CC BY
31
ELECTRONICS. RADIO ENGINEERING
INVESTIGATION OF ANALOG DOWNCONVERSION IMPERFECTIONS DEGREE OF INFLUENCE ON THE RECEPTION QUALITY OF THE ZERO-IF RECEIVERS
DOI 10.24411/2072-8735-2018-10057
Elena R. Khasianova,
MTUCI, Moscow, Russia, [email protected]
Keywords: quadrature downconverters, IQ-imbalance, DC-offset, simulation, QAM-modulation, direct conversion receivers.
In addition to the analytical approach of analyzing the analog quadrature conversion imperfections effect on the signal reception quality changing, computer simulation methods are often used in the study of complex type signals. This approach allows obtaining a visual estimation of noise immunity for real signals under the I/Q-imbalance, DC offset and phase noise influence. Quantitative evaluation of amplitude-phase imbalance and DC-offset influence level on the sensitivity decreasing and reception quality degradation has been made in the article be means of simulation. The phase noise attendance is also taking into account. Studies were carried out for M-QAM signals having indices 4...256 (M-QAM). This is related to the fact of them widespread using in modern mobile, satellite and digital television systems. The researches allow defining the cases in which I/Q-imbalance and DC offset compensation is mandatory required and in which it is sufficient to use modern methods of signal coding.
The asymmetry of quadrature channels of modern analog frequency converters, according to the specification of the world's leading manufacturers, is 0.2-2% and 0.2-2.0° (in amplitude and phase, respectively). Thus, even with the use of modern technology for the production of integrated circuits, not all conversion errors can be reduced to an acceptable level. Collected in the literature results do not include studies for signals with high modulation orders (QAM-128, QAM-256). In addition, there is no quantitative assessment of the level of influence of amplitude-phase imbalance on the sensitivity of a direct conversion receiver. This article is devoted to the solution of these problems.
Для цитирования:
Хасьянова Е.Р. Исследование степени влияния погрешностей аналогового квадратурного преобразования на качество приема сигнала в приемниках с нулевой промежуточной частотой // T-Comm: Телекоммуникации и транспорт. 2018. Том 12. №3. С. 82-88.
For citation:
Khasianova E.R. (2018). Investigation of analog downconversion imperfections degree of influence on the reception quality of the zero-if receivers. T-Comm, vol. 12, no.3, pр. 82-88.
Nom-3-2018-1.qxd 24.04.2018 19:11 Page 83 ^ >
I
ELECTRONICS. RADIO ENGINEERING
-e-
1. Introduction
The Zero-IF receiver architecture has several advantages over the superheterodyne scheme: the simplicity of the structure, the possibility of providing broadband reception and multi-standard operation, the absencc of the image rcccption channels, the elimination of the complex selectivity path at the intermediate frequency. However, the design features of the radio-frequency (RF) path of the direct conversion receiver result in the appearance of the static and dynamic DC offset, which limits the sensitivity of the radio receiver, an l/Q-imbalance of the quadrature channels leading to BFR degradation, the occurrence of a carrier frequency offset due to a mismatch between the oscillators of the receiver and the transmitter. The asymmetry of quadrature channels of modem analog frequency converters, according to the specification of the world's leading manufacturers, is 0.2-2% and 0.2-2.0° (in amplitude and phase, respectively). Thus, even with the use of modern technology for the production of integrated circuits, not all conversion errors can be reduced to an acceptable level. Collected in the literature results do not include studies for signals with high modulation orders (QAM-128, QAM-256). in addition, there is no quantitative assessment of the level of influence of amplitude-phase imbalance on the sensitivity of a direct conversion receiver. This article is devoted to the solution of these problems.
2. RF channel simulation model for the investigation of the quadrature down con version imperfections influence level on the reception quality
The general structure of the simulation model is shown in Figure 1. Some blocks were added to it during the simulation: coding or interference from a neighboring channel, depending on the researching task.
OER raw >M 'A RraiMK
trtrt.T.)
№
Fig. 1, The general structure of the simulation model using due to the investigation of the quadrature dow neon vers ion imperfections influence level on the reception quality
AWGH complex
e*p !Z(i>|
Fig. 2. The Phase Noise power spectral density £ ( ^ simulation principle with the real synthesizers values consideration
To select the phase noise simulation prototype, a short analysis of the modem frequency synthesizers was carried out (Table 1). Based on its results, the frequency synthesizer 1508PL9T was selected as the prototype. The experimental value bf the phase noise PSD was recorded in the MTUC1 laboratory (Figure 3).
Table 1
Phase noise power spectral density of the different synthesizers
Model GHz St. GHz 1.0. dBc/Hz
100 Hz 10 kHz 100 kHz 1 MHz
A [>4360-0 [2) 2400...2725 2,6 -80,5 ■82,5 -109,0 -132,5
HMC840 [31 2620...2830 2,6 -110,0 -103.0 -103,0 -130.0
AIM350 141 0,14...4,40 2,2 - -97,0 -108,0 -136,0
1308PI.'It [51 3--.4.3 2,9 -92,2 -111.4 -117,8 -113,3
liSSi
m
i -- i i
DC-ofTsel, 1Q-imbalance and phase noise are sequentially introduced in the "Errors introduction block" in according to them models [1]. Dynamic DC-offset is modeling by the slow varying sinusiode, which is additively appended into the receiving signal.
The main goal in phase noise modeling in the framework of the problem is to approximate the spectral power density of the model to the SPM of real frequency synthesizers. To this end, the formation of noise is carried out in the frequency domain and the AWGN is multiplied with the spectral power density of the phase noise of one of the modem frequency synthesizers (Figure 2).
Fig. 3. Lff) figure of (he 1508PL9T. The oscillation frequency is 2,9 GHz [6]
BFR characteristic vs. the different Eb/No (dB) was chosen as the receiving signal quality parameter.
Theoretical BER dependence from Eb/No can be computed as [7]:
i* 1
4M 2
g (2.
1 2(M-1) )
(1)
T-Comm Vol.12. #3-2018
-е-
ELECTRONICS. RADIO ENGINEERING
-e-
For this modulation type, as can be seen from the graphs of Figure 4.2a, the difference between the BER values of the ideal and the 1/Q imbalance cases is almost two orders, which is already a more significant value than the case of the QAM-4 signal. DC offset has a significant effect on the signal reception quality and requires the compensation.
The constructed curves allow us to determine that if QAM-64 modulated signal at amplitude/phase mismatch of g = 1 dB and phi = 2° is taking into account, BER=4 * ]0° becomes possible when the signal-to-noise ratio for bit is 21 dB. In the absence of imbalance, a sufficient value of Eb/No is 17 dB.
QAM-128 signals are more sensitive to I/Q imbalance. The limiting value of the amplitude imbalance for them is g = 0.5 dB. But even with this value, the difference in the BER value can be 1.5...2 orders. For example, at a signal-to-noise ratio of 21 dB, the BER is ideally 1,178 * 10°, while for a case with an amplitude imbalance of 0.5 dB and a phase imbalance of 2 degrees, this value is 7.65 * 10 The threshold value of the DC offset is 10 mV.
In the ease of QAM-256 modulation, with a signal-to-noise ratio of 20 dB and 1/Q-imbalancc parameters g=l dB, and <p=2 the BER value is 0,00659, When the amplitude imbalance is increased to I dB, the BER degrades to 0.02585. In the ideal case, the bit error probability is 0.0005. The threshold value of the DC offset is 10 mV.
3,2. Estimation of the quadrature downconversioR errors influence on the signal reception quality with employing of the coding
To assess the effect of 1/Q imbalance in the case of structured redundancy, eonvolutional coding and turbo codec blocks are introduced into the model. The simulation results for QAM-16 modulated signals (Figure 5.1a), QAM-64 {Figure 5.1b), QAM-128 (Figure 5.1c) and QAM-256 (Figure 5.Id) are presented below.
It can be noted that in the ease of the eonvolutional coding and in the presence of amplitude-phase imbalance observed against the background of phase noise, the necessary signal-to-noise ratio to achieve BER =1*10-6 can, depending on the type of modulation, be increased by 0.5 ... 1,2 dB compared with the ideal case.
To investigate the correcting power of turbo codes, a turbo encoder based on the parallel connection of recursive eonvolutional encoders with a code constraint length of 3 and a random interleaving matrix is used. The number of iterations in decoding is chosen to be 7, since the studies show that with a further increase in the number of iterations, an insignificant improvement in BER is possible with a change in the signal-to-noise ratio [8]. The maximum of a posteriori probability (MAP) algorithm is chosen as the decoding algorithm.
The QAM-16 signal simulation shows that with using of turbo decoding, the energy losses are about 0.4 dB (Figure 5.2), This indicates their good resistance to the values of amplitudephase imbalance, which corresponds to the modem manufacturers' datasheets.
3.3. Estimation of the adjacent channel selectivity changing due to the quadrature dow neon version imperfections in the Direct conversion receivers
To summing up the adjacent channel selectivity, a situation with several frequency-spaced identical signals was simulated. The principle of modeling is the follow ing: for each type of modulation (QAM-4 ... QAM-256), the values of the signal-to-noise ratio and the central frequency are chosen so that adjacent channels do not interfere with each other (Table 2). For QAM-256, this value is 25 dB, 23 dB for QAM-128, for QAM-64 -20 dB, for QAM-16 the required signal-to-noise ratio is 15 dB, for QAM-4 - 11 dB,
Quantitative estimates of the BER changes were obtained in the presence of amplitude imbalance for the signal with QAM-16 ... QAM-256 modulations (Figure 6) as a results of the simulation. The conductcd experiments confirm the need to compensate the amplitude imbalance due to improving the selectivity of the adjaccnt channel. The graphs for QPSK are not given, since in this case the l/Q-imbalance does not have a significant effect on the signal reception quality.
10°
iff7
m
E----
* 1
:*-1
1
/ 1
Ifj
Jir 7
Fig. 5.2. BER curvcs changing from ]/Q-imbalance prcscnce when using turbo decoding
9. HE
Fig. 6. BER changes due to the 1/Q-imbalance and the adjaccnt channel
interfering signal inlluence of the signals with QAM-16 (red color), QAM-64 (blue color), QAM-128 (magenta color) and QAM-256 (green color) modulations
-е-
4. Conclusion
1. DC oft set has a significant effect on the signal reception quality and requires mandatory compensation. For example, achieving a BER of 1 * Iff* for the signal with QAM-64 modulation and small DC offset of 10 mV in the rcceivcr requires an increase in the signal noise ratio by 1.8 dB, by 2.7 dB for QAM-128 and by 6.7 dU for the signal with QAM-256 modulation. On the basis of these observ ations and experimental results on the boundary values searching for DC offset values, it can be concluded that the residual value of the DC drill alter the compensation algorithm operation should not exceed 0,003 V.
2. When considering the effect ol'l/Q-iuibalance on the signal reception quality with di lie rent types of modulation, the obvious conclusion is a more pronounced effect of the amplitude imbalance on the BFR value than the phase one. But in practice, due to the presence of phase imbalance, it is possible that the carrier synchronizing system of the receiver in the steady-state mode will work with some shift depending on the value of A(p, which will result in inaccuracy of the carrier reconstruction for QPSK modulation. Also, in the case of widely used OFDM systems, phase imbalance requires the compensation, because of its effect, the orthogonality of the subcarricrs is broken.
4. The use of convolutional coding reduces the influence of the I/Q-im balance on the degradation of the BER value. However, even w ith its use, the BER values of the ideal and imbalance cases differ by a factor of 3 for QAM -16 modulation, 4.5 times for the QAM-f>4 modulation. For QAM-128 and QAM-256, the difference is 6 and 7 limes, respectively. Certainly, these values will be altered in some way if the encoding parameters are changed. But this is beyond the scope of this study
5. The presence of amplitude imbalance affects on the selectivity of the receiver along the adjacent channel. The conducted experiments show that the BFR at ail amplitude imbalance of 1 dB and the presence of interference from adjacent channels, which in the absencc of unbalance do not affect the reception quality, degrades by 1,3,4 or 5 orders for the signals with QAM-16, -64 , -128 and -256 modulation, respectively. Such degradation indirectly indicates a decrease in the interference cancellation level of a neighboring channel.
References
1. Khasianova E.R. (2017). The modern state analysis of digital signal processing applying due to the quadrature dow neon vers ion inaccu-rates compensation at the zero-if receivers. T-Cnnmj, vol, 11, no.l I, pp. 84-88,
2. h Lip:/.\v wav. ana log.com/med ia'eri/technic a l-doc Li men (alio n/'data-sheets/AD E4360-0.pdf.
3. htlp://www.analog.coin/media/en/technical-docLinienlalion/data-s heets/hmc 84 0, pd f.
4. hup: //w ww. ana log, com/med i a/en/techn ic a I -docum enta l io n/data-shcets/A D F-'4350. pd f,
5. http: //m u I ticore. ru/index. ph p?id=656.
6. hnp i'/www . :ac i ;>c omp. ru/j oom/images/storage/prodixrtiv'e M 5 0 B p ! pn2950int.gif.
7. J, Proakis, (2000). Digital communications. Translation edited by D. Klovsky. Moscow: Radio and Communications. 800 p.
8. I. Nebaev. (2013). Turbocodes effectiveness analysis in the systems of data processing and transmission* PhD dissertation synopsis. SUT. ¡6 p.
ИССЛЕДОВАНИЕ СТЕПЕНИ ВЛИЯНИЯ ПОГРЕШНОСТЕЙ АНАЛОГОВОГО КВАДРАТУРНОГО ПРЕОБРАЗОВАНИЯ НА КАЧЕСТВО ПРИЕМА СИГНАЛА В ПРИЕМНИКАХ С НУЛЕВОЙ ПРОМЕЖУТОЧНОЙ ЧАСТОТОЙ
Хасьянова Елена Равыловна, Московский Технический Университет Связи и Информатики, Москва, Россия,
Аннотация
При исследовании сложных типов сигналов, помимо аналитического подхода к анализу влияния погрешностей аналогового квадратурного преобразования на качество приема сигнала, часто используют методы компьютерного моделирования. Такой подход позволяет получить наглядную оценку помехоустойчивости для реальных сигналов при воздействии 1/О-разбаланса, дрейфа постоянной составляющей и влиянии фазовых шумов. В статье, с помощью имитационного моделирования, проведена количественная оценка уровня влияния амплитудно-фазового разбаланса на снижение чувствительности и ухудшение качества приема сигналов. Исследования проводятся для сигналов с М-КАМ модуляцией, имеющей индексы 4...256. Это связано с широким распространением таких типов модуляции в современных системах мобильной, спутниковой связи и системах цифрового телевидения. Проведенные исследования позволяют определить, в каких случаях компенсация требуется, а в каких достаточным является применение современных способов кодирования сигнала.
Ключевые слова: квадратурные преобразователи частоты, разбаланс квадратурных составляющих, дрейф постоянной составляющей, имитационное моделирование, КАМ-модуляция, приемники с нулевой промежуточной частотой.
О
Литература
1. Khasianova E.R. The modern state analysis of digital signal processing applying due to the quadrature downconversion inaccurates compensation at the zero-if receivers. T-Comm, 2017, vol. 11, no.11, pp. 84-88.
2. http://www.analog.com/media/en/technical-documentation/data-sheets/ADF4360-0.pdf.
3. http://www.analog.com/media/en/technical-documentation/data-sheets/hmc840.pdf.
4. http://www.analog.com/media/en/technical-documentation/data-sheets/ADF4350.pdf.
5. http://multicore.ru/index.php?id=656.
6. http://www.radiocomp.ru/joom/images/storage/products/ebl508pl9t/pn2950int.gif.
7. Прокис Дж. Цифровая связь / Пер. с анг. под ред. Кловского Д.Д. М.: Радио и связь, 2000. 800 с.
8. Небаев И.А. Анализ эффективности турбокодов в системах обработки и передачи данных / Автореферат диссертации на сосискание ученой степени к.т.н. СПбГУТ,2013. 16 c.
T-Comm Том 12. #3-2018
¥
