Wise big block interleaving in GNURadio


#1

Hi everybody,
I need to do interleaving on very big blocks of data, I have tried to
demultiplex my stream into several (6048) output streams in order to be
able to act on it and then re-multiplex it to something serial.

unfortunately this seems to be too much, could anybody tell me what is
the wise way to do interleaving on big blocks in gnuradio?
(or point me towards some useful example?)

the error i get is:

gr_vmcircbuf_sysv_shm: shmget (2): No space left on device
gr_buffer::allocate_buffer: failed to allocate buffer of size 32 KB
terminate called after throwing an instance of ‘std::bad_alloc’
what(): St9bad_alloc
./run_tests: line 37: 9751 Aborted $file
FAIL: run_tests

Many Thanks
vincenzo


#2

Hello,

What is the stop band of the FPGA CIC filter? How do I figure out the
attenuation at a given frequency outside my bandwidth?

I’m decimating by 250.

Thanks,
Hans

----- Original Message -----
From: “Vincenzo P.” removed_email_address@domain.invalid
To: “gnuradio mailing list” removed_email_address@domain.invalid
Sent: Wednesday, May 16, 2007 3:13 PM
Subject: [Discuss-gnuradio] wise big block interleaving in GNURadio

Internal Virus Database is out-of-date.
Checked by AVG Free Edition.
Version: 7.5.467 / Virus Database: 269.6.5/792 - Release Date: 5/6/2007
9:01 PM


This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.


#3

On Wed, May 16, 2007 at 04:00:24PM -0700, Hans G. wrote:

Hello,

What is the stop band of the FPGA CIC filter? How do I figure out the
attenuation at a given frequency outside my bandwidth?

I’m decimating by 250.

Thanks,
Hans

It’s a fourth order CIC.
See http://users.snip.net/~donadio/cic.pdf for the transfer function.

Eric


#4

Hans:

I don’t think you really want the CIC transfer function but the transfer
function of the convolution of the half band filters and the CIC since
that is the total transfer function through the FPGA. Am I right?

Bob

Hans G. wrote:


AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL,
TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair
“If you’re going to be crazy, you have to get paid for it or
else you’re going to be locked up.” Hunter S. Thompson


#5

Dear Hans,

I think you meant the frequency response of CIC+ HBF filters. However,
the USRP CIC decimation filter has 4 stages. I did MATLAB analysis for 4
stages CIC filter with decimation rate of 250 (as you requested) without
the HBF. The spectrum analysis result is attached. The first plot is for
entire range [0 to Fs/2]. The second one is zoomed near the interested
band. You will see a huge theoretical gain (more than 190 dB) because of
CIC bit growth.

Firas

Eric B. removed_email_address@domain.invalid wrote: On Wed, May 16, 2007 at 04:00:24PM
-0700, Hans G. wrote:

Hello,

What is the stop band of the FPGA CIC filter? How do I figure out the
attenuation at a given frequency outside my bandwidth?

I’m decimating by 250.

Thanks,
Hans

It’s a fourth order CIC.
See http://users.snip.net/~donadio/cic.pdf for the transfer function.

Eric


#6

Hello,

Thank you for the graphs. I wasn’t aware that there was a HBF after the
CIC. Yes, I’d like to know the frequency response of the whole usrp
from input to output when I decimate by 250.

Thank you to everyone for the help so far,
Hans
----- Original Message -----
From: Firas abbas
To: Hans G.
Cc: gnuradio mailing list ; Eric B.
Sent: Wednesday, May 16, 2007 11:19 PM
Subject: Re: [Discuss-gnuradio] CIC stop band

Dear Hans,

I think you meant the frequency response of CIC+ HBF filters.
However, the USRP CIC decimation filter has 4 stages. I did MATLAB
analysis for 4 stages CIC filter with decimation rate of 250 (as you
requested) without the HBF. The spectrum analysis result is attached.
The first plot is for entire range [0 to Fs/2]. The second one is zoomed
near the interested band. You will see a huge theoretical gain (more
than 190 dB) because of CIC bit growth.

Firas

Eric B. removed_email_address@domain.invalid wrote:
On Wed, May 16, 2007 at 04:00:24PM -0700, Hans G. wrote:
> Hello,
>
> What is the stop band of the FPGA CIC filter? How do I figure out
the
> attenuation at a given frequency outside my bandwidth?
>
> I’m decimating by 250.
>
> Thanks,
> Hans

It's a fourth order CIC.
See http://users.snip.net/~donadio/cic.pdf for the transfer 

function.

Eric


_______________________________________________
Discuss-gnuradio mailing list
removed_email_address@domain.invalid
http://lists.gnu.org/mailman/listinfo/discuss-gnuradio



Internal Virus Database is out-of-date.
Checked by AVG Free Edition.
Version: 7.5.467 / Virus Database: 269.6.5/792 - Release Date:
5/6/2007 9:01 PM


#7

Hello,

I can give you a precise MATLAB based simulated USRP frequency response.
All
what I need is the HBF coefficients. So this is an open invitation to
our
dear Matt to give us his designed USRP HBF coefficients or frequency
response.

Firas,

Hans G. wrote:

From: Firas abbas
stages CIC filter with decimation rate of 250 (as you requested) without
>
See http://users.snip.net/~donadio/cic.pdf for the transfer function.

Internal Virus Database is out-of-date.


Discuss-gnuradio mailing list
removed_email_address@domain.invalid
http://lists.gnu.org/mailman/listinfo/discuss-gnuradio


View this message in context:
http://www.nabble.com/wise-big-block-interleaving-in-GNURadio-tf3768964.html#a10666406
Sent from the GnuRadio mailing list archive at Nabble.com.


#8

On 5/17/07, Eng. Firas removed_email_address@domain.invalid wrote:

Hello,

I can give you a precise MATLAB based simulated USRP frequency response. All
what I need is the HBF coefficients. So this is an open invitation to our
dear Matt to give us his designed USRP HBF coefficients or frequency
response.

Reading the documentation in the Verilog source file is helpful:
http://gnuradio.org/trac/browser/gnuradio/trunk/usrp/fpga/sdr_lib/hb/halfband_decim.v

The coefficient ROM is located here:
http://gnuradio.org/trac/browser/gnuradio/trunk/usrp/fpga/sdr_lib/hb/coeff_rom.v

Firas

Brian


#9

Eng. Firas wrote:

Hello,

I can give you a precise MATLAB based simulated USRP frequency response. All
what I need is the HBF coefficients. So this is an open invitation to our
dear Matt to give us his designed USRP HBF coefficients or frequency
response.

Firas,

The coefficients are here:

http://gnuradio.org/trac/browser/gnuradio/trunk/usrp/fpga/sdr_lib/hb/coeff_rom.v

Matt


#10

Thank you Brian & Matt

Using Matt HBF coefficients given by the verilog file [coeff_rom.v], I
did
the HBF frequency response analysis by MATLAB FDATOOL. Attached is the
USRP
HBF result frequency response. As we can see, there is a 6 dB passband
gain.
In my next step, I will try to provide the frequency response of the
cascaded USRP CIC+HBF sections.

Regards,
Firas

http://www.nabble.com/file/8495/USRP%20HBF.JPG USRP HBF.JPG

Matt wrote:

Hello,

I can give you a precise MATLAB based simulated USRP frequency
response. All
what I need is the HBF coefficients. So this is an open invitation to
our
dear Matt to give us his designed USRP HBF coefficients or frequency
response.

Firas,

The coefficients are here:

http://gnuradio.org/trac/browser/gnuradio/trunk/usrp/fpga/sdr_lib/hb/coeff_rom.v

Matt

Brian P. wrote:

Brian


Discuss-gnuradio mailing list
removed_email_address@domain.invalid
http://lists.gnu.org/mailman/listinfo/discuss-gnuradio


View this message in context:
http://www.nabble.com/wise-big-block-interleaving-in-GNURadio-tf3768964.html#a10667330
Sent from the GnuRadio mailing list archive at Nabble.com.


#11

I’m using 4rx_0tx.rbf.

Thanks for your help,
Hans

----- Original Message -----
From: “Eric B.” removed_email_address@domain.invalid
To: “Hans G.” removed_email_address@domain.invalid
Cc: “gnuradio mailing list” removed_email_address@domain.invalid
Sent: Thursday, May 17, 2007 8:04 PM
Subject: Re: [Discuss-gnuradio] CIC stop band

Hans,
Checked by AVG Free Edition.
Version: 7.5.467 / Virus Database: 269.7.1/807 - Release Date: 5/16/2007
6:05 PM


This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.


#12

On Thu, May 17, 2007 at 08:06:22AM -0700, Hans G. wrote:

Hello,

Thank you for the graphs. I wasn’t aware that there was a HBF after the CIC. Yes, I’d like to know the frequency response of the whole usrp from input to output when I decimate by 250.

Thank you to everyone for the help so far,
Hans

Hans,

Are you using the 2rxhb_2tx.rbf (the default) or the 4rx_0tx.rbf
fpga configuration? If it’s the 4rx version, there is only the CIC,
no halfband. The default 2rxhb_2tx has the halfband after the CIC.

Eric