Newbie DSP Questions

This is merely a historical archive of years 2008-2021, before the migration to mailman3.

A maintained and still updated list archive can be found at https://lists.osmocom.org/hyperkitty/list/baseband-devel@lists.osmocom.org/.

Thomas Tsou ttsou at vt.edu
Tue Jan 3 19:11:42 UTC 2012


On Sat, Dec 31, 2011 at 5:16 PM, Avner Ezra <avner.ezra at yahoo.com> wrote:
> A) Sometimes when I scan a network bandwidth like GSM1800 using kalibrate, I
> see some channels like 820, 538, etc. When I re-scan, I cannot find them.
> Does this mean that kalibrate finds a channel when a mobile handset has a
> live conversation or sms send or receive in progress? or what?

No, it most cases it means a weak signal. Kalibrate performs
per-channel energy detection prior to offset calculation against the
FCCH, which is always active. The variance is probably just noise.

> B) I want to wideband capture for example 125 ARFCN. It needs 25MHz
> bandwidth which USRP N210 can handle and stream it easily. Even AFAIK I can
> double this number and capture 250 ARFCN using single N210 (50MHz, in
> USRPN210 data sheet it says it's capable of streaming 50MHz wide signals).

To capture 50MHz you need to 8-bit samples to avoid saturating the
Ethernet bus. So yes it's possible, though I'm not sure about the
current state of device or API support for the feature.

> How can I wideband capture 125 ARFCNs? I tried to do it using:
> ./uhd_rx_cfile.py -f `arfcncalc -a 512 -b 1800 -d` --samp-rate=25000000 -N
> 200000000 -g 70 b1.cfile
>
> What I understood and decided to write such command above:
>
> B-1) arfcncalc calculates frequency of first GSM1800 channel (512 ARFCN)
> which is start point (in above command)
> B-2) Sampling Rate is the bandwidth I want to capture, in our case it's
> 25MHz means 125 ARFCN which each ARFCN has 200kHz bandwidth
> B-3) 200M samples will be received (-N parameter)
> B-4) Gain value is 70, means it will boost antennas to maximum power to
> receive signals, I think USRPN210 max. gain is 80
> B-5) My decimation rate here using 25M sampling rate and USRP N210 which has
> 100MHz ADC, will be 4. So if I decided to read cfile I have to use 4 as
> decimation rate.

You want the centre frequency of the spectrum you want to capture -
not the start point. Also, the maximum gain value will also vary
depending on the daughterboard. Otherwise, you are correct.

> C) How can I seperate and process 200khz by 200khz channels in wideband
> captured file?

The brute force approach is to individually frequency translate each
channel down to baseband, low-pass filter, and perform sample rate
conversion. The more efficient method it to construct a multirate
channelizer. GNU Radio has blocks for either approach, though
construction may not be straightforward.

  Thomas




More information about the baseband-devel mailing list