1 / 35

Continuous Time Domain Filters

Continuous Time Domain Filters. A few very simple active filters. C2. K. R1. R3. C4. 2nd order Sallen and Key low-pass filter. Continuous Time Domain LPF. Generic Architecture. Check Stability!. C. R. R. C. 2nd order Sallen and Key low-pass filter. Continuous Time Domain LPF.

grazia
Download Presentation

Continuous Time Domain Filters

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Continuous Time Domain Filters A few very simple active filters. ESINSA

  2. C2 K R1 R3 C4 2nd order Sallen and Key low-pass filter Continuous Time Domain LPF Generic Architecture Check Stability! ESINSA

  3. C R R C 2nd order Sallen and Key low-pass filter Continuous Time Domain LPF Practical realization ESINSA

  4. R4 C5 R1 R3 C2 2nd order Multiple FeedBack low-pass filter Continuous Time Domain LPF Generic Architecture Stable! ESINSA

  5. R C R R C 2nd order Multiple FeedBack low-pass filter Continuous Time Domain LPF Practical realization ESINSA

  6. 10 Sallen and Key MFB 0 R*C = 1.0e-3 -10 1st order -20 -30 2nd order -40 -50 1 10 100 1000 10000 159.2 Hz Continuous Time Domain LPF ESINSA

  7. C R R R C C R C Continuous Time Domain HPF LPF HPF  2nd order Sallen and Key filter ESINSA

  8. C C R R R R R C C C Continuous Time Domain HPF LPF HPF  2nd order Multiple FeedBack filter ESINSA

  9. C R R C Differential LP Filters Single-ended Differential  Inappropriate 2nd order Sallen and Key low-pass filter ESINSA

  10. C C R R R R R R C/2 C R R R C Differential LP Filters Single-ended Differential  2nd order Multiple FeedBack low-pass filter ESINSA

  11. Continuous Domain Filters It is important to evaluate the sensivity to electrical parameters, (absolute values, matchings) especially if high Q is expected. Parameters mismatches introduce transfer function errors. Opamp gain and slew-rate have the most dangerous effects. Slew-rate, being a non linear effect, makes the analysis more difficult. Extend the analysis outside the bandwidth of the filter! ESINSA

  12. Continuous Domain Filters DO NOT UNDERESTIMATE THE NOISE ANALYSIS. ESINSA

  13. Noise! Continuous Domain Filters Input Noise Filtered Noise ESINSA

  14. Input Noise Filtered Noise dB Noise! freq Continuous Domain Filters Output Noise! Input Noise ESINSA

  15. Continuous Domain Filters Kickback Noise Still a Risk ! ESINSA

  16. Interfacing SWC modules ESINSA

  17. Interfacing SWC modules At the input of a SWC module, the signal bandwidth must be limited below the Nyquist frequency of the module. AntiAliasing filter At the output of a SWC module, the signal must be smoothed to reconstruct a continuous-time domain signal. Smoothing filter ESINSA

  18. Signal Noise Brick wall AA filter Simpler AA filter It is difficult to build linear and accurate brick wall filter Practical antialiasing filter has a more limited frequency range. AA Filter freq Nyquist Sampling ESINSA

  19. AA Filter This is the first challenge to solve. A quasi brick wall filter could use the bandwidth more efficiently. It has a very complex structure with many poles and zeroes. It uses a lot of active devices. It will not be very flat and will not generally respect the phase of the signal. Distortion, noise, power dissipation, size, … are also prohibitive. Spread of the transfer function is a killer. ESINSA

  20. AA Filter Oversampling is often a good solution. Antialiasing filters are then simpler and safer! But this is not using the bandwidth very efficiently. AA freq Nyquist Sampling Signal spread ESINSA

  21. AA Filter Oversampling? SWC modules must run faster! SWC modules are not necessarily able to be run at much higher frequencies. At least they consume much more power. ESINSA

  22. Antialiasing filter Simple LPF filter Continuous Time Domain High order LPF SWC filter SWC Module AA Filter • A solution: An antialiasing filter could be a cascade of 2 LP filters. • First filter is a loose continuous time domain LPF. • Second filter is a high order, oversampled, precise, SWC LPF. @ FS * OSR @ FS ESINSA

  23. AA Filter • We have now to build a loose low order LPF in the • continuous time domain. • Specifications: • As steep as possible • Flat response transfer at low frequencies • Linear phase at low frequencies • Very precise gain • Very low distortion • High TSNR • Low power, small size, ... ESINSA

  24. AA Filter The Sallen and Key and the Multiple FeedBack LP filters make good building blocks for antialiasing filters. They are not very steep (2nd order) but uses only one Opamp. Higher order versions exist. They could be advantageously cascaded with other filters. They request a fair OSR. ESINSA

  25. AA Filter • Expected qualities of an antialiasing filter: • it has good rejection of out of band frequencies • it respects signal integrity • spread of passive devices does not lower the yield • appropriate settling time • low power • low cost ESINSA

  26. AA Filter • Very often, the antialiasing filter is expected also to: • have a high input impedance • build an isolation of input from IC feedback noise • accurately sample the input signal • makes a programmable gain or attenuation • makes the single-ended to differential conversion • etc… • (It makes the life of the analog designer more interesting) ESINSA

  27. Signal Nyquist Sampling Brick wall Smoothing filter Simpler Smoothing filter It is difficult to build linear and accurate brick wall filter Practical smoothing filter has a more limited frequency range. freq Smoothing Filter ESINSA

  28. Smoothing filter Signal freq Nyquist Sampling Smoothing Filter Again, oversampling is often a good solution. Smoothing filters are then simpler and safer! But this is not using the bandwidth very efficiently. ESINSA

  29. Smoothing Filter • A solution: A smoothing filter could be a cascade of 2 LP filters. • First filter is a high order, oversampled, precise, SWC LPF. • Second filter is a loose continuous time domain LPF. Smoothing filter SWC Module High order LPF SWC filter Simple LPF filter Continuous Time Domain @ FS @ FS * OSR ESINSA

  30. Smoothing Filter The Sallen and Key and the Multiple FeedBack LP filters make again good building blocks for smoothing filters. They are not very steep (2nd order) but uses only one opamp. Higher order versions exist. They could be advantageously cascaded with other filters. They request a fair OSR. ESINSA

  31. Smoothing Filter • Expected qualities of an smoothing filter: • it has good rejection of out of band frequencies • it respects signal integrity • spread of passive devices does not lower the yield • low power • low cost ESINSA

  32. Smoothing Filter Smoothing filters are difficult: the continuous LP filter is not sampling the end of the phase but filter the complete waveform. This waveform is not necessarily perfect! ESINSA

  33. Smoothing Filter • Very often, the smoothing filter is expected also to: • have a (very) low output impedance • makes a programmable gain or attenuation • makes the differential to single-ended conversion • etc… • (It makes again the life of the analog designer more interesting) ESINSA

  34. Smoothing Filter Be very careful when an output amplifier is driving an external load. As this impedance is fairly unknown, it is dangerous to use an Opamp essential in the construction of the transfer function of a filter as an output driver. ESINSA

  35. Smoothing Filter  Modified feedback  Modified filter C R External Load R R C Interaction between Opamp and Load ESINSA

More Related