Re: [Discuss-gnuradio] Re g : spectrum sensing code

[sumitstop]

Hi Sreeraj,

Thanks for the reply.

sreeraj r wrote:
> 
> 
> 
>>Question - 1 
>>
>>I was working with the UHD version of spectrum sensing code.
>>
>>I have USRP2 with me. I learnt that USRP2 can achieve 50MHz of RF
bandwidth
>>with 8 bit samples and 25MHz of RF bandwidth with 16 bit samples.
>>
>>I wanted to know how does this information translates while passing
min_freq
>>ans max_freq parameters to the spectrum sensing code. Does it mean
(max_freq
>>- min_freq) = 25MHz.
>>
>>Please correct me if I am wrong :teeth:
> 
> As far as I understand spectrum sensing code is used for wideband spectrum
> analysis, that is 
> to scan across wide RF bands which USRP is not able to analyze at a time.
> So obviously 
> max_freq - min_freq > 25 MHz .
> 
> e.g Conider you have to analyze the spectrum from 400 MHz to 500 MHz. You
> can keep
> min_freq = 400 MHz and max_freq = 500MHz. If you have configured USRP
> sample rate 
> to get 25 MHz bandwidth, ideally one round of scan will be complete by 4
> tuning steps.
> (practically 6 steps as the code is using 25% overlap to reduce
> non-linearity in DDC response).
> 
> Firas posted a detailed explanation of the code in the mailing list long
> back. You can find it 
> here http://www.ruby-forum.com/topic/174437.
> @@@@@@@@@@@@@@@@@@@@@@@@@@
> I already went through the code explanation by Firas. So according to your
> saying :
> To scan 400MHz to 500MHz I have to take 4 tuning steps. Center frequency
> of tuning steps shall be 412.5MHz, 437.5MHz, 462.5MHz and 487.5 (for the
> time being lets ignore the FFT overlaping)
> And this tuning step size is calculated according to : freq_step = 0.75 *
> usrp_rate 
> where USRP rate is the sampling rate we give.
> I have a basic doubt. Since the processing bandwidth at a time is 25MHz,
> so shall I pass sampling rate = 2*bandwidth i.e. for this case 2*25Mhz =
> 50MSPS (Pl correct if I am wrong)
>>Question - 2
> 
>>Why the minimum center frequency is set as follows :
>>
>>self.min_center_freq = self.min_freq + self.freq_step/2 
> 
>>Why it simply doesn't take the minimum frequency as we provided. Setting
the
>>max_center_freq I understood somehow.
> 
> e.g. Scan range 10MHz to 40 MHz
>      USRP1 B/W = 8MHz
>      First tuning step is with centre freq 14Mhz (10+8/2 ideally) which
> will cover the bandwidth from 10 to 18 Mhz
> @@@@@@@@@@@
> Thanks , I was so intuitive :) 
> @@@@@@@@@@@
>>Question - 3
> 
>>In non UHD version of spectrum sensing code the adc_rate was fixed fro
>>USRP-1 i.e. 64M and that was used to calculate usrp_rate while in UHD
>>version of the same code usrp_rate is calculated by the sampling rate we
>>give.
> 
>>I was wondering how does the code manages to work with devices of
different
>>sampling rates i.e. USRP-1 with 64M USRP2 with 100M 
> 
> sampling rate ----> Bandwidth  ----> feq_step 
> 
>>Context : I am trying to scan WLAN channels 1, 6, 11 continuously. So I
have
>>to keep working with a total bandwidth of 76 MHz 
> 
> Check the tuning delay of the device you are using and make sure that you
> are not losing any data during the sweeping process.I may be wrong
> somewhere. Please go through Firas's explanation which will make things
> clear.
> @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
> Any standard process you know to check the tuning delay.
> 
> 
> regards
> Sreeraj Rajendran
> _______________________________________________
> Discuss-gnuradio mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
> 
> 


-----
Sumit Kr.
Research Assistant
Communication Research center
IIIT Hyderabad
India
-- 
View this message in context: 
http://old.nabble.com/Reg-%3A-spectrum-sensing-code-tp34330101p34335158.html
Sent from the GnuRadio mailing list archive at Nabble.com.