As an example, let’s assume that we are running a program which after using address 2600H we would like it to use some data stored in address 3800H. We could do this by loading the 16-bit number 1200H into the X register and instruct the microprocessor to go to the address written as PC+IX (program counter and index X). In this case, the microprocessor goes to address 2600H+1200H=3800H to retrieve the data to be used. This jump from 2600H to 3800H is called an offset or we may say that the index X register has provided an offset of 1200H. In this example, index register Y could have been used equally well.
Index registers are often used to input data previously entered into a table or from another part of the memory. The user of a program may be asked to type in a number. This number would be entered in RAM but the program that is going to use this data is stored in ROM. A suitable offset in one of the index registers could shift the address sufficiently to bring it into the RAM area of the memory map. These registers are also used in situations where the program seeks a response from the user. The screen may show:
We could implement this by storing the response in one of the index registers and, as the printer performs each copy, the number stored is decreased by one (decremented) until it reaches zero. Every time that the counter is decremented, we instruct the microprocessor to check the state of the Z flag. When the count finally reaches zero, the Z flag will be set and the printer can be stopped.
R counter (R)
This is a simple counter in which the lower seven bits are used to count the number of instructions that have been carried out by the microprocessor while it has been running the present program. It is reset to zero whenever the micro is restarted.
Figure 8.7 shows the remaining parts of a microprocessor. All these parts are inside the single block marked ‘microprocessor’ (µP) in Figure 7.4. These include all the facilities for connecting the device to external circuits and a few other essentials.
Figure 8.7 The Z80180 microprocessor
Data buffers
The data buffer is an 8-bit register to store the information being provided by the external data bus. Once the buffer stores the binary information then tri-state buffers can lock-in the information disconnecting the external data bus so it can be used for carrying other data. This process is called ‘latching’. As a general rule, all inputs to a chip from a bus are latched in. This is because the buses take time to set up new voltages on their lines and need to get started on this job as soon as possible.
The address and data buses
As with all microprocessors, there are three buses, the address, data and control. As we saw in the previous chapter, the control bus is a loose collection of connections that send signals around the system to make connections and control the operation of each area as necessary. Here we will concentrate on the address and data buses in the Z80180.
The Z80180 has a (sort of)16-bit address bus and an 8-bit data bus.
In the smaller microprocessors, it was usual practice to make the address bus twice the width of the data bus but there is no reason why this has to be the case. The choice of address bus width will depend on the anticipated size of the memory that needs to be addressed in the final designs. Unusually, in this microprocessor there are two conflicting requirements: it has to remain compatible with the 16-bit address bus of the Z80 but we would like to be able to address 1 MB which requires 20 address lines.
These requirements are met by the block shown as MMU or memory management unit.
Memory management unit (MMU)
This circuit includes an 8-bit register that holds address information that can be controlled by the software being used. These 8 bits are combined with the top four address lines, A12–A15, of the address bus. The lower half of these 8 bits are added to the top 4 bits of the address bus and the upper 4 bits are used to expand the useable memory by adding the four extra address lines, A16–A19, to allow a total of twenty lines to be addressed – see note below. The process is shown in Figure 8.8.
Figure 8.8 Expanding the address bus
The Z80180 family is manufactured in three different packages. The original 40 pins of the old Z80 has increased to 64 pins in the dual-inline (d.i.l.) package similar to that shown at the top of Figure 1.6. They also make a 68 pin plastic leaded chip carrier (PLCC) version which is just like the square package in Figure 1.6. Finally, there is an 80 pin surface mount type.
Surface mount devices do not have pins that go through holes in the printed circuit board but have connectors that come out horizontally and then soldered onto the surface of the board. This method allows significant space savings.
The smaller number of pins on the d.i.l. layout has limited the number of address pins to nineteen and therefore the maximum memory that can be accessed is now 219 or 512k (really 524288 since a ‘k’ in this case is taken as 1024 as we met in Chapter 2).
Address buffers
These are tri-state buffers just like the data buffers except they are oneway devices. The microprocessor sends addresses out along the address bus but no address information can come in this way. If we need to send an address into the microprocessor then we are loading information and all information whatever it is goes into the data bus which, as we have seen, can send any data out to the external circuits or accept any data into the microprocessor.
Clock generator
As we saw in Chapter 7, a microprocessor needs a square wave signal to keep all the internal operations in step otherwise it will all end in chaos with data and instructions moving at the wrong times and getting jumbled up.
In this particular microprocessor the clock speed can be set to values between 6 MHz and 33 MHz. The clock signal originates from either a crystal or an external signal source – but never at the same time. Using an external signal can be useful where the microprocessor is used as part of a larger installation and this would allow the microprocessor timing to be governed by the surrounding circuitry. There are three nice things about increasing the frequency of a crystaclass="underline" they get smaller, lighter and cheaper. The clock circuitry includes a divide-by-two circuit to allow the use of a double-frequency crystal with the above-mentioned benefits.
You may remember our sad little story of our microprocessor-based system controlling the printer while all around the office was burning. It stopped its main program and went off to phone the fire services, alert the maintenance staff, activate sprinklers etc.
Fire is not the only hazard our little friend can safeguard us from. It could warn for other things like the paper running out in the printer or data corruption on the telephone cable.
Now, we would not want it to send for more paper to deal with a possible problem with the telephone cable so the microprocessor needs to be told what the problem is and what to do about it. The different programs for each of these problems are stored in a group of addresses in a ROM chip. As an example, we could load the programs at these addresses:
0800–0855H = paper supplies
0870–08A8H = telephone data.
Both of these programs have 08 as the high byte of the address. This value would be stored in the interrupt vector register. The low bytes 00 and 70 are supplied by the sensing device. As soon as the telephone data is corrupted it is noticed by a sensor, and it sends a signal to interrupt the microprocessor together with the 70H.