EECS 247 Lecture 8: Filters: Continuous-Time & Switched-Capacitor© 2009 H.K. Page 1EE247 Lecture 8• Continuous-time filter design considerations– Active bandpass filter design (continued)• Gm-C bandpass filter using simple diff. pair (continued)– Various Gm-C filter implementations• Performance comparison of various continuous-time filter topologies• Switched-capacitor filters– Emulating a resistor by using a switched capacitor– Tradeoffs in choosing sampling rate– Effect of sample and hold – Switched-capacitor network electronic noise EECS 247 Lecture 8: Filters: Continuous-Time & Switched-Capacitor© 2009 H.K. Page 2Summary Lecture 7• Automatic on-chip filter tuning (continued from previous lecture)– Continuous tuning (continued)• DC tuning of resistive timing element– Periodic digitally assisted filter tuning• Systems where filter is followed by ADC & DSP, existing hardware can be used to periodically update filter freq. response• Continuous-time filter design considerations– Monolithic highpass filters– Active bandpass filter design• Lowpass to bandpass transformation•Example: 6thorder bandpass filter• Gm-C bandpass filter using simple diff. pairEECS 247 Lecture 8: Filters: Continuous-Time & Switched-Capacitor© 2009 H.K. Page 3Linearity of the Source-Coupled Pair CMOS Gm-Cell• Note that max. signal handling capability function of gate-overdrive voltage() ()()()2435ii111324iiGS th GS thimax GS thrms3GSth in3a 2 5aˆˆIM3 v v ............4a 8aSubstituting for a ,a ,....ˆˆvv325IM3 ............32 1024VV VV2ˆv4VV IM33ˆIM 1% & V V 1V V 230mV≈+⎛⎞⎛⎞≈+⎜⎟⎜⎟−−⎝⎠⎝⎠≈−××=−=⇒ ≈EECS 247 Lecture 8: Filters: Continuous-Time & Switched-Capacitor© 2009 H.K. Page 4Dynamic Range for Source-Coupled Pair Based Filter()31% & 1 230rmsGS th inIMVVVVmV=−=⇒ ≈• Minimum detectable signal determined by total noise voltage• It can be shown for the 6thorder Butterworth bandpass filter fundamental noise contribution is given by:2ointgintgrmsnoisermsmax36kTvQCAssuming Q 10 C 5pFv 160 Vsince v 230mV230x10Dynamic Range 20log 63dB160x103μ−−≈==≈==≈EECS 247 Lecture 8: Filters: Continuous-Time & Switched-Capacitor© 2009 H.K. Page 5Simplest Form of CMOS Gm-Cell•Pros– Capable of very high frequency performance (highest?)– Simple design• Cons– Tuning affects max. signal handling capability (can overcome)– Limited linearity (possible to improve)– Tuning affects power dissipationRef: H. Khorramabadi and P.R. Gray, “High Frequency CMOS continuous-time filters,” IEEE Journal of Solid-State Circuits, Vol.-SC-19, No. 6, pp.939-948, Dec. 1984.EECS 247 Lecture 8: Filters: Continuous-Time & Switched-Capacitor© 2009 H.K. Page 6Gm-CellSource-Coupled Pair with Degeneration()()dsVsmalleffM3 M1,2mdsM1,2 M3mdsM3effdsCWox2VVI2VVgs thddsds2LIWdVVCggs thds oxVLds1g12ggfor g gggμμ⎡⎤−=−⎣⎦∂−=≈∂=+>>≈M3 operating in triode mode Æ source degenerationÆ determines overall gmProvides tuning through varing Vc (DC voltage source)EECS 247 Lecture 8: Filters: Continuous-Time & Switched-Capacitor© 2009 H.K. Page 7Gm-CellSource-Coupled Pair with Degeneration•Pros– Moderate linearity– Continuous tuning provided by varying Vc– Tuning does not affect power dissipation• Cons– Extra poles associated with the source of M1,2,3 Æ Low frequency applications onlyRef: Y. Tsividis, Z. Czarnul and S.C. Fang, “MOS transconductors and integrators with high linearity,”Electronics Letters, vol. 22, pp. 245-246, Feb. 27, 1986 EECS 247 Lecture 8: Filters: Continuous-Time & Switched-Capacitor© 2009 H.K. Page 8BiCMOS Gm-CellExample• MOSFET operating in triode mode (M1):• Note that if Vdsis kept constant Æ gmstays constant• Linearity performance Æ keep gm constant as VinvariesÆ function of how constant VdsM1can be held– Need to minimize gain @ node X• Since for a given current, gmof BJT is larger compared to MOS- preferable to use BJT• Extra pole at node X could limit max. freq. B1M1XIoutIsVcm+VinVbVarying Vbchanges VdsM1 Æ Changes gmM1Æ adjustable overall stage gm()M1mM1 B1xmminCWox2VVI2VVgs thddsds2LIWdCgVox dsVLgsVAggxVμμ⎡⎤−=−⎣⎦∂==∂==EECS 247 Lecture 8: Filters: Continuous-Time & Switched-Capacitor© 2009 H.K. Page 9Alternative Fully CMOS Gm-CellExample• BJT replaced by a MOS transistor with boosted gm• Lower frequency of operation compared to the BiCMOS version due to more parasitic capacitance at nodes A & BAB+-+-EECS 247 Lecture 8: Filters: Continuous-Time & Switched-Capacitor© 2009 H.K. Page 10• Differential- needs common-mode feedback ckt• Freq.tuned by varying Vb • Design tradeoffs:– Extra poles at the input device drain junctions– Input devices have to be small to minimize parasitic poles• Results in high input-referred offset voltage Æ could drive ckt into non-linear region• Small devices Æ high 1/f noiseBiCMOS Gm-C Integrator-Vout+Cintg/2Cintg/2EECS 247 Lecture 8: Filters: Continuous-Time & Switched-Capacitor© 2009 H.K. Page 117thOrder Elliptic Gm-C LPFFor CDMA RX Baseband Application-A++B-+ --A++B-+--A++B-+ -+A-+B-+ --A++B-+ --A++B-+ --A++B-+ -VoutVin+C-• Gm-Cell in previous page used to build a 7th order elliptic filter for CDMA baseband applications (650kHz corner frequency)• In-band dynamic range of <50dB achievedEECS 247 Lecture 8: Filters: Continuous-Time & Switched-Capacitor© 2009 H.K. Page 12Comparison of 7thOrder Gm-C versus Opamp-RC LPF+A-+B-+ -+A-+B-+ -+A-+B-+ -+A-+B-+ -+A-+B-+ -+A-+B-+ -+A-+B-+ -VoutVin+C-• Gm-C filter requires 4 times less intg. cap. area compared to Opamp-RCÆFor low-noise applications where filter area is dominated by Cs, could make a significant difference in the total area• Opamp-RC linearity superior compared to Gm-C• Power dissipation tends to be lower for Gm-C since OTA load is C and thus no need for bufferingGm-C Filter++--++--inVoV++--++--++--+-+-+-+-Opamp-RC FilterEECS 247 Lecture 8: Filters: Continuous-Time & Switched-Capacitor© 2009 H.K. Page 13• Used to build filter for disk-drive applications• Since high frequency of operation, time-constant sensitivity to parasitic caps
View Full Document