Unit 8 “Modems”

1. Read and translate the following international words:

To communicate; to concentrate; to electrify; standard; technique; to manipulate; diagram.

II. Read the text and find English equivalents for the following Russian words and word combinations:

Обрабатывать информацию; двойной язык; связка единиц и нулей; более общий термин; произвольно; в обратном направлении; единственный недостаток.

What is a modem and how does it work?

Most people are aware of the fact that computers process information in binary language. If computer A wants to communicate with computer B, A will have to send a bunch of ones and zeros to B, in some meaningful (to them) order, across some medium. But how is it actually done?

Let us concentrate on the communications part. Taking things in order, we first must have something with which to connect the two machines. We'll run a wire between the two computers.

Next, we need to send a bunch of ones and zeros across that wire. We'll worry about making the stream meaningful later. First we electrify the wire. We'll pick a standard voltage of +5V DC. We'll make that equal to a "one." A "zero" will be represented by OV DC, or ground. So, by applying and removing voltage on the wire we have ran between the two machines, we can send across a bunch of ones and zeros. It looks like it

We did say something about making the stream meaningful, though. What if we want to send across fifteen ones in a row? For convenience let's refer to ones and zeros by the more generic term "bits." We can arbitrarily assume a standard transmission time for each bit of one second. So if machine A wants to send the stream "101101110100," it would take twelve seconds and look something like this:

Since the machines have agreed on sending a new bit once per second, all machine В has to do is wait until A has begun transmission, wait another 1/2 second (to ensure that it isn't trying to read the value of the line while machine A is changing it), and start checking the voltage on the wire each second thereafter, as shown by the hash-marks in the following diagram:

So now В knows what A is sending. How does В know when A is ready to send bits? How does it know when A is finished sending?

We need another wire.

This new wire will hang out at OV DC until A is ready to send data, then A will "take it high," or put +5V DC on it. When В notices that the new wire is now carrying 5 volts, it waits 1/2 second and starts reading bits. It continues to read a new bit every second until A takes the new wire low again (drops the voltage on the wire to OV DC), then stops.

We really are only looking at half the problem. A can talk to B, but В is going to have to talk to A, if only to let A know В has received the data. We can add two more wires providing the same functions but in the opposite direction.

This is very inefficient, but it would work.

Requiring four wires to implement bi-directional (sometimes called "full duplex") communication would not win an award from anyone.

We could do it with two.

Say we chose, instead of+5V DC and OV DC, +5V DC and -5V DC. We made the machines agree that once A took its line high, В would wait a predetermined amount of time and start reading. Knowing the amount of time В would wait, A would wait 1/2 second less and start transmitting. Communication in the other direction would work the same way.

We could do it with one.

The above paragraph describes a technique that works just as well with one wire as with two, with a small exception. Whoever takes the line high becomes the transmitter, the other machine becomes the receiver. The only drawback is we can only operate in half-duplex mode, since both machines can't be manipulating the voltage on the same wire at the same time.

In reality there are many ways for machines to decide what is a one and what is a zero, and many more ways of sending said ones and zeros all over the place. In reality even an inexpensive modem can handle data transmission rates 19,200 times faster than our example.