Making Custom Oscilloscope Measurements

Making Custom Oscilloscope MeasurementsUsing MATLAB and VB

April 22nd, 2015

Signal Path Block Diagram - from Acquisition to Processing

Amp

DigitizedWaveform

ADC

Analog-to-Digital Converter Display

Processing

AcquisitionMemory

Trigger Circuit

SIMPLIFIED OSCILLOSCOPE BLOCK DIAGRAM

Analog Waveform

Three Possible Communication Configurations

Configuration 1: Remote Control Scope From Computer

Computational results reside in MATLAB software

Scope Platform

GPIB /ENET

Computer Platform

Scope Software MATLAB

4


Configuration 2: MATLAB Software Installed on Scope

Scope Platform

Scope Software MATLAB



Configuration 3: MATLAB Software Integrated Within Scope Application

MATLAB-assisted dynamic collaboration as the oscilloscope and MATLAB share native capabilities.

Scope Platform

DRIVER

Scope Software

MATLAB

INLINE PROCESSING


Scope Platform

Scope App

Computer Platform

Configuration 1: Remote Control Scope From Computer

GPIB /ENET

Scope Platform

Scope AppMATLAB

Configuration 2: MATLAB Installed on Scope

Scope Platform

Scope App

Configuration 3: MATLAB Fully Integrated Within Scope Application


INLINE PROCESSING

MATLAB-assisted DSO shares native capabilities


MATLAB

MATLAB

Workflow Model for Incorporating A New Algorithm

ApplEngr

ProdMngr

Contacts applications

Collects data, logs report, contacts product manager

EngrMngr

DsnEngr

Evaluates request, prioritizes projects

Assigns resources, places project in queue

Already moved on to a different project, no longer needs measurement

Six months later, implements inline measurement into firmware

Releases firmware with feature included

Notifies local rep

Delivered to end user

End User

Traditional Model:

SalesEngr

Requires a new feature; contacts local

representative with request

Time Scale: 6 - 9 months

Inline Processing Model:

End User

Incorporates new inline measurement today, and can begin to use it immediately

Time Scale: Immediate

Proprietary measurements can remain classified and do not need to be shared with scope vendor or the outside world

How Is A Measurement Selected?

Traditional parameters added to the measurement

How Is A Custom Measurement Selected?

Allows user-defined in-line custom measurements

Custom parameters are added to the measurement list just like traditional parameters

Fully-Integrated Custom Measurement

Add your own customized measurement or math function

And the result is fully integrated into the scope process as a measurement or trace

Inline Custom Measurement

Custom MATLAB parameter finds the time elapsed for half-life of the damped sine

The value 3.149 is the number of cycles that have occurred when the signal reaches 50% of its peak amplitude

Real-Time Modification of Custom Measurement

MATLAB dialog box displays MATLAB responses

On-scope editor allows loading, saving, and real-time modification of script algorithms

Arbitrary Waveform Generated on Scope

OutResult.Samples = InResult.SamplesstartData = 0endData = OutResult.SamplesnewNumPoints = endData - startData

ReDim newDataArray(OutResult.Samples)unscaledData = InResult.DataArray(False)Randomize TimerPi = 3.14159 : TwoPi = 2.0 * PiWaveLength = 0.02 ' as a fraction of the timebaseWL = WaveLength * newNumPointsWN = TwoPi / WLTimeConstant = 0.1 ' as a fraction of the timebaseTC = TimeConstant * newNumPointsA = 30000Noise = 100.0

For i = WL To endData - 1Y = A * (exp(-i/TC)) * sin(WN * i)Y = Y + Noise * (rnd+rnd+rnd+rnd+rnd+rnd-3.0)newDataArray(i) = YNext

OutResult.DataArray(False) = newDataArray

The damped sinusoid wassimulated mathematically(scope simulates but does not output waveform)

Waveform now exists inscope memory buffer

Peak Decay Ratio

Custom parameter finds the peak decay ratio of the damped sinusoid. The 0.82 reading indicates each peak in the waveform is 0.82 the amplitude of the adjacent previous peak.

Application Example:MATLAB Digital Filtering

MATLAB Butterworth Filter

Swept Sine

Low Pass Butterworth

MATLAB Butterworth Filter

MATLAB Signal Processing Toolbox supports up to 500th order Butterworth filter

This MATLAB filter was implemented with only 5 lines of code

Wavelet Transforms

WformOut = WformIn1(1:2:end)-WformIn1(2:2:end);

Because of vectorization in MATLAB, a Haar Wavelet can be performed usingonly one line of code:

Applications for Wavelets include filtering, data compression, and signal processing

Notch Filter

Swept Sine

Notch Filtered Swept Sine

Example Mixing Mathematical Waveforms with Input(Real signal captured by the scope with noise added algorithmically)

Pure Mathematical Gaussian Noise

Input Signal

Input Signal + Noise

Automatic Leveling Filter

Filter removes slow mean variations by repositioning waveform data

User-designed FFT Windowing

Input signal

Custom window applied to data

FFT is now performed on custom window

Any FFT windowing algorithm can be used with FFT

Application Example:MATLAB Inverse FFT

Inverse FFT performed with one line of source code

WformOut = ifft(WformIn1 + j*WformIn2);

Application Example:Graphic Programming with MATLAB

Processing Web Building parameters or math functions from basic building blocks

Processing Web Building parameters or math functions with MATLAB operators

Simulink Blocksets Can Be Used for Realtime Simulation

Application Example:Display Techniques

3-D Display of Data

Surface mesh shows progressionof square wave over time

Rotating the Surface Viewing AngleImage vantage point can rotate on 3 axes while viewing this Swept Sine

Custom Display Techniques

Waterfalls and 3-D Data Rotation

Custom Graphical User Interfaces

Crest Factor is computed and displayed via a custom MATLAB graphical user interface

Application Examples:Custom Timing Measurements

IEEE Risetime Definition

EMC Risetime Custom Definition


IEEE Risetime Definition


Custom EMC Pulse Measurement

Crossing Point Between Two Waveforms

The editor is provided in the scope application

In case of error, the status will be set to error and gives the number of the concerning line.

Possible to write the code in a separate editor and import it

Export the code to a file

When saving the setup, the code will be saved internally for further processing, so no need to re-write it.

Crossing point time between two waveforms

Takes two internal parameters as input and gives one as an output

The user provides the crossing level in the scope’s time@lvl parameter

This combination uses the scope’s internal parameters as input

Application Example:Waveform Correlation

Waveform CorrelationCustomization allows the scope to apply the spreadsheet's

correlation capability for waveform analysis

Waveform and measurement results can be "pushed" directly into a spreadsheet

Measurements on spreadsheet-generated trace

Trace generated

from spreadsheet

Acquired

Trace

Application Example:Decision Process Math Operators

Dynamic Automation Controls

Input Waveform

Static Result

Continuously Sampled Output

Dynamic Automation Controls

MATLAB Dynamic Feedback During Testing

h=actxserver('LeCroy.XStreamDSO');vertoffset = get(h.Acquisition.C2.VerOffset,'Value')if (h.PassFail.Q1.Out.Result.Value) == 0 %% mask test fails vertoffset = get(h.Acquisition.C2.VerOffset,'Value') if (vertoffset > 0) %% trace offset is positive set(h.Acquisition.C2.VerOffset,'Value',(vertoffset + 0.005)); else set(h.Acquisition.C2.VerOffset,'Value',(vertoffset - 0.005)); endend

3

2

1 Custom scripts can use Automation Controls to act as a feedback mechanism to modify scope settings based on measurement results

This example uses Automation Controls to automatically modify scope settings and move waveform clear of mask region

= automation controls

Mask test pass/fail result queried and scope channel vertical offsetmodified using Automation

Measurement on Trace 1 is Gated by Pulses of Trace 2

This example shows how a measurement definition can be dynamically determined by the properties of the waveform data itself. In this example, the portion of data where the measurement takes place on Trace 1 is dependent on the area marked between the two pulses of Trace 2.

Script has not begun computing average becausePass/Fail condition has not yet been True

Script begins averaging F8

Averaging continues on acquisitions only when Pass/Fail condition is True

Averaging pauses whenever Pass/Fail condition is False

Averaged waveform takes shape of C1 after 287 sweeps. 79 averages have taken place because Pass/Fail condition was True 79 out of 287 sweeps

Application Example:Linear Regression / Slope Intercept

Oxygen Depletion in a Muscle Fiber

Experiment to measure oxygen depletion in muscle fiber

Nitrogen is injected into container to displace oxygen

Trend line records oxygen level as a function of time

Long-term trend line can be extracted as a waveform

Long-term trend line of oxygen vs. time

Linear Regression to Determine Slope Intercept of Oxygen Depletion at t=0

Application Example:Precision Burst Timing

A ground-fault circuit interruptor (GCFI) is a device that disconnects a circuit whenever it detects that the electric current is not balanced between the energized conductor and the return, and is used to prevent injury caused by electric shocks.

The measurement goal is to determine the exact amount of time that the 60 Hz cycle is present before the GFCI disables the output, and determine the start, stop, and duration of the ground fault circuit interrupter tripping time.

Ground fault circuit interruptor

GFCI Test Panel

Ground Fault Circuit Interrupter Tripping Time Measurement

The exact start and stop time of each burst is variable.

The burst length includes non-integer portions of the waveform cycles.

Ground Fault Circuit Interrupter Tripping Time Measurement Challenge

Non-integer portions of the waveform are not measured by standard measurements (such as period) by oscilloscopes

Timing measurement example using IEEE pulse parameters

GFCI Burst measurement requirement

The use of cursors is not desirable because the burst length varies for each acquisition and the replacement of manual cursors by the operator significantly reduces overall test time

Burst length varies from acquisition to acquisition and includes non-integer cyclical values. The exact size and shape of the non-integer portion is also variable.

Example GFCI burst containing 6+ cycles

Example GFCI burst containing 5+ cycles

Zoom at start of burst showing fine placement of cursor on start time

Zoom at end of burst showing fine placement of cursor on stop time

The GFCI burst length varies for each acquisition. The manual placement of cursors and zooms (both of which need to be repositioned for every acquisition by the operator), is time-consuming and significantly reduces overall test throughput.

Scripting reports accurate burst start and stop time with 1 ns precision on 100 ms time capture window

Scripting dynamic repositions labels and dynamically zooms on every acquisition

Measurements-on-statistics shows range, median, mode, etc. of cumulative burst times

Script only 416 lines of code, covers all corner case scenarios, calculates burst times, and dynamically repositions zooms and labels

Application Example:RF Signal Processing

Hilbert Transform on 10 GHz RF Burst using In-line MATLAB

Hilbert Transform on 10 GHz RF Burst using In-line MATLAB

Hilbert Transform on Burst using In-line MATLAB

Getting Started Resources

Sample MATLAB Routines can be downloaded from LeCroy’s Website

Pre-written math and measurement samples on Teledyne LeCroy’s web site www.teledynelecroy.com/MATLAB

http://www.lecroy.com/

Application Note: Implementing MATLAB Filters on Waveform Data

Application Note: Decoding NRZ Data with MATLAB

Application Note: Interfacing with Simulink

Application Note: MATLAB COM object programming examples

Making Custom Oscilloscope MeasurementsQuestions?

Technology

Making Custom Oscilloscope Measurements