Min. decimation value

GNU Radio
1

Hello everyone,

My understanding is that the decimation value for the fpga has to be
between [4, 256]. In usrp_spectrum_sense.py, when I set the decimation
values to 16 or 8 (spanning 4Mhz and 8Mhz chunk at once), then
m.data[0] …m.data[255] print out fine. However, when I set the
decimation to 4 (i.e. we will be observing a chunk of 16 Mhz), then
all the values data[0] …data[255] are ZERO.

I am trying to understand why is this the case … is it that the USB
is not fast enough for decimation value of 4? I get some "u0"s printed
out even when I run for decimation values of 16 and 8 … is this a
cause of concern?

I appreciate the help.

Thanks,

Shravan.

2

On Sun, Dec 03, 2006 at 04:03:08PM -0600, Shravan Rayanchu wrote:

is not fast enough for decimation value of 4? I get some "u0"s printed
out even when I run for decimation values of 16 and 8 … is this a
cause of concern?

decim = 4 is only valid if you’re using 8-bit I&Q across the USB.
Otherwise it won’t fit. Use decim = 8.

Eric

3

Hi Eric,

I get it now, 64/4 Msps = 16 Msps = 32MBps ==> each sample has to be
2 bytes i.e. 8 bit I&Q.

Can we make usrp_spectrum_sense.py to run with 8 bit I&Q instead of 16
bit I&Q? I mean, If I want to be able to observe a chunk of 16Mhz at
once (if possible, even more), is there something I can do?

Thanks much!

Shravan

4

Shravan Rayanchu wrote:

is not fast enough for decimation value of 4? I get some "u0"s printed
out even when I run for decimation values of 16 and 8 … is this a
cause of concern?

If you use a decimation value of 4, you need to switch to 8-bit
samples. The bus isn’t fast enough for 16-bit samples with a decimation
of 4.