AnsweredAssumed Answered

Save waveform to rear usb

Question asked by AdrianP on Jul 10, 2017
Latest reply on Jul 12, 2017 by algoss

Like • Show 0 Likes0
Comment • 7

I have remote programmed the DSOX1102G scope to setup the time-base and 2 channels. This was successful.

I am using a C library on our bespoke PC/operating system.

I cannot get the waveform to capture correctly via the read USB connection to my PC. I only need to capture about 1000 points not the full detail but. Can I use SAVE:WAVeform commands to do this as a CSV and if so how? (Using setup and WAVeform:DATA only captures the first 1000 points.)

NOTE - I am ising this scope and a replacement for an old TEK scope and all my post processing algorithms need to remain the same for backwards compatibility. I am only looking for the peak values of my waveform and their relationship, I do not need the precision that a point by point waveform capture from its entire memory will give.

Best Regards

Adrian

No one else has this question

Outcomes

7 Replies

algoss Jul 10, 2017 5:21 AM
Mark Correct
Correct Answer
I'm sorry, but I can't figure out what you want. As near as I can make out, you want to capture and save 1000 points, and the scope is capturing 1000 points.

What does "but not the full detail" mean? What do you want to capture?

Are you having problems saving? Do you want to save to a separate USB memory stick, or are you trying to offload the data through the USB to your controlling PC

Al
Like • Show 0 Likes0
Actions
- AdrianP @ algoss on Jul 10, 2017 6:34 AM
  Mark Correct
  Correct Answer
  Hello Al
  
  Thanks for the reply.
  
  I want to get waveform information for channel 1 via the rear USB to the controlling PC. I have tried to setup and capture using WAVEFORM and ACQUIRE command with little success ( I know my inbuilt driver woks because I can setup the channels and time-base great.
  
  I am wanting to capture an echo waveform from a transducer under test and have an algorithm that analyses the raw scope data to determine the peak relationship of the echo in order to determine if the transducer has a reverse polarity.
  This algorithm works great to post process data from other scopes gathered in a csv format.
  
  I only need to to get the entire trace/picture from the scope trace in a csv format for 1000 points. I don't really need the full waveform trace data in its native sample resolution. But if I have to do that, so be it.
  
  Best Regards
  Adrian
  Like • Show 0 Likes0
  Actions
algoss Jul 10, 2017 7:20 AM
Mark Correct
Correct Answer
OK, you are headed in the right direction. The correct command to use is "WAVeform:DATA".

From the Programming Manual: The :WAVeform:DATA query returns the binary block of sampled data points transmitted using the IEEE 488.2 arbitrary block data format. The binary data is formatted according to the settings of the :WAVeform:UNSigned, :WAVeform:BYTeorder, :WAVeform:FORMat, and :WAVeform:SOURce commands. The number of points returned is controlled by the :WAVeform:POINts command.

It is important to check and/or modify the settings controlled by the other ':WAVeform' commands. For instance, ':WAVeform:POINts' controls how many values get transmitted, and choices other than the default 1000 are available. The default formats are 'block data', but an ASCII format is available. I don't have a DSOX1000 scope available to test that format, but it should be easy enough to figure out. I don't know if speed is important, but the block data will be faster.

If you haven't seen it, the Programming Guide is available here: http://www.keysight.com/upload/cmc_upload/All/1000_X-Series_prog_guide.pdf There is a good C programming example starting on page 872, and there is code for offloading, and decoding, block data starting on Page 877.

Let us know how this works out for you.

Al
Like • Show 0 Likes0
Actions
- AdrianP @ algoss on Jul 10, 2017 10:55 PM
  Mark Correct
  Correct Answer
  Thanks for the reply.
  
  I have got as far as you mentioned already but do not get data that matches the screen waveform.
  
  The waveform I see on screen is attached.
  
  The code I am using is below, the TMC_send etc are our library code. The trail code below seems to capture data in a tab/space delimited format but when I plot out in xls it is just a low voltage noise waveform and not what I see in the above image.
  
  //TMC_send( addr, rp_sets);
  //timebase
  TMC_send( addr, ":TIMebase:MODE MAIN\n");
  TMC_send( addr, ":TIMebase:POSition 3.4E-4\n");
  TMC_send( addr, ":TIMebase:SCALe 1E-5\n");
  //channel 1
  TMC_send( addr, ":CHANnel1:PROBe 1\n");
  TMC_send( addr, ":CHANnel1:SCALe 0.3\n");
  TMC_send( addr, ":CHANnel1:OFFSet 5.3E-1\n");
  TMC_send( addr, ":CHANnel1:INVert OFF\n");
  TMC_send( addr, ":CHANnel1:COUPling DC\n");
  TMC_send( addr, ":CHANnel1:DISPlay ON\n");
  TMC_send( addr, ":CHANnel2:BWLimit ON\n");
  //Channel 2
  TMC_send( addr, ":CHANnel2:PROBe 1\n");
  TMC_send( addr, ":CHANnel2:SCALe 0.1\n");
  TMC_send( addr, ":CHANnel2:OFFSet -2.8\n");
  TMC_send( addr, ":CHANnel2:INVert ON\n");
  TMC_send( addr, ":CHANnel2:COUPling DC\n");
  TMC_send( addr, ":CHANnel2:DISPlay OFF\n");
  TMC_send( addr, ":CHANnel2:BWLimit OFF\n");
  //Trigger
  TMC_send( addr, ":TRIGger:MODE EDGE\n");
  TMC_send( addr, ":TRIGger:SOURCe CHANnel2\n");
  TMC_send( addr, ":TRIGger:SWEep NORMal\n");
  TMC_send( addr, ":TRIGger:SLOPe POSitive\n");
  TMC_send( addr, ":TRIGger:LEVEL 6.8E-1\n");
  TMC_send( addr, ":TRIGger:HOLDoff 6.0E-5\n");
  TMC_send( addr, ":TRIGger:HFReject ON\n");
  
  //Capture waveform
  mpl_pause("Next - capture wave from scope");
  xypc( 5, 5, "Capture waveform from scope...");
  delay( 500);
  TMC_send( addr, ":WAVeform:DATA?\n");
  int n = TMC_read2( addr, sbuff, 1000,5); // blocking
  if( n < 1 )
  {
  M_log( "TDS echo read failed: %s", strerror( errno));
  memset( sbuff, 0, 1000);
  }
  else
  {
  //Save waveform
  mpl_pause("Next - save wave from scope");
  xypc( 5, 5, "saving waveform from scope...");
  delay( 500);
  char path[50];
  sprintf( path, "/sml/var/mpl/log/keysight2_rp_wave");
  FILE *f = fopen( path, "w");
  if( !f) { M_log("Failed to open file path %s", path);}
  for( int i=0; i<1000; i++)
  fprintf( f, "%d %d\n", i, sbuff[i]);
  fclose(f);
  }
  xypc( 5, 5, "Complete");
  Like • Show 0 Likes0
  Actions
  - algoss @ AdrianP on Jul 11, 2017 6:11 AM
    Mark Correct
    Correct Answer
    One thing came immediately to mind:
    
    It appears that your display scale is set to 1e-5 (10 usec/division). If you are running at full sample rate, 2GHz, then 1000 points is 500 nsec which would be 1/20 of one division. Your sample rate may be slower, but you will probably not be capturing the full screen, with 1000 points.
    
    Sample rate is not controllable with this scope, but you can find out what it is with the ':ACQuire:SRATe?' query. Once you have done that you can adjust the number of samples you offload. At a rough guess, you may need to get 200,000 points. Once you have done that, you can decimate the data to get down to a lower number of samples.
    
    Al
    Like • Show 0 Likes0
    Actions
    - AdrianP @ algoss on Jul 12, 2017 1:57 AM
      Mark Correct
      Correct Answer
      Hello
      
      I have now managed to solve the issues. I needed to change ASCii to BYTE and use an unsigned int as my sbuff datatype within my code.
      
      All works brilliantly now.
      
      Thanks
      Adrian
      Like • Show 0 Likes0
      Actions
      - algoss @ AdrianP on Jul 12, 2017 6:18 AM
        Mark Correct
        Correct Answer
        I'm glad it all worked for you.
        
        If this was an Infiniium Scope, I would encourage you to use WORD rather than BYTE.
        
        The Infiniiums linearize the 8-bit A/D output to 16 bits during calibration. They also do some DSP to improve the magnitude and phase of the data, so there are more than 256 possible output values.
        
        Al
        Like • Show 0 Likes0
        Actions

Save waveform to rear usb

Outcomes

Related Content

Recommended Content