Example - - PDF document

example
SMART_READER_LITE
LIVE PREVIEW

Example - - PDF document

Example https://www.bsse.ethz.ch/bel/research/cmos-microsystems/microelectrode-systems/ _jcr_content/par/fullwidthimage/image.imageformat.fullwidth.340125254.png Prof. Andreas Hierlemann's Group Simulating noise (Discrete samples) Background on


slide-1
SLIDE 1

Example

https://www.bsse.ethz.ch/bel/research/cmos-microsystems/microelectrode-systems/ _jcr_content/par/fullwidthimage/image.imageformat.fullwidth.340125254.png

  • Prof. Andreas Hierlemann's Group

Simulating noise (Discrete samples) Background on Fourier Transforms Reference: https://web.eecs.umich.edu/~fessler/course/451/l/pdf/c5.pdf Accessed on 07 August 2018 Computing PSD Spectral leakage and Windowing Discrete Fourier Transform Discrete Time Fourier Transform Fourier Transform

Modules Page 2

slide-2
SLIDE 2

Spectral leakage and Windowing Example 1: Fs = 100Hz, Fin = 2Hz, N = 100 Multiplication in time Convolution in Frequency To ensure no spectral leakage

Modules Page 3

slide-3
SLIDE 3

Repeating the above sequence of N (100) -samples Example 2: Fs = 100Hz, Fin = 2.1Hz, N = 100

Modules Page 4

slide-4
SLIDE 4

Repeating the above sequence of N (100) -samples Example 2: Fs = 100Hz, Fin = 2.1Hz, N = 100 Windowing What is the problem? Spectral leakage Asynchronous sampling a. Signal frequency is unknown b. The signal may not lie is the desired frequency bin ( Fin/Fs ≠ m/N) 1. Periodic repetition by DFT introduces kinks --> leakage a. Additive noise 2. What is the solution? Basic idea: Less weight to signal in the ends: See signal through a window. Examples of windows R T H

Modules Page 4

slide-5
SLIDE 5

See Window_Example.m and PSDCalculator2. H B

Modules Page 5

slide-6
SLIDE 6

Examples of pulsed signals in sensor ICs

Lopez, Carolina Mora, et al. "An implantable 455-active-electrode 52-channel CMOS neural probe." IEEE Journal of Solid-State Circuits 49.1 (2014): 248-261. Braga, Leo Huf Campos, et al. "A Fully Digital 8× 16 SiPM Array for PET Applications With Per-Pixel TDCs and Real-Time Energy Output." J. Solid-State Circuits 49.1 (2014): 301-314.

Module-2e: Front-End for Detecting Pulses

26 July 2018 11:58 Modules Page 1

slide-7
SLIDE 7

How to define SNR?

  • What is optimum SNR?
  • What is the right bandwidth?
  • Since we know pulse shape, can we improve SNR further?
  • When signals are random pulses in time:

Case 1: Sufficient SNR

Koutsos, Ermis, Vlad Cretu, and Pantelis Georgiou. "A Muscle Fibre Conduction Velocity Tracking ASIC for Local Fatigue Monitoring." IEEE Trans. Biomed. Circuits and Systems 10.6 (2016): 1119-1128.

Higher SNR case

Modules Page 2

slide-8
SLIDE 8

See Matlab example: Amplitude_Thresholding.m for noise (sigma = 0.1 and 0.2) Receiver Operating Characteristic Higher SNR case

Modules Page 3