# About the latest submission of usrp\_spectrum\_sense.py **URL:** \ **Category:** GNU Radio\ **Created:** [July 25, 2013, 11:13pm UTC](https://www.ruby-forum.com/t/about-the-latest-submission-of-usrp-spectrum-sense-py/235954 "2013-07-25T23:13:22Z")\ **Posts on this page:** 4\ **Page:** 1
**Author:** ![addis\_a](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/a/c4cdca/32.png) [@addis\_a](https://www.ruby-forum.com/u/addis_a)\ **Post date:** [July 25, 2013, 11:13pm UTC](https://www.ruby-forum.com/t/about-the-latest-submission-of-usrp-spectrum-sense-py/235954/1 "2013-07-25T23:13:22Z")
Dear all, I checked your latest submission of “usrp\_spectrum\_sense.py” and I don’t understand why there is a division of the value of each bin by the sampling rate. Also I don’t understand why there is a subtraction of the noise floor from the power db value to calculate an actual db value. Kindly explain those modifications. I am working with a USRP N210 and when I run uhd\_fft to display an analog TV channel I get values around -60’s db for the central frequencies of both the audio and video signals. When I used the latest version of “usrp\_spectrum\_sense.py”, I got positive values for power\_db and when I removed the deduction of the noise floor, I got readings almost near the noise floor for all bins (including the ones at the central frequency of the audio and video signals). I hope you can point me to the right direction and correct my misunderstanding of the code. Thank you for your time. ## Kind regards, Ahmed Saeed. Ahmed Saeed Research Assistant Department of Computer Science and Engineering Egypt-Japan University of Science and Technology (E-JUST) email: removed\_email\_address@domain.invalid mobile: +201112389430 web: [http://wrc.ejust.edu.eg/people/saeed](http://wrc.ejust.edu.eg/people/saeed) ---
**Author:** ![Ahmed\_Saeed](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/a/f17d59/32.png) [@Ahmed\_Saeed](https://www.ruby-forum.com/u/Ahmed_Saeed)\ **Post date:** [July 27, 2013, 2:10pm UTC](https://www.ruby-forum.com/t/about-the-latest-submission-of-usrp-spectrum-sense-py/235954/2 "2013-07-27T14:10:41Z")
Hi Ahmed, Dividing the bin power by the sample rate in the calculation for noise\_floor\_db and power\_db (lines 286 and 287 of [uhd: enhanced version of usrp\_spectrum\_sense.py · gnuradio/gnuradio@cc12277 · GitHub](https://github.com/gnuradio/gnuradio/commit/cc12277f64207ff79a40094268c843e6b3e2d2d3#L0R286)) was done so that the power level stays constant if the sample rate is increased. This is required because the bin statistics function sums together more samples if the usrp\_rate is increased and therefore increases the bin value result. The noise\_floor\_db value is subtracted from the power\_db value to give a relative signal strength above the noise floor to make it easier to set the squelch and catch signals: noise\_floor\_db = 10_math.log10(min(m.data)/tb.usrp\_rate) power\_db = 10_math.log10(m.data[i\_bin]/tb.usrp\_rate) - noise\_floor\_db Regards, Mike – Mike J. M0MIK BSc MIET Email: removed\_email\_address@domain.invalid Web: [http://scanoo.com](http://scanoo.com) ---
**Author:** ![Ahmed\_Saeed](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/a/f17d59/32.png) [@Ahmed\_Saeed](https://www.ruby-forum.com/u/Ahmed_Saeed)\ **Post date:** [July 27, 2013, 2:13pm UTC](https://www.ruby-forum.com/t/about-the-latest-submission-of-usrp-spectrum-sense-py/235954/3 "2013-07-27T14:13:37Z")
Hi Mike, Thank you for your reply. So what if I want to sense with the sensing threshold set by the FCC for white spaces at -114 dBm or -107 dBm. How does the value of “power\_db” relates to the value I should compare against given the FCC thresholds ? Kind regards, Ahmed Saeed. On Sat, Jul 27, 2013 at 1:35 PM, Mike J. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) wrote: > ## The noise\_floor\_db value is subtracted from the power\_db value to give a > > > understand why there is a division of the value of each bin by the sampling > > noise floor for all bins (including the ones at the central frequency of > > Department of Computer Science and Engineering – Ahmed Saeed Research Assistant Department of Computer Science and Engineering Egypt-Japan University of Science and Technology (E-JUST) email: [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) mobile: +201112389430 web: [http://wrc.ejust.edu.eg/people/saeed](http://wrc.ejust.edu.eg/people/saeed) ---
**Author:** ![Ahmed\_Saeed](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/a/f17d59/32.png) [@Ahmed\_Saeed](https://www.ruby-forum.com/u/Ahmed_Saeed)\ **Post date:** [July 27, 2013, 2:39pm UTC](https://www.ruby-forum.com/t/about-the-latest-submission-of-usrp-spectrum-sense-py/235954/4 "2013-07-27T14:39:19Z")
Hi Ahmed, You will need to use a calibrated rf source to compare your USRP N210 SDR receiver against and adjust the power\_db result (without subtracting the noise floor result from it) to match. I assume it will be a simple case of adding or subtracting ‘X’ db’s to the power\_db result for each calibration power level tested against. Unfortunately I haven’t any calibrated rf signal sources to hand so am unable to confirm how easy it will be to calibrate. Regards, Mike – Mike J. M0MIK BSc MIET Email: removed\_email\_address@domain.invalid Web: [http://scanoo.com](http://scanoo.com)