Various simultaneous L-matching networks are analyzed and the substantial impact of component non-idealities on the matching performance is addressed. In broadband impedance matching, the influence of parasitics is analyzed for different topologies using real and imaginary impedance equations. The observation made from this analysis is that the parasitic effect on broadband impedance matching is not significant.
Introduction
In all the design approaches, the passive components are considered ideal and their parasitic impacts are never addressed. In such cases, the complete system requires more hardware and software resources and the design will be more time-consuming. Therefore, there is a need for generic analytical approach for the design of multi-band simultaneous matching networks that account for their inherent parasitics in the design phase itself.
Aim and Motivation
But the quality factor of the on-chip passives is in the range of ten, which leads to significant influence from the parasitics associated with them. The effects of component idealities in RF circuits and the need for parasitic-aware design are presented in [14]. This control algorithm can be implemented in on-chip or software control with external interfaces.
Literature Survey
Contribution of Thesis
Introduction
Possible simultaneous dual-band circuits derived from narrow-band L-matched circuits are shown in Fig. However, due to the advantages of circuit simplicity and fewer components, L-matched networks are preferred [18]-[19]. Performance matching of networks deteriorates due to the inherent parasitics of the passive components and is observed in the literature [6], [13].
Parasitic Aware Technique For Concurrent Dual-Band Network
Equivalent Impedance method
In the first step of the proposed approach, a narrowband low-pass L-matching network is derived using the parasite detection method as reported in [20]. By separating the real and imaginary part of the load impedance and equating it with the original counterpart, we obtained the following quadratic equations. Dividing the real and imaginary parts of the input impedance and equating these to their double source impedance leads to two quadratic equations as given below.
Parasitic Absorption method
The corresponding S11 plots using these components are shown in Fig. 7 shown for high to low and low to high matching networks, respectively. Therefore, the proposed Equivalent Impedance method offers significant improvement in matching performance both for off-chip and on-chip packaging. The obtained component values for CC-DB-BP L matching circuit to match from 500 to 50 and 5 to 50 are given in Table IV.
Validation with Real Components…
The result shows that the tuning is improved over the lossy component circuit and the frequency drift is also close to zero even at low Q cases. Furthermore, the proposed methods are validated using real components in the next subsection. onchip) and Agilent ADS (off-chip) simulation tools. Coilcraft inductors from the CCI 0402CS series with a quality factor of approximately 53 and GJM15 series capacitors from Murata libraries are used respectively. On-chip spiral inductors (Q=5) and MIM capacitors from the UMC 0.18um RF CMOS process are used for on-chip validation.
Matching performance and insertion loss are plotted and verified in terms of input reflection (S11) and forward gain (S21). The values obtained for dual-band matching using off-chip components (Coilcraft and Murata) are listed in Table V. The component values obtained for on-chip matching (UMC 0.18um RF CMOS) are listed in Table VI.
From result plots (Figs. 10 and 11), it is observed that the frequency shift due to inherent parasitics of the real off-chip components was significant and this influence is almost eliminated with the proposed method. On the other hand, the insertion loss is more in on-chip implementation (Figs. 12 and 13). This is probably because on-chip inductors have higher internal resistance, which makes the loss unavoidable.
Further in the next section, the proposed method is applied to design quad band L-matching networks.
Parasitic Aware Quad-Band Matching
Equivalent impedance method for Quad Band network
The initial steps of this design method are followed in the same manner as explained in Section III-B. Further, the transformation is also performed to a different frequency. Component values for quad-band networks using equivalent impedance method at 450MHz, 900MHz, 1.2GHz and 2.4GHz. The derived component values for different frequency sets and quality factors are listed in Tables VII and VIII.
CC-QB HtoL L Matching Network S11 response with derived components using the Equivalent Impedance method at 900 MHz, 1.8 GHz, 2.4 GHz, and 5 GHz. S11 response of LtoH L-matching CC-QB network with derived components using equivalent impedance method at 900 MHz, 1.8 GHz, 2.4 GHz and 5 GHz. CC-QB HtoL L Matching Network S11 response with derived components using the Equivalent Impedance method at 450 MHz, 900 MHz, 1.2 GHz, and 2.4 GHz.
S11 response of the CC-QB LtoH L-matching network with derived components using the equivalent impedance method at 450 MHz, 900 MHz, 1.2 GHz, and 2.4 GHz.
Parasitic Absorption method for Quad Band network
Component values for quad-band network using the parasitic absorption method at 900 MHz, 1.8 GHz, 2.4 GHz and 5 GHz. Obtained components for high-to-low and low-to-high tuning using the quad-band absorption method are listed in Table IX and X.
Validation with Practical Components
The practical implementation of the proposed method for quad-band matching is done in the same way as for dual-band. The practical components (same as dual-band) from the UMC 0.18um RF CMOS process (on chip) and Coilcraft and Murata (off-chip) libraries are also used here. The quality factor of external capacitors is more frequency dependent due to the complexity of the internal parasitic properties (equivalent circuit).
This will cause the match to be imperfect when the desired bands are widely separated. On the other hand, the natural resonant frequency (FSR) of high value on-chip inductors is quite low. Therefore, designing for the same frequencies in an on-chip and off-chip scenario with the practical components available today will be a difficult task.
Although it is a trade-off in frequency selection, ISM450, GSM900, GPS1200 and ISM2400 frequency bands are chosen for validation, which will provide better performance in off-chip scenarios. The obtained components for off-chip and on-chip categories are listed in Table XI. The results show that the proposed method provides better performance in off-chip scenarios.
This type of response may be due to the lower natural resonant frequencies of high value inductors.
Conclusion
Introduction
A cascaded LC section with source impedance R1 and load impedance R2 is converted to a bandpass by replacing the inductor with a series LC branch and the capacitor with an LC shunt, as shown in Figure 17. A simplified expression after equalization and neglecting a small-value multiplier or divider is given as ,. It can be seen from the equation that the first term in the numerator and denominator are independent of frequency.
Parasitic effect of capacitor is not very prominent and therefore not considered in the load impedance equation. The output of the high-pass network is given as input to the low-pass network, and its output is given to another high-pass network as input. The improvement in matching performance from the parasitic aware technique is less compared to narrowband matching shown in Fig. 16.
The observation from the analysis of broadband impedance matching network is that the effect of component parasitic is less compared to narrowband impedance matching. Also the impact of component parasitics of broadband impedance matching network is analyzed and showed that the impact is less compared to narrowband.
Broadband matching using filter design approach
Transformation of LC-section to Bandpass netwrok
Band Streatching technique
Conclusion
Introduction
Cascade of LC sections
Cascade of two LC-sections
Both the capacitor and the inductor are lossy, generally the inductor loss dominates the total loss in an adaptation network. In this section, we analyze the effect of an inductor with a finite quality factor on the reflection coefficient of the broadband matching network. The terms multiplied or divided by have a range of 10-3, hence the impact of.
For example, for a broadband matching network with a source impedance of 50Ω and a load impedance of 20Ω, the actual impedance is for ideal values at the desired frequency. As Q increases the first and second terms in the numerator, the first, second and third terms in the denominator will yield negligible values and thus can be neglected. Off-chip, the simplified equation gives exactly the same results as the full equation.
Ri1 matches the intermediate resistance Ri2, and this Ri2 matches the load impedance R2, as shown in the figure. After we design a bandpass filter with Rin_eq terminated at both ends, Rin_eq on the input side must be transferred to Zs with an impedance inverter as shown in Figure 13. As a result, the other series inductor, L2(1-n)/ n2, becomes small enough to be implemented either as a simple line in the layout or as a connecting wire.
The use of asymmetric LC bandpass filters as input matching networks for broadband impedance matching allows for simpler unequal impedance matching structures, some voltage gain and stronger high frequency attenuation. The effect of interference in broadband matching using asymmetric network is smaller, it is verified by considering an example. 199 ), "Motorola Application Note: Impedance Matching Networks Applied to RF Power Transistors", http://www.rfwireless.rell.com/pdfs/AN 721 D.pdf.
Cascade of three LC section
Parasitic impact on Filter networks
Impedance inverting technique
Since we have made Rin_eq less than Zs, the inverting factor n in Figure 13 is less than unity, so the inductance L2(n-1)/n, which is connected in series with L1, is a negative value, resulting in smaller gate inductance Lg. In general, Rin_eq is not very small compared to Zs, so n is very close to unity according to the definition in eq s n R Z.
Analysis of parasitic impact on Asymmetrical filters
This thesis proposed two methods, Equivalent Impedance method and Parasitic Absorption method, for the design of simultaneous multi-band L-matching networks including the component non-idealities. The proposed methods have also been validated in a practical environment for different impedances and on-chip and off-chip cases. 5] Parssinen, A., “Multimode-multiband transceivers for next generation of wireless communications,” Solid-State Device Research Conference (ESSDERC), 2011 Proceedings of the European, vol., no., pp Sept.
Yu, “Dual-band 2.45/6 GHz CMOS LNA using matched network based on dual resonant transformer,” IEEE Trans. 9] Nieuwoudt, A.; agheb, T.; Nejati, H.; Massoud, Y., “Numerical Design Optimization Methodology for Wideband and Multiband Inductively Degenerate Cascode CMOS Low-Noise Amplifiers,” Circuits and Systems I: Regular Papers, IEEE Transactions on , vol.56, no.6, p. June 2009.
Conclus ion
