Re: Determining peak instantaneous power in waveform

[Ron Economos]

I have a block that may be useful for you. It just prints the peak IQ level. It's usually placed just before a hardware sink block.

https://github.com/drmpeg/gr-iqlevels

The output looks like this:

peak real = +1.631239e-01, -1.961090e-01, 0, 0
peak imag = +2.698523e-01, -1.962542e-01, 0, 0

It's printing the maximum I and Q level in both positive and negative directions and the total number of times they exceeded +/- 1.0.

I also have an offline program that does PAPR calculations. It just takes a float IQ file as input.

https://github.com/drmpeg/dtv-utils/blob/master/papr.c

The output looks like this:

Peak magnitude = 0.869561
average power = 0.040009514, peak power = 0.756135583 @ 121986288

Maximum PAPR = 12.764364
percentage above 0 dB = 36.79375350
percentage above 1 dB = 28.40112150
percentage above 2 dB = 20.49609423
percentage above 3 dB = 13.58977407
percentage above 4 dB = 8.10234174
percentage above 5 dB = 4.22610082
percentage above 6 dB = 1.86289176
percentage above 7 dB = 0.66361316
percentage above 8 dB = 0.18129208
percentage above 9 dB = 0.03537920
percentage above 10 dB = 0.00439378
percentage above 11 dB = 0.00032525
percentage above 12 dB = 0.00001121

peak real positive = 0.840840, peak imaginary positive = 0.798450
peak real negative = -0.791407, peak imaginary negative = -0.821772

peak real positive @ 1703752736, peak imaginary positive @ 1624430697
peak real negative @ 1846543928, peak imaginary negative @ 2753521145

Ron

On 5/17/22 12:44, John Ackermann N8UR wrote:

> Hi --
>       
>       
>       I have broadband HF data recorded off the air (ham radio bands 384
>       kHz wide) and am trying to figure out what sort of gain settings I
>       need to play it back on a USRP without distortion.  My worry is
>       the way multiple fairly strong signals could add together in phase
>       to create momentary spikes way higher than the visible average (I
>       think this is called crest factor in digital signals?).
>       
>       
>       Is there a (relatively) straightforward way to measure this with
>       Gnuradio blocks?  If so, can anyone point me in the right
>       direction to build a GRC flow graph to read an IQ file and print
>       out the peak instantaneous power seen in the file, relative to
>       100% (ie, I don't need calibrated numbers, just to know how to set
>       the gain in Gnuradio for undistorted playback).
>       
>       
>       Thanks!
>       
>       John
>       
>       
>