Instead they have been replaced with semiconductor ICs which can store a vast amount of data in a very small footprint. Once an active line is selected the 4-bit data encoded by the diodes will appear as a binary output at the data lines D 1-D 4.ĭiode matrices are rarely used in modern times as they are difficult to scale up due to the physical constraints of the hardware. The data can be addressed (in this example using a decimal addressing system) by enabling one of the push switches to select the active line of diodes (you can think of each line or row of diodes as a read-only register).
Diagram of a 4-bit Diode Matrixįigure 2 shows the layout of intersecting diodes, which are used in combination with a pull-down resistor configuration to 'encode' the data permanently. Figure 2 shows the construction of a ROM diode matrix using the data values presented in figure 1. Therefore, in its most basic form, it can be constructed as a simple diode matrix - a popular technique for creating ROMs in the 1960s and 70s.
This type of memory is also referred to as 'non-volatile', as it retains its data after it is powered off.Īs ROMs are used for the permanent storage of data, it doesn't require a mechanism for continually writing data to the memory array. In other words - you cannot continually write new data to the memory - the information it contains is fixed before operation. As per its name, it is a type of memory which is readable only. Read Only Memory is the simpler of the two major memory types. Of course, this is a simplified and generalised representation of how memory is structured, and to understand some of the complexities and differences in memory structures we must first understand the two major memory types: 'Random Access Memory' and 'Read Only Memory'. Each register, or memory cell, can store a value of 4-bits (values stored in memory are often referred to as 'words' - in this case a 4-bit word), and can be addressed by a unique 3-bit binary value. Basic Memory: Addressing an array of 8 x 4-bit registersįigure 1 shows a simple memory structure.
A computer's memory, as we know it today, can contain a vast array of registers or 'memory cells', which can be individually referenced (addressed) to recall the value stored within them. In our previous note: From Logic Gates to Registers: Exploring the 74HC173 we explored how registers - a fundamental digital electronics circuit - can be made from basic logic gates to remember, or store, a binary value. Perhaps the most important part of this excerpt is the idea that "memory is equivalent to thousands of registers". Excerpt from 'Digital Computer Electronics' by P.Malvino The memory is equivalent to thousands of registers, each storing a binary word." Figure 1. The memory is therefore one of the most active parts of a computer, storing not only the program and data but processed data as well.
During a computer run, the control section may store partial answers in the memory, similar to the way we use paper to record our work. " The memory of a computer is where the program and data are stored before the calculations begin. To answer these questions lets first define, in modern terms, what we mean by a store, or the memory of a computational machine: Of course, in the 21st century we'd recognise this as 'computer memory', but in the 19th century this really was a groundbreaking idea.īut how does one go about mechanising memory? How does a thing remember other things? How did Babbage's ideas turn into reality? The latter - a store - would retain the numbers and instructions required to define the successive stages in computation. Charles Babbage, the famous 19th century English mathematician and polymath, once said that for a machine to perform the functions of a human computer it must possess three things: a unit capable of performing the operations of arithmetic, a built-in power of judgement and a store.