF
CS
RAM
4K 4K 8K
Address Bus A11 . . . A0
Data Bus D7 . . . D0
From MPU
From MPU
A12


Figure3.15: Memory Interfacing
Copyright 2008 by Dr.K.K.Thyagharajan 18

3 3 . . 3 3 I I n nt t e e r r f f a a c c i i n ng g l l o o w w s s p pe e e e d d m me e m mo o r r y y

To avoid bus contention, the RD or WR signal of the microprocessor is used
along with the address decoding to produce the CS signal for the memory. The RD or
WR signal starts approximately in the middle of the T
2
state and available up to the end
T
3
state. If the microprocessor operates at 3 MHz, these signals are available for
approximately 1.5 T state or 500 ns (1.5 / 3 MHz = 0.5 micro seconds) and hence the
memory access time should be less than 500 ns. If the memory access time is more than
this value the data cannot be read. If a low speed memory has access time greater than
this value, Wait cycles are introduced between T
2
and T
3
. In such a case, the CS select
signal need not be delayed until RD or WR signals are issued by the microprocessor. The
memory can decode the address as soon as it gets the address from the microprocessor. In
this case the total delay introduced will be equal to the sum of the delays introduced by
the decoder, wait cycles and the delay introduced due to memory access time. For
example 2716 EPROM has access time of 450 ns and therefore it can be directly
interfaced with 3 MHz 8085, but one wait cycle is necessary if it is interfaced with a
microprocessor operating at 5 MHz. A low speed memory can also be interfaced with a
high speed processor by using low frequency crystal but this will slow down not only the
memory read/write operation but all activities of the processor.

The READY pin of 8085 processor is used to introduce wait states. For normal
operations of the processor this pin should be connected to logic high. When this pin is
made low, the processor will enter into wait state until it is made high again. A special
circuitry called Wait State Generator as shown in figure 3.16 is required for this purpose.
This circuit will introduce a wait state equal to one T-state of the processor cycle as
shown in figure 3.17. The ALE signal fed to the clock input of the first D-type flip-flop
will transfer the logic high input (+5V) placed on its D
0
pin to Q
0
during its positive edge.
The Q
0
output is connected to D
1
input of the second flip-flop. So, the logic high output
on Q
0
will be transferred to the output Q
1
of the second flip-flop during the positive edge
of T
2
state. The complement value of Q
1
is available on Q
1
pin of the second flip-flop.
This will reset the first flip-flop i.e. Q
0
will become logic 0 as shown in figure 3.17. At
the same time since Q
1
has been connected to the READY pin of the processor, the
processor enters into wait state. It will remain in that state until Q
1
becomes high. This
extends the length of the RD signal as shown in figure 3.17 and allows memory to take
one more clock period time to place the contents on the bus. The logic 0 placed on D
1
pin
during this time will be transferred to Q
1
in the positive edge of the next clock cycle (T
3

state). This will make Q
1
logic high and hence the READY signal becomes high and the
processor continues its work. It should be noted that wait state can be introduced in any
machine cycle of 8085 except in bus idle cycle. The 8085 checks the READY signal at
the second T-state of every machine cycle.

Copyright 2008 by Dr.K.K.Thyagharajan 19
CLK
R
Q
Q
D
CLK
R
Q
Q
D
+5V
ALE
0 0
0
1 1
1
CLKOUT from 8085
To READY
Pin of 8085
D Flip-Flop D Flip-Flop


Figure 3.16: Wait State Generator



T
1
CLK
41H
00H
Low-Order address
4FH
High-Order Memorry Address
A
15
A
8
AD
7
AD
0
ALE
T
2
T
3
T
WAIT
Data
Q0
Memory Read
Q1
T
1
Memory Read
READY/
RD
Delayed Data
EXTENDED READ SIGNAL
POSITIVE OR LEADING EDGE OF THE CLOCK CYCLE
+VE EDGE


Figure 3.17: Timing Diagram with Wait State
Copyright 2008 by Dr.K.K.Thyagharajan 20

3 3 . . 4 4 I I n nt t e e r r f f a a c c i i n ng g M Mo o r r e e t t h h a a n n 6 6 4 4 K K M Me e m mo o r r y y

The 8085 microprocessor can address only 64K memory at a time, if you want to
interface more than 64K memory; you can use a jumper or an IO port bit as shown in
figure 3.18. the chip select signal CS3 will select RAM 2 if the jumper connects B and C
and it will select RAM 3 if the jumper connects A and B. Similarly CS4 will select RAM
4 if the port bit is zero and RAM 5 if the port bit is one. The main difference between
jumper selection and port bit selection is the manual selection in the first case and
automatic or program activated selection in the second case.

A14
A15
IO/M
CS1
CS2
CS3
CS4
+5V
Y0
Y1
Y2
Y3
Y4
Y5
Y6
Y7
A
B
C
G1
G2
G3
74LS138
CS CS
RAM 4
RAM 5
CS
RAM1
CS
CS
RAM 2
RAM 3
CS
EPROM
A0 - A13 Address Bus
Figure 3.4: Memory Interface Diagram
16K 16K
16K
16K
16K
16K
PORT BIT
A B C


Figure 3.18: Interfacing more than 64 K Memory


Summary

A microprocessor has address space for memory. Any memory chip that is being
interfaced with processor must fit into this memory space. A memory map diagram
shows the address boundary for each memory chip interfaced with the processor and an
address decoder decides the address range for each chip. There are two types of address
decoding viz absolute address decoding and linear select address decoding. When high
speed memories are interfaced with slow processors bus contention may occur and when
a low speed memory is to be accessed by a fast processors wait state must be introduced
using external circuitry.