AnsweredAssumed Answered

Agilent infiniium 54831D MSO - Capture using Matlab

Question asked by Gerald on Feb 7, 2011
Latest reply on Feb 16, 2011 by algoss

Like • Show 0 Likes0
Comment • 9

hi,

I want to measure the current on channel 3!

The time across the screen should be 10ms, resolution is 10us/DIV. I need to record 1000 samples therefore.
10us/DIV * 1000 = 10ms screen.

Does the following configuration work therefore? How to set :WAVEFORM:POINTS and :ACQuire:POINts?? what :WAVEFORM:POINTS:MODE will i need??

...
fprintf (oszi, ':TIMebase:RANGe 10.00E-03');
...
fprintf (oszi, ':CHANnel3:UNITs AMPere');
...
fprintf (oszi, ':TRIGger:EDGE:SOURce CHANnel1;SLOPe POSitive');
...
fprintf (oszi, ':STOP'); % is this necessary???
fprintf (oszi, ':ACQuire:MODE RTIMe;AVERage OFF;POINts 2000');
fprintf (oszi, ':WAVEFORM:SOURCE CHANnel3');
fprintf (oszi, ':WAVEFORM:POINTS:MODE BINary');
fprintf (oszi, ':WAVEFORM:BYTeorder LSBFirst');
fprintf (oszi, ':WAVEFORM:POINTS 1000');
fprintf (oszi, ':WAVeform:DATA?');

how can i plot the measurement data?

Bye

No one else has this question

Outcomes

9 Replies

algoss Feb 9, 2011 6:09 AM
Mark Correct
Correct Answer
Here are some cut and paste sections from some MATLAB code I've written. They were written for the newer scopes, so the syntax might have changed. Your mileage may vary. I have added a few non-MATLAB comments

% Set the initial instrument parameters
Always stop the scope
fwrite(inst, ':stop;:cdis');
fprintf(inst, ':acquire:points %i', depth);
fprintf(inst, ':acquire:srate %d', sampleRate);
fwrite(inst, ':waveform:view all');
fwrite(inst, ':acquire:interpolate off');
fwrite(inst, 'WAV:SOUR CHAN1');

fwrite(inst, '*cls');
fwrite(inst, ':single');

This works if you are in triggered mode, not in auto mode. It makes sure that the scope is ready to give you data
ter = 0;
while(ter ~= 1)
    fwrite(inst, ':ter?');
    stringTer = fscanf(inst, '%s');
    ter = str2double(stringTer);
end

fprintf(inst, ':wav:format WORD');
fprintf(inst,':WAVEFORM:BYTEORDER LSBFirst');

fwrite(inst, sprintf(':wav:data?'));
data(1:depth) = binblockread(inst, precision);

% read last character, this is a dummy read.
    C = fread(inst,1);

% Get the vertical parameters
fwrite(inst, ':wav:yor?');
yor = str2double(fscanf(inst))
fwrite(inst,':wav:yinc?');
yinc = str2double(fscanf(inst))

% Convert the data into actual vertical (voltage) values
data = data * yinc + yor;

There are similar commands to get the Horizontal information. Plotting is left as an exercise to the user, but I know from experience that plot(data) should work, at least to verify that you have the correct data
Like • Show 0 Likes0
Actions
Gerald Feb 9, 2011 12:53 PM
Mark Correct
Correct Answer
Hello,

Thanks for you help.

I tested the following exaple: http://www.mathworks.com/matlabcentral/ ... -interface
Can you please test this? It seems that there is an problem in capturing when channel 1 has an offset...Could you also verify this?
like.: fprintf (oszi, ':CHANnel1:OFFSet 2E+00'); or fprintf (oszi, ':CHANnel1:OFFSet -2E+00');

Bye
Like • Show 0 Likes0
Actions
algoss Feb 10, 2011 5:02 AM
Mark Correct
Correct Answer
I don't have a 54831D, or access to one, so I have no way to test it.

You say that there's a "problem in capturing when channel 1 has an offset", but you don't say what the problem is.

Have you tried using the Interactive I/O capability in the Agilent I/O libraries? You can type SCPI commands, and see the results.
Like • Show 0 Likes0
Actions
Gerald Feb 10, 2011 7:57 AM
Mark Correct
Correct Answer
The problem is that the captured curve looks different...not the same as on the oscilloscope..somehow flipped and moved...

i think you can test this example with any other agilent oscilloscope! can you give a try?

cu
Like • Show 0 Likes0
Actions
algoss Feb 11, 2011 2:01 PM
Mark Correct
Correct Answer
OK, I tried it, and yes, it's not correct. I stepped through the code and saw multiple places where there were problems. I won't be able to look more closely at it for several days, as I'll be traveling. Since the problem is in the code, not in the scope, you should send your comments to Vinod on the Mathworks site, including screenshots, etc,.
Like • Show 0 Likes0
Actions
Gerald Feb 16, 2011 12:37 AM
Mark Correct
Correct Answer
Hello Algoss,

Did you already have time to take a look more closer? Is there another Matlab Code for Capturing which works fine...can you provide a link?

Bye Gerald
Like • Show 0 Likes0
Actions
Gerald Feb 16, 2011 1:08 AM
Mark Correct
Correct Answer
Hello!

I found the bug, its a sign problem!

its in:
waveform.RawData = binblockread(oszi,'uint16');
you should use:
waveform.RawData = binblockread(oszi,'int16');

the problem occurs if the signal is above and below the zero offset!!

can you verify this? does it work if you make this change?

bye Gerald
Like • Show 0 Likes0
Actions
algoss Feb 16, 2011 4:59 AM
Mark Correct
Correct Answer
Gerald,

I just got back to my office, and got my scope set up. I have not had a chance to check out the script again. I saw a few problems with it the last time I looked at it.

Have you tried to contact Vinod through the Mathworks File Exchange? He wrote the code and will provide support for it.

AG
Like • Show 0 Likes0
Actions
algoss Feb 16, 2011 11:05 AM
Mark Correct
Correct Answer
OK, I found 3 problems in the code...

Line 36 was: fprintf(visaObj,':WAV:POINTS 1000');
Should be: fprintf(visaObj,':ACQ:POINTS 1000');

Since 1000 is the default depth after an autoscale, you wouldn't see this as a problem unless you changed the depth before you ran the code. It would never switch to 1000
--------------------------
You found this next one:
Line 62 was: waveform.RawData = binblockread(visaObj,'uint16');
Should be: waveform.RawData = binblockread(visaObj,'int16');
---------------------------
Line 113 was: fprintf(visaObj,':DISPLAY:DATA? BMP, SCREEN, GRAYSCALE');------ Or was it GREYSCALE? I don't remember
should be: fprintf(visaObj,':DISPLAY:DATA? BMP, SCREEN');

'GRAYSCALE' may have been a keyword at one time, but no longer is.

AG
Like • Show 0 Likes0
Actions