The fundamentals of Digital modulation for radio
In this article, I will attempt to walk you through some of the basics and also take a look at the genesis and evolution of digital radio communication. The average radio amateurs who come from various academic backgrounds and may not have an electronic or electrical communication leaning often find it difficult to relate themselves to the stereotypical explanations of the various nuts and bolts of the digital communication model. The objective is to keep this narrative intuitive and simple without dwelling into the associated mathematics as much as possible. Other than the use of digital modulation methods to implement digital voice communication and file transfer of documents, images, etc, the most common digital modulation methods have been used very effectively to implement text mode communication.
It is sometimes felt, especially amongst a significant section of the amateur radio community that the digital text and data modes of radio communication uses special RF modulation methods in the transceiver in comparison to what is done to produce analog radio transmissions. This is not entirely true. Let us try to clear the mist and examine the important fundamentals as we go along.
Is it Digital Carrier modulation or Sub-carrier modulation?
In a typical amateur radio digital communication scenario, the transceiver used is essentially the same analog transceiver that we use for regular radio-telephony. The RF carrier modulation to generate the transmitter output feeding into the antenna system remains the same. Hence, one can safely conclude that the eventual carrier still resorts to the standard analog modulation methods. So, now the question is, what is it that makes it digital?
It is the payload at the input of the transmitter modulator that contains digitally coded information. Rather than a proportionately time-variant analog payload like a varying voltage that follows the voice as in the case of radio-telephony, the input to the transmitter, in the case of the digital payload has only a few pre-defined discrete levels. For instance, in the case of a binary digital data stream, there are only two discrete levels of signal voltage transition.
How can we efficiently modulate a digital data stream comprising of these discrete voltage levels? Although theoretically, we could directly modulate the digital data directly in a typical SSB transmitter, that will only produce a train of pulsed RF bursts from the transmitter. This would result in a not so efficient modulation method. The modulation type would be restricted to On-Off Keying (OOK) or what we might call Amplitude-Shift Keying (ASK).
To make digital modulation more flexible, robust, and immune to noise and data stream corruption, the modern digital modulation methods resort to other means. Typically, we choose to replace the audio (AF) input to the transmitter that would normally come from a microphone in the case of radio-telephony with a specially processed audio frequency signal that we call the digitally modulated AF sub-carrier.
Digitally modulated AF Sub-carrier
The method of sub-carrier modulation provides us with a great amount of flexibility to tailor the overall performance characteristics of each and every different digital modulation modes in vogue as of date. Some of these sub-carrier modulation techniques focus on increasing the data-carrying capacity within a limited channel bandwidth, while others might sacrifice the channel capacity in favor of higher reliability and communication robustness. The choice lies entirely with the designer of the modulation mode.
Due to the vast flexibility offered by sub-carrier modulation, today we have at our disposal hundreds of modulation variants. Many of these have attained widespread acceptance and have been duly standardized. For instance, as we are all familiar, we have PSK31, PSK63, RTTY, AMTOR, PACTOR, MFSK, MSK, MT63, GMSK, QPSK, QAM, M-ary modes, Contestia, Hellschrieber, Olivia, FT4, FT8, WS65, FSK441, etc, etc… and the list goes on. Each of these digital modulation protocols have their unique strengths and limitations. Therefore, they are used under different types of communication environments and conditions. One thing that they all have in common is that they all resort to sub-carrier modulation of an AF tone and restrict their bandwidth to the narrow baseband channel bandwidth of a typical analog radio communication transceiver.
The data-carrying capacity of a digital channel is determined by its bandwidth and the desired SNR in a noisy communication environment at the receiver end. This is well established and governed by Shannon’s mathematical theory of communication or more precisely the Shannon-Hartley theorum. Going into the details of Shannon’s theory is beyond the scope of this article since it would involve quite a bit of mathematics but for the moment it would suffice to know that there is a trade-off between the quantum of data (information) or the rate of data carried by a communication medium and the specified bandwidth. In the case of amateur radio, where the typical bandwidth of the AF baseband determines the limits of the sub-carrier bandwidth and consequently the upper limit of data-carrying capacity.
What are the options for digital sub-carrier modulation?
After having introduced the concept of sub-carrier modulation for digital radio, let us examine as to what are the parameters of a typical AF sub-carrier to achieve desired results. An unmodulated AF sub-carrier is just a pure homogeneous sinusoidal (Sine) wave at a discrete audio frequency. The sub-carrier by itself carries no information. The information (data) needs to be loaded (modulated) onto the sine-wave sub-carrier so that it can piggy-back itself onto the sub-carrier and ride along. It is like riding on horseback where the sub-carrier is the horse and the digital data (in our case) is the rider or the payload.
The question is, how do we achieve the above-cited objective? In principle, the idea is pretty simple… A homogenous electrical sine-wave (sub-carrier) has three distinct parameters that could be varied in consonance with the (binary) data stream to achieve modulation. These variations that are modulated at the transmitter are replicated at the receiver after due process of demodulation. Thereafter, another round of demodulation of the obtained sub-carrier yields a replica of the original data stream structure that was used while transmitting.
What are these three parameters of the electrical sine-wave? They are the amplitude, the frequency, and the phase. modulating any of these three parameters effectively modulates the sine-wave and allows it to carry the information payload as piggy-back.
Physical electronic hardware circuits, as well as mathematical algorithms that run on computers, may both be used to obtain the necessary parametric modulation of the sinusoidal sub-carrier. For the purpose of typical amateur radio digital sub-carrier modulation, it has now more-or-less become a standard practice to use computer software to this rather than to use dedicated electronic hardware. To get further clarity, let us examine a few typical digital radio modes. For instance, the well known RTTY modulates the frequency parameter of the sub-carrier to produce frequency deviations from its original value in accordance with the binary voltage levels of the RTTY code. This results in what is known as frequency-shift keying (FSK). Similarly, PSK31 (binary phase-shift keying BPSK) which is also very common, instead of varying the frequency per se, varies the phase of the sub-carrier. This is termed as phase-shift keying (PSK). In both these cases, the amplitude of the sub-carrier remains constant and hence they are often called constant Envelope modulation modes.
The average or RMS power of the modulated signal remains constant during transmission. These modes are far more immune to unwanted degradation of data integrity because the radio noise typically alters the amplitude of the RF far more than the phase or frequency of the signal. In the case of FSK or PSK, even though the RF transmission channel might have been subjected to the onslaught of a noisy propagation environment, it is easier to segregate and eliminate the noise at the receiver end and retrieve the original quality sub-carrier. Amplitude shift keying (ASK) is rarely used for machine-generated digital modes, however, the traditional CW morse code transmission over radio uses ASK (OOK).
Before we move on, let us also recognize the fact that digital data for radio transmission may not always be binary. Multi-level or M-ary data streams are also often used in some of the methods to pack more information by enhancing the effective data throughput. Though these methods have their own merits, they also usually have a downside of being less immune to noise interference.
The Digitization ProcessTypically, the regular text-mode digital transmission does not normally require to undergo any digitization process. The reason is that the native text data is already in digital format. For instance, a text message would be coded as per the ASCII or UTF8 format or any other variant. The PSK31 uses a code table called Varicode format. The TX-side encoder and the RX-side decoder simply use the Varicode lookup table for translation to form characters.
However, though pictures and data files may already be available in digital format if they reside on a computer, there may be instances in the case of a picture coming from an analog source like a camera that would need to b digitized before being subjected to sub-carrier modulation for transmitting over the radio. In the same way, digital voice (DV) will also require analog-to-digital conversion (ADC) on the transmitter side and a reciprocal digital-to-analog conversion (DAC) at the receiver.
These processes of converting between analog and digital equivalent or vice-a-versa, whenever needed, are done by means of various digitizing methods. The process involves sampling and quantization of the analog signal for ADC, while DAC at the receiver end may often employ a bit-weighted integrator of some form. Dedicated, high-performance hardware integrated chips, as well as efficient and fast software algorithms, are available for the purpose.
The sharp-cornered square pulses of a digital bit stream are not ideal for use in sub-carrier modulation for radio transmission. The sharp edges resulting from fast rise and fall time of the pulses generate strong harmonics that would require wider transmission bandwidth. To mitigate the situation Pulse shaping techniques are employed to optimize bandwidth requirement. These shaped pulses convey the same information, yet they use smaller bandwidth. The pulse edge roll-off (rounded) characteristics determine bandwidth. Some common methods of digital pulse stream shaping that are employed along with their typical applications are as follows…
- Root-raised Cosine (used in QPSK)
- Half-Sinusoid (used in FSK)
- Gaussian pulse shaping (used in GMSK)
How does digital modulation retain better data integrity?
Unlike analog signal modulation that essentially depends only on its inherent robustness to negotiate a hostile and noisy radio communication environment, digital modulation methods may be designed in ways to provide additional checks and balance and also quite often have some means of detecting and correcting error that might get introduced during propagation. This adds an extra dimension to digital modes and makes it attractive for radio communication.
For instance, a text data stream bing transmitted over radio may be designed to briefly pause after transmitting a few characters to allow the receiver to verify its integrity. To do this, very often an extra bit of data is introduced by the transmitter side known as the Parity Bit. The RX side does a quick checksum calculation on the received bitstream (packet) and compares it with the parity bit. If any bit changed its character during propagation, then the parity-bit checksum error occurs.
Now, the receiving side would revert back to the transmitter through Automatic Repeat Request (ARQ) and ask it to re-send the segment of the corrupted data string. The TX re-sends the data which again undergoes a parity test at the receiver. If all is fine, the RX asks the TX to send the next segment of the data stream. This process goes on continuously in the background ensuring received data integrity. This process is called Hand-Shaking. Though the received data is more reliable, the downside is that it takes additional time to conduct handshaking. This results in a reduction in effective transmission speed.
What I cited above is a very basic method. The modern digital signal transmission protocols have evolved extensively. Without going into their details in this introductory article on digital modulation, it might suffice to mention that various methods using Forward Error Correction (FEC) that do not require aggressive handshaking have been created. FEC, within its boundaries, can successfully correct a certain amount of error without referring back to the transmitter for a re-transmission. Other methods like Cyclic Redundancy Check (CRC) is often used in larger file transmissions by embedding a special code along with the data block. Hybrid schemes using a combination of several methods may also be used.
Though not applicable to amateur radio, for extreme conditions requiring uncompromising reliability where ARQ methods may not be possible as in the case of deep-space communication over millions of kilometers in space, robust self-correcting algorithms like Viterbi-decoded Convolutional codes, Golay codes, and Reed-muller codes have been used. In some instance, the concatenated Reed–Solomon–Viterbi (RSV) code have also been used. Nowadays, the Turbo code or Low-Density Parity Check (LDPC) code has been found to be extremely robust. These codes allow a high degree of data integrity in a very noisy communication environment. Their performance nearly approaches the Shannon limit of maximum channel capacity… However, that’s another story for another day.
A typical digtal radio communication transmission model?
So far we have tried to understand the basic principles behind the process of digital signal modulation by the way of its implementation using a low frequency (AF) sub-carrier within the bandwidth limits of the baseband of a typical radio communication channel. However, we now need to focus on the next step of the process that would render it viable for transmission over a radio frequency communication medium.
Please recall what we mentioned earlier. We had stated that despite the fact that the information payload at the input of the transmission chain is digital (binary) in nature, the final modulated RF that is produced by a transmitter and fed to the antenna is analog. This fact is vital and is vital and is the cornerstone of the fundamental principle of typical digital radio communication as used almost invariably, especially in the context of amateur radio.
As we have discussed so far, the original incoming digital (binary) data stream carrying the information payload is applied to act upon an AF sinusoidal carrier called sub-carrier to alter either one or more of its basic characteristics in consonance with the prevailing state of the digital data stream. This modulation of the sub-carrier may be in the form of either varying its amplitude, frequency, or phase to produce ASK, FSK, or PSK sub-carriers respectively. However, the sub-carrier cannot be effectively transmitted or propagated over a radio communication medium.
So far so good… What happens next? Now comes the time to do the rest of the magic. At this stage, we must recognize that the composition of the modulated sub-carrier is similar to any other baseband AF signal. Barring the fact that the payload content of the baseband signal rather than being voice-related AF voltage as in the case of radiotelephony, in this case, contains AF voltage swings related to digital data information. The nature of the variation of spectral composition observed on the modulated sub-carrier is quite similar to the baseband signal of a radiotelephony system.
Therefore, beyond the baseband stage, the transmitter can safely treat both analog radiotelephony signals as well as the digital sub-carrier in an absolutely similar way. As a consequence, all the technical signal handling, processing, and final modulation methods further down the transmitter chain can be (is) identical for both analog and digital radio transmission modes. The reverse but identical process is applicable in the case of the receiver on the other side of the radio communication circuit.
The bottom line is that the core transmitter, receiver, or transceiver used either for analog or digital radio communication is identical and usually the same. The transceiver architecture does not matter. A typical radio transceiver, whether it is AM, SSB, FM, or whatever can work equally well with both analog or digital baseband input signals. Digital radio only requires an additional pre-processing stage to covert (by sub-carrier modulation) a digital data stream to an equivalent AF baseband signal to conform to the specifications of a regular transceiver.
The associated illustration above clearly cites the underlying principle. An amateur radio operator making a debut into the world of digital communication need not feel overwhelmed or intimidated. The commonly used digital text communication in amateur radio is as simple to work as regular CW or SSB voice. The only additional factor is AF sub-carrier generation using external pre-processing software that runs on any PC. All that needs to be done is to interface a regular transceiver audio input and output using a simple cable to a PC that runs the digital mode software.
The radio rigs in the ham shack and the station antenna, all remain the same. Thereafter, one is good to go. Due to far narrow modulation bandwidths of typical digital modes, one needs far less transmitter power to establish communication using these modes. Moreover, some of these modes are highly optimized for working under weak-signal conditions thus rendering communication feasibility even under quite adverse propagation conditions.
The evolution of amateur radio digital communication modes
Over the last few decades, various digital radio communication protocols have evolved. The advent of newer digital signal processing methods and the enhancement in the processing power of a typical PC have accelerated the pace of development of digital radio communication.
In my view, digital communication has undergone generational changes over many years. I would like to classify them into three distinct generations. Though some might like to say that CW morse telegraphy was the first digital mode, I humbly beg to differ. In my view, although CW uses OOK (a variant of ASK), it is a human decodable mode. There is no digital signal processing involved in CW either for encoding or decoding. Traditionally, CW is sent using a key and the fist and decoded by the human ear and the brain. On the other hand, all digital text or data modes require a machine (e.g. a PC) to encode and decode them. Hence they are not human modes. They are machine-to-machine communication modes. It is this aspect that sets them apart.
I will cover various digital radio modes and introduce the readers to the finer aspects that could highlight and distinguish these modes and their suitability under different conditions. A set of articles related to digital modes for radio communication will be progressively added to the website… Stay tuned!