Optimising digital RF communication systems

Alignment techniques can dramatically improve (or degrade) performance of the radio equipment you use.

With the rapid expansion of digital radio technologies like DMR, dPMR, P25, and NXDN in both the public safety and commercial land mobile radio markets, many radio operators, engineers and technicians are facing system performance issues that are difficult to analyse. These digital systems operate very differently in comparison to analog FM systems. Understanding how these digital systems work will enable RF professionals to improve their digital radio system operation through the use of better alignment techniques and proper digital verification methods.

Most RF professionals are aware of coverage studies that are related to transmit power variations. Figure 1 shows typical coverage maps of 100 W (50 dBm), 50 W (47 dBm) and 25 W (44 dBm).

Given equivalent performance factors except for power level, we can see that the coverage area becomes smaller as the transmitter power level is decreased. This is expected and would hold true for both analog and digital systems.

Figure 1. Coverage study with various site transmitter power settings.

With digital systems, there are additional factors that can cause issues with coverage. Coverage is not only affected by the RF power level, but also the quality of the transmitted digital signal. Just because a radio is digital, it does not mean the signal is necessarily good. There are performance metrics of a digital RF transmitter that must be measured to quantify the quality of the digital signal. While digital radios transmit modulation using digital methods, in most radios, the quality of this signal is directly related to proper FM deviation alignment. In addition to sufficient RF level, the receiver must see a good quality digital signal to decode the data properly.

FM deviation alignment is critical to the performance of a radio in the digital mode. This requires precise and accurate measurement to ensure that the radio meets its published specifications. Alignments, when done correctly, will enhance your digital radio system range and performance. Poor FM deviation alignments can create coverage reductions of 3 to 6 dB. This is equivalent to the coverage loss seen in the transmit power reductions shown in Figure 1.

How does digital radio technology work?

Many of today’s land mobile radio technologies rely on digital modulation based on 4 level frequency shift keying (FSK) to represent the digital ones and zeros used to transmit information over the air. Digital radios use a voice coder/decoder (vocoder) to convert analog voice to digital data. These systems can carry both vocoded audio and data (like short messaging services) using the same modulation method or air interface. The data rates are slow compared with digital systems used in cellular networks like LTE and 4G. This is because almost all land mobile radio frequencies use very narrow bandwidth channels of 6.25 kHz, 12.5 kHz, or 25 kHz in order to maximise the number of users within a given band.

Some technologies utilise time division multiple access (TDMA) techniques to divide these channels further into slots that have multiple users per RF channel. DMR for example, accommodates two users in a 12.5 kHz RF bandwidth channel. Figure 2 shows a typical block diagram of a digital radio. Here you can see the use of the vocoder, the channel coder and a modulator that converts the digital data to 4FSK modulation over the air.

Figure 2. A typical digital radio block diagram.

4FSK modulation means that there are four allowed frequency deviation states of the RF carrier (also called symbols) that represent the digital data. If there are four specific deviation states, then each of these states (or symbols) are called ‘dibits’ and represent two bits of data: 00, 01, 10 or 11. It is important to understand that these symbols (deviation states) must be precisely synchronised with the symbol clock within the radio. The amount of deviation at the symbol clock time dictates whether that symbol is 00, 01, 10 or 11.

How close the symbols are in relation to their specified deviation state at the symbol clock time affects the quality of the transmitted signal. This is critical to the performance of the system for both range and voice quality. A digital measurement, called symbol deviation, tells us how closely a transmitter’s signal is meeting the specification for correct deviation at the symbol clock time.

Figure 3 shows a P25 C4FM digital signal with deviation levels after demodulation (through an FM discriminator) on an oscilloscope. These deviation points can be plotted at the symbol time using a distribution display that shows the four frequency states and the corresponding ‘dibits’.

Figure 3. P25 C4FM modulation showing the ‘dibit’ and the corresponding frequency deviation of each symbol.

What can go wrong?

With analog FM technology, our human ear can discern a voice, even in the presence of a high level of noise, at ranges that can sometimes exceed those of digital systems. Digital radios typically exhibit very good voice quality up to the point that the receiver can no longer handle the number of bit errors. Then the call is abruptly dropped. We’ve all experienced cell phone calls where simply moving a few feet can cause the voice quality to quickly degrade or the call to be lost entirely. This is due to excessive bit errors, or errors caused when the signal has so much impairment, that the receiver can no longer determine if the symbol is 00 or 01.

The differences from how analog FM works, as noted above, are why digital systems must be maintained and aligned correctly to ensure the best possible voice quality and coverage range if they are to match the performance of analog FM systems. Digital radio receiver performance is tested using bit error rate, or BER, measurements to determine the receiver sensitivity instead of SINAD as used with analog FM. Besides digital receiver BER performance, the quality of the digital transmission also affects the system range and voice quality. The transmitter must modulate the carrier with 4FSK to the precise symbol deviation at the symbol clock time to correctly impart the digital data onto the carrier. How close these symbols are to the ideal frequency deviation points at the symbol clock time greatly affects the performance of the system.

If the transmitter symbol deviation is not precisely where it should be, then the receiver will have a difficult time determining what the symbol value is intended to be. When the receiver cannot accurately determine the value of the symbol, then it will produce an output with a high BER. This will cause problems with voice quality, data integrity and overall range of the radio system.

As you can imagine, transmitting a poorly modulated 4FSK signal may work during a short range ‘radio check,’ since the transmitted signal has sufficient signal strength where the receiver can decode even a poorly modulated signal. However, over any significant distance, the RF air interface between radio transmitters and receivers will cause additional impairment of any RF signal. Effects such as noise, interference, multipath and inter-symbol interference (ISI) will degrade the quality of the received signal. That is why a ‘Test 1-2-3’ voice check at short range does not accurately replicate real world RF scenarios where the quality of the signal is significantly degraded.

Figure 4 shows a ‘good’ DMR modulated signal versus a poorly modulated signal. The good signal shows a transmitted DMR signal at full power where the symbol deviation points are clearly aligned to the clock producing a clean transmitted DMR signal. The other displays a signal that has been degraded through improper alignment, yet is still at full power. You can easily see why the receiver will have difficulty decoding the poor signal once the quality is further impaired over its propagated distance.

Figure 4. A good DMR modulated signal versus a poorly modulated signal.

The effect of analog alignment on digital radios

Symbol deviation and modulation fidelity or FSK error performance are greatly affected by P25, DMR, dPMR, and NXDN radios’ analog FM deviation settings that are created during the alignment of the radio. Accurate FM deviation alignments are required to produce a good digital signal when the radio is transmitting in digital mode. In order to validate that the analog FM deviation alignments were made correctly, the radio needs to be tested for symbol deviation and modulation fidelity, or FSK error, in a digital mode. Having the proper test equipment is vital in meeting these requirements.

The accuracy of these alignments is significantly affected by the accuracy of the deviation meter used for FM modulation measurements. Both the absolute and the relative accuracy of the meter over the modulating frequency range are important factors for correct FM deviation measurements. In addition, once an alignment has been completed on a radio, you must have the ability to measure the digital modulation performance (symbol deviation and modulation fidelity, or FSK error). This will verify that the analog alignment actually produces a good digital transmitter signal. It is only through testing the radio transmitter’s symbol deviation and modulation fidelity, or FSK error, that you can verify whether it will work properly on the radio network. Figure 5 shows the relationship between an analog FM receiver’s sensitivity (measured as SINAD) and a digital radio receiver’s sensitivity (measured as BER) for both a signal with good modulation fidelity and symbol deviation versus one with poor modulation fidelity and symbol deviation.

Figure 5. Analog FM receiver sensitivity compared with digital receiver sensitivity.

All RF signals, both analog and digital, lose amplitude and are more susceptible to noise as the radio wave propagates over a given distance. However, the effect between analog and digital is substantially different. With analog FM signals, the noise is directly demodulated and present at the speaker. However, with digital 4FSK signals, the noise degrades the quality of the symbols, making recovery of the data more difficult. This is the same issue caused by improper FM deviation alignment.

The digital radio receiver will try to decode the bits until the receiver’s forward error correction can no longer correct for the symbol errors, resulting in a rapid roll off of voice quality and range. To visually see how noise impacts a digital signal, Figure 6 shows a view of a demodulated P25 C4FM digital signal with low noise and with high noise.

Figure 6. A good P25 C4FM signal at two different signal strengths.

As you can clearly see, a weak signal becomes more difficult for the receiver to decode due to the noise that is present, and this is further complicated when the transmit signal has other impairments, such as those from improper deviation adjustment.

Understanding FM deviation meter operation

FM deviation measurements play an important part in the radio’s digital performance. For this reason, we need to understand how a FM deviation meter operates. This will help us to understand the issues surrounding its use in digital radio deviation adjustments. The meter is a peak responding AC voltmeter that is measuring the audio of the demodulated output from the test receiver. There are two different types of FM deviation measurements that are common in the alignment of digital radios: absolute and relative. Radio deviation alignment is very important as the radio creates its digital signal based on how these deviation settings are made. Improper deviation alignment will result in poor digital signal quality.

Absolute measurements

Absolute measurements are measurements that must reference an absolute value. These are typical when specifying a radio performance in hard numbers. RF power, RF frequency error, audio frequency error and deviation are examples of absolute levels or readings referenced to an absolute standard that are taken when measuring a radio.

When measuring the absolute deviation level of a transmitter, it is important to consider all of the factors that might impact that measurement. FM deviation meters are peak reading. Therefore, other frequency components, besides the primary frequency being tested, will combine and add to the deviation level displayed on the meter. Noise is the largest additive contributor to deviation measurements. Most test receivers or modulation meters that use a super-heterodyne receiver will have controls for setting the audio frequency (AF) and intermediate frequency (IF) filters to limit the amount of noise that is allowed to pass through the receiver and effect the measurement.

To understand this, we need examine how the filter settings affect the measurements taken. One typical measurement for P25 radios is the setting of the deviation level of a 1.2 kHz signal (1200 Hz audio tone) to 2.83 kHz. This is a common adjustment made on many digital radios. The accuracy of this alignment will directly affect the symbol deviation of the transmitter when it transmits a digital signal.

Relative measurements

Another type of measurement is a relative measurement where the deviation of one rate (or tone) is compared to the deviation of another rate. Most digital radios require this type of alignment where a low audio frequency tone (for example 80 Hz) is transmitted and measured followed by a higher audio frequency tone (for example 3 kHz or 6 kHz). The high tone deviation is then adjusted to be the same level as the low tone. The goal is to ensure maximum flatness across the audio range that the radio uses to modulate the digital signal.

The IF and AF filter settings for this type of measurement should be configured for the highest rate to be measured. This setting should be the same for the low-rate tone when making relative measurements to reduce any error that may occur due to differences in filter response and settings. This particular measurement will be impacted by the flatness of the measuring device.

If the instrument does not have a flat response between the two test tones, the radio will be misaligned by the amount of ‘slope’ that the measuring device has. As an example, a 0.4 dB slope on the measuring device from 80 Hz to 6 kHz will cause approximately 5% error in the adjustment.

From this we can see how complex FM deviation measurements for digital radios are compared to legacy analog FM measurements. Conversations with RF technicians indicate that filter settings and the accuracy of FM deviation meter specification are rarely understood. These filters are typically not set properly in the field for testing or alignment and the accuracy of the test equipment FM deviation meter is overlooked. This can and will adversely affect digital radio performance.

*Mike Fortna is Senior Product Marketing Engineer for Cobham (now Viavi) AvComm, and (at the time of writing) Rob Barden is Director of Product Marketing. Ken Weipert (Senior System Technologist, Motorola Solutions) and Neil Horden (Chief Consultant, Federal Engineering) also contributed to this article, which is an edited version of a much longer Cobham AvComm white paper. The full version can be read at http://www.vicom.com.au/images/syncfiles/whpvacoba01_151111_1_b.pdf

Please follow us and share on Twitter and Facebook. You can also subscribe for FREE to our weekly newsletter and bimonthly magazine.