Tonight is the first of what will be a multi-part session on Digital Signal Processing (DSP). Unlike Lee's impedance series we won't be building a set of concepts leading to DSP. DSP isn't a single concept; rather it is a catch-all term for several distinct and only loosely related methods. My approach to teaching DSP will be to break it down by how it is used in Ham Radio rather than by theoretical construct.
Also, DSP is used in many fields today besides Ham Radio but we will limit our discussions to Ham Radio. Of course, if you have knowledge you would like to share about non-Ham uses of DSP, you are welcome to share it but I won't be going into non-Ham uses myself.
Let's start with a very basic idea of what DSP is. First of all, the signal that is being processed, usually starts out analog and ends up analog. Whether it is the analog audio signal received by your microphone that ends up as an analog radio wave transmission, or an analog radio wave reception that ends up as an analog audio wave coming out of the speaker, it is still analog at both ends and digital only in the processing circuitry. The exception to this is the ever growing list of digital modes that start off as digital information, in the form of characters, are converted to analog for radio transmission, and end up as characters again.
Once you are representing the signal digitally you can do your digital processing. Probably the most well known use is to create better filters than you can with analog circuits. Other uses are to create displays showing signals on a frequency line so that you can tune to the signal you want, or even just point and click in some cases. Less well known but equally important is that DSP is used to convert the audio signal from your microphone to the SSB, AM or FM that is sent out. Probably the ultimate use of DSP is the Software Defined Radio. I will go into these uses in detail in subsequent sessions.
Right now I am going to go into more detail on the conversion between Analog and Digital. As I said earlier, in order to do digital signal processing, you must convert the signal from analog to digital, then do your processing, then convert it back to analog again. These steps are known as Analog to Digital Conversion (ADC) and Digital to Analog Conversion (DAC). The same abbreviations are used for the circuits that perform the steps. Usually these circuits are combined into a single Integrated Circuit, or chip, so you will commonly hear about an ADC or a DAC chip. These conversions are done by the time-slice method. In this method, voltage measurements are taken of the analog wave at regular time intervals. You will need several measurements per wavelength in order to accurately describe the analog signal. The rule of thumb for for a simple sine wave is that the frequency of taking voltage measurements should be at least twice as high as the frequency of the wave you are measuring. Keep in mind that complex waveforms can be thought of as the sum of sine waves of different frequencies and amplitudes. So to accurately represent a complex waveform your sample frequency must be at least twice as high as the highest frequency sine wave that is a component of your waveform. Typically when the voltage is measured it is stored as a 16 bit binary value. Allowing for positive and negative that gives 32,768 voltage levels from the smallest measurable above zero to the maximum. This affects the dynamic range of your system, that is, how much difference you can have between the smallest and largest amplitudes. In Software Defined Radios where a PC is an integral part of the radio, you can use floating point processing to greatly increase the number of voltage levels represented and thus increase dynamic range at least for internal processing. Ultimately it must be converted back to analog through a DAC which will probably be only 16 bit.
As we continue on and go into more detail about digital signal processing, keep in mind that for voice communication we always start with an analog audio signal, have an analog wave traveling through space and end with an analog audio signal on the other end. It is only the processing internal to the sending radio and receiving radio that works in digital.
Wednesday, May 20, 2009
DSP 1: Introduction to Digital Signal Processing for Ham Radio, Bob, no. 52
Posted by Bob Helling at 4:56 PM
Subscribe to: Post Comments (Atom)
Post a Comment