AnsweredAssumed Answered

Analyzing Impedance

Question asked by mduden on Jan 4, 2016
Latest reply on Jan 11, 2016 by mduden

Like • Show 0 Likes0
Comment • 12

Hi everyone, I'm trying to write matlab program that sends sine waves to any circuit and evaulates desired component's impedance in phase-frequency and amplitude-frequency.

Now, I'm able to communicate with my 33522b waveform generator and I can make sweep on it via matlab. However, I have a problem with getting data in terms of phase-frequency and I don't have any clue about it.

Is there any guide or helpful matlab script about it ? How can i adjust my horizontal axis to frequency domain ?

Thanks in advance.

No one else has this question

Outcomes

12 Replies

algoss Jan 5, 2016 7:20 AM
Mark Correct
Correct Answer
Interesting...
- You cross-posted to two forums. I have deleted the other post.
- You don't indicate what scope you are using or what you have tried.

Most Keysight (and Agilent) scopes have an FFT or Spectrum function available. Have you looked at those?

Is this for a school project?

Al
Like • Show 0 Likes0
Actions
- mduden @ algoss on Jan 5, 2016 9:23 AM
  Mark Correct
  Correct Answer
  Yeah, it's kinda school project, I'm trying to develop a program for our electronic lab to save time for repeating measurement at circuits.
  
  I'm using 33522B Waveform generator which on I'm sending sine wave whose values decided by user to circuits. Here's the code.
  
  +function userTest(start,stop,amp)+
  
  +fgen = visa('AGILENT','GPIB0::10::INSTR');+
  +set (fgen,'OutputBufferSize',100000);+
  +fopen(fgen);+
  
  +fprintf (fgen, '*IDN?');+
  +idn = fscanf (fgen);+
  +fprintf (idn)+
  +fprintf ('\n\n')+
  
  +fprintf (fgen, '*RST');+
  +fprintf (fgen, '*CLS');+
  
  +fprintf(fgen,'SOURCE1:FUNCTION SIN');+
  +fprintf(fgen,'SOURCE1:VOLT:UNIT VPP');+
  +fprintf(fgen,convertStrAmp(amp)');+
  
  +fprintf(fgen,convertStartFrq(start));+
  +fprintf(fgen,convertStopFrq(stop));+
  
  +fprintf(fgen,'SOURCE1:SWEEP:SPACING LINEAR');+
  
  +fprintf(fgen,'SOURCE1:SWEEP:TIME 20');+
  
  +fprintf(fgen,'TRIGGER1:SOURCE IMM');+
  
  +fprintf(fgen,'SOURCE1:SWEEP:STATE ON');+
  
  +fprintf(fgen,'OUTPUT1 ON');+
  
  +fprintf(fgen, 'SYST:ERR?');+
  +errorstr = fscanf (fgen);+
  
  +% error checking+
  +if strncmp (errorstr, '+0,"No error"',13)+
  +errorcheck = 'Waveform generated without any error \n';+
  +fprintf (errorcheck)+
  +else+
  +errorcheck = 'There is a error \n';+
  +fprintf (errorcheck)+
  +end+
  
  +fclose(fgen);+
  +delete(fgen);+
  
  +end+
  
  Now I want to measure phase and amplitude of an element in a circuit (R,L,C etc.). By the way the instrument is MSO-X 2014A. Actually, I thought that I can use
  
  MEASure: FREQUENCY and MEASure: PHASE command. The problem I can't hold the value on the matlab and plot it.
  
  Unfortunatley, I couldn't find a command for spectrum in 4000's series programmer guide.
  
  Also, sorry for repeating posts. I've realized my mistake for first post and I couldn't delete it.
  Like • Show 0 Likes0
  Actions
  - algoss @ mduden on Jan 5, 2016 10:30 AM
    Mark Correct
    Correct Answer
    Maybe you mistyped, but I don't understand why you looked in the 4000 series programming guide, when you have a 2000 series scope.
    
    In any case, both the 2000 and the 4000 series scopes have an FFT function. If you want to use it, you will have to turn it on and unload the data from it.
    
    *That's probably not the correct way to do this measurement, though*. What you probably want to do is set the source to a given frequency (right now you are sweeping it) and amplitude, then read the current magnitude of the input and output, along with the phase between the two signals. Both of these measurements will be noisy, so you will probably want to make all measurements multiple times and average the results. Then you'll want to set the source to a new frequency and amplitude and make another set of measurements. You can't just sweep and make one measurement. You have to synchronize the source and the scope.
    
    Al
    Like • Show 0 Likes0
    Actions
    - mduden @ algoss on Jan 5, 2016 10:38 AM
      Mark Correct
      Correct Answer
      Ok, got what the problem is now. Thanks a lot !
      
      I couldn't find the 2k-X series guide by the way.
      Like • Show 0 Likes0
      Actions
      - algoss @ mduden on Jan 5, 2016 10:46 AM
        Mark Correct
        Correct Answer
        "I couldn't find the 2k-X series guide by the way."
        
        Hmmm... We try to make that pretty easy.
        
        I searched on "MSOX2014" and got to the Product page. I selected the 'Documents' tab, then refined the list by clicking on 'Manuals'. There were 6 in the list, including the Programming Guide in pdf and chm formats.
        
        Good luck with your application...
        
        Al
        Like • Show 0 Likes0
        Actions
        mduden @ algoss on Jan 6, 2016 10:17 AM
        Mark Correct
        Correct Answer
        Well, I came with a solution. Here's the deal. But it gives a timeout error.
        
        +function userTest(start,stop,amp,data_points)+
        
        +fgen = visa('AGILENT','GPIB0::10::INSTR');+
        +set (fgen,'OutputBufferSize',100000);+
        +fgen.Timeout=10;+
        +fopen(fgen);+
        
        +%MSO-2014A+
        
        +mso = visa('AGILENT','GPIB0::7::INSTR');+
        +set (mso,'OutputBufferSize',100000);+
        +mso.Timeout=10;+
        +fopen(mso);+
        
        +%Query Idendity string and report+
        +fprintf (fgen, '*IDN?');+
        +idn = fscanf (fgen);+
        +fprintf (idn)+
        +fprintf ('\n\n')+
        +%Query Idendity string and report+
        +fprintf (mso, '*IDN?');+
        +idn = fscanf (mso);+
        +fprintf (idn)+
        +fprintf ('\n\n')+
        
        +%Clear and reset instrument+
        +fprintf (fgen, '*RST');+
        +fprintf (fgen, '*CLS');+
        +%Clear and reset instrument+
        +fprintf (mso,'*RST');+
        +fprintf (mso,'*CLS');+
        
        +fprintf(mso,':ACQuire:TYPE NORMal');+
        
        +%Select the waveform shape, amplitude and offset+
        +fprintf(fgen,'SOURCE1:FUNCTION SIN');+
        +fprintf(fgen,'SOURCE1:VOLT:UNIT VPP');+
        +fprintf(fgen,convertStrAmp(amp)');+
        
        +%Select the frequency boundaries of the sweep+
        +fprintf(fgen,convertStartFrq(start));+
        +fprintf(fgen,convertStopFrq(stop));+
        
        +%Select the sweep mode+
        +fprintf(fgen,'SOURCE1:SWEEP:SPACING LINEAR');+
        
        +%Set the sweep time in seconds.+
        +fprintf(fgen,'SOURCE1:SWEEP:TIME 5');+
        
        +%Select the sweep trigger source+
        +fprintf(fgen,'TRIGGER1:SOURCE IMM');+
        
        +%Enable Sweep State.+
        +fprintf(fgen,'SOURCE1:SWEEP:STATE ON');+
        
        +%Enable output.+
        +fprintf(fgen,'OUTPUT1 ON');+
        
        +a = start:(stop-start)/data_points:stop-1;+
        
        +b = 1:1:data_points;+
        +c = 1:1:data_points;+
        
        +for j=1:1:data_points+
        
             fprintf(mso,'MEASure:VAMPlitude CHAN2');
            b(j) = str2double(query(mso,'MEASure:VAMPlitude?'));
            fprintf(mso,'MEASure:PHASe CHAN1,CHAN2');
            c(j) = str2double(query(mso,'MEASure:PHASe? CHAN1, CHAN2'));
            pause(5/data_points)
        
        
        
        +end+
        
        +%Enable Output+
        +% fprintf(fgen,'OUTPUT ON'); % turn on channel 1 output+
        +% Read Error+
        +fprintf(fgen, 'SYST:ERR?');+
        +errorstr = fscanf (fgen);+
        
        +% error checking+
        +if strncmp (errorstr, '+0,"No error"',13)+
        +errorcheck = 'Waveform generated without any error \n';+
        +fprintf (errorcheck)+
        +else+
        +errorcheck = 'There is a error \n';+
        +fprintf (errorcheck)+
        +end+
        
        +%closes the visa session with the function generator+
        +fclose(fgen);+
        +delete(fgen);+
        +fclose(mso);+
        +delete(mso);+
        +% disp(size(a));+
        +% disp(size(b));+
        +subplot(1,2,1);+
        +stem(a,b);+
        +xlabel('frequency [hz]');+
        +ylabel('magnitude [v_{pp}]');+
        +subplot(1,2,2);+
        +stem(a,c);+
        +xlabel('frequency [hz]');+
        +ylabel('phase [pi]');+
        
        +end+
        
        Edited by: mduden on Jan 6, 2016 11:16 AM
        Like • Show 0 Likes0
        Actions
        algoss @ mduden on Jan 7, 2016 5:36 AM
        Mark Correct
        Correct Answer
        If you're getting a timeout error, then it's time to do some debugging to figure out which command isn't working.
        
        Before you do that, you need to go back to basics and STOP SWEEPING!
        
        - You might get some sort of a curve with this code, but it won't be calibrated to frequency. You will get results at "random" times, but you won't know what frequency each measurement is captured at. You need to know that.
        - Because your frequency will be changing constantly, measuring amplitude and phase one after the other, and then averaging will give you strange, and bad results. Each measurement will be made at a different frequency, and, as I said above, you won't know what those frequencies are, or how much they change.
        - Knowing the output amplitude is great, but you need to know the input amplitude as well. Your circuit may have varying input impedance which will cause that amplitude to change as well.
        - I'm pretty sure that you are using the MEASURE commands and queries incorrectly, but I don't have a 2000 series scope to test with. All of my programming is done with Infiniium scopes. You should spend some time with the Interactive I/O feature of the Keysight I/O Libraries Suite. It allows you to fine-tune your commands. It might also be useful to download the free Command Expert SW. It will also help you get the code right, creating sequences that you can test interactively, then export directly to MATLAB scripts.
        
        You are trying to emulate a VNA, and to do that you need to learn how a VNA works. There's some basic information here: http://www.keysight.com/main/redirector.jspx?action=ref&cname=EDITORIAL&ckey=1000001258-1%3Aepsg%3Atcn&lc=eng&cc=US&nfr=-11143.0.00 It's more than you need, but they talk about testing amplifiers and filters.
        
        Al
        Like • Show 0 Likes0
        Actions
        mduden @ algoss on Jan 7, 2016 8:35 AM
        Mark Correct
        Correct Answer
        Ok, got your point. I'm sending sweeping frequencies in 5 second. So, if I select 200 data points, I have to make a 5/200 pause time for each measurement. At each measurement I'm adding new value the data array. Is it bad algorithm to do that ?
        
        I have attached some pictures. I've measured 22uF capacitor.
        
        Edited by: mduden on Jan 7, 2016 9:34 AM
        Attachments
        2016-01-07.png58.1 KB
        Like • Show 0 Likes0
        Actions
        algoss @ mduden on Jan 7, 2016 10:05 AM
        Mark Correct
        Correct Answer
        Your data looks wrong to me.
        - You don't say if the 22uF cap is in series or in parallel. In either case both graphs look wrong.
        - It's unclear if there is any termination on the output, or if it's just probed.
        - You need to convert your phase display to unwrapped phase
        - Just putting a delay in your code doesn't guarantee that the scope will make the measurements at exact intervals. It takes time for the scope to get the command, parse it, and get the data back. There are varying delays in this process.
        
        Your code should be as follows:
        - set a frequency and amplitude
        - run the scope in single mode some number of times and measure the input and output amplitudes, along with the phase each time
        - do it again
        
        You should also run a simulation to see what your results should look like.
        
        Al
        Like • Show 0 Likes0
        Actions
        mduden @ algoss on Jan 8, 2016 3:51 PM
        Mark Correct
        Correct Answer
        Ok, thank you very much. Especially thank you for your patience. I have one more question :)
        
        Considering your step and 150 ohm resistor and 22uF capacitance in series, is it possible to create complex number array (impedance) Z = R + jX for capacitance in a for loop ? For loop will measure voltages at every frequency as you mentioned, not sweeped.
        
        R = resistor value (DC), V(c) / I, I can measure current value at every measurement in the loop from the other resistor 100 ohm.
        
        X = reactance = 1/(2*pi*f*c), will measure at every frequency.
        
        at the end of the loop, it will create Z array and I think I can plot unwrap(angle(Z)) and abs(Z) for magnitude and phase of impedance.
        Like • Show 0 Likes0
        Actions
        algoss @ mduden on Jan 11, 2016 4:13 AM
        Mark Correct
        Correct Answer
        In theory, it is possible to do what you want. In practice, it's very hard.
        
        Every device, whether it's a resistor, capacitor or inductor, has resistance, capacitance, and inductance, both in series and in parallel. The same thing applies to even the wires and adapters that you use to make the measurement. You have to take into account the complex impedance of the signal source, and the probe, that you are using to make the measurement.
        
        There are very smart people who work very hard to write SW to make these calculations. There are commercial SW packages, some very expensive, that will take your results and turn them into a best fit model.
        
        Take a look at this scope probe manual, especially the Performance Plots and Spice Models.
        http://www.keysight.com/main/redirector.jspx?action=ref&cname=EDITORIAL&ckey=890537&lc=eng&cc=US&nfr=-34023.536894991.00
        
        I know the guys who created this data. I know how to interpret the data they create, but there's no way I could create these graphs and models. They have been doing this for years, and still admit that they don't know as much as they should.
        
        Al
        Like • Show 0 Likes0
        Actions
        mduden @ algoss on Jan 11, 2016 4:38 AM
        Mark Correct
        Correct Answer
        Ok, I got it, thanks for your messages and help. I'll consider this.
        Like • Show 0 Likes0
        Actions

Analyzing Impedance

Outcomes

Attachments

Related Content

Recommended Content