e022020.pdf

MICRO PROCESSOR

Microcontroller

Basics Course (2)

part 3: port characteristics and port accesses

By B. Kainka

In the ﬁrst instalment of this basics course, we introduced the assembler

that we will be using. Now we come to the first practical applications

using the processor ports. First, though, we have to take a look at the

physical characteristics of the port connections.

Anyone who already has a bit of experience

with digital electronics knows that there are

many different types of outputs. Each type

has quite speciﬁc characteristics, regardless

of whether it is a TTL, CMOS, tri-state, open-

collector or open-drain output, and if you

want to connect something to the output you

need to be aware of these characteristics.

In any case, the Port 1 outputs of an 8051

microcontroller do not fall in line with any of

the know logic families, but instead employ a

rather unique solution. These ports are what

is known as ‘quasi-bidirectional’, which

means that they can be used as inputs or out-

puts without having to be specially switched

over. You should keep in mind that the ports

of a microcontroller represent a sort of door to

the outside world. Depending on the task to

be performed, inputs or outputs are needed.

Some microcontrollers use tri-state buffers

that must be switched to the high-impedance

state to allow them to be used as inputs. This

naturally requires a special switching signal or

special instructions to switch the data direc-

tion. This is not necessary with an 8051, since

all ports can be used as both inputs and out-

puts without any switching.

resistor alone defines the internal

resistance of the port. Consequently,

it is certainly possible to connect any

desired logic output here or change

the signal level by means of a switch

connected to ground. Even a logic

input with high input impedance,

such as that of a CMOS IC, will not

have any difficulty recognising a

High state.

The situation is quite different

when the output is conducting and

thus forces the signal level to be

Low. In this case, the port imped-

ance is relatively low. Anyone who

attempts to force the output level of

the port to a High state when it is in

this state can only have bad inten-

tions, since he or she is trying to

force the microcontroller into the dig-

The port in detail

A glance at the detailed circuit diagram of a

port ( Figure 1 ) shows how a quasi-bidirec-

tional port is built. There is a single FET with

a pull-up resistor located at the output. In the

High state, the FET is cut off and the pull-up

Elektor Electronics

2/2002

MICRO PROCESSOR

speaking, the pull-up resistor con-

sists of two (in fact, three) FETs. One

of them provides the normal, very

weak pull-up current. A second one

provides a much stronger current,

but it is active only for a short inter-

val when the port is switched from

the Low state to the High state (or

when an address must be output).

The port pin thus changes state very

quickly, even with a certain amount

of capacitive loading. However, it can

still be used as an input and can be

actively pulled to ground, since any

possible short circuit lasts less than

a microsecond.

A quasi-bidirectional port can also

directly drive an LED, but only if

series resistor for the LED is con-

nected to V cc rather than ground.

Figure 2 shows how an LED and a

switch can be connected. For the ﬁrst

program of the previous instalment of

our course, the LED must be con-

nected to one of pins P1.4 through

P1.7 for it to be illuminated, since

only these pins are switched Low. In

this case, the switch should be con-

nected to one of pins P1.0 through

P1.3, since they are in the High state

and can thus act as inputs. By the

way, here you can operate the switch

as often as you like, but nothing will

happen, since we must ﬁrst write a

program that polls the input and

evaluates the result.

V CC

Listing 1.

A program loop with port outputs.

Read

Latch

Internal

Pull Up

Arrangement

;ﬂash2.asm, fast loop

#include 8051.H

.org 0000H

Int.Bus

main mov a,#0Fh ;1 a = 15

mov P1,a ;1 P1 = a

mov a,#0F0h ;1 a = 240

mov P1,a ;1 P1 = a

sjmp main ;2

.end

Bit

Latch

Write

Latch

CLK

Read

010208 - 3 - 11

port value 0Fh and then outputs the port

value F0h. When hexadecimal notation is

used, the first character must always be a

numeral. This is why the listing shows a ‘0’

in front of the second value, which is thus

‘0F0h’ instead of ‘F0h’.

Another feature of this listing is that the

Instead, we have added an ‘include file’

( 8051.h ) that contains all important defini-

tions, including much more than just Port 1.

Besides this, the starting address is explicitly

specified using the .org statement. The

microcontroller always starts its programs at

address 0000h following a reset. Finally, the

critical change is that the loop has been

expanded. It now encompasses the entire pro-

gram, which is executed over and over again.

This small program helps answer a quite

important question, which is how fast such a

microcontroller can actually run programs. All

we have to do is to touch the input probe of

an oscilloscope to one of the port pins. Here

we will see a rectangular waveform with a

period of around 150 kHz. This can also be

demonstrated using a radio. A short piece of

wire attached to the port pin can serve as an

antenna. In the long-wave band, you will ﬁnd

the signal at around 150 kHz. The fifth har-

monic can be received at roughly 750 kHz in

the medium-wave band. This gives us more

insight into the fundamental significance of

the EMC directives. Whenever high frequen-

cies and steep edges occur, special provisions

must be taken to prevent the circuit in ques-

tion from acting as a transmitter.

The number of instruction cycles is shown

in the listing comments. An instruction cycle

takes 12 oscillator clock cycles, which means

it has a period of (12 ÷ 11.059 MHz) =

1.095 µs. Most instructions have a duration of

one instruction cycle, while the jump instruc-

tion takes two cycles. In total, the sum of the

instruction times is six instruction cycles.

This means that the loop takes 6.51 µs, which

gives a frequency of 153 kHz.

There is yet another interesting observa-

Figure 1. Internal structure of a port.

ital ‘happy hunting ground’. We thus

have the following rule: outputs

must never be connected to other

outputs. You must always remember

that if you want to use a port pin as

an input, it must be in the High

state! This is always the case fol-

lowing a reset.

The internal circuit diagram of the

port pin contains yet another simpli-

ﬁcation, since the pull-up resistor is

in fact also a FET. Consequently, it

acts like a constant-current source.

This naturally raises the question of

how large this constant current is. A

measurement on the AT89S8252

shows that a short-circuit current of

around 19 µA ﬂows when the port is

pulled to ground. This is not very

much. It is far less than what is

needed to drive an LED, for example.

Is it actually enough to change the

signal level on a lead reasonably

quickly? Actually, it isn’t, since every

lead has an intrinsic capacitance that

must ﬁrst be charged or discharged.

Here again there is a refinement in

the internal circuit of the port. Strictly

Our ﬁrst program loop

After this introduction, it’s time for

something practical. What we want

to do is to automatically switch the

output levels on the port pins. To do

so, we will modify the program from

the first instalment of the course to

obtain the program shown in List-

ing 1 . This program ﬁrst outputs the

V CC

P1.x

GND

LDR

010208 - 3 - 12

Figure 2. Port connections for inputs and outputs.

2/2002

Elektor Electronics

MICRO PROCESSOR

tion that we can make with this program. We

can use it to watch the two pull-up FETs

working ‘live’. To do so, we connect a 33-kΩ

resistor from one of the port pins to ground.

With this resistance value, the smaller FET is

no longer able to pull the signal level to V cc ,

but the larger FET can still do so. In the oscil-

logram, we can now see how long (or better,

how short) the current level is increased. The

initial edge is very fast. This is followed by a

high plateau with a duration of around 100 ns

(one quarter of a clock period), and ﬁnally the

voltage drops to a low level.

The port evidently has yet another charac-

teristic that is not mentioned in most data

sheets. The smaller FET is also divided into

two and behaves differently when the port

state is High than when it is Low. With a load

resistance of 6.8 kΩ , we see a second kink in

the curve at a voltage of roughly 1.5 V. Evi-

dently, the current that flows in the region

above this input voltage is greater than the

current that ﬂows below this level. Altogether,

this acts like a form of current feedback and

results in a certain amount of input hystere-

sis. It can be easily measured using a multi-

meter. At a voltage greater than 1.5 V, the port

supplies up to 200 µA, but below this voltage

the current is only 10 µA. Thanks to this port

behaviour, connecting a simple resistance to

an input always results in an unambiguous

input state. It is even possible to connect a

potentiometer or an LDR, which will then be

read with well-deﬁned hysteresis. In the orig-

inal Intel and Philips data books, the division

into three FETs can still be seen, but the data

sheets for the more recent 8051 derivatives

from Atmel neglect this detail.

Figure 4 shows a small circuit for our ﬁrst

experiment with input connections for a port.

Here P1.0 can be set to zero either by means

of a switch or by a shining a sufficiently

bright light on the LDR. A matching program

must read the port state and switch on the

LED attached to port 1.1 only when a High

signal is found at the input. Here we must

take into account the fact that the LED is con-

measurement cable

P1.0

1...100k

ca. 10pF

oscilloscope

GND

010208 - 3 - 13

Figure 3. Measuring the port signal under load.

nected to V cc , which means that it

has an inverted function; it is ‘on’

when P1.1 is in the Low state. S2

allows us to provide feedback

between the output and input of our

test setup. What happens if we actu-

ate S1 and S2 together and thereby

short the P1.1 output to ground?

Nothing serious, since a connection

to ground is always allowed for a

quasi-bidirectional port. The only

thing is that in this case the LED is

always on.

the lines with the label ‘ON’, it is

necessary to jump over this part of

the program. An additional jump

instruction thus leads to the desti-

nation ‘OFF’, where the loop is

closed by jumping back to the very

beginning of the program.

This small example shows several

interesting results. For one thing, we

have practically built a logical

inverter. A Low state on the input

results in a High state on the output,

and vice versa. Since the LED (Low

= on) and the input switch (on =

Low) have inverting functions, the

overall operation is inverting. This

means that if you press S1, the LED

goes off, while if you release the

switch the LED goes on again.

This experiment will show the

input hysteresis of the port if a resis-

A conditional jump

The assembler program for Listing 2

reads the port state at P1.0 and then

executes a conditional jump. The

instruction jb (jump if bit set)

belongs to the group of special

instructions for single-bit process-

ing. Although most instructions

work with byte values, these

instructions evaluate, set or clear

individual bits. Here ‘P1.0’ refers to

a single port pin, while ‘P1’ in the

ﬁrst example refers to the complete

port with all eight of its leads. The

header ﬁle ( 8051.h ) deﬁnes all nec-

essary bit addresses, such as P1.0

and P1.1. You can view this ASCII

ﬁle using a simple text editor.

In the program, if a ‘1’ state is

read at P1.0 in the first line, a jump

to the destination ‘ON’ is executed.

At this location, port P1.1 is placed

into the Low state by means of clr

P1.1 (‘clr’ = ‘clear’), which switches

on the LED. On the other hand, if a

Low state is read at P1.0 in the ﬁrst

line of the program, the jump is not

executed. Instead, the program quite

properly executes the next following

instruction. Here we find setb

P1.1 (‘setb’ = ‘set bit’). This is an

instruction to enable port P1.1,

which effectively means that the

LED is switched off. In order to pre-

vent this from being counteracted by

Figure 4. Port loaded with 33 k

Listing 2.

Responding to an input state

;ﬂash3.asm, input/output

#include 8051.H

.org 0000H

main jb P1.0,ON ;P1.0 = ?

setb P1.1 ;P1.1 = 1

sjmp OFF

ON clr P1.1 ;P1.1 = 0

OFF sjmp main

.end

Figure 5. Port loaded with 6.8 k

Elektor Electronics

2/2002

MICRO PROCESSOR

If we attempt something like this

using a logic gate, in this case an

inverter, the result can look much dif-

ferent. Usually the signal assumes

an average voltage level, by means

of which the gate reveals its ana-

logue roots, since it actually an

inverting dc amplifier. A sequential

logic circuit (such as a microcon-

troller), by contrast, allows no possi-

bility of analogue behaviour. Unam-

biguous yes or no decisions are

always taken.

Listing 3. A counting loop.

;ﬂash4.asm port outputs

#include 8051.H

.org 0000H

V CC

main mov a,#00

next mov P1,a ;1

mov r1,#255 ;1

loop djnz r3,loop ;2 * 255

inc a ;1

sjmp next ;2

.end

P1.1

P1.0

LDR

GND

on. The program also contains a second

counting loop to cause everything to run a bit

slower. The objective is to allow at least the

lowest-frequency signal on port P1.7 to be

directly observed using an LED.

The counting loop uses a register (r1).

There are eight registers in total (r0 through

r7). It does not matter here exactly which reg-

ister is used. The register is loaded with a

value of 255, which means that in this case a

decimal number is used. We could have just

as easily written this as 0FFh. The actual

counting loop employs the complex assem-

bler instruction djnz (decrement and jump if

not zero). During the first pass through the

loop, the value in r1 is reduced to 254. Since

this is greater than zero, a jump is made back

to loop , where the same instruction is again

executed. During this pass, the value is

reduced to 253. After a total of 255 passes

through the loop, the value of zero is reached.

Now the jump is no longer executed, but

instead, the program continues with the next

following line. In this manner a total of 255

loop cycles, amounting to 510 instruction

cycles or 558 µs, are consumed. If we add the

remaining instructions, we have a total

period of approximately 564 µs, or a frequency

of 1.77 kHz. This is thus the clock frequency

at the input to the counter chain. Conse-

quently, on P1.0 we find a square-wave sig-

nal with a frequency of 885 kHz, on P1.1

443 kHz and so on, down to around 7 Hz on

P1.7. This frequency can be observed using

an LED. It is worthwhile to perform a number

of additional experiments using different loop

parameters. The highest frequency results

when the starting value is 1. By the way, the

greatest reduction in the clock rate can be

achieved with a value of 0, rather than 255,

since when 0 is decremented it rolls over to a

value of 255, resulting in 256 passes through

the loop.

Counting loops

The ﬁnal program for this instalment

uses only outputs. In this case, we

want to have all eight pins of port 1

output symmetrical square waves at

different frequencies. Our model for

this (in digital electronics) is an 8-

stage binary counter. A clock signal

is applied to the input, and its fre-

quency is divided by exactly 2 for

each stage of the counter. The micro-

controller generates the clock signal

itself using its program. The divider

chain can be obtained very simply

by incrementing a binary number.

This can be done either by using an

addition instruction or an inc (incre-

ment) instruction. The instruction

inc a increases the value stored in

the accumulator by 1 each time. The

program must execute this instruc-

tion repeatedly in a loop and repeat-

edly output the value of the accumu-

lator to the port.

The program shown in Listing 3

ﬁrst loads an initial value of zero into

the accumulator. Following this, the

accumulator value is output to the

port, the value in the accumulator is

incremented, a jump is made to

next for the next port output and so

010208 - 3 - 16

Figure 6. A circuit with an input and an

output.

tor potentiometer or LDR is con-

nected to the input. There is a defi-

nite gap between the switch-on and

switch-off brightness of the light

falling on the LDR.

If S2 is closed, feedback gives rise

to a rapidly alternating succession of

signal levels. The only visible effect

of this is that the brightness of the

LED is reduced, but an oscilloscope

reveals the true situation and shows

a sequence of fast rectangular sig-

nals. Naturally, this is exactly what

should be expected, since the input

state is only polled once at the

beginning of the program. With feed-

back, the microcontroller will always

find the output state resulting from

the previous pass through the loop

at this location, and that is exactly

the opposite of the previous input

state. The microcontroller thus has

no other choice than to repeatedly

switch the port state back and forth.

(010208-3)

This completes our introduction to assembler.

The next instalment will start with BASIC-52.

010208 - 3 - 17

Figure 7. Output signals on P1.0 through P1.7.

2/2002

Elektor Electronics

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: