Last Updated on October 13, 2005 by Ralph Brandi
A frequent complaint that SW listeners and DXers have had for years has been that receiver manufacturers don’t listen to what hobbyists say they need and want in receivers. I’m pleased to say that the R. L. Drake Company has invalidated that argument with their R8 series of communications receivers.
Permalink | Comments off
Last Updated on December 8, 2005 by Ralph Brandi
During an international broadcasting conference in 1997, Thomcast, a unit of the French company Thompson, demonstrated its “Skywave 2000” system for conference attendees. This was the first public broadcast of digital short-wave radio that actually used the ionosphere to bounce the signal back to earth. This month we will learn about that experiment and some of the technology behind it.
Every morning during the exhibition, Radio France International was digitally broadcast from the TDF (TéléDiffusion de France) SW station located in the center of France (Issoudun) and received at the Thomcast exhibit at the IBC show in Amsterdam.
This transmission used a standard Thomcast TRE 2355 pulse duration modulated, 500 kW, SW transmitter without any modifications. The digital program was supplied by a Skywave 2000 coder/modulator which is based on Thomcast’s proprietary modem design.
Skywave 2000 was tested under real HF environmental conditions. The digital program was transmitted on 6175 kHz and fit in a standard 10 kHz ITU channel.
The test was performed in the presence of multipath reflections from both the E and F layers (incident angles of 17° and 40°). The vertical plane antenna pattern did not discriminate between the two reflections in order to provide a demonstration that stressed the multipath discrimination capability of the equipment and processing software.
Multipath reflections are a concern to digital system designers. The paths will have slightly different delay times because one signal travels farther than the other. This delay can cause errors in the reception of digital signals. Multipath reflections cause a degradation known as “intersymbol interference”. Simply, this effect can be pictured as a smearing of the digital transitions so the detector has trouble accurately figuring out which state was sent.
In part 3 of this series I mentioned Shannon’s law of information transmission theory. Shannon defined the maximum data rate that could be sent over a channel as a function of the signal to noise ratio and the bandwidth. Communications engineers have been working for the past 50 years to develop modulation and coding techniques which allow their systems to approach closer to Shannon’s ideal capacity limit. They found, just as Shannon predicted, you can send more information in a given bandwidth by using advanced modulation and coding techniques if a high enough signal to noise ratio can be maintained.
One technique engineers have developed to get closer to Shannon’s ideal is known as Quadrature Amplitude Modulation (QAM). QAM is used in telephone modems like the kind you may have on your computer. QAM allows engineers to squeeze 56 kbps digital data rates down a phone line originally designed to only pass 3 kHz analog voice signals. Of course telephone line signal to noise ratio and phase stability are normally much better than those experienced on short-wave radio links. Today, engineers are working to design a system that can tolerate the errors that the short-wave propagation medium causes.
Here is how QAM works in principle. Before I get into this too deeply, I should clearly define some terms. The signal is characterized and the bandwidth is determined by the number of signaling intervals or transitions that are transmitted each second. Each interval is called a baud so the number of baud per second can be described as the “baud rate”. The units are transitions per second. The term “bps” stands for bits per second. So, bps is equal to the bauds times the number of bits per baud.
A now obsolete definition of a baud was that a baud was identical to the number of bits per second. Using this old definition, “baud rate” would have units of bits per second per second, a forever increasing number of bits per second. Clearly “baud rate” was an incorrect usage back when one baud was equal to one bit per second.
This definition evolution over the past 40 years has caused much confusion among technical people trying to communicate with each other about digital signaling. It all depends upon when somebody last read up on the subject. The terms “bit rate” and “baud” are often confused because early modems transmitted only 1 bit per baud. An early 1200 baud telephone modem also transmitted 1200 bps.
These days, we need higher speeds. So the objective is to try and “pack” as many bits as you can into a baud. A modem operating at 9600 bps is still only transmitting at 1200 baud. But it is “packing” 8 bits into each baud. Thus 9600 bps equals 1200 baud times 8 bits per baud. One technique for “packing” multiple bits into a baud is called quadrature amplitude modulation.
QAM uses simultaneous amplitude modulation of two carriers that are 90 degrees apart in phase to transmit multiple bits per baud. An unmodulated signal exhibits only two possible states, ON or OFF, allowing us only to transmit a zero or a one. (Morse code transmission using ON/OFF keying is an example of this.) With QAM, it is possible to transmit many more bits for each amplitude state.
This scheme works by adding amplitude modulated cosine and sine waves. These two components, being 90° out of phase, are said to be in quadrature. Hence the name Quadrature Amplitude Modulation. By simultaneously amplitude modulating both signals, we can transmit more bits for every transition or baud. One convenient way to represent the possible states is to use a constellation pattern diagram such as the one shown in Figure 1 where amplitude is represented by the distance away from the axes’ intersection and phase is the angle measured counterclockwise from the 0 degree reference. By convention the zero angle reference is normally to the right.
Figure 1: A signal with this phase/amplitude structure is called 8 QAM because it has the ability to signify 8 different states.
In this pattern we see that the states are represented at different amplitudes and phases. Points on the 0, 90, 180, and 270 degree axes can each have two possible amplitudes resulting in eight different states. With eight unique states, it is possible to transmit 3 bits in every state as shown in Table 1.
Table 1 shows a table of amplitude and phase states which can generate 8QAM and the corresponding binary bit pattern each state might convey. For example, if the modulated signal is of amplitude 1 at 0 degrees, the receiver decoder senses three zeros (000).
| Amplitude | Phase | Bit Pattern |
|---|---|---|
| 1 | 0 | 000 |
| 2 | 0 | 001 |
| 1 | 90 | 010 |
| 2 | 90 | 011 |
| 1 | 180 | 100 |
| 2 | 180 | 101 |
| 1 | 270 | 110 |
| 2 | 270 | 111 |
Table 1: Possible states and the resulting bit patterns for 8 QAM
If the modulated signal starts as a 1 at 90 degrees, then goes to a 2 at 270 degrees, then a 1 at 0 degrees, and then a 1 at 180 degrees, the resulting bit pattern is (010111000100). In only 4 bauds, twelve bits have been transmitted.
Today, modern communications needs require modulation schemes exhibiting more dense constellation patterns such as the one shown in Figure 2.
Figure 2: Four bits per baud 16 QAM
Figure 2 depicts a 16 state constellation pattern resulting in the transmission of four bits in every baud. The number of states grow exponentially to the number of bits transmitted. This is a disadvantage since transmitting eight bits per baud would require 256 possible states resulting in a very dense constellation pattern.
As the constellation pattern becomes more dense, the receiver has more trouble distinguishing which position is the real one that was sent. Amplitude noise or phase disturbances, caused by the transmission path, have more impact as the pattern becomes more dense. So one consequence of increasing the packing of more bits into a baud is that the signal to noise ratio must be increased to prevent errors from rising excessively. Forward error correction techniques can also help. Shannon was right. You can trade signal to noise ratio and coding efficiencies for an enhanced data rate.
During the Thomcast experiment, different transmission modes were tested:
Thomcast is now working on investigating different types of existing AM transmitters and how they work with the Skywave 2000 technique. Their objective is to develop a plug-in exciter that broadcasters can simply connect to their existing transmitters. They are also working on the implementation of an improved source coder to increase the audio quality.
In the next part of this series we will look at how the bit error rate varies with transmission channel degradation. Until then stay tuned.
Permalink | Comments off