Giriş Çıkış Uyumlu, Yüksek Dinamik Aralıklı, Si-ge Tranzistorlu Rf Karıştırıcı Tasarımı

Abstract

Gilbert karıştırıcı yapısında dayanarak 0.25um SiGe teknolojisi ile üretilmiş tranzistorlar kullanılarak 2.15 GHz sıklığında çalışan bir etkin karıştırıcı devresi tasarlanmıştır. Girişindeki simetrik sürücü devresi birbirleriyle bağlantılı çalışan OB(ortak baz)‘lı ve OE(ortak emetör)‘lü devreler kullanılarak gerçekleştirilmiş ve OB’lı devreye gerilimden-seri kapasitif bölmeli edilgen ileri-besleme uygulanarak girişte geniş bantlı bir empedans uyumu sağlanmıştır. Karıştırıcının benzetimleri AWR yazılım ortamında yapılmıştır. Devrenin DC besleme gerilimi 5V olup, -2.5 dBm düzeyindeki yerel osilatör sürümü ile OIP3 değeri ve Dönüştürücü kazancı sırasıyla, 22.4 dBm 3 dB olarak bulunmuştur. Devrenin benzetim sonucu gürültü sayısı ise 11.75 dB’dir. Tezde önerilen karıştırıcı devresi, X Bandında çalışan bir alıcı dizgesinde ikinci karıştırıcı katı olarak kullanılmak üzere, Gc=3dB, OIP3≥20 dBm ve NF≤12 dB isterlerini karşılayacak şekilde tasarlanmıştır. Devre temel olarak Gilbert Karıştırıcı devresi olmakla birlikte, girişte herhangi bir balun devresi kullanılmamakta ve dengesiz-dengeli dönüştürücü (BALUN) işlevi, her ikisi de geçiş iletkenliği kuvvetlendiricisi katı olarak kullanılan OB ve OE devrelerinin birlikte etkileşimli olarak çalıştırılmasıyla gerçeklenmiştir. OE’lü devreden OB’lı devreye gerilimden-seri kapasitifbölmeli edilgen zıt ileri-besleme uygulanarak girişte geniş bantlı bir empedans uyumu sağlanmıştır. Ara-Sıklık (AS) katı ise, dengeliden-dengesize dönüşüm sağlayan ve çıkış empedansı RL=50Ω yüke uyumlu yüksek dinamikli kuvvetlendiriciden oluşmaktadır. Böylece, giriş ve çıkışta hiçbir edilgen balun devresi kullanılmadan karıştırıcı devresi tasarlanmıştır. Bunun sonucu olarak da, giriş ve çıkış katlarına yüksek empedans uyumu gereksinimi olan RS ve AS süzgeçleri dışarıdan doğrudan bağlanabilmektedir. Empedans uyumunu sağlamak amacıyla, kapasitif gerilim bölücülü gerilimden-seri ileri-besleme devresi kullanılmıştır. Bu türden ileri-beslemeli empedans uyum devresi herhangi bir ek gürültü katkısı sağlamadığı için, diğer geri besleme kullanan yöntemlere göre büyük üstünlük sağlamaktadır. Anahtarlayıcı tranzistorlarında kullanılan anahtarlama gerilimindeki ufak bir sapma, anahtarlama zamanını modüle ederek farksal akım dalga şeklinde bozukluğa neden olmaktadır. Bu nedenle devre elemanları, anahtarlayıcı tranzistorlar hızlı konum değişikliği yapacak şekilde seçilmişlerdir. Anahtarlayıcı tranzistorlarından çıkışa gelen kırpışım gürültüsü, tranzistor akımı ile doğrudan orantılıdır. Gilbert hücresi üzerinden gelen akım gürültüsünü azaltmak için, dirençlerin ve anahtarlama tranzistorlarının üzerinden geçen akım,Gilbert Hücresine bağlantı noktasına konan dirençler üzerinden DC akım katkılanarak azaltılmıştır.

An active mixer which is operating at 2.15 GHz and based on the Gilbert Mixer topology is designed by using 200 GHz fT, 0.25 SiGe technology. Mixeris simulated with AWR software environment. The circuit exhibits an Output IP3 of 24.7 dBm and the conversion gain of 3 dB with the local oscillator power of -2.5 dBm. DC power supply voltage is5V. The noise figure of the circuit is simulated as 11.75 dB. Mixers are key components in both receivers and transmitters. Mixers translate signals from one frequency band to another. The output of the mixer consists of multiple images of the mixers input signal where each image is shifted up or down by multiples of the local oscillator (LO) frequency. The most important mixer output signals are usually the signals translated up and down by one LO frequency. The voltage conversion gain is the ratio of the root mean square voltages of the IF and RF signals. The power conversion gain is the ratio of the power delivered to the load and the available RF input power. The 3rd order intercept point (IP3) is the point where the third-order term as extrapolated from small-signal conditions crosses the extrapolated power of the fundamental. Basically mixers are classified as active and passive mixers. Although passive mixers have conversion gain less than one (lossy) and has noise figures bigger than active mixers, they present less intermodulation distortion (the bigger OIP3 values). An important advantage of passive mixers over their active counterparts is their much lower output flicker noise. However, the low gain of passive mixers makes the 1/f noise contribution of the subsequent stage critical. Additionally mixers are categorized as single-balanced and double-balanced topologies. Double-balanced topologies are more preferable than single-balanced mixers because of their isolation between ports, suppressing the unwanted signals and linearity specialities. One advantage of double-balanced mixers over their single-balanced counterparts stems from their rejection of amplitude noise in the LO waveform. The proposed mixer circuit is designed for to be used as the second mixer stage in a RF receiver IC. The performance requirements are specified as; 3 dB Conversion Gain, Output Third Order Intercept Point (OIP3) of minimum 20 dBm, and the noise figure of maximum 12dB. The mixer is consists of a modified Gilbert Cell topology where the input voltage –current converting stage is realized by using CB and CE stages which operate interactively to provide matching at the input. The IF output amplifying stage with reasonably high dynamic range is arranged as the combination of the inverting and non-inverting stages to provide the power doubling at the output. Gilbert Cell is the most common double balanced mixer switching topology which is used in various RFIC applications. It is quite convenient to obtain the high gain with high dynamic range and low power consumption by using this circuit. The modified version of this circuit which is presented in this work eliminates the use of any passive baluns at the both ports and provides very broadband impedance matching at the both ports. At the input stage, an active balun topology is used. The feed-forward connection from the CE to a CB stage with capacitive voltage divider is used for to provide impedance matching at the input. This technique provides highly broadband input matching with the additional advantages of good linearity and less noise-figure than the other techniques. For a certain RF input power, all element values and ratio of capacitors which is used as voltage divider are calculated. At the switching transistors, a small change on the switching voltage can modulate the switching time and creates distortion on the differential current waveform. Therefore, the circuit elements are selected as to provide the rapid tranition of the toggling transistors. The first significant part of the total noise figure is coming from the switching transistor. Second, the noise source is transistors at the gain stage and the third is the source resistors. Noise is measured using the noisefigure (NF) definition, which is a measure of how much noise the mixer adds to the signal relative to the noise that is already present at the input signal. The noise figure of 0 dB is ideal, meaning that the mixer adds no noise. The NF of 3 dB implies that the mixer adds an amount of noise equal to that alreadypresent in the signal. For a mixer alone, a NF of 10-12 dB is typical.There are there different frequency bands which must be considered during mixer NF analysis. Firstly, transistors and resistors at the circuit produce noise at intermediate frequency (IF). Some of these noise, for example IF noise produced by collector resistors occurs at the output. Secondly, noise produced at RF and the image frequency, mixes with local oscillator at the mixer and they seen at frequency IF at the output. Collector shot noise at the gain stage is an example for this type of noise. Lastly, noises produced by LO can be transmitted to the IF output. Noise power which is transmitted to the IF output is not constant through the LO period. At the higher LO values, the dominant noise comes from the gain stage transistors. It is an expected behavior because LO causes the differential pair transistors to switch between saturation and cut off regions. At both of two stages there is no gain at the transistors, they contribute very little noise. Also, gain through the input is maximum in these conditions. So a sharply switching high amplitude LO signal is needed to obtain a high signal-to-noise ratio. However, during finite fall and rise time, for square or sinusoidal local oscillator signal, local oscillator has zero crossings. During this time, switching transistors operate at active region. In this region, transistors act as amplifiers and noise produced at switching transistors and local oscillator, such as thermal noise due to base resistors and collector shot noises become dominant. Flicker-noise is one the critical issues in the direct conversion mixers. There are two major mechanisms that generate the flicker noise of the switching pair devices. The first one is the direct mechanism, due to the finite slope of the switching pair transitions. In order to decrease flicker noise in the direct mechanism, the size of the switching pairs needs to be increased, and large switching devices increase the parasitic capacitance of the switching pairs, resulting in the flicker noise indirectly translating to the output. The second mechanism that generates flicker-noise is the indirect mechanism, flicker-noise mainly depends on the tail capacitance (Cp) at the node between the LO switches and RF transconductance stage. In order to decrease the flicker-noise in CMOS active mixers, the bias current of the local oscillator (LO) switches should be small enough to lower the height of the noise pulses. The static current injection technique was proposed to reduce the bias current of the LO switches. However, the impedance of the LO switches as seen from the RF stage is increased as we reduce the bias current of the LO switches. In addition, RF leakage current flows through the injection circuit, which decreases conversion gain and also allows more RF current to be shunted by the tail capacitance (Cp) at the node between the LO switches and RF transconductance stage. The shot- noise coming to the output from the switching transistors, which is stated as (I_n^2 ) ̅=2qI_Cis proportional with the transistor collector current. However the noise coming to the output from the input stage, which is stated as (V_n^2 ) ̅=2kT/g_m =2kTV_T/I_C is inversely proportional with the collector current. To reduce the shot- noise coming from the Gilbert Cell, the currents which flow through the load resistors and switching transistors is decreased by injecting a current to the connection point of Gilbert Cell through resistors. In this way, collector current which flows through input transistors remains constant, and collector current which flows through switching transistors decreases. So, the collector currents become independent from each others. Noise figure is decreased by this current reduction technique. To obtain the same conversion gain, load resistors of the Gilbert Cell are increased with the current reduction ratio. Therefore,because of this increased resistor value, decreasing in the noise figure is not as high as prediction. All noise sources which are both thermal and shot at the circuit is determined and calculated. The noise components of interest lie in the RF range before downconversion and in the IF range after downconversion. Noise sources at the circuit are RF and IF noise coming from switching transistors, source noise, noise coming from active balun topology including thermal noise of resistors and the shot noise of transistors.Another shot noise mechanism in active mixers arises from the finite capacitance at tail node of Gilbert Cell. Noise voltage which is at emitter of switching transistor creates a noise current because of parasitic capacitor Cp. The source of noise voltage at this node is shot- noise at the switching transistors. This noise contains both RF and IF components. To estimate the input-referred noise voltage, for each source of noise, a conversion gain is determined to the IF output; the magnitude of each noise is multiplied by the corresponding gain and add up all of the resulting powers, thus the total noise at the IF output is obtained; the output noise is divided by the overall conversion gain of the mixer to refer it to the input. Noise figure is the ratio of this total noise power and noise power coming from the source. The terms in the total noise power statement is weighted bearing in mind that when all switching transistors are on, noise coming from amplifying stage is cancelled due to differential output. On the other hand, noise produced at switching transistors can be negligible when input stage and Gilbert Cell act as a cascade amplifier. So, the time when all switching transistors are on was calculated and noise figure equation is obtained. Lastly, different ratios of current reduction is tried and circuits simulated. Simulations shows that when the ratio of current reduction increases, the noise figure decreases. Input reflection coefficient has a decreasing behavior by decreasing the collector current. Output third order intercept point shows great rise with decreasing the collector current but operating point changes. Optimum values are obtained with %50 current reduction and -2.5dBm local oscillator power.