Hi,
In gnuradio 3.1.2, we can actully get 16MHz bandwidth when using 8-bit
per sample (8 bit for I and Q respectively) and decimation = 4, the
lowest decimation value defined in the software.
Does anyone know whether it is possible to set demication rate to be 2
to double the bandwidth when running usrp_rx_cfile.py to record the
incoming signal?
Thanks.
Xiaowei Wei
On Thu, Sep 04, 2008 at 12:56:13AM +0000, WeiXiaowei wrote:
Hi,
In gnuradio 3.1.2, we can actully get 16MHz bandwidth when using
8-bit per sample (8 bit for I and Q respectively) and decimation =
4, the lowest decimation value defined in the software.
Does anyone know whether it is possible to set demication rate to be
2 to double the bandwidth when running usrp_rx_cfile.py to record
the incoming signal?
No, decimation of 2 does not work with the standard fpga image.
With sufficient effort, it’s probably possible to get 4-bit I & Q
working, which could allow 32MHz bandwidth.
Eric
Eric B. wrote am 2008-09-04 05:23:
On Thu, Sep 04, 2008 at 12:56:13AM +0000, WeiXiaowei wrote:
Does anyone know whether it is possible to set demication rate to be
2 to double the bandwidth when running usrp_rx_cfile.py to record
the incoming signal?
No, decimation of 2 does not work with the standard fpga image.
With sufficient effort, it’s probably possible to get 4-bit I & Q
working, which could allow 32MHz bandwidth.
The limiting factor is USB bandwidth. USB hast 480 Megabit/sec gross
data rate, which translates to a effective maximum of 32 Megabyte/sec to
transfer between the host and the USRP. The USRP samples at a rate of 64
Msamples/sec with 12 bit/sample complex, which are sent as 2x16
bit/sample = 4 bytes/sample.
(32 Mbyte/sec)/(4 byte/sample)=8 Msamples/sec. This is exactly what a
decimation factor of 8 yields.
With 8 bit/sample real, that is 16 bit/sec complex, you can double you
sampling rate, but loose 4 bit of resolution.
Patrick
Engineers motto: cheap, good, fast: choose any two
Patrick S.
Student of Telematik, Techn. University Graz, Austria