usrp-users@lists.ettus.com
Discussion and technical support related to USRP, UHD, RFNoC
View all threadsUSRP N210 Spectrum Sensing Output Issue
Hi All,
I have been working on USRP N210 devices to sample RF spectrum for purposes
of identifying available whitespace that can be used to implement WRAN
network.
The USRP devices are equipped with WBX Daughterboard capable of operating
range of 50-2200 MHz. My setup antenna is VERT900 for this experiment.
Running spectrum_sensing code within 1710-1990 MHz, I have seen that for
some frequencies, the output power(dB) and noise_floor(dB) signals does not
make sense. I ran system calibration few months on this device. The
highlighted data below belong to one center-freq. At this freq, power
signal shows +28dB to +30dB with noise_floor of -105.4dB. The next center
freq shows +6dB to +9dB and noise level at -81dB.
2016-01-26 22:16:56.634181 center_freq 864125000.0 freq 869181250.0
power_db 29.7757859896 noise_floor_db -105.402901134
2016-01-26 22:16:56.634238 center_freq 864125000.0 freq 869187500.0
power_db 30.6689845596 noise_floor_db -105.402901134
2016-01-26 22:16:56.634298 center_freq 864125000.0 freq 869193750.0
power_db 29.6794337041 noise_floor_db -105.402901134
2016-01-26 22:16:56.634346 center_freq 864125000.0 freq 869200000.0
power_db 28.9447909474 noise_floor_db -105.402901134
2016-01-26 22:16:56.634392 center_freq 864125000.0 freq 869206250.0
power_db 28.6664248343 noise_floor_db -105.402901134
2016-01-26 22:16:56.634438 center_freq 864125000.0 freq 869212500.0
power_db 29.1072308423 noise_floor_db -105.402901134
2016-01-26 22:16:56.634483 center_freq 864125000.0 freq 869218750.0
power_db 29.0124014759 noise_floor_db -105.402901134
2016-01-26 22:16:57.084144 center_freq 864875000.0 freq 865175000.0
power_db 6.42345734409 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084251 center_freq 864875000.0 freq 865181250.0
power_db 8.23228471111 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084394 center_freq 864875000.0 freq 865187500.0
power_db 7.6820290985 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084486 center_freq 864875000.0 freq 865193750.0
power_db 9.43910247372 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084553 center_freq 864875000.0 freq 865200000.0
power_db 7.60343080472 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084616 center_freq 864875000.0 freq 865206250.0
power_db 6.2934181926 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084707 center_freq 864875000.0 freq 865212500.0
power_db 6.60803587184 noise_floor_db -81.7362911404
Issue here is the high noise_floor and power signals not clear to me.
I would be grateful if anyone can help me decouple these issues.
Thank you,
Roland
So the point here is that usrp_spectrum_sense of course can't actually
measure noise floor; it's bound to be estimating it based on the sample
observation:
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
For a lot of reasons, "min(m.data)", that's the minimum FFT magnitude
squared bin, is a rather hand-wavey approach to estimating receiver
characteristics, but it's the "easiest" and hence, somewhat comparable,
approach to giving you an estimate.
Now, for some reasons, you get -105.4.. dB on all these frequencies;
that is not a coincidence, or at least would be an extremely unlikely
one; it implies that the minimum detected bin always had exactly the
same energy.
Hence: what is your RX gain?
Best regards,
Marcus
On 01/27/2016 06:46 AM, Roland Awusie via USRP-users wrote:
Hi All,
I have been working on USRP N210 devices to sample RF spectrum for
purposes of identifying available whitespace that can be used to
implement WRAN network.
The USRP devices are equipped with WBX Daughterboard capable of
operating range of 50-2200 MHz. My setup antenna is VERT900 for this
experiment.
Running spectrum_sensing code within 1710-1990 MHz, I have seen that
for some frequencies, the output power(dB) and noise_floor(dB) signals
does not make sense. I ran system calibration few months on this
device. The highlighted data below belong to one center-freq. At this
freq, power signal shows +28dB to +30dB with noise_floor of -105.4dB.
The next center freq shows +6dB to +9dB and noise level at -81dB.
2016-01-26 22:16:56.634181 center_freq 864125000.0 freq 869181250.0
power_db 29.7757859896 noise_floor_db -105.402901134
2016-01-26 22:16:56.634238 center_freq 864125000.0 freq 869187500.0
power_db 30.6689845596 noise_floor_db -105.402901134
2016-01-26 22:16:56.634298 center_freq 864125000.0 freq 869193750.0
power_db 29.6794337041 noise_floor_db -105.402901134
2016-01-26 22:16:56.634346 center_freq 864125000.0 freq 869200000.0
power_db 28.9447909474 noise_floor_db -105.402901134
2016-01-26 22:16:56.634392 center_freq 864125000.0 freq 869206250.0
power_db 28.6664248343 noise_floor_db -105.402901134
2016-01-26 22:16:56.634438 center_freq 864125000.0 freq 869212500.0
power_db 29.1072308423 noise_floor_db -105.402901134
2016-01-26 22:16:56.634483 center_freq 864125000.0 freq 869218750.0
power_db 29.0124014759 noise_floor_db -105.402901134
2016-01-26 22:16:57.084144 center_freq 864875000.0 freq 865175000.0
power_db 6.42345734409 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084251 center_freq 864875000.0 freq 865181250.0
power_db 8.23228471111 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084394 center_freq 864875000.0 freq 865187500.0
power_db 7.6820290985 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084486 center_freq 864875000.0 freq 865193750.0
power_db 9.43910247372 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084553 center_freq 864875000.0 freq 865200000.0
power_db 7.60343080472 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084616 center_freq 864875000.0 freq 865206250.0
power_db 6.2934181926 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084707 center_freq 864875000.0 freq 865212500.0
power_db 6.60803587184 noise_floor_db -81.7362911404
Issue here is the high noise_floor and power signals not clear to me.
I would be grateful if anyone can help me decouple these issues.
Thank you,
Roland
USRP-users mailing list
USRP-users@lists.ettus.com
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
Marcus,
Thank you for the timely response.
For your question, "Hence: what is your RX gain?". My earlier sample was
generated with the gain set to "None" by default. However, I ran another
tests this morning with gains set to 3.5 and 10 but the results show no
significant change.
I hope this will help narrow down options.
Thank you in advance.
Roland
On Wed, Jan 27, 2016 at 2:04 AM, Marcus Müller usrp-users@lists.ettus.com
wrote:
So the point here is that usrp_spectrum_sense of course can't actually
measure noise floor; it's bound to be estimating it based on the sample
observation:
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
For a lot of reasons, "min(m.data)", that's the minimum FFT magnitude
squared bin, is a rather hand-wavey approach to estimating receiver
characteristics, but it's the "easiest" and hence, somewhat comparable,
approach to giving you an estimate.
Now, for some reasons, you get -105.4.. dB on all these frequencies; that
is not a coincidence, or at least would be an extremely unlikely one; it
implies that the minimum detected bin always had exactly the same energy.
Hence: what is your RX gain?
Best regards,
Marcus
On 01/27/2016 06:46 AM, Roland Awusie via USRP-users wrote:
Hi All,
I have been working on USRP N210 devices to sample RF spectrum for
purposes of identifying available whitespace that can be used to implement
WRAN network.
The USRP devices are equipped with WBX Daughterboard capable of operating
range of 50-2200 MHz. My setup antenna is VERT900 for this experiment.
Running spectrum_sensing code within 1710-1990 MHz, I have seen that for
some frequencies, the output power(dB) and noise_floor(dB) signals does not
make sense. I ran system calibration few months on this device. The
highlighted data below belong to one center-freq. At this freq, power
signal shows +28dB to +30dB with noise_floor of -105.4dB. The next center
freq shows +6dB to +9dB and noise level at -81dB.
2016-01-26 22:16:56.634181 center_freq 864125000.0 freq 869181250.0
power_db 29.7757859896 noise_floor_db -105.402901134
2016-01-26 22:16:56.634238 center_freq 864125000.0 freq 869187500.0
power_db 30.6689845596 noise_floor_db -105.402901134
2016-01-26 22:16:56.634298 center_freq 864125000.0 freq 869193750.0
power_db 29.6794337041 noise_floor_db -105.402901134
2016-01-26 22:16:56.634346 center_freq 864125000.0 freq 869200000.0
power_db 28.9447909474 noise_floor_db -105.402901134
2016-01-26 22:16:56.634392 center_freq 864125000.0 freq 869206250.0
power_db 28.6664248343 noise_floor_db -105.402901134
2016-01-26 22:16:56.634438 center_freq 864125000.0 freq 869212500.0
power_db 29.1072308423 noise_floor_db -105.402901134
2016-01-26 22:16:56.634483 center_freq 864125000.0 freq 869218750.0
power_db 29.0124014759 noise_floor_db -105.402901134
2016-01-26 22:16:57.084144 center_freq 864875000.0 freq 865175000.0
power_db 6.42345734409 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084251 center_freq 864875000.0 freq 865181250.0
power_db 8.23228471111 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084394 center_freq 864875000.0 freq 865187500.0
power_db 7.6820290985 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084486 center_freq 864875000.0 freq 865193750.0
power_db 9.43910247372 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084553 center_freq 864875000.0 freq 865200000.0
power_db 7.60343080472 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084616 center_freq 864875000.0 freq 865206250.0
power_db 6.2934181926 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084707 center_freq 864875000.0 freq 865212500.0
power_db 6.60803587184 noise_floor_db -81.7362911404
Issue here is the high noise_floor and power signals not clear to me.
I would be grateful if anyone can help me decouple these issues.
Thank you,
Roland
USRP-users mailing listUSRP-users@lists.ettus.comhttp://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
USRP-users mailing list
USRP-users@lists.ettus.com
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
Hi Roland,
The yellow highlighted set of lines are all from the same dwell having
the same centre frequency and would have been derived from the same FFT
data. It stands to reason that the reported noise floor would be the
same as it comes from the weakest bin, as mentioned by Marcus
The next set of lines are presumably from the next dwell and I'd expect
to be able to work out the sample rate you are using if I could remember
how much overlap there is between dwells, but I can't remember. They
also share a common noise floor power level as previously described.
This is to be expected.
If I remember correctly, and things haven't changed, the power_db values
that are shown are with respect to the reported noise floor and are not
absolute or relative to full-scale on the ADC. The values should all be
greater than the squelch argument to usrp_spectrum_sense, and their
reference levels for the squelch and power_db are equivalent. It seems
to me that all of these 'hits' are of similar absolute power (whatever
that is).
I pottered around with usrp_spectrum_sense some time ago when I
contrived this hacktastic monstrosity:
https://github.com/g0hww/onsense
If it is the significant difference in noise floor levels that is your
main issue, is it possible that you are suffering from intermod problems
resulting from out of band signals overloading the USRP/WBX and
dynamically changing the noise floor level? You can probably tell this
by observing the state of the band of interest using the GUI waterfall
and FFT sinks in GRC or something like gqrx.
Tune your graphical receiver (gqrx or some GRC generated flow-graph) to
the centre frequencies shown in your output from usrp_spectrum_sense,
using the same gain levels and sample rate that produced the quoted output.
If you see the noise floor varying significantly without any very strong
signals in the observed spectrum, with signals that are only present
when the noise floor lifts, it is probably overload/intermod from out of
band signals causing it. If there are very strong signals in-band then
they too could be overloading your receiver.
Your best bet might be to use a physical band-pass filter to limit the
amount of RF energy reaching your USRP, though reducing the gain in the
receiver or inserting a physical attenuator might also help. If you can
identify any very strong out-of-band signals that come close to -20dbFS
or thereabouts, you could also try physical notch or band reject filters
to take them out specifically.
As it happens, I have only just solved my own intermod problems here.
These issues plagued my attempts to do anything really interesting with
onsense, which I shall now start playing with again now they are gone.
I now have a custom dual notch filter (kindly made by g8fek.com) to cut
a pair of broadcast FM stations down by 20Db and a tuneable 40dB notch
filter set to take out a local pager transmitter on 153.350Mhz, both
installed before my mast-head BBA-M variable gain pre-amp (also from
g8fek.com). Now my pre-amp is not overloaded by these signals and does
not produce intermod, but my USRP B-100 with WBX will still produce
intermod with enough gain set for weak signal listening unless I insert
a bandpass filter in its feeder. This does affect the 868MHz band for me.
In fact there are 5 independent receivers (3 conventional analogue
receivers and 2 SDRs) all receiving the same output from the BBA-M
preamp via a multi-coupler, and they will still all overload and produce
intermod to some degree from the same out-of-band signals (even with the
notch filters in front of the preamp, as the notched signals are also
amplified), unless I reduce the receiver's gain or provide some
band-pass filtering for their feeds from the multi-coupler.
The effect of the intermod is to scatter many wideband FM signals all
over the receiving range of the WBX, which come and go when the pager
transmitter fires up. With a bandpass filter in-line, the spectrum
received with my USRP/WBX looks nice and clean, with only the expected
signals present.
Apologies if this went on a bit, and I could be wrong about intermod
being the cause of your issues. Hope this helped though.
Cheers,
Darren, G0HWW
On 27/01/16 16:55, Roland Awusie via USRP-users wrote:
Marcus,
Thank you for the timely response.
For your question, "Hence: what is your RX gain?". My earlier sample was
generated with the gain set to "None" by default. However, I ran another
tests this morning with gains set to 3.5 and 10 but the results show no
significant change.
I hope this will help narrow down options.
Thank you in advance.
Roland
On Wed, Jan 27, 2016 at 2:04 AM, Marcus Müller
<usrp-users@lists.ettus.com mailto:usrp-users@lists.ettus.com> wrote:
So the point here is that usrp_spectrum_sense of course can't
actually measure noise floor; it's bound to be estimating it based
on the sample observation:
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
For a lot of reasons, "min(m.data)", that's the minimum FFT
magnitude squared bin, is a rather hand-wavey approach to estimating
receiver characteristics, but it's the "easiest" and hence, somewhat
comparable, approach to giving you an estimate.
Now, for some reasons, you get -105.4.. dB on all these frequencies;
that is not a coincidence, or at least would be an extremely
unlikely one; it implies that the minimum detected bin always had
exactly the same energy.
Hence: what is your RX gain?
Best regards,
Marcus
On 01/27/2016 06:46 AM, Roland Awusie via USRP-users wrote:
Hi All,
I have been working on USRP N210 devices to sample RF spectrum for
purposes of identifying available whitespace that can be used to
implement WRAN network.
The USRP devices are equipped with WBX Daughterboard capable of
operating range of 50-2200 MHz. My setup antenna is VERT900 for
this experiment.
Running spectrum_sensing code within 1710-1990 MHz, I have seen
that for some frequencies, the output power(dB) and
noise_floor(dB) signals does not make sense. I ran system
calibration few months on this device. The highlighted data below
belong to one center-freq. At this freq, power signal shows +28dB
to +30dB with noise_floor of -105.4dB. The next center freq shows
+6dB to +9dB and noise level at -81dB.
2016-01-26 22:16:56.634181 center_freq 864125000.0 freq
869181250.0 power_db 29.7757859896 noise_floor_db -105.402901134
2016-01-26 22:16:56.634238 center_freq 864125000.0 freq
869187500.0 power_db 30.6689845596 noise_floor_db -105.402901134
2016-01-26 22:16:56.634298 center_freq 864125000.0 freq
869193750.0 power_db 29.6794337041 noise_floor_db -105.402901134
2016-01-26 22:16:56.634346 center_freq 864125000.0 freq
869200000.0 power_db 28.9447909474 noise_floor_db -105.402901134
2016-01-26 22:16:56.634392 center_freq 864125000.0 freq
869206250.0 power_db 28.6664248343 noise_floor_db -105.402901134
2016-01-26 22:16:56.634438 center_freq 864125000.0 freq
869212500.0 power_db 29.1072308423 noise_floor_db -105.402901134
2016-01-26 22:16:56.634483 center_freq 864125000.0 freq
869218750.0 power_db 29.0124014759 noise_floor_db -105.402901134
2016-01-26 22:16:57.084144 center_freq 864875000.0 freq
865175000.0 power_db 6.42345734409 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084251 center_freq 864875000.0 freq
865181250.0 power_db 8.23228471111 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084394 center_freq 864875000.0 freq
865187500.0 power_db 7.6820290985 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084486 center_freq 864875000.0 freq
865193750.0 power_db 9.43910247372 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084553 center_freq 864875000.0 freq
865200000.0 power_db 7.60343080472 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084616 center_freq 864875000.0 freq
865206250.0 power_db 6.2934181926 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084707 center_freq 864875000.0 freq
865212500.0 power_db 6.60803587184 noise_floor_db -81.7362911404
Issue here is the high noise_floor and power signals not clear to me.
I would be grateful if anyone can help me decouple these issues.
Thank you,
Roland
_______________________________________________
USRP-users mailing list
USRP-users@lists.ettus.com <mailto:USRP-users@lists.ettus.com>
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
_______________________________________________
USRP-users mailing list
USRP-users@lists.ettus.com <mailto:USRP-users@lists.ettus.com>
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
USRP-users mailing list
USRP-users@lists.ettus.com
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
Hi Darren,
Thank you for your useful suggestions. Certainly, this is another approach
to resolving this issue. I really appreciate it.
Marcus suggested earlier to work around RX gain which I will be revisiting
to see why changing the gain did not correlate with the power signal
strength.
I will update on my latest finding later.
Thank you,
Roland
On Wed, Jan 27, 2016 at 11:32 AM, Darren Long via USRP-users <
usrp-users@lists.ettus.com> wrote:
Hi Roland,
The yellow highlighted set of lines are all from the same dwell having
the same centre frequency and would have been derived from the same FFT
data. It stands to reason that the reported noise floor would be the
same as it comes from the weakest bin, as mentioned by Marcus
The next set of lines are presumably from the next dwell and I'd expect
to be able to work out the sample rate you are using if I could remember
how much overlap there is between dwells, but I can't remember. They
also share a common noise floor power level as previously described.
This is to be expected.
If I remember correctly, and things haven't changed, the power_db values
that are shown are with respect to the reported noise floor and are not
absolute or relative to full-scale on the ADC. The values should all be
greater than the squelch argument to usrp_spectrum_sense, and their
reference levels for the squelch and power_db are equivalent. It seems
to me that all of these 'hits' are of similar absolute power (whatever
that is).
I pottered around with usrp_spectrum_sense some time ago when I
contrived this hacktastic monstrosity:
https://github.com/g0hww/onsense
If it is the significant difference in noise floor levels that is your
main issue, is it possible that you are suffering from intermod problems
resulting from out of band signals overloading the USRP/WBX and
dynamically changing the noise floor level? You can probably tell this
by observing the state of the band of interest using the GUI waterfall
and FFT sinks in GRC or something like gqrx.
Tune your graphical receiver (gqrx or some GRC generated flow-graph) to
the centre frequencies shown in your output from usrp_spectrum_sense,
using the same gain levels and sample rate that produced the quoted output.
If you see the noise floor varying significantly without any very strong
signals in the observed spectrum, with signals that are only present
when the noise floor lifts, it is probably overload/intermod from out of
band signals causing it. If there are very strong signals in-band then
they too could be overloading your receiver.
Your best bet might be to use a physical band-pass filter to limit the
amount of RF energy reaching your USRP, though reducing the gain in the
receiver or inserting a physical attenuator might also help. If you can
identify any very strong out-of-band signals that come close to -20dbFS
or thereabouts, you could also try physical notch or band reject filters
to take them out specifically.
As it happens, I have only just solved my own intermod problems here.
These issues plagued my attempts to do anything really interesting with
onsense, which I shall now start playing with again now they are gone.
I now have a custom dual notch filter (kindly made by g8fek.com) to cut
a pair of broadcast FM stations down by 20Db and a tuneable 40dB notch
filter set to take out a local pager transmitter on 153.350Mhz, both
installed before my mast-head BBA-M variable gain pre-amp (also from
g8fek.com). Now my pre-amp is not overloaded by these signals and does
not produce intermod, but my USRP B-100 with WBX will still produce
intermod with enough gain set for weak signal listening unless I insert
a bandpass filter in its feeder. This does affect the 868MHz band for me.
In fact there are 5 independent receivers (3 conventional analogue
receivers and 2 SDRs) all receiving the same output from the BBA-M
preamp via a multi-coupler, and they will still all overload and produce
intermod to some degree from the same out-of-band signals (even with the
notch filters in front of the preamp, as the notched signals are also
amplified), unless I reduce the receiver's gain or provide some
band-pass filtering for their feeds from the multi-coupler.
The effect of the intermod is to scatter many wideband FM signals all
over the receiving range of the WBX, which come and go when the pager
transmitter fires up. With a bandpass filter in-line, the spectrum
received with my USRP/WBX looks nice and clean, with only the expected
signals present.
Apologies if this went on a bit, and I could be wrong about intermod
being the cause of your issues. Hope this helped though.
Cheers,
Darren, G0HWW
On 27/01/16 16:55, Roland Awusie via USRP-users wrote:
Marcus,
Thank you for the timely response.
For your question, "Hence: what is your RX gain?". My earlier sample was
generated with the gain set to "None" by default. However, I ran another
tests this morning with gains set to 3.5 and 10 but the results show no
significant change.
I hope this will help narrow down options.
Thank you in advance.
Roland
On Wed, Jan 27, 2016 at 2:04 AM, Marcus Müller
<usrp-users@lists.ettus.com mailto:usrp-users@lists.ettus.com> wrote:
So the point here is that usrp_spectrum_sense of course can't
actually measure noise floor; it's bound to be estimating it based
on the sample observation:
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
For a lot of reasons, "min(m.data)", that's the minimum FFT
magnitude squared bin, is a rather hand-wavey approach to estimating
receiver characteristics, but it's the "easiest" and hence, somewhat
comparable, approach to giving you an estimate.
Now, for some reasons, you get -105.4.. dB on all these frequencies;
that is not a coincidence, or at least would be an extremely
unlikely one; it implies that the minimum detected bin always had
exactly the same energy.
Hence: what is your RX gain?
Best regards,
Marcus
On 01/27/2016 06:46 AM, Roland Awusie via USRP-users wrote:
Hi All,
I have been working on USRP N210 devices to sample RF spectrum for
purposes of identifying available whitespace that can be used to
implement WRAN network.
The USRP devices are equipped with WBX Daughterboard capable of
operating range of 50-2200 MHz. My setup antenna is VERT900 for
this experiment.
Running spectrum_sensing code within 1710-1990 MHz, I have seen
that for some frequencies, the output power(dB) and
noise_floor(dB) signals does not make sense. I ran system
calibration few months on this device. The highlighted data below
belong to one center-freq. At this freq, power signal shows +28dB
to +30dB with noise_floor of -105.4dB. The next center freq shows
+6dB to +9dB and noise level at -81dB.
2016-01-26 22:16:56.634181 center_freq 864125000.0 freq
869181250.0 power_db 29.7757859896 noise_floor_db -105.402901134
2016-01-26 22:16:56.634238 center_freq 864125000.0 freq
869187500.0 power_db 30.6689845596 noise_floor_db -105.402901134
2016-01-26 22:16:56.634298 center_freq 864125000.0 freq
869193750.0 power_db 29.6794337041 noise_floor_db -105.402901134
2016-01-26 22:16:56.634346 center_freq 864125000.0 freq
869200000.0 power_db 28.9447909474 noise_floor_db -105.402901134
2016-01-26 22:16:56.634392 center_freq 864125000.0 freq
869206250.0 power_db 28.6664248343 noise_floor_db -105.402901134
2016-01-26 22:16:56.634438 center_freq 864125000.0 freq
869212500.0 power_db 29.1072308423 noise_floor_db -105.402901134
2016-01-26 22:16:56.634483 center_freq 864125000.0 freq
869218750.0 power_db 29.0124014759 noise_floor_db -105.402901134
2016-01-26 22:16:57.084144 center_freq 864875000.0 freq
865175000.0 power_db 6.42345734409 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084251 center_freq 864875000.0 freq
865181250.0 power_db 8.23228471111 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084394 center_freq 864875000.0 freq
865187500.0 power_db 7.6820290985 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084486 center_freq 864875000.0 freq
865193750.0 power_db 9.43910247372 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084553 center_freq 864875000.0 freq
865200000.0 power_db 7.60343080472 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084616 center_freq 864875000.0 freq
865206250.0 power_db 6.2934181926 noise_floor_db -81.7362911404
2016-01-26 22:16:57.084707 center_freq 864875000.0 freq
865212500.0 power_db 6.60803587184 noise_floor_db -81.7362911404
Issue here is the high noise_floor and power signals not clear to
me.
I would be grateful if anyone can help me decouple these issues.
Thank you,
Roland
_______________________________________________
USRP-users mailing list
USRP-users@lists.ettus.com <mailto:USRP-users@lists.ettus.com>
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
_______________________________________________
USRP-users mailing list
USRP-users@lists.ettus.com <mailto:USRP-users@lists.ettus.com>
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
USRP-users mailing list
USRP-users@lists.ettus.com
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
USRP-users mailing list
USRP-users@lists.ettus.com
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com