Hi again,
I recently found out that the Bandwidths of the XCVR 2450 Daugtherboard
are as
follows for the RX and TX side respectively:
Bandwidths (Hz):
RX: 15M, 19M, 28M, 36M; (each +-0, 5, or 10%)
TX: 24M, 36M, 48M
Can someone elaborate on these values please and how to decide upon
which one to
use? As I can understand, this is dependent on the specific application,
however, I am puzzled as to why they are different since these values
correspond
to the bandwidth allowed in the bandpass filter prior to
transmission/reception.
Thank you once more for your time.
Best Regards,
Konstantinos
On Wed, Sep 28, 2011 at 5:18 AM, Konstantinos
[email protected]wrote:
Can someone elaborate on these values please and how to decide upon which
one to
use? As I can understand, this is dependent on the specific application,
however, I am puzzled as to why they are different since these values
correspond
to the bandwidth allowed in the bandpass filter prior to
transmission/reception.
These filters are programmable low-pass filters which operate at
baseband.
In general, you just need to set them wide enough for the signals you
wish
to deal with.
Matt
On 28/09/2011 11:19 AM, Matt E. wrote:
values correspond
Discuss-gnuradio mailing list
[email protected]
Discuss-gnuradio Info Page
My impression is that, except in special circumstances, the analog
filters are set “appropriately”, if the hardware has settable filters.
The analog bandwidth should generally be set to be somewhere less than
Fs/2, for whatever Fs is in use on the motherboard–
64Msps on the USRP1/E100 and 100Msps on the USRP2/N2XXX family.
Since the FPGA normally does digital filtering, as long
as the analog bandwidth is set to less than Fs/2, you shouldn’t need
to adjust the filters, and I think UHD does this automatically.
Thanks for your reply.
However, I am a little bit confused since in your reply you mention that
these
are programmable low-pass filters which operate at baseband, while in
the code
the associated comment says that this is the bandwidth of the bandpass
filter on
the board. Secondly, why are these baseband filters? These are on the
daughterboard so I would expect them to operate on the intermediate
band, since
the daughterboard converts complex IF to Real RF. Please let me know
where I get
this wrong.
Thank you for your time again.
BR,
Konstantinos
On 28/09/2011 2:32 PM, Konstantinos wrote:
Thank you for your time again.
BR,
Konstantinos
Most of the daughtercards convert to/from complex-baseband to/from real
RF. The analog filters operate at baseband, and are therefore
low-pass filters.
There are exceptions–the TVRX and TVRX2 use a non-zero IF. But all the
other daughtercards use complex-baseband.
So, in complex baseband, if you want a total bandwidth of 40MHz, what
you’re actually asking for is -20Mhz to +20Mhz. Which means that
I and Q are both low-pass filtered to 20MHz.
Thanks for your reply.
I think I understand, however it would be helpful if you could clarify
where on the chain the complex IF to Real RF conversion takes place. I
was reading from various sources that the motherboard is the one that
handles the conversion from complex IF to complex baseband and vice
versa and not the daughterboard. This is because in the motherboard
there exist DUCs/DDCs that handle this conversion (Complex IF to Complex
Baseband/ Complex Baseband to Complex IF). The input/output to the FPGA
is complex IF/complex Baseband, dependent on the direction of flow. Now,
from your comments it seems that the daughterboard converts from complex
baseband to Real RF (and the opposite). Clearly, I am missing something
in here. Are we referring to the low pass filter directly after the
complex mixer?
Please allow me to add one more comment. As detailed in various
datasheets that can be found online prior to transmission/reception from
the daughterboard there is a bandpass filter and then the rest of the
chain follows; which essentially performs a downconversion for further
processing. Is this true? I initially thought that the aforementioned
Bandwidth values correspond to this filter.
Best Regards,
Konstantinos