 Modern data processing systems like these use thousands of magnetic cores. These cores are small, rugged, and they consume less power than most of the other devices that are used in these systems. You'll also find magnetic cores in automatic equipment employed for communication, fire control, cryptography, and high-speed computation. What are magnetic cores? They are tiny rings of nickel alloy or other magnetic materials. They have replaced vacuum tubes for many important functions in data processing systems. Cores come in a variety of sizes and in two basic types. The ferrite type of core is molded from compounds that magnetize easily and is used to provide data processing systems with their memories. The metallic ribbon type consists of a non-magnetic spool wound with a thin ribbon of highly magnetic alloy. This type is used in operating circuits of ADP systems. Both types of magnetic cores are remarkably adaptable to the language of data processing. Many ADP systems use the binary system of numbering, which employs just two digits, zero and one. These two digits are called binary digits or bits for short. A bit is either a zero or a one, nothing else. Why do we use binary? Because a bit can be represented by any bistable device and a magnetic core is a bistable device. Bistable because it can be magnetized in two directions. When the magnetic lines of force point in one direction around the core, they are a positive magnetic flux and the core is said to hold a binary one. If the direction of magnetism is reversed, the flux becomes negative and the core holds a binary zero. However, remember that this arbitrary definition of positive magnetic flux representing binary one and negative magnetic flux representing binary zero may be reversed in some systems. So far we've been establishing the function of magnetic cores. They store binary digits or bits, zero or one. Now we're going to see how they do it and also how they transfer bits from one core to another. To explain how cores store binary digits, we must know how they are constructed. Around the core is wrapped a conductor. Current flowing through the winding produces a magnetic field around the conductor. Because the core is made of highly permeable material, it forms a path for the magnetic lines of force. Because it is ring shaped, the path it forms is a closed loop. The conductor here is wound so that current flowing in this direction produces a positive flux, setting the core in the binary one state. Current flowing in the opposite direction produces a negative flux, which sets the core in the binary zero state. To show the polarity of the winding in a core diagram, we mark one side with a black dot. By definition, current that flows into the winding from the dot side sets the core to zero. Current flowing into the non-dot side sets the core to binary one. Remember, non-dot current for one and dot current for zero. Any winding around the core obeys the laws of electromagnetism. The magnetizing force it exerts is proportional to the amount of current flowing through it measured in amperes times the number of turns in the coil. Or, as it is usually expressed, the ampere turns. Since the number of turns in this winding is constant, we can simplify the rule here. The magnetizing force exerted by the winding on the core is proportional to the winding's current. Suppose we increase the amount of current. Will the total magnetism of the core also increase? It will, but not proportionately, as we'll see on this graph. We're going to take a minute and trace the magnetization curve of a core. Because in this way we can point up some of the characteristics that make cores so useful. We measure the magnetism of the core in terms of its flux density, which is the number of magnetic lines of force per square inch. The horizontal axis will represent the magnetizing force of the winding. Remember that the magnetizing force is the cause and the flux density is the effect. Initially, a relatively small amount of magnetizing force produces a relatively high flux density. The curve on the graph will begin like this. But as we increase the magnetizing force, the curve flattens out because we are adding proportionately less to the flux density. Further increases in magnetizing force will have no appreciable effect on flux density. The symbol for magnetizing force is the letter H. The symbol for flux density is the letter B. The relationship between magnetizing force and flux density for any given core material is represented by its BH curve. This is the point of saturation. The core is holding just about all the magnetism it can hold. We've been applying a positive magnetizing force here which produced a positive flux density and brought the core to a state of positive saturation. If we now reduce or stop the magnetizing force for all practical purposes, there is no change. The core material retains most of its magnetism. Here's how we plot this fact on the graph. The BH curve is brought back to the zero line on the horizontal scale. Watch the difference in flux density on the vertical scale near the plus B. It's negligible. This then is the amount of remaining magnetism. This amount is known as the residual magnetism of the core material. The magnetism that remains after the force that brought it about is removed. When non-dot current sets the core to the binary one state, the residual property keeps it that way indefinitely. Suppose there is a second pulse of non-dot current. It will have no practical effect on the core because the core is already saturated with positive magnetism. But if we send current through the other way into the dot side of the winding, then of course the core state will change. To plot the effect, we'll extend our graph. With dot current, we're applying a negative magnetizing force represented by minus H. Negative flux density is symbolized by minus B. We need a certain amount of negative magnetizing force to overcome the existing positive residual magnetism in the core. The amount needed is called the coercive force. At this point, the material contains no magnetism. The flux density is zero. Adding more negative magnetizing force produces a negative flux density. And again, the core is brought to a state of saturation, negative saturation. If we reduce or stop the magnetizing force now, there will be little effect on the flux density of the core. The BH curve returns to zero on the horizontal magnetizing force scale, but it remains at almost the saturation level on the vertical flux density scale. In other words, there is residual magnetism in the negative direction too. Therefore, when dot current is fed into the core, magnetism builds up and the core is set to binary zero. Residual property keeps it there. To switch the core, we apply a positive magnetizing force. Enough first to cancel out the negative residual magnetism. That is to act as a coercive force in the opposite direction. And then, by adding more positive magnetizing force, we can bring the core material back to a state of positive saturation. We have now plotted the positive and negative magnetization of a core. The original S-shaped curve is produced only when we are first magnetizing the core from a completely un-magnetized state. From then on, we are concerned only with the rectangular loop, which tells us the limits of saturation that can be reached with dot and non-dot current. Also, the amount of magnetism that will remain in the core after the current stops. The word hysteresis comes from a Greek verb, meaning to lag behind. A hysteresis loop, therefore, shows how much magnetism remains behind after the force that produced it is gone. The shape of the loop indicates the ability of a particular core material to retain magnetism. Shallow rounded loop, low residual magnetism. The core of this type obviously requires less force during a switch from one state to the other. Rectangular loop, high residual magnetism. A greater amount of force needed. If a core with a rectangular hysteresis loop is in the binary zero state, we're going to have to apply a substantial pulse of non-dot current in order to set it to binary one. A lesser pulse just won't make the switch. Now, how do we get information out of a core? By means of a second winding. Let's follow the transfer of information from input to output. Assume the core is in a state of positive residual magnetism. In the binary one state. To change the core to binary zero, we send current into the dot side of the input winding. Dot current switches the core to binary zero, reversing the direction of the magnetic lines of force. If we diagrammed it, we'd see the positive residual field collapse entirely and a negative field build up to saturation. Anytime a magnetic field builds up or collapses across a conductor like this winding, the movement of the lines of force induces a voltage in the winding and current flows. Observe the direction of current flow out the top of the winding. The way these two conductors are wound, a pulse of current sent into the bottom of the input winding results in a pulse of current going out of the top of the output winding. We used a dot to keep track of the polarity of the input winding and a similar dot will establish the polarity of the output winding. This gives us another rule about the dots in a magnetic core diagram. When dot current moves into the input winding, it causes the current to move out the dot side of the output winding provided the input current produces a change in flux. For example, our core is now set at binary zero, so a second pulse of dot current will have no practical effect. The core is already at the level of negative residual magnetism. There is no change in flux and therefore no output current. However, if we send current in the opposite way into the non-dot side of the input winding, the direction of flux in the core is reversed, switching it to binary one and produces a current that flows out of the non-dot side of the output winding. To repeat, a non-dot current input results in a non-dot current output. Remember these two rules. Dot current in results in binary zero and dot current out. Non-dot current in results in binary one and non-dot current out. Now, suppose we want to send non-dot current into the core, setting it to the binary one state, but we don't want to send current out, at least not yet. By connecting in a second core on the right, we can see more clearly what the problem is. Now, the first core's output winding is directly connected to the input of the second core. This input has been wound, so its polarity is the same as that of the input of the first core. The dot side is below. In a typical core circuit, current flows in only one direction, into the input of the first core. We'll take the non-dot side for our example. We'll assume that initially, the left core is set to zero and the right core to one. All we want to do is store a binary one in the left core. Watch what happens with the circuit as we have it. Current flows into the non-dot side of the first core's input. The first core is switched to one as we wanted, but this produces current to flow out of the non-dot side of the first core's output along the circuit and into the dot side of the input winding on the second core, switching that core to zero, which we did not want. If we are going to control the transfer of information from one core to another, we need something in the circuit that will block the flow of the current and do it selectively. The device commonly used is a crystal diode, which has a high resistance to current flow in one direction, the back resistance, and practically no resistance at all in the forward direction. Any unidirectional device or arrangement will serve the purpose and will symbolize this with a general symbol for one direction flow. The arrow points in the direction in which it allows the current to flow. Current coming the other way meets the back resistance of this device and is blocked. With such a device in position in the circuit, the second core is effectively isolated from the first. We'll create the same conditions we had before. The first core is initially set to binary zero and the second to binary one. We send non-dot current into the input winding of the first core, switching it to one. But due to the back resistance of the electronic device, no current can flow against the arrowhead. No current, no flux change, no change in the second core from binary one to binary zero. Thus we have changed the binary state of the first core and retained the binary state of the second core by the use of this unidirectional device. How can we switch either core to zero? We've said that current flows into the input of the first core on the non-dot side only and this would be typical in a core circuit. The device will allow current to flow through it in its forward direction only, as we've set it up here. That means that current can flow into the non-dot side of the second core's input but not into the dot side. What we need for dot current is a shift winding on each core. Wound so the dot side is at the left in our diagram. The shift winding shifts or transfers bits from one core to the next, as these two will demonstrate, but we'll complete our diagram first by giving the second core an output winding that is identical with the first core's output. And for this example, we'll assume that both cores are in the zero state. All right, we'll follow the events in their proper sequence. A non-dot input sets the first core to one. Nothing happens to the second core. Current flow in this direction sets a core to zero, but that's where this core is already. We can hold the binary one in the first core as long as we wish. When we're ready to transfer it out, we send a pulse of current into the dot side of the shift winding. And three things happen. The first core switches to zero. Current flows out of the dot side of the first core's output through the one-way device and into the non-dot side of the second core's input. And the second core switches to one. We have now transferred a binary one from the first core to the second and controlled the time of the transfer to the exact moment we wanted. The binary one may be stored here, or again, a shift pulse may be applied to make the transfer to the next core along the line, thus clearing this core back to binary zero. However, there is one final problem to consider here. A shift pulse will induce a voltage not only in the output winding, but also in the input winding. Current will flow out of the dot side of the input back toward the first core. The one-way device won't stop it because it's going around the circuit in the direction it allows. This backward flow of information would, if unchecked, switch the first core, interfering with its operation. It is checked, however, by the design of the output winding. The output winding of the first core has many turns in contrast to the input winding of the second core, which has relatively few turns. The ratio may be as high as 12 to 1. Consequently, the small current produced in the second core's input meets a high impedance when it flows back to the first core's output. There just isn't enough current to provide the force necessary to switch the first core. We have been diagramming the single diode transfer loop, the simplest and most basic of the circuits that transfer bits from one core to other cores. The diagram we have been building up is a schematic, but there's an easier way to represent these loops. Logical diagrams eliminate the complex details found in schematics by concentrating on what the magnetic core circuit does. It's logic for reason for being there. Only a few symbols are needed for logical diagrams, and they're easily understood. A circle represents a core. An arrow pointing into the circle is an input. An arrow pointing away, an output. A binary digit inside the core indicates its state, 0 or 1. If the bit is placed in front of an input arrow, it shows the state to which that input will set the core, in this case 1. A bit placed behind an output arrow means that when the core is switched to this state, an output will be induced. Here the core must be switched to 1 for an output to occur. If we remember that initially, the core is either at 0 or 1, we'll see that these arrows and bits tell us four things about this circuit. If the core is initially 0, the input switches it to 1, and an output occurs. If the core is initially 1, there will be no switch. Therefore, no output, since there will be no change in the flux of the core. Other symbols used for logical diagrams are just as simple. For example, if we wish to show an input that will set the core to 0, using direct current, the arrowhead is closed. In our discussion, we've been dealing with pulse current, shown by open-head arrows. Double arrowheads mean an override input, an input that will hold the core to the indicated state, despite the simultaneous presence of other inputs. The letter T is used to designate time. This override pulse will hold the core to 1 at a moment in time identified as T1. Other letters serve to label individual inputs and outputs where several may be present. We can relate an input to an output with the capital and small form of the same letter. Thus here, where the core is initially 0, an input Q comes in at time T1 and switches it to 1, an output occurs. This output is identified by small Q. Now let's take another look at the logical diagram of the single diode transfer loop. It gives us the whole story very quickly. An input pulse sets the first core to binary 1, unless it's there already. The core stays at 1 until a shift pulse comes in at time T1 and switches the core to 0. This switch induces an output pulse in the first core's output winding, which causes the second core's input winding to set that core to 1. In effect, a binary 1 has been transferred from the first core to the second, where it's stored until at time T2, another shift pulse clears the second core back to 0. We can draw logical diagrams like this of any magnetic core circuit, and we depend on them for a clear understanding of the work done by the circuit. In this picture, we've been dealing with the basic properties of magnetic cores, memory units, and a variety of operating circuits. Later, you'll be seeing how they work. They have demonstrated their efficiency and reliability in dozens of applications. They conserve space and power. They operate with precision at high speed. In fact, the development of the modern computer has been due in large part to the techniques devised for using these remarkable cores.