火曜日, 1月 24, 2006

On RF basics

I am reading a book by Andrew Miceli titled 'Wireless technicians handbook' by artech house publishers. It has a very good review in the beginning which is more of what I was looking for. Software radio started getting over my head, and when you are spending one hour per sentence at Wikipedia its time to rethink my self-educational approach to wireless.

It occured to me that their are other people less fortunate than me to have free access to most books in print or online (at least concerning tech.) The prices at the bookstore for some of these specialized books are prohibitive to say the least, and I remember not long ago, spending hours at Barnes n Noble, or the UW engineering stacks, because I didn't quite know how to access information freely (so to speak).

So I am including an excerpt from this book, without the express permission of its author--Which now days sounds like some henious crime. I think they can stick copyrights up their arse.(that is until the send me a letter and tell me to remove it B-)

"In order to understand the complexities of today's digital wireless formats, it is essential that technicians and engineers have a very sound foundation in basic RF and digital principals. The technicians who completely understand these fundamentals have a tremendous advantage over their peers. This will explain many of the common concepts of RF as well as review some commonly used digital modulation techniques.

Domains: Time and Frequency
A word used in the electronic world quite often is domain. If we look at the basic XY chart, the X component is called the domain. Essentially it is the basis for the graphical representation of the item--in our case, an RF signal. In RF we often talk of looking at the amount of power in a signal in two domains (latter we will see more domains): the time domain and the frequency domain. Understanding these domains is useful in understanding exactly what an RF signal is.

If you have used an oscilloscope than you have seen the time domain. If we look at the signal on the oscilloscope from left to right, what we are looking at is the time span. From bottom to top is the amount of voltage. RF, like your house current, travels in waves in the time domain.

In this domain, we can see that we can measure the position of the wave in degrees as it moves from point to point in time, much the same way we can measure an angle moving around a circle. We call this measurement phase, and, like a circle a signal can have a phase of up to 3600. The amount of time it takes for the signal to repeat itself, or get back to it's original phase state, is called the period, which is a time measurement (time per one cycle). If we inverse this period we end up with the frequency (cycles per 1 second), in which the unit of measurement is Hertz (Hz).

Examining and measureing the phase of a signal is quite important in communications. When we speak of a phase measurement, it is generally with respect to a second signal. For instance, this book will discuss multipath signals (pp) (cool constellation diagram)

[it is amazing to me how much real information you can turn up when you scroogle with -com after whatever it is you are trying to find]

In RF design we are able to seperate signals with different frequencies to different applications so they will not interfere with each other. This is seen much easier in the second domain--the frequency domain. (somewhat unrelated link).

If you have ever used a spectrum analyzer, then you have seen the frequency domain. In the frequency domain, we measure the amount of power at each frequency. What is important to understand is that the power of a signal is shown in a frequency domain. In the frequency domain, it is easy to see how we can divide applications by a specific frequency.. For instance in North America, the cellular phone system is assigned a specific band (around 800-900 MHz) in which to operate, while frequency modulation (FM) radio stations are assigned different frequencies. With advanced mobile phone service (AMPS), the analog cellular phone system used in North America, each mobile phone conversation is then subdivided and then assigned its own frequency band, which is actually 30 KHz wide. (mac spectrum analyzer)

Power
Once we understand the domain in which we are looking at a signal, of coarse it makes sence to look at the power component (in other words, the "Y" component of the graph) in the various domains. If you are like most technicians, you are probably very familiar with the oscilloscope. Oscilloscopes usually measure a signal's voltage. In RF and microwave applications, we generally measure a signals power with one of two units of measure: the watt or the dBm.

While power is always voltage times current (which gives us the wattage), the levels in RF are often so small that it is much easier to represent the wattage logarithmically. The dBm is an absolute measurement--0 dBm always represents 1 mW. When we use dB, it is a relative measurement--in other words a 3-dB increase would represent a doubling of power, regardless of the starting power level. A very easy way to convert between the two is to remember that 1 mW equals 0 dBm. Then, every 3-dBm increase or decrease is a doubling or halving of the power, respectively. You can see why you must have a good understanding of dBm vs watts--the difference in power from 37dBm to 40 dBm is quite significant, while the difference from 0 dBm to 3-dBm is but 1 mW.

We also sometimes represent power levels in terms of their signal-to-noise (S/N) ratio. This is important, as many times the ability to recieve a signal is not necessarily tied exclusively to the absolute power of the signal at the recieve antenna, but the amount of power relative to the noise floor. For instance if the noise floor in a particular area is particularly high, due to interfering signals or just thermal noise, the S/N ratio will be lower, representing this condition. With the same amount of power in another area, the noise floor might be much lower, meaning a higher S/N ratio and a better condition to try to recieve a signal.

When it comes to spread spectrum systems (which will be discussed later), a modification of the S/N ratio is performed to create a different measurement, called
Eb/No (pronounced 'ebno' by those in the field).Eb/No is often called the digital S/N ratio. It is essentially the S/N ratio with what is called the processing gain factored in. This is the benifit derived from using a larger bandwidth to transmit than is actually needed.

1 Comments:

Blogger Mad Russian the Natural Philospher said...

Stu,

I love this stuff on rf.

About security, freedom, et al...
You have no idea what we have to go through as workers here at the airport. Management is watching every minute, we have cameras everywhere, we have badges with the predecessor to rfid embedded in them which record the exact location and time when we pass through different parts of the airport. The airlines will implement rfid on the baggage tags as soon as the cost per tag reaches $0.05 each (currently at $0.25 each). Fortunately we are allowed a few minutes on the computer during our break (with close tracking of course).

I live with a fascist hierarchy you have no idea about (the price to gain some small measure of economic freedom).

Jim.;)

6:08 午後  

コメントを投稿

Links to this post:

リンクを作成

<< Home