Greetings,
from a spectrum of 25Mhz (GSM-900) i would like to sample as
many 200khz bands as possible. Since the data rate of the usb
bus is only 8Mhz, i cannot sample all of this 25mhz wide band.
Still 8Mhz would be enough to transfer 40 distinct 200khz wide
bands, which are spread over the 25Mhz band. As far as i understand,
i would have to use 1 daughterboard, 1 ADC and 40 DDCs for this
to work. Is it possible to build an FPGA image that does that?
If not, is it possible on the USRP2?
Thanks for helping
Sascha
Sascha Krissler wrote:
Greetings,
from a spectrum of 25Mhz (GSM-900) i would like to sample as
many 200khz bands as possible. Since the data rate of the usb
bus is only 8Mhz, i cannot sample all of this 25mhz wide band.
Still 8Mhz would be enough to transfer 40 distinct 200khz wide
bands, which are spread over the 25Mhz band. As far as i understand,
i would have to use 1 daughterboard, 1 ADC and 40 DDCs for this
to work. Is it possible to build an FPGA image that does that?
If not, is it possible on the USRP2?
You can just bring the whole 8MHz in a single block then split it on the
PC side. It’ll be horribly CPU expensive but if you code a special CUDA
block that would work.
With two daughter board you could bring 4MHz for each and monitor both
up and downlink frequency in s synchronized fashion for 20 ARFCNs.
Sylvain
You can just bring the whole 8MHz in a single block then split it on the
PC side. It’ll be horribly CPU expensive but if you code a special CUDA
block that would work.
The analysis filterbank block should easily do the above task (only a
couple of multiplies for the prototype filter and a 32/64 point FFT).
–
Sebastiaan H.
Radar Remote Sensing Group, University of Cape Town, South Africa
Tel: +27 72 950 9370