In the 1960s, electronics started to awake from its slumber that had used thermionic valve technology that was recognizably similar to circuits that had been built for thirty years. The pace of progress was gentle. The first semiconductor material was developed and the transistor came into use in just a few years. The photographic process used to design and produce the transistor quickly led to simple integrated circuits and the microprocessor.
No sooner had the microprocessor and the associated memories arrived in 1971 than it became obvious that the microprocessor was always accompanied by other circuits, like input/output devices, memory and timing circuits so it would be a good move to combine them into a single device.
We had a choice – we could keep everything general and universal and call it a microprocessor or design it for a single purpose and call it a microcontroller.
The multipurpose devices went into computers and even here we had a choice. Computers were either ‘microcomputers’ where price was a significant feature and these microprocessors had some built-in ROM and RAM. Soon, however, speed became the main feature as the prices began to fall and we could afford to equip our homes with computing power which equalled many offices of just a few years previously and the microprocessors became expensive and fast. Speed headlines drive the publicity machines as home and office computers became faster and faster. They sold in their millions. Meanwhile the single-purpose devices, really the descendents of the early microcomputers, were developed further and made really cheaply, sold by the billion and were never mentioned. They power the pocket calculator, video recorders, cameras, microwaves, washing machines and greetings cards that play music – in fact almost anything vaguely electronic.
Just a thought
The microcontrollers outnumber the population of the world many times and as mentioned earlier we are likely to be sharing our homes with, possibly, fifty of them. They are in every essential industry – food production, transport, communications, research, weaponry, power generation, medicine, heating and air conditioning – there is little that we rely on that does not use a microcontroller. If they learn to communicate independently of us, they may develop their own agenda. Now there’s a thought.
Once we have learned to drive, most vehicles are easily recognized as being very similar. We are happy with the general idea and can concentrate on the minor differences. Microcontrollers are much the same. Having already become familiar with the basic building blocks of the simple microprocessor in Chapter 8, we can move very easily into the microcontroller. It is not surprising then to find that all microcontrollers are basically very similar.
To give an overall impression of the range of microcontrollers available we are going to look at three popular ranges. The first is the 8051, probably the most widely used microcontroller, over twenty years old and continuously developed by many different companies and showing no signs of fading away. The next is from the AVR family produced by the Atmel Corporation, one of the leaders in this field. From this range we look at the AT90S/LS2343 one that is small, modern and RISC. The final one will be explored in Chapter 16.
Probably the transition between the microcomputer to the microcontroller occurred with the Intel 8048 as we saw in Chapter 11. The 8048 added on-chip RAM, ROM and a timer so it could be used as a single purpose device such as controlling a keyboard – it was, in fact, a microcontroller.
With the experience gained by using this, it became apparent that there was a significant market for a microcontroller. In 1980, Intel launched the 8051 which, twenty-three years later, is alive and well. In fact very well indeed. It is probably the most popular microcontroller ever. It is made by about 44 suppliers. These suppliers have often added some extra features to make versions or ‘variants’ as they are called particularly suitable for specific jobs. There are at least 92 variants all compatible with the original code. Even within variants, there are a series of options that lifts the total number of members of the 8051 family to several hundred.
Numbering
The device numbering is not very obvious as many microcontrollers are available from several different suppliers with their own product code. They then produce a group of basically similar devices with minor changes like different operating voltages or differing amount of RAM and ROM on-board memory – these groups are referred to as ‘families’.
The family is given a name which often has little connection with the product codes. For example, Intel’s 8051 family has the family name of MCS51. This contains the 803X, 805X, 875X and the low power versions bXC45X and the 8XCX52. As usual the X refers to any figure or letter in that position.
The situation is further confused (or possibly simplified) by referring to all of them as ‘the’ 8051.
The block diagram of the 8051
The block diagram shown in Figure 15.1 is the family portrait of the 8051 family. There are some features that differ between the family members, principally the memory configuration – some versions have less memory and some have none at all. However, we will look at the operation of our ‘middle-of-the-road’ version and worry about the individual differences later.
Figure 15.1 8051 block diagram
The 8051 pinout
As in all micro and digital chips, a line over a pin designation indicates that it is active low or, put more simply, to use this feature we need to apply zero volts.
The pinout shown in Figure 15.2 looks, at first glance to be rather complicated due to the dual use of many pins. This is a common feature of microcontrollers as a method of reducing the number of pins to be used. The more pins, the more expensive and the larger the device. This is bad news for a device often destined to be embedded within another circuit.
Figure 15.2 Pinout of the 8051
There are variants available that provide the increased number of pins so that there is a separate pin for each function.
Reset
Regardless of what program is being run, we must always be able to gain control of the microcontroller just as we must with a microprocessor. The procedure is just the same. The microcontroller has a reset pin which, in the 8051, is taken from the bus control block and, in normal operation must be held to zero volts. When a positive voltage over 2.0 V is applied to it the microcontroller immediately returns to its startup memory location, which in this case is 0000H. We can arrange this to occur automatically when the power is switched on but we should also provide a reset switch to gain control of the system at any time without removing the power. This is the ‘reset’ switch which we use when our computer locks up and ignores us.
When changing microcontrollers, remember to check the polarity of the reset voltage. Compare this circuit in Figure 15.3 with the one shown in Figure 8.7.
Figure 15.3 The reset switch
Clock input
As we have seen in Chapter 7, we are going to need a clock signal. The original design of the 8051 called for a 12 MHz crystal though later versions can run at 33 MHz, 40 MHz or even 44 MHz. As an alternative, we can use a ceramic resonator or an external signal. The clock input is shown in Figure 15.4 using a crystal. To use an external signal throw away the crystal and the capacitors then apply the external signal to the Xtal1 pin and leave Xtal2 disconnected.
Figure 15.4 The clock
Ports
We have four ports numbered 0 to 3. In the standard 8051 ports 0, 2 and 3 are dual purpose and port 1 is just an input/output port. In some variants, all ports are dual purpose.
With regard to the memory, we can use the on-chip memory, which keeps the circuit as simple as possible and makes the embedding of the device at its most convenient.
The on-chip memory may be masked ROM or an EPROM which we met in Chapter 6. To use external memory obviously increases the chip count but can allow up to 64 kB of memory, which is the standard size for 8-bit microprocessors.
The external memory is accessed by taking the ‘external access’ (EA) pin to a logic zero. To switch the external ROM on, the output enable (OE) pin of the external memory is taken low by the program store enable (PSEN) pin of the 8051. In a similar way, an external RAM is accessed for reading and writing by the read (RD) and write (WR) pins applying a logic zero voltage to the output enable (OE) and read/write (R/W) pins respectively.
Interrupts
The 8051 has a total of five interrupt signals. Two of these can be externally generated and three are of internal origin. Interrupts are discussed more generally in the chapter dealing with interfacing but basically, when an interrupt occurs, the program which is running at the time is interrupted and another code sequence is ran, called an interrupt service routine (ISR). When the ISR is compete, the microcontroller returns to its original program and continues as if the interruption has not occurred. There is a different ISR for each interrupt which must be pre-loaded in specific memory addresses, all ready to go if needed.
What if two interrupts occur at the same time?
The interrupts are checked continuously in what we call the polling order. Starting from the top, the order in the 8051 is:
External Interrupt 0
Timer 0
External Interrupt 1
Timer 1
Serial port
In addition, each interrupt can be given two levels of priority so if two interrupts occur, the one with the high priority will be handled first. If two have the same priority, it is decided by the polling order.
Timer/counters
A timer or counter is a series of bistables or flip-flops that change state once for every input signal, thus one of these circuits would divide the input frequency by a factor of two. If this signal is then fed into the next bistable, the output is ¼ of the original frequency. The next circuit would have an output of 1/8, then 1/16 and so on.
There are two timers, T0 and T1. These can be programmed to divide by 256, 8192 or 65536 and will generate an interrupt signal upon completion that can be detected by the software. One of the modes allows the timer/counter in 8-bit (divide by 256) mode to reload and start counting again each time continuously.
The input signal being counted can originate from an external circuit so it counts the number of incoming pulses, or it can use an internal signal which is actually 1/12 of the clock frequency in use. As mentioned above, it generates an interrupt signal when it reaches its maximum value. We can preload the timer with a number to start counting from. This will allow the interrupt to be generated after any required number of events, or time interval.
Serial port
Since the microcontroller normally handles data eight bits at a time, it is operating in parallel but two receive or transmit serial data we have perform serial/parallel conversions. This is always achieved by using a shift register working under the control of a clock signal. The working of a shift register is described in Chapter 17.
The serial port is able to transmit and receive data, at the same time, this is referred to as a full duplex system. As the name suggests, the bits of data are moved one after the other in a continuous stream. Pin 10 is called RXD or receive data and pin 11 is the TXD, transmit data but, as is often the case, things are not quite that simple.
It can operate in three ways, or modes numbered 0, 1, 2 and 3.
Mode 0
This is the case that spoils the simple RXD, TXD as, in this mode only, the TXD pin is actually used as a clock signal and the RXD is used to receive or transmit data. The clock frequency is fixed at 1/12 of the onboard oscillator frequency and this, of course, determines the speed at which the data is transferred via an 8-bit shift register.
Mode 1
This mode also sends data in 8-bit lumps but its frequency is adjustable and operates as an 8-bit UART (universal asynchronous receiver transmitter – see more about this in Chapter 17). The 8 bits are increased to 10 bits buy adding a logic 0 to indicate the start of the group and a logic 1 to mark the end of the group. This is the normal format used in RS232 transmissions. Unfortunately, the output voltages do not comply with the RS232 standards so an external chip must be added to do the voltage conversion. Some suitable chips are discussed in Chapter 17.
Mode 2
This is very similar to Mode 1 except the number of bits transmitted is increased from 10 bits to 11. The extra bit can be used as a parity bit which is used to check for transmission errors, The pattern is Start bit (0), 8 data bits, parity bit and stop bit (1). (Have a look at Chapter 17 again.) The transmission rate can be 1/32 or 1/64 of the onboard oscillator frequency.
Mode 3
This is an 11-bit transmission with a programmable baud rate. The baud rate is near enough the same as the transmission rate measured in bits per second.
Watchdog timer
When a microcontroller is embedded in equipment it may find itself used in areas where electrical interference is a problem. This, or a software problem can cause the microcontroller to lock up by getting into an endless loop. The watchdog timer will reset the microcontroller after a period of time, about 20 milliseconds, unless it is told not to. Stuck in a software loop, it no longer generates this ‘don’t reset me’ signal and so the microcontroller is reset and escapes from the loop.
The watchdog timer is only fitted into some of the 8051 variants but is available as a stand-alone chip but is commonplace in newer designs.
When we want to leave the microcontroller in a continuous loop, we have a choice of ensuring that the loop contains the required software code or disabling the watchdog timer before the loop is started.
Taking the AT90S/LS2343 as an example, we can see a really basic microcontroller with minimal complexity yet having many useful features that make it inexpensive, small and comparatively fast. The RISC design and the width of the registers allows the vast majority of instructions to be executed in a single clock cycle and all the others, apart from five, are completed in two cycles.
In Figure 15.5, we have a block diagram. It is quite a relief to see that mostly it is quite familiar after looking at the earlier 8-bit microprocessors and the 8051 considered in previous chapters. Already we see that most devices are a combination of one or two innovative ideas added to a standard mix.
Figure 15.5 AVRAT90S/LS2343
Inputs and outputs
One notable feature of the block diagram, we see only five input/outputs shown as PortB 0 to 4 so we have only five connections for data which is a small number until we look at the sister version, the AT90S/LS2323 which has only three – PortB 0 to 2! Yet they are called 8-bit devices and in previous 8-bit micros we have grown to expect at least one and sometimes two 8-bit input/output connections. The answer is just that the ‘8-bit’ description refers to the internal data bus width.
The PortB pins can be used to send data in either direction, they can be used as inputs or outputs. As is common with other microprocessors and microcontrollers, the direction of data movement through each of the PortB lines is individually controlled by a data-direction register. Loading a ‘one’ into the data-direction register will make the corresponding PortB line into an output, on the other hand a ‘zero’, of course, will change it into an input.
Memories
This microcontroller has three memories, or four if we count the general purpose registers. The loading of the memory storage areas employs a serial transfer of data and is achieved by the SPI (serial programming interface).
One ROM storage area is achieved by a Flash memory organized as 1k×16. The program instructions are all either 16 or 32 bits wide and can be cleared and reprogrammed whilst remaining in circuit. It can be cleared and reprogrammed at least 1000 times. The contents of the Flash memory cannot be changed by the program being executed by the CPU and so is free from accidental corruption. The data is held in an EEPROM, which again can be cleared and reprogrammed electronically without removal from the device. It only holds 128 bytes of memory but is able to go through at least 100 000 cycles.
Data corruption can occur in the EEPROM if the supply voltage is reduced too far but this effect can be avoided by any one of the following three methods. The first we have already mentioned – use the Flash memory for critical data. The other two methods are ways to detect the reduction in voltage and immediately put the microcontroller into a safe condition. This is often referred to as ‘brown out protection’. An external circuit detects the falling voltage and applies a low voltage to the reset pin which effectively switches the chip off until the supply voltage recovers. The alternative is to put the microcontroller into a power-down sleep mode which is a power saving mode which has the effect of preventing any decoding or execution of any instructions – which, of course, precludes any ‘writes’ to the EEPROM.
It also has 128 bytes of SRAM (Static RAM) for the temporary storage of data and 32 8-bit general purpose registers that can be connected two at a time with the ALU (arithmetic and logic unit) which is the heart of the ‘brain’ within the microcontroller.
Clock
The AT90S/LS2343 has an internal RC oscillator which runs at 1 MHz, 4 MHz or 10 MHz depending on the version in use. It is one of the few of the micro devices that does not make use of an external crystal although it can use an external clock pulse. This external clock pulse only requires a single pin and hence we have an extra pin to use as an output.
Interrupts
There are only three interrupts. The first, and highest priority, is the reset which is activated by a low voltage applied to pin 1, a power-on reset or a signal from the watchdog. The next is an external interrupt request as described in a moment. Lastly, an overflow from the timer/counter circuit.
Pinout and package
These are shown in Figure 15.6 and we can see that it is available as an 8-pin DIL (dual in line) package which has two lines of pins and also the surface mount version, plastic gull wing SOIC (small outline IC).
Figure 15.6 AT90S/LS2343 pinout
Pin 1 – Active low reset. Must go low for at least 50 ns.
Pin 2 – External clock signal input or PortB 3
Pin 3 – PortB 4. All lines can sink 20 mA and therefore are able to power LEDs directly. Sinking means that the LED or other load is connected between the positive Vcc supply and a low voltage output at the port.
Pin 4 – Ground.
Pin 5 – PortB 0 or MOSI. In serial programming mode, MOSI is the serial data input.
Pin 6 – PortB 1 or MISO/INT0. In serial programming mode, MISO is the serial data output. This pin can also act as the external interrupt described in the previous paragraph.
Pin 7 – PortB 2 or SCK/T0. In serial programming mode, SCK is the serial clock input. This pin can also provide the timer/counter0 clock input.
Pin 8 – Vcc. The LS version requires a positive supply voltage that remains in the range 2.7–6.0 V.
Sleep modes
When the microcontroller is not being used, it can switch off some of its circuitry to save power. Sleep modes are employed in all modern microcontrollers and make an enormous difference to the overall life of an intermittently used device. This enables sealed units in toys and greetings cards to remain active for months or years.
Idle-mode
To see the benefits, this microcontroller has a normal operating current drain of 2.4 mA but when switched to the ‘idle’ mode, the current falls to 0.5 mA. It does this by stopping the CPU activity but allows the timer/counter, watchdog and interrupts to remain operational. This is about an 80% power reduction but we can do a lot better than that otherwise my musical socks would have stopped long ago.
Power-down mode
In this mode, only the external interrupt and watchdog (if enabled) continue to work and current falls to less than 1 microamp, which is a really impressive reduction in power. The microcontroller can be aroused from its sleep only by one of the following: an external reset, the watchdog (if enabled) or INTO external interrupt.
This is another modern RISC development and has many features that are similar to the AVR that we have just looked at. The AT90S/LS2343 was chosen as representative of the very small and basic microcontrollers found embedded in many products. This PIC16F84A example from Microchip Technology is a mid-range device which is larger and more capable than the AVR.
The PIC series ranges from a really simple 8-pin, 4 MHz microcontroller on a level with the AT90S/LS2343 that we have just considered up to a 40-pin 25 MHz device. As mentioned the PIC16F84A is a mid-size version that has 18 pins and runs up to 20 MHz.
The PICs are designed for easy use and are becoming increasingly popular as the first step into the world of microcontrollers. Microchip Technology provides a PICSTART™ PIC development system that provides, at a very reasonable price, an assembler, compiler, EPROM and EEPROM programmer, all the hardware manuals and even a sample PIC to play with. It should be mentioned that other companies have similar systems compatible with the PIC series and for other microcontrollers like the AVR and 8051 series.
It has proved to be such an easy, off the shelf, starting point that to many people ‘PIC’ is not only their first choice but is becoming used as a generic term for any microcontroller.
The general layout of the PIC16F84A is shown in Figure 15.7.
Figure 15.7 PIC16F84A
Supply voltage
The DC supply voltage must remain in the range 2.0–5.5 V for it to work happily though similar versions such as the PIC16F84 can run between 2.0 and 6.0 V.
Sleep mode
To save the current drain, a software instruction can put the PIC into a sleep mode. The supply current is very dependent on the clock frequency and is normally between 1 and 20 mA and when put to sleep the drain is reduced to approximately 1 µA. To obtain the lowest possible sleep current we should hold all I/O pins at VDD or VSS and disable any external clock and hold the master clear pin in a logic high state – not in the reset state. The purpose of all this is to prevent any voltages from floating up and down. If it did so, it would switch and the technology used results in very low currents drawn except at the moment of switching during which it causes a really high spike of current so the higher the frequency, the more often this spike occurs – hence the increased current.
The microprocessor will wake if: a reset occurs by a logic low voltage being applied to the MCLR pin, a wakeup pulse arrives from the watchdog unit (if it is enabled) or an ‘EEPROM write complete’ signal.
Memory
As usual, we have two blocks of memory. One is the program memory and the other is for data. For maximum speed, they each have a separate bus connection so that both memories can be accessed during a single clock cycle.
Program memory
The program memory is situated in the flash memory which is organized as 1028×14. All instructions in the PIC16 series use 14-bit instructions.
The reset vector points to address 0000H and the interrupt vector is 0004H so address locations 0005H to 03FFH are available for us to hold our programs.
Data memory
The data area is subdivided into two areas, the FSR (file select register) and the GP (general purpose registers) as shown in Figure 15.8.
Figure 15.8 Register file map
The SFR (special function register)
This register controls the operation of the CPU and involves such things as the input and output ports, EEPROM address and data, timer, program counter and that sort of housekeeping.
All the register files are 8-bits wide and are arranged in two banks called bank 0 and bank 1. We have to instruct the microcontroller as to which bank is to be used and this is done by using special instructions to access some of the page 1 registers. Those accessible are indicated in Figure 15.8. Microchip Technology are planning to remove the choice of using the OPTION and the two TRIS registers and suggest that the STATUS register is used instead. This does not affect the use with this chip but it will ensure that upgrading in the future will not require any modifications to the software.
I/O (input/output) ports
All outputs can source or sink 25 mA and can therefore power significant external circuits without further power amplifiers being required. Sinking means that the load is connected between the positive Vcc supply and a low voltage output at the port and sourcing is connecting the load between a positive output on the pin to the ground.
PortA and TRISA registers
PortA is a 5-bit wide bi-directional port, each line being individually controlled so some of the lines can be inputs whilst the others are outputs. The choice of input or output is made by loading a 0 (output) or a 1 (input) into the appropriate bit of the data direction register TRISA.
In common with other devices, when it first starts at power-on, the port is set as an input. This provides a safer option that running the risk of random information being sent out to whatever it is connected to.
PortB and TRISB registers
PortB is a 8-bit wide bi-directional port, each line being individually controlled using TRISB in the same manner as in PortA. Each of the PortB pins have a weak internal pull-up which can be switched on or off by the RBPU of the option register. The pull-ups are disabled when the port is being used as an output and also during power switchon. Any of the Pins RB4–RB7 that just happen to be configured as an input have an interrupt-on-change feature that can be useful. If any one or more of these pins have changed logic state since they were last read, it causes an RB port change interrupt. This interrupt can be used to wake the microcontroller from its ‘sleep’ mode.
Status register
This is very similar to the one we met when looking at the Z80 in Chapter 8.
The bits are:
Bit 0 is the C(carry/borrow) bit. 1 = a carry out from the MSB (most significant bit) of the result otherwise it is cleared to zero. For ‘borrow’ the values are reversed. Subtraction is carried out using the two’s-complement method that we met in Chapter 4.
Bit 1 reflects the carry situation that last occurred from the 4th bit of the result. This is also called the half-carry bit.
Bit 2 is the zero flag. It is set to ONE when the result of the last arithmetic or logic operation is ZERO. Be careful not to misread this.
Bit 3 goes to 0 after running the SLEEP instruction.
Bit 4 goes to 0 when a watchdog time-out has occurred.
Bit 5 is used to select between the two memory banks. It is cleared to 0 to access Bank 0 and set to 1 if we need access to Bank 1.
Bit 6 and bit 7 – not used. It will be used in the future so by programming them for 0, future compatibility will be assured. This may save a lot of time if our program is used on an upgraded version.
Option register
As the name suggests, it offers a series of options. One example is the control of the prescaler.
The Prescaler
Two functions are affected by the prescaler, they are the timer, TMR0 (timer zero) and the watchdog timer. Each of these circuits provides an output pulse after the count overflows and restarts from zero. In the case of the watchdog, the time interval is about 18 ms. If a longer time interval is needed, we have three alternatives. We can simply switch the watchdog off but, of course, we lose the benefits of the watchdog if the microcontroller gets caught in a loop. A simple way is to place the software code CLRWDT (clear watchdog timer) in the program at anytime before the end of the countdown so it is reset for another 18 ms and we can repeat this as necessary. Lastly, we can use the prescaler to reduce the frequency of the incoming pulses and hence increase the time before the output signal is generated.
Bits 0, 1 and 2 PS2:PS0 (prescaler rate select bits) provide eight alternative pulse rates for use by the watchdog or TMR0. Setting the three inputs to 000 will provide a clock signal which is equal to an external signal or chip oscillator. Changing the setting to 001 will halve the frequency (or double the time). Increasing the setting to 010 will decrease the frequency to one quarter of the original. As the count increases, the frequency will halve for each count until we reach the maximum bit value of 111 which will cause the watchdog time to increase by a factor of 128. As it happens, the TMR0 has a divide by two circuit built in all the time so a setting of 000 will halve the frequency and the maximum count will reduce the frequency by a factor of 256.
Bit 3 PSA – (prescaler assignment bit) – Unlike most other PICs, the prescaler can be applied to the TMR0 or the watchdog, but NOT both, so the option register controls the choice. Bit 3 of the option register is set to 0 to prescale the TMR0 and a 1 selects the watchdog.
The other bits
Bit 4 T0SE (TMR0 source edge select). This controls the moment of timing the clock input. A ‘0’ increments the count on the low-to-high transition and a ‘1’ increments on the high-to-low transition.
Bit 5 T0CS (TMR0 clock source select). This decides where the clock pulses come from. The choices are a ‘0’ to an internal clock as on the CLKOUT pin and a ‘1’ counts the transitions on the RA4/T0CKI (timer zero clock input). This option allows pulses to be generated by any external source like cans of beans moving along the conveyor belt or the revolutions of an engine.
Bit 6 INTEDG (interrupt edge select). This is similar to bit 4 except we are controlling the moment at which an interrupt signal is recognized of the RB0/INT pin. A ‘0’ uses the falling voltage edge and a ‘1’ sets the rising edge.
Bit 7 RBPU (PortB pull-up enable). This controls the ‘pull-ups’ on the PORTB output. A ‘0’ allows each line to have its own pull-up enabled or switched off as required. A ‘1’ switches them all off. A pull-up circuit is shown in Figure 15.9. When the switch is closed by applying a ‘0’ state to this pin it connects the output to the positive supply via a current limiting resistor. This ensures that in the absence of any input data the port will be pulled up to a positive state so that it doesn’t wander about applying unpredictable inputs. The higher the value of the resistance used, the easier an incoming voltage finds it to bring the voltage down and this is referred to as a ‘weak’ pull-up like this one. Likewise, a reduction in the resistance value will make the port input more determined to stay high and this is referred to as a ‘strong’ pull-up.
Figure 15.9 A pull-up circuit
PCL (program counter low)
PC is the program counter that keeps track of the instruction being executed at the time. It is a 13-bit register divided into PCL for low byte and PCH for high byte. Its behaviour is common with other microcontrollers and microprocessors as described in Chapter 8.
Stack
This register stores the return address when interrupts occur. It stores up to eight 13-bit addresses. Again, there are more details in Chapter 8.
Register 6–1: PIC16F84A configuration word
This is a special register which is just memory location 2007H which can only be accessed during programming. Unprogrammed pins are read as ‘1’.
It is a 13-bit word with the following options:
Bit 13–4. This protects the program code from being read after the programming is complete. 1 = no protection 0 = all program memory is code protected.
Bit 3 Power-up timer enabling. This allows a nominal 72 ms delay when first switching the microcontroller on to allow power supplies to settle. The delay can be activated by loading bit 3 with ‘0’, no delay = ‘1’.
Bit 2 watchdog timer, disable with ‘0’, enable with ‘1’.
Bits 1–0 oscillator selection bits
00 = LP oscillator
01 = HS oscillator
10 = XT oscillator
11 = RC oscillator
Oscillators
The crystal or ceramic resonator options are shown in Figure 15.10 and the crystal oscillators are divided into the frequency ranges:
LP low power crystals 32 kHz–200 kHz.
XT crystal/ceramic resonator 100 kHz–4 MHz.
HS high speed crystal/ceramic resonator 4 MHz–20 MHz.
RC resistor/capacitor.
Figure 15.10 Oscillator options
A note on ceramic resonators
These offer an alternative to crystals for oscillators. They are generally lighter, more shock resistant and are usually somewhat smaller and cheaper to buy. The downside is that they do not have quite the frequency accuracy or stability and in many respects they are positioned somewhere between an RC network and the quartz crystal.
An interrupt summary
The PIC16F84A has four sources of interrupts:
(i) External interrupt on the RB0/INT pin.
(ii) TMR0 overflow.
(iii) PortB change detector pins RB7–RB4
(iv) EEPROM data write complete interrupt. This is used only when we are loading new programs into the EEPROM.
In each case, choose the best option.
1 An 8-bit microcontroller has:
(a) an 8-bit data bus.
(b) eight ports.
(c) an 8-bit address bus.
(d) eight internal registers.
2 Compared with the Pentium 4, a microcontroller:
(a) is faster.
(b) consumes less current.
(c) is more complex.
(d) is larger.
3 watchdog circuit prevents:
(a) damage due to excessively high supply voltages.
(b) an overrun in timer/counter circuits.
(c) burglars.
(d) wastage of power while the microcontroller is not being used.
4 It is NOT essential for a microcontroller circuit to include:
(a) registers.
(b) a reset pin.
(c) some memory.
(d) a crystal.
5 The term ‘ISR’ refers to an:
(a) immediate service register.
(b) internal system reset.
(c) interrupt service routine.
(d) instruction set register.