Fullscreen HTML5Med Res (??MB)HTML4

ii I AMATEUR RADIO By GARRY CHATT, VK2YBX Amplitude companded sideband: a new technique for greater spectrum efficiency New technology is always of interest to the radio amateur. This month, we take a look at a new technique called amplitude companded sideband. It offers greater spectrum efficiency, improved speech quality and the ability to add selective calling from the base station. Many countries are now having problems catering for the growing demand for radio communications within a limited frequency spectrum. And although it has a much smaller population than the USA or Europe, Australia will also reach a situation in the near future where spectrum congestion becomes a real problem. A possible answer to this problem is a new mode of transmission called "amplitude companded sideband" , or ACSB for short. The technique promises greatly improved spectrum efficiency. In fact, ACSB will offer up to six times the number of discrete radio channels over existing FM 2-way systems. For this reason, it will be 10d8 PER DIVISION VERTICAL 2kHZ PER DIVISION HORIZONTAL Fig.1: frequency spectrum for a typical ACSB transmission. The 3.1.lcHz pilot tone is 1.3.lcHz above the channel frequency. 64 SILICON CHIP of interest to both commercial users and amateur operators alike. Basically, ACSB is a form of SSB (single sideband) transmission with speech compression and frequency pre-emphasis during transmission and speech expansion and deemphasis on reception. A further refinement is the use of a pilot tone. The basic disadvantage of SSB is that no carrier is broadcast to act as a frequency or amplitude reference in the receiver. This means that the SSB receiver must regenerate a carrier to the exact frequency produced by the transmitter in order to demodulate the incoming signal. This is no problem at HF but becomes increasingly difficult at VHF and above. Assuming reasonable transmitter and receiver fre-quency stabilities of plus and minus two parts per million (ppm), each could be off centre frequency by ± 300Hz at 150MHz, resulting in a worst case (and unacceptable) 600Hz frequency mismatch between transmitter and receiver. ACSB eliminates this problem by transmitting a low-power, fixedfrequency pilot tone. Current ACSB equipment uses a 3. lkHz pilot tone transmitted at a power level of 10% of the total PEP output. This pilot tone is generated accurately to within 1Hz in both the transmitter and receiver. At the receiver end, automatic frequency control (AFC) circuits adjust the synthesiser frequency so that the received pilot tone matches the internal 3. lkHz reference. As a result, the recovered audio is reproduced within 1Hz of its original frequency. DD~~~~RWCf:CND J::ssT AMPLIFICATION/ PRE-EMPHASIS/ COMPRESSION/ LIMITING A:OIO \ UPPER CENTRE ,r:i"'D, % 5. '/E1 VOICE 10.2369 MHz 3100Hz l0 •241MH z 10.2431 MHz 10 4 MHz 10MHz 2431 RF BAND PASS FILTER 10.24MHz LOW PASS ALTER USB CRYSTAL FILTER -----t / MliR 152MHz 0038 RF LINEAR AMPLIAERS 'X, 10.24MHz ~------------1 PILOT OSCILLATOR 3100Hz SYNTHESISER 10.24MHz ALC DETECTOR REFERENCE I"\., ALC CHANNEL SELECTOR CHANNEL DISPLAY Fig.2: block diagram of an ACSB transmitter. The audio filter limits the audio bandwidth to about 2.5kHz so that audio components are kept well clear of the pilot tone. Only the upper sideband is transmitted. The pilot tone also serves as a signal strength reference at the receiver. This enables the AGC (automatic gain control) circuitry to keep the gain at the desired level over a wide range of signal strengths, regardless of pauses in speech. This eliminates listener fatigue due to the "gain pumping" produced by voice-actuated AGC systems used by conventional SSB systems. Squelch The presence or absence of the pilot tone is also used to operate the receiver squelch system. This eliminates random opening of the squelch circuits by noise. In addition, low deviation phase modulation can be added to the pilot tone for signalling purposes. The most common application of this is the use of CTCSS which allows selective calling by the base station. The frequency spectrum of a typical ACSB transmission is shown in Fig.1, along with the standard emission limitations for a 12.5kHz channel. In this case the centre frequency is 152.0025MHz. Only the upper sideband is transmitted. The suppressed carrier is 1.BkHz below the centre frequency at 152.0007MHz, while the 3.lkHz pilot tone is 1.3kHz above the_ channel frequency at 152.0038 MHz. The top line of the graph corresponds to 25W PEP (peak envelope power). The amplitude linearity of the final RF amplifier is the key element in determining the amount of occupied bandwidth which occurs as a result of intermodulation distortion. This means that all amplifier stages must be linear enough to reproduce variations in signal amplitude. All power amplification stages in ACSB must therefore operate in either class A or AB. Although there are some pronounced differences between some of the circuits used in FM radio systems and those used in ACSB, there is nothing mysterious about the circuitry used. Basically, all the receiver IF (intermediate frequency) and RF circuits operate in their linear regions. Most of the audio circuits also operate linearly. Strictly speaking, the audio compressor, limiter and expander circuits are the only circuits not operating in the linear mode. However, these circuits are designed to produce minimal harmonics and intermodulation products. The modulation signal is trans- lated to or from the RF channel frequency using mixers (including a balanced modulator and product detector). Frequency multipliers cannot be used in a modulated signal path due to the non-linear effect on the modulation. However, the local oscillator system can use multipliers since it does not carry modulation. The ACSB transmitter Fig.2 shows the block diagram of an ACSB transmitter. Let's see how it all works. Low level audio from the microphone is first applied to an amplifier/compressor circuit. Compression amplifiers have more gain at low input levels than at higher levels. In this circuit, the gain selfadjusts quickly enough to increase the level of weak voice syllables, so that the average voice level at the output of the compressor is increased relative to the peak level and the overall dynamic range is reduced. For example, a 2:1 compressor would reduce a 60dB dynamic range at the input to a 30dB dynamic range at the output and increase the average voice level by as much as 15dB. Thus, the major benefits of amplitude compression JULY 1988 65 CENTRE 1E 152.0038MHz AUDIO POWER ~ vie-v-t.J.,J 10.24MHz SQUELCH LOGIC SQUELCH CONTROL LOCK CT 10.24MHz _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ____, SYNTHESISER REFERENCE AFC CHANNEL DISl'LAY Fig.3: block diagram of the receiver circuit. A bandpass filter separates out the pilot tone which is then compared with a locally generated 3.1kHz pilot tone to derive AFC information. As well, the separated pilot tone drives the AGC circuits. are reduced dynamic range (which allows easier control of amplitude linearity), improved modulation efficiency (which allows maximum RF power output over a wider range of audio input levels), and improved signal to noise ratio in the receiver when an expander is used. Audio peak limiting may also be used to ensure that a desired peak audio level is not exceeded. An audio peak limiter tends to increase the compression ratio at higher voice input levels and has some affect in increasing the average modulation level. Audio peak limiters must be preceded by audio pre-emphasis to minimise in-band harmonics. Audio pre-emphasis is used to boost the level of high frequencies compared to the bass frequencies which are more predominant in speech signals. This tends to flatten the average output across the passband of the transmitter, helps reduce intermodulation spreading at the RF output, and improves the recovery of high frequency voice signal at the receiver. The preemphasis in an ACSB transmitter is typically between 6dB and 12dB per octave across the audio band. The compressed speech signal is 66 SILICON CHIP fed to a sharp cut-off low pass filter, which limits the audio bandwidth to about 2.5kHz. This ensures that the audio components are kept well clear of the 3. tkHz pilot tone. The pilot tone, which is crystal generated, is then added to the filtered audio. A technique called "pilot pumping" is usually used to reduce the pilot level when the voice power exceeds the maximum pilot power level. This reduction in pilot level during voice peaks helps reduce intermodulation spreading at the transmitter output, and allows the ACSB receiver to detect the level of pilot tone and increase its gain proportionately. This serves to expand the voice dynamic range at the receiver and keeps the receiver quieter during gaps in received speech. Pilot pumping can be achieved by using another audio compressor and/or as a result of feedback ALC (automatic level control) action. Transmitter IF stages The balanced modulator mixes the processed audio plus pilot tone and the IF carrier oscillator together. The resultant output from the modulator includes a DSB (dou- ble sideband) signal "wrapped" around the carrier oscillator frequency. Because the circuit is balanced, this carrier is suppressed by about 30dB. The upper sideband is produced from the sum of the carrier oscillator and the audio frequencies. The lower sideband is produced from their difference. The DSB (double sideband) signal is passed through a buffer amplifier which is also used as an ALC control point. The IF filter is typically an 8-pole unit with steep skirts. This has low ripple and a narrow bandwidth since it is only required to pass frequencies from 300Hz to 3.3kHz above the IF. Thus, the output of the filter contains only the USB (upper sideband) ACSB signal which is then amplified to drive the RF mixer. RF stages The ACSB IF signal is combined with the local oscillator (LO) at the mixer. The mixer is usually a double balanced type so that little output occurs at either the IF or LO frequencies. Among the signals present at the RF mixer output are the sum and the difference of the IF and LO frequencies. The desired This mobile ACSB transceiver features 16 channels and a power output of 25 watts PEP. that as the PA output reaches its compression point, any further increase in drive level reduces gain in the previous stages. The ACSB receiver output is the sum of the IF and LO frequencies, which is an USB ACSB signal. The lower sideband is discarded. The Local Oscillator is generated by a frequency synthesiser, similar to those used in synthesised FM equipment, and has a resolution of 5kHz. As in most commercial and amateur equipment these days, frequency selection is made by a channel switch and PROM (programmable read only memory) circuitry. A stable master reference oscillator is essential, and typical circuits use 10.240MHz as the reference frequency with a stability of ±2ppm. After some amplification, the mixer output is fed to an RF bandpass filter which removes all undesired mixer products. Several gain stages are then used to increase the amplitude of the signal to about lO0mW to drive the RF power amplifier. After linear amplification, the RF output is passed through a low pass filter to suppress any harmonics which may be present, and then fed to the antenna socket. Safety circuitry in the form of ALC protects the power amplifier (PA) from being overdriven. The PA output is sampled, rectified, filtered, and fed back to one or more of the earlier IF or audio stages. The ALC threshold is set so Fig.3 is a block diagram of the receiver circuit. The received signal first passes from the antenna to a selective RF amplifier. The tuned circuits in this amplifier reject the input signal at the image frequency (twice the IF below the desired frequency) and pass the correct input signal. The same synthesiser local oscillator and mixer circuit used in the transmitter is operated in reverse for the conversion of the RF input frequency to the IF (intermediate frequency). From there, the signal is fed to the first IF amplifier which in turn feeds an IF delay circuit and noise blanker. A separate IF amplifier is used to magnify noise pulses and these are used to trigger a noise gate during strong impulse noise periods. The IF delay circuit ensures that the noise pulse arrives at the blanking gate at exactly the same time the gate is turned o'n. The IF signal then passes through a crystal filter (as used in the transmitter) and two subsequent amplifying stages. The IF amplifiers have AGC applied to them so that the average signal level at the product detector is held fairly constant for a wide range of input signals. This AGC is derived from the recovered pilot tone. The product detector uses the carrier oscillator to mix with the IF to produce baseband audio. The baseband audio contains both the audio passband and the 3. lkHz pilot tone. Two separate audio circuits are used here to process these signals. The pilot tone bandpass filter strips the pilot tone from the baseband audio. It is then used to drive the AGC circuits. The pilot tone filter output is also fed to an amplitude limiter which preserves the pilot phase and frequency. This signal is compared with the locally generated 3. lkHz pilot tone and is used to derive AFC information. This AFC is routed back to the synthesiser, where the first local oscillator frequency is adjusted so that the recovered pilot tone is set exactly to 3. lkHz. This guarantees that the recovered speech will be correctly demodulated. In repeater systems, the received pilot tone may have CTCSS phase modulation applied to it. An additional discriminator or phase locked loop can be used to demodulate this information which can then be used for repeater control or squelch control. The voice audio is sent to a fast acting AGC amplifier which has its gain controlled by the pilot tone amplitude. This AGC almost entirely eliminates amplitude fluctuations caused by multipath propagation and vehicle movement. The AGC circuit also acts to reexpand the amplitude compression encoded at the transmitter in the form of pilot amplitude pumping. The fade-corrected signal passes to an amplitude expander which restores the dynamic range of the voice signal. This audio is then subject to de-emphasis. Both deemphasis and amplitude expansion provide a great reduction in channel noise and help to improve the overall signal to noise ratio. The expanded and de-emphasised audio is then fed to the squelch circuit, which feeds the receiver volume control and audio amplifier. Footnote: most of the developmental work on ACSB has been carried out by Stephens Engineering Associates Inc, 7030 220th Street, S.W. Mountlake Terrace, WA 98043, USA. The author wishes to acknowledge the use of public domain information provided by Stephens Engineering, USA. ~ JULY 1988 67