How Computers Work
Part 3: Chips
Those of you old enough to remember things like transistor radios or Macintosh tube amps for your stereo can picture this with ease: Everything in that big old chassis has been made into a “chip” or IC (integrated circuit). Some chips are as small as a finger nail, others are as big as a cracker or cookie. We call this a solid state component.
Transistors or diodes were among the first solid state components and we talked about the diode in part one when we looked at the power supply. A diode consists of two layers (called N and P) of semi-metallic material (hence we call them a semiconductor) each of which is laced (or “doped”) with a special material like arsenic. One of these layers has a lot of free electrons or can deliver them, the other one needs them. A diode with just two layers passes electricity only in one direction, which is how the rectifier diode in a power supply changes the alternating wall current (it moves in two directions) into a single direction direct (or DC) current the computer needs.
A transistor, on the other hand, has three layers and is made as either P – N – P or N – P – N.
In the beginning we didn’t think big so we made these devices one by one. Then one day some smart engineer put forth the idea that you can create a big surface area coated with various layers of P and N in various configurations, put little connecting wires to these areas and make many transistors or diodes on a single device. This was the birth of the IC or integrated circuit.
These devices are basically made photographically. Engineers draw out the layers like a landscape on paper. The thickness of the various N and P layers and their configurations determine the use of the specific area (different transistor values do different jobs). Then a device akin to your television picture tube coats a silicon wafer (or chip) with thin coatings of semiconductor material using an electron beam.
Transistors work by switching current from one direction to another based on where you apply power (voltage). We call the ability to do this process a “flip-flop”
Computer chips are far more complex then just a little radio or amplifier, but the heart and soul of a computer is the concept of a flip-flop. A door or gate that opens or closes.
A Japanese company in the late 1960’s came to an American chip maker with an idea to create an integrated circuit (IC) that would do math calculations, replacing the mechanical adding machine. After discussing things, the Americans suggest a far more complex chip than just a simple adder. One that was capable of doing more functions down the line. The result was called the 4040 and it became one of the first pocket calculator chips using 4 and 8 bit technology.
A bit is a single flip-flop. Four bits means that four flip-flops are linked together to work in tandem. In an 8 bit chip it means that 8 flip-flops work together. The reason this is important is that computers work on binary number principals. The value of one and zero or on and off. A flip-flop. Where flip is one (or on) and flop is zero (or off). If you connect a string of these together with some built-in instructions for doing math then you can deal with numbers larger than one and zero. An 8 bit flip-flop can handle up to the number 255 or go from –127 to plus 128. If you join several banks of these together or break large numbers into smaller pieces then you can work with values up to several million if there are instructions for doing this process.
That is what a smart chip was about. Holding values in memory, breaking numbers into smaller pieces or working with other banks of 8 bit flip-flops at the same time for big number crunching.
Each of these banks that do actual processing are called registers. They are connected to a set of instructions for doing simple math: Addition, subtraction, multiplication and division. You can even add in complex functions like a square root or the value of PI (3.1415 which is used to compute circles).
So the first calculator has an “instruction set” for doing math and several banks of tandem 8 bit registers for dealing with values. These were connected to a keypad and visual display. A button on the keypad connected several wires to the flip flops on the chip, which gave power to the individual transistors who then either did a “flip” or a “flop.” When you pressed a key like + it connected to the “instruction set” which told the registers how to move (which is either a flip, a flop or a shift where they pass a value to the left or right bit).
Think of your hands and fingers. We often count on our fingers. Each finger is a BIT and each hand is a BANK or REGISTER of multiple bits. Putting a finger up is a FLIP and down is a FLOP. You can count large numbers by using one hand to do the counting and the other hand to remember how many times you counted. This allows you to compute up to the number 25 using just 10 fingers
Also added to the chip was a process dating back some years called “Boolean logic or algebra” which works on the basis that things can be explained by a simple yes or no comparison expressed as: AND, OR, NOT and XOR. To better understand this think of how you do an internet search with a tool like Overture to find a name like Issues Magazine. Overture’s internal logic looks for any instance of “issues” AND “magazine” even if the are not together. If you change this to “issues” OR “magazine” then any page with either word would be returned.
So now we had a set of instructions that allowed us to do math (called the math instruction set) and to make logical comparisons (the logic set).
The next generations was the 8080 and this used 8 and 16 bit technology and they added a far more complex instruction set that allowed the chip to actually do custom work via instructions from the user, returning an answer to the user. This chip also had more registers or places to store, hold and process information. This, then, became the first true computer chip and was used in the Altair computer of the 1970s and later any computer calling itself an “80” machine or based on a Z80 chip, such as the TRS-80 from Radio Shack.
Next issue we will continue to look at the computer chip and how is has evolved into the modern machine we now take for granted.