An Analog LO Harmonic Suppression Technique for SDR Receivers

Description

Existing System:

SOFTWARE-DEFINED radios (SDRs) are reconfigurable systems that are able to tune to any wireless standard with variable carrier frequency, modulation type, and bandwidth. This concept is inspired by the increased demand for integrated multi standard wireless transceivers. These transceivers can support a variety of radio standards such as Global System for Mobile communications (GSM), Wi-Fi, positioning systems (GPS), Bluetooth, 3G/4G cellular, and Zigbee. The ideal SDR receiver includes a wideband analog-to-digital (A/D) converter followed by a digital signal processor (DSP) that accomplishes the radio functions. However, wideband data converters require very high-power consumption. Hence, in order to efficiently realize the SDR concept, mixed-signal and digital techniques can be utilized.

Reported techniques to realize mixed-signal or discrete-time (DT) SDR receivers usually include a mixing operation before A/D conversion. Mixing can be continuous or discrete in time. In  the receiver includes a low-noise amplifier (LNA) and a continuous-time mixer, and the sampling and conditioning of the  . sampled signal are realized after the down conversion at zero-IF. Several approaches use RF sampling with DT mixing. One example of this approach is proposed in where the RF signal voltage is oversampled and multiplied by the oversampled local oscillator (LO) waveform for down conversion. Nowadays, direct sampling of an RF signal can be realized in highly scaled CMOS. Different RF-sampling receiver architectures are proposed in working at different sampling frequencies, with different subsequent DSP techniques.

This paper presents an SDR front end with high-Q RF filtering resilient to out-of-band blockers and proposes a low complexity analog technique for attenuation of LO harmonics. This technique can be mixed with already available DT mixing techniques to improve their harmonic rejection.

Disadvantages:

• It presents the SDR receivers for high complexity analog technique for attenuation of LO harmonics.
• Achieves a lower performance of SDR receivers.
• Efficiency is Low.

Proposed System:

Software Defined radios (SDRs) are reconfigurable systems that are able to tune to any wireless standard with variable carrier frequency, modulation type, and bandwidth.  This concept is inspired by the increased demand for integrated multi standard wireless transceivers.  This paper proposes the SDR front end with high-Q RF filtering resilient to out-of-band blockers and low complexity analog technique for attenuation of LO harmonics. The transconductance can be modulated by the (LNTA) low-noise transconductance amplifier with a raised-cosine signal.  Thus, the effective attenuation of the LO harmonics is achieved. Removing any RF pre filtering causes several problems includes wideband noise and large blockers that can directly leak into the receiver.  Therefore, the proposed system presents the SDR receivers for low complexity analog technique for attenuation of LO harmonics.  This achieves a better performance of SDR receivers.  Finally, this technique is implemented in the VHDL and synthesized in the XILINX and shown the comparison in terms of area, power and delay comparison.

This section overviews the SDR concept and discusses its limitations due to the blockers and interferers at the LO harmonics. The traditional narrowband radios include fixed high Q external surface acoustic wave (SAW) filters that remove out-of-band blockers, images, and interferers. Blockers can desensitize the receiver, while images and interferers can be down converted to IF or baseband corrupting the desired signal. In a portable battery-powered handheld wideband SDR device, it is not easily possible to use SAW filters. Hence, the research focus in recent years has been noticeably focused on SAW-less wideband receivers. These receivers are software-defined for various standards, bandwidths, and frequencies. Removing any RF pre filtering causes several problems: wideband noise and large blockers that can directly leak into the receiver. Moreover, during down conversion, any unwanted signal located in LO harmonic frequencies can be easily aliased with the wanted signal at the baseband. Therefore, there are two main issues in the design of any wideband receiver front end, as depicted in Fig. 1, which must be considered: out-of-band blockers high-Q filtering to avoid gain compression, and LO harmonics rejection of the mixer to avoid aliasing.

For a typical receiver, a wanted channel signal needs to be received at very low amplitude. Therefore, the receiver’s LNA must operate at a relatively-high gain. For instance, if the gain of the LNA is equal to 20 dB, a 0-dBm out-of band blocker results in a 632-mV blocker (assuming a 50- system). This is amplified to 6.3 V and completely compresses the receiver. Hence, without external filtering, on-chip high-Q filtering is required.

Four techniques are proposed in the literature to address blockers: translational filtering, mixer-first technique, current-mode systems, and voltage-mode system with RF filtering. Recent work presenting a voltage-mode system with RF filtering  includes a linear LNA that amplifies the wanted signal and rejects blockers. The high output impedance of the following switching mixer amplifies the wanted channel, while the blocker suppression is realized by designing this output impedance to be low outside of the wanted channel, where the blocker exists.

Due to the nature of mixing a signal by a square wave, any undesired signal at the harmonics of the LO frequency fLO also down convert to baseband, with a conversion gain of 4/kπ , where k is the odd harmonic number. This conversion gain is −9.5, −14, and −17 dB at the third, fifth, and seventh harmonics, respectively. Therefore, the undesired signal at the harmonics of fLO is not rejected enough by the receiver.

In order to realize a wideband receiver with harmonic suppression, different architectures have been developed. Almost all of these techniques are based on N-path filtering to realize the rejection at the harmonic frequencies. Fig. 2 shows the general architecture of a wideband receiver [19]. Although N-path techniques can deliberately reject the harmonics of fLO, the non idealities in the clock and the timing errors, including finite slopes and process–voltage–temperature-dependent mismatch errors of buffers and gates, can limit the amount of rejection. In the classical one-stage harmonic rejection mixer [20], three square-wave LOs with phases of 0°, 45°, and 90° are generated and applied to three paths that are weighted by “1: √2: 1.” It can be shown that this rejects the third and fifth LO harmonics. The main drawback of this approach is that the irrational number √2 is not trivial to be accurately implemented. Moreover, the harmonic rejection ratio (HRR) is very sensitive to both gain and phase mismatches, i.e., 1° phase and 1% gain errors limit the HRR to 35 dB [19]. A two-stage technique is proposed, to improve HRR, where the irrational ratio “1: √2: 1” is realized in two steps, such that it achieves a weighting ratio of 29: 41: 29 that can be realized using integer numbers. It is shown that this ratio represents only 0.03% error, corresponding to an HRR of 77 dB.

In order to improve the harmonic rejection, this paper proposes an analog technique that can be integrated with the already-proposed sampling mixer technique. The proposed technique is applied to a single-path receiver architecture that results in simplified clocking and reduced power consumption, as compared to traditional N-path architectures. In this technique, the output current of the low-noise transconductance amplifier (LNTA) is modulated before the switching mixer. This technique results in a notch at the third harmonic and at least 30-dB rejection at the fifth and seventh harmonics. Fig. 3(a) shows the block diagram of a basic single-path software-defined receiver comprising an LNA with a sampling mixer, and Fig. 3(b) shows the conceptual block diagram of the proposed SDR receiver comprising an LNTA with an embedded time-based analog modulating element, a sampling mixer, RC filtering, and sampling.

The  objective  of  the  truncated  multiplier design is to compute  P  MSBs of the product with a maximum  truncation  error  of  no  more  than  1  ulp, where 1 ulp = 2−P.The  FIR  filter  design  in  this  brief  adopts the  direct  form  is shown in figure.  where  the  MCMA module  sums  up  all  the  products  aˆi ×  x[n  −  i]. Instead  of  accumulating  individual  multiplication for each product, it is more efficient to collect all the PPs  into  a  single  PPB  matrix  with  carry-save addition  to  reduce  the  height  of  the  matrix  to  two, followed by a final carry propagation adder. In order to avoid the sign extension bits, we complement  the  sign  bit  of  each  PP  row  and  add some  bias  constant  using  the  property  s¯  =  1  −  s, where  s  is  the  sign  bit  of  a  PP  row,  as  shown  in Figure.  All  the  bias  constants  are  collected  into  the last  row  in  the  PPB  matrix.  The complements of PPBs are denoted by white circles with over bars.

In the proposed truncated multiplier design in  FIR  filter  implementation,  it  is  required  that  the total  error  introduced  during  the  arithmetic operations is no larger than one ulp. compares the two approaches.  In the  removal  of unnecessary  PPBs  is  composed  of  three  processes: deletion,  truncation,  and  rounding.  Two  rows  of PPBs  are  set  undeletable  because  they  will  be removed at the subsequent truncation and rounding.

  Advantages:

• DT mixing techniques of LO harmonic is used to improve their harmonic rejection.
• It presents the SDR receivers for low complexity analog technique for attenuation of LO harmonics.
• Achieves a better performance of SDR receivers.
• Efficiency is high.