ICEMEASURE

ICE module test routines (NextMidas Only)

<function>    Function to perform ( BAND, CHAR, NPR, ... )
<out>         Type 3000 output data file
<in1>         Type 1000/2000 scalar frequency domain input file

This primitive will perform signal processing algorithms on waveforms received
from an ICE module. The input data is assumed to be the magnitude of a
transformed waveform. Passing in complex or time domain waveform will result
in unexpected behavior. 

The valid functions are:

  BAND <out> <in1> <fs> <fbw> <thresh>
    Characterize a band-limited frequncy spectrum.

    Additional Arguments:
      fs - Sampling rate of the waveform (in Hertz)
      fbw - Fractional bandwidth the waveform occupies (perc. b.w.)
      thresh - Value for comparison (in dB)

    Outputs:
      Creates a type 3000 data file with the following sub-records:
          NF - Median Noise Floor value inside signal bandwidth (in dB)
          IN - Percentage of in-band bins above NF+thresh
          OUT - Percentage of out-of-band bins above NF+thresh

    Example:
      wave tmp1 shape=white
      firkais tmpfir band .5 .3 ,,, 10
      fdfilter tmp1 tmpfir tmp2
      fft/mag tmp2 ftmp2
      icemeas band flist ftmp2 1 .3 3
      list flist /bg
      fcalc fplot ftmp2 log 10 *
      plot fplot type=line

  CHAR <out> <in1> <fs> <fbw> <thresh> <freq1>
    Characterize a frequency band with one active signal.

    Additional Arguments:
      fs - Sampling rate of the waveform (in Hertz)
      fbw - Fractional bandwidth the waveform occupies (perc. b.w.)
      thresh - Value for comparison (in dB)
      freq1 - Frequency of the tone (in Hertz)

    Output:
      Creates a type 3000 data file with the following sub-records:
        SNR - Signal-to-Noise Ratio in dB
        SFDR - Spur-free dynamic range in dB
        NF - Median Noise Floor in dB (adjusted for FFT size)
        THRESH - Percentage of bins above NF+thresh
        SIG - Signal Amplitude in dB
        MAXSPUR - Max. Spur Amplitude in dB
        MAXSPURFREQ - Max. Spur Frequency

    Example:
      wave tmp freq=10e6 elem=128k delta=5e-9
      fft/mag tmp ftmp 32k bh92 
      icemeas char flist ftmp 200e6 1 3 10e6
      list flist /bg
      fcalc fplot ftmp log 10 * 
      plot fplot type=line

  HARM <out> <in1> <fs> <thresh> <freq1> <freq2> <order>
    Characterize a frequency band with two active tones.

    Additional Arguments:
      fs - sampling rate of the waveform (in Hertz)
      thresh - Value for comparison (in dB)
      freq1 - Frequency of the 1st signal (in Hertz)
      freq2 - Frequency of the 2nd signal (in Hertz)
      order - Number of harmonics and intermods to consider 

    Output:
      Creates a type 3000 data file with the following sub-records:
        SNR - Signal-to-Noise Ratio in dB
        S1S2 - Peak amplitude delta between the 2 signals in dB
        SFDR - Spur-free dynamic range in dB
        NF - Median Noise Floor in dB (adjusted for FFT size)
        THRESH - Percentage of bins above NF+thresh
        MAXSPUR - Max. spur amplitude in dB
        MAXSPURFREQ - Max. spur frequency in Hertz
      Optional subrecord (based upon order):
        HNF1FREQ - Nth order harmonic frequency from 1st signal in Hertz
        HNF1POW - Nth order harmonic power from 1st signal in dB
        HNF1DBC - Nth order harmonic power below carrier from 1st signal in dB
        INAFREQ - Nth order intermod 1 frequency in Hertz
        INAPOW - Nth order intermod 1 power in dB
        INADBC - Nth order intermod 1 power below carrier in dB

    Example:
      home fat ice
      %gentones tmp 100e6 128k 1 5k 4 5k 64 f1 f2 
      home restore
      fft/mag tmp ftmp 32k bh92 
      icemeas harm flist ftmp 100e6 6 f1 f2 2
      list flist /bg
      fcalc fplot ftmp log 10 *
      plot fplot type=line

  MSE <out> <in1> <in2> <segments>
    Determine the mean squared error between two waveforms. One file should be
    a templated or 'clean' version of the waveform. Specify the number of 
    (frequency) segments to partition the transformed waveform for increased
    resolution. This is particularly useful for 'noise-like' waveforms. Each
    (frequency) segment will get averaged together for comparison. Note, the
    two signals must have the same transform size (NFFT).

    Additional Arguments:
      in2 - Templated/ideal comparison signal
      segments - Number of frequency partitions

    Output:
      Creates a type 3000 data file with the following sub-records:
        AVERAGE - Average mean-squared error between two signals per segment
        WORST - Worst mean-squared error between signals (single sample)

    Example:
      wave twave1 elem=33000 amp=512
      wave nwave elem=33000 shape=white
      fcalc twave2 twave1 nwave + 
      thin twave1 mwave1 232
      thin twave2 mwave2 ,, 32k
      fft/mag mwave1 fwave1 nfft=1k win=bh92 navg=10
      fft/mag mwave2 fwave2 nfft=1k win=bh92 navg=10
      icemeasure mse flist fwave1 fwave2 10
      list flist

  NOISE <out> <in1> <fs> <fbw> <thresh>
    Characterize a noise filled frequency spectrum.

    Additional Arguments:
      fs - Sampling rate of the waveform (in Hertz)
      fbw - Fractional bandwidth the waveform occupies (perc. b.w.)
      thresh - Value for comparison (in dB)

    Output:
      Creates a type 3000 data file with the following sub-records:
        NF - Median Noise Floor with adjustment for FFT size (in dB)
        THRESH - Percentage of bins above NF+thresh
        MAXSPUR - Max. Spur Amplitude (in dB)
        MAXSPURFREQ - Max. Spur Frequnecy (in Herz)

    Example:
      calc xdel 1 1k /
      wave mwave elem=32k shape=white delta=xdel
      fft/mag mwave fwave nfft=1k win=bh92 navg=10
      icemeasure noise flist fwave 1k .5 3
      list flist /bg
      fcalc fplot fwave log 10 *
      plot fplot type=line

  NOTCH <out> <in1> <fs> <nbw> <fbw>
    Characterize the carrier suppression algorithm on the L-Band module. The
    input waveform is expected to be the frequency domain output of the L-Band
    tuner module. Note, use a small percentage (~2%) to catch if the algorithm
    is suppressing too much of the waveform.

    Additional Arguments:
      fs - Sampling rate of the waveform (in Hertz)
      nbw - Bandwidth of the notch (in Hertz)
      fbw - Fractional bandwidth the waveform occupies (perc. b.w.)

    Output:
      Creates a type 3000 data file with the following sub-records:
        MEDIAN - Median Signal Level across the bandwidth of interest (in dB)
        NOTCH - Median level of the Notch (in dB)
        TONE - Amplitude of the carrier leak-through (in dB)

    Example:
      set rate 1000000
      set card PIC1
      set flags "IOM1=LB2D|MUXCLK=P|MBITS=-16|RFOPTS=(DCS)/TC=OFF/ACQ/RCVR"
      set port TUNER1
      picd/flags=force|^flags reset card
      pic create fileo ci 4m rate
      picd/flags=^flags/port=^port set card rffreq 900
      picd/flags=^flags/port=^port set card rfbw 50
      picd/replay=-2/port=^port/flags=^flags acquire card fileo
      fft/mag fileo ftmp 64k navg=10 win=bh92
      icemeasure notch flist ftmp rate 4e3 .02
      list flist

  NPR <out> <in1> <freq1> <bw1> <freq2> <bw2> <freqn> <bwn>
    Characterize a Noise Power Ratio test signal. This function will calculate
    the amplitude difference between the two passbands. It will also calculate
    the notch depth.

    Additional Arguments:
      freq1 - First passband starting offset (percentage)
      bw1 - First passband width (percentage)
      freq2 - Second passband starting offset (percentage)
      bw2 - Second passband width (percentage)
      freqn - Notch starting offset (percentage) 
      bwn - Notch width (percentage)

    Output:
      Creates a type 3000 data file with the following sub-records:
        SIGNAL - Amplitude delta amplified spectra (in dB)
        NOTCH - Notch depth (in dB)

    Example:
      home fat ice
      %gennpr/cmplx ci_npr 10k 32k 100e6 .5 .3 .05 512 
      home restore
      fft/mag ci_npr fwave win=bh92 navg=10
      icemeasure npr flist fwave .1 .4 .64 .24 .59 .02
      list flist /bg
      fcalc fplot fwave log 10 *
      plot fplot type=line

  SONET <out> <in1> <in2> <skip2> <frame> <cor> <off>
    Look for correlations between two input files and determine BER

    Additional Arguments:
      in2 - Second input file
      skip2 - Number of samples to skip at the front end of 2nd input
      frame - Frame size of OC signal
      cor - Number of frames to correlate over
      off - Number of samples into the first input (in1) signal where 
            correlation was found (output result) 

    Output:
      Creates a type 3000 data file with the following sub-records:
      ERRORS - Number of errors
      PROC - Number of points processed
      BER - Bit Error Rate

    Example:
       home fat ice
       wave inp1 elem=16k shape=white amp=2k
       wave inp2 elem=64k shape=sin amp=4k
       mergefile inp2 inp1
       icemeas sonet flist inp1 inp2 0 1k 64 off
       list flist /bg
       say "Found match ^off samples in"
       home restore

  UDP  <out> <in1> <levels>
    Validate input ramp file 

    Additional Arguments:
      levels - Amplitude scale (2^levels) 

    Output:
      Creates a type 3000 data file with the following sub-records:
      MAXDELTA - Largest delta between samples
      ERRORS - Number of errors in file

    Example:
      home fat ice
      set levels 10
      calc amp_ 2 levels **
      calc freq_ 1 2 levels 1 + ** /
      wave mfile si 8k saw amp_ freq=^freq_
      icemeas udp flist mfile levels 
      list flist /bg
      home restore

Switches: NONE

Widgets: NONE

See Also:  ICEFAT, PIC, PICD