This is a public Discussion Forum  public

Topic

    steveratcliff
    Import .csv scan files from TTi PSA1302
    Topic posted April 21, 2013 by steveratcliffCompetent, last edited November 15, 2013 by AdminProficient 
    1148 Views, 13 Comments
    Title:
    Import .csv scan files from TTi PSA1302
    Description:

    Sorry to ask this question again but I haven't found a functional answer. Just received my PSA1302 scanner and can't seem to import scan files into WWB6.6.0.41 on my MBP Unibody i7 2.66GHz OSX 10.7.5

    It has been repeatedly stated by Sam (and others?) that the .csv file needs to be two columns, the first being MHz and the second being dBm. The first column must hold an integer value in relation to kHz. I would interpret this as having no more than three decimal places, though a modified file posted by Sam labeled "Spectrum CSV File-Modified" has between 4 and 6 decimal places on all values. During my troubleshooting I have taken the .csv and removed all unnecessary information from the spreadsheet such as device info, reference values, offsets, etc. I have attempted to scan with different RBWs: 15kHz, 280kHz and 1MHz are the options with this scanner. I have also tried different step sizes, everything from 1kHz (.001 MHz) up to 1MHz. All of this to no avail!

    I've attached a non-functioning scan file (470-698) from the TTi PSA1302. I have modified it to have no scanner information in the file, only the info pertinent to WWB. Error message that I receive is, "This is the wrong type of file. Please select a valid scan file."

    I have a scan file from a PSA1301 (Crystal Springs.csv) that imports just fine into WWB6. It appears the same in Excel: two columns, the first in MHz with between 2 and 4 decimal points, the second in dBm.

    What am I doing wrong? Any thoughts, Sam or advice from other TTi users?

    Thanks!

    -Steve

    File 1:
    File 2:

    Answer

    • glasstronomic
      Hi Steve,

      I opened those files in Notepad, rather than Excel, to see the actual data and formatting.

      Crystal Springs.csv is laid out correctly:
      550.0000,-71.7
      550.1563,-70.9
      550.3125,-71.3
      550.4688,-71.2
      550.6250,-70.9
      550.7813,-71.4
      550.9375,-71.1
      551.0938,-70.7
      551.2500,-71.4
      551.4063,-71.4
      551.5625,-71.1
      551.7188,-71.6
      551.8750,-71.3
      552.0313,-71.3
      552.1875,-71.1
      552.3438,-71.2
      552.5000,-71.7
      552.6563,-71.9
      etc.

      470-698.CSV doesn't list all frequency entries to a consistent number of decimal places, and omits decimal zeroes from integers. Also, it displays as a single long line of text in Notepad, rather than two columns, although it shows two columns in Excel and when I paste it here. Hmmm, very odd... Perhaps the ASCII returns are goofed up?
      470,-76.6
      470.844,-75.7
      471.689,-76.5
      472.533,-75.2
      473.378,-75.8
      474.222,-75.4
      475.067,-75.6
      475.911,-76
      476.756,-75.7
      477.6,-73.6
      478.444,-75.9
      479.289,-75.3
      480.133,-73.5
      480.978,-75.1
      481.822,-75.3
      482.667,-66.3
      483.511,-59.8
      484.356,-74.1
      485.2,-72.9
      486.044,-75.8
      486.889,-74.3
      487.733,-75.4
      etc.
    • steveratcliff
      Thanks Jason,

      That's interesting. I should have checked it in text edit. It does seem to import into IAS just fine...
      Any suggestions about changing the output format from the scanner or importing it into my mac? Hoping to not have to go through and put in each 'Return' manually!

      -Steve
    • glasstronomic
      Hi Steve,

      Hmm. Here's a shot in the dark... Can you run Windows in Boot Camp or Parallels? Perhaps editing your file in Excel or Open Office in Windows might resolve the "return" issue when saving as a CSV.
    • buhmanb
      I spoke to one of the developers. it is true that some end-of-line (eol) termination can confuse the file parsing. But that isn't the root cause of this issue.

      See: http://en.wikipedia.org/wiki/Newline for some general background on line termination.

      The file 470-698.CSV - uses a carriage return (ASCII 13) as the eol terminator.
      The file Crystal Springs.csv - uses carriage return + line feed (ASCII 13 + 10) as the eol terminator.

      There are utilities that can be used to convert the line termination. (i.e. unix2dos, dos2unix, OpenOffice, perl/python scripts, etc.)

      I have converted the 470-678.CSV to use the standard CRLF termination, and the file still does not open.

      WWB6 is confused by the step size in the file. The developers are working on improving the import functionality. Check the manual of the scanner, see if you can control the step size to something like 25 kHz or 125 kHz.

      Hope that helps!
    • steveratcliff
      Brian,

      I checked the scanner and confirmed that it is set to 25kHz step size (at least currently, it may not have been for the file you opened). I deleted the returns and re-entered them manually in Text Edit (mac). Again, deleted all extraneous information from the top of the file. Still no love with WWB6 :-( The TTi .csv files all seem to open just fine in IAS. Should I be using the PC version of WWB6?

      Is anyone else out there successful importing files from the new PSA1302 scanner?

      Thanks,

      -Steve
    • buhmanb
      Can you attach a copy of the problematic file? (click "More Reply Options" then Attach Files)
      Attach the copy of the file as you got it from the TTi PSA1302, not after you've edited it. That way we can try to figure out the easiest way to get the data into WWB6.

      I don't have access to that device, maybe Sam or someone else on the team does.

      What are the steps you use creating this file?
      i.e. Are you using scanner manually, then saving to the internal drive, or external flash, then copy to the Mac as a .csv file?
      Or using some additional software from TTI to collect and save the file?
    • steveratcliff

      Thanks Brian,

      Attached are 3 files. These are peak scans of 470-530MHz performed on and saved in the internal memory of my PSA1302. After they have been saved, I connect my scanner as a USB device to my laptop and transfer the files across. These files were all scanned with a step size of 25kHz (.025MHz in the language of the scanner).
      File A has an RBW of 15kHz
      File B has an RBW of 280kHz
      File C has an RBW of 1MHz

      Happy to send you any file with any modifications to settings that you would like if it helps to sort this out for users of this scanner.

      Thanks!

      -Steve

    • buhmanb
      Here is what the file 470-530A.CSV looks like: (the others are similar)

      Model Type: PSA1302,
      Start Frequency: 0470.0000 MHz,
      Stop Frequency: 0530.0000 MHz,
      RBW: 15kHz,
      Video Filter: On,
      Reference Level: -20dBm,
      Fixed Offset: Off,-010.0dB,
      Graticule Scale: 10dB/div,
      Graticule Shift: 00,
      0470.0000,-076.1
      0470.2220,-076.4
      0470.4440,-075.9
      0470.6670,-076.1
      ...
      0529.5560,-076.3
      0529.7780,-075.7
      0530.0000,-075.6

      So there are 2 things about the file that is confusing the current version of WWB6:
      1. The extra "header" text
      2. The step size is 0.222 (222 kHz)

      One thing the developers told me is that WWB6 attempts to re-sample the data to 25 kHz steps so the data can be used for Frequency Coordination. It appears that the algorithm to re-sample is getting confused by the 222 kHz step sizes.

      I used a spreadsheet to round the data to 25 kHz points, and removed the header. The data does import.
      I'll talk to the developers and see what they can do to ignore the header, and better "re-sample" the data.

      As a work-around - can you see if you can get the scanner to save at consistent steps of 25 kHz, and remove the header.
      I know WWB6 will "re-sample" other steps, but 25 is an obvious step size that should be a starting point.

      Thanks for sending the files!
    • steveratcliff
      Brian wrote:
      I used a spreadsheet to round the data to 25 kHz points, and removed the header. The data does import.


      Success! Thanks Brian. It's kind of a pain in the butt but it's a solution for now. Don't know why when I set it to scan in 25kHz steps it doesn't actually do that...

      Brian Wrote:
      I'll talk to the developers and see what they can do to ignore the header, and better "re-sample" the data.

      That would be fantastic.

      Thanks!

      -Steve
    • steveratcliff
      Brian,

      Another interesting thing that I seem to find with the import function is that the file name format seems to be very important to WWB6. Default file names from the TTi are in a format TRACE001. When importing into WWB I find that the difference in file name between TRACE015 and TRACE 015 can be the difference between a successful import and an unsuccessful one. This is exacerbated by the file naming system of the TTi scanner which doesn't allow spaces. Instead I am forced to use an underscore or dash. I have found that WWB does not seem to like that format. Again, a file named "HALL_B1" is rejected by WWB whereas "HALL B1" is accepted.

      Cheers,

      -Steve
    © 2009-2014 Shure Incorporated All rights reserved.
    • Privacy
    • Powered by RightNow