From Digital to RF Debugging in the Time and Frequency Domain. Embedded Systems Conference 2015 May 6-7, 2015

Transcription

1 From Digital to RF Debugging in the Time and Frequency Domain Embedded Systems Conference 2015 May 6-7, 2015

2 Agenda In this seminar we ll discuss ı The challenges of debugging mixed domain embedded systems ı Frequency domain analysis and FFT basics ı Time gating: The key to unraveling the mysterious correlation between time and frequency. ı Frequency Domain Mask: Triggering in the frequency domain ı Live Example: Debug of noise issue on digitallycontrolled attenuator chain

3 Agenda In this seminar we ll discuss ı The challenges of debugging mixed domain embedded systems ı Frequency domain analysis and FFT basics ı Time gating: The key to unraveling the mysterious correlation between time and frequency. ı Frequency Domain Mask: Triggering in the frequency domain ı Live Example: Debug of noise issue on digitallycontrolled attenuator chain

4 The Challenge of Debugging Embedded Systems Complex Interactions ı Baseband digital, RF and analog signals are interdependent Feedback control of RF by microcontroller Low speed serial busses Critical timing relationships Interference between RF and digital signals EMI related issues and interference problems

5 Agenda In this seminar we ll discuss ı The challenges of debugging mixed domain embedded systems ı Frequency domain analysis and FFT basics ı Time gating: The key to unraveling the mysterious correlation between time and frequency. ı Frequency Domain Mask: Triggering in the frequency domain ı Live Example: Debug of noise issue on digitallycontrolled attenuator chain

6 Frequency Domain Analysis Using an FFT for Spectrum Analysis Different techniques exist for viewing multiple domains on a single instrument Performing an FFT (Fast Fourier Transform) on a captured analog waveform enjoys the following benefits: ı Every analog channel is also an RF channel ı Inherent synchronization between time and frequency ı Wide frequency band capture ı Time gating can be used to locate causal effects The quality and speed of measurement relies on ı A good input system Dynamic Range with clean ADC and low noise ı HW based processing for fast FFT updates even with deep memory ı Spectrum Analyzer Like setup hiding complexity of acquisition setup

7 Frequency Domain Analysis Fourier Transform Concept A waveform can be separated into the sum of orthogonal functions Most common are sine and cosine waves The FFT breaks waveforms down into these constituent components These are represented as a set of complex numbers of magnitude and phase This is known as the spectrum

8 Frequency Domain Analysis Digital Down Conversion ı Conventional method: Calculate FFT over part or all of the acquisition e.g. 25,000 point FFT for 1 GHz cf and 100 KHz RBW and 100 MHz span Conventional Method: ı Improved method: Calculate FFT only over span of interest f C = center frequency of FFT e.g point FFT for 1 GHz cf and 100 KHz RBW and 100 MHz span Improved Method: Benefit: Faster processing of FFT.

9 ı Conventional method: FFT over complete acquisition Frequency Domain Analysis FFT Implementation first aquisition FFT 1 second aquisition FFT 2 third aquisition FFT 3 ı Improved method: FFT can be split in several FFTs and also overlapped first aquisition FFT 1 FFT 2 FFT 3 FFT 4 FFT 1 FFT 2 FFT 3 50% overlapping FFT 4 Benefit: Faster processing Faster display update rate Ideal for finding sporadic/intermittent events

10 Frequency Domain Analysis Spectrum Measurement is a Function of Time Glitches time f 1 f 2 f 3 f 4 f 5 f 6 f 7 Measurement frequency

11 Agenda In this seminar we ll discuss ı The challenges of debugging mixed domain embedded systems ı Frequency domain analysis and FFT basics ı Time gating: The key to unraveling the mysterious correlation between time and frequency. ı Frequency Domain Mask: Triggering in the frequency domain. ı Live Example: Debug of noise issue on digitallycontrolled attenuator chain

12 Time Gating A key to debug Signal characteristics change over the acquisition interval Gating allows selection of specific time intervals for analysis

13 Time Gating A key to debug Tg f 1 T g

14 Time Gating A key to debug Frequency spectrum is often a function of time Locking of a PLL EMI caused by time domain switching Time gating allows the user to select a specific portion of the waveform for frequency domain analysis Window limits frequency resolution

15 Agenda In this seminar we ll discuss ı The challenges of debugging mixed domain embedded systems ı Frequency domain analysis and FFT basics ı Time gating: The key to unraveling the mysterious correlation between time and frequency. ı Frequency Domain Mask: Triggering in the frequency domain. ı Live Example: Debug of noise issue on digitallycontrolled attenuator chain

16 User definable mask can be placed in the spectrum window Set for stop on failure If a violation of the mask occurs it will stop the entire scope acquisition (including time and frequency domain) Frequency mask Frequency Domain Mask Triggering in the Frequency Domain

17 Agenda In this seminar we ll discuss ı The challenges of debugging mixed domain embedded systems ı Frequency domain analysis and FFT basics ı Time gating: The key to unraveling the mysterious correlation between time and frequency. ı Frequency Domain Mask: Triggering in the frequency domain. ı Live Example: Debug of noise issue on digitallycontrolled attenuator chain

18 Live Example Debug of noise issue on digitally-controlled attenuator chain ı The 5 ATTEN bits show the digital signal that sources a digitally-controlled attenuator chain that controls the signal strength at the RFOUT port. ı The ATTEN bits form a 5-bit word which is 3dB per LSB.