ATSC demod tips

GNU Radio
1

I just got a USRP, and I’m learning how it works.

I’m having trouble getting the ATSC demod to work.

The WFM demod works OK, so I know the antenna, etc works.

I know I’m tuning the right frequency (KQED with a center
frequency of 569MHz), because I can see the signal with usrp_fft.

I got a few underruns during the “usrp_rx_cfile”. About 10 in 10
seconds. The raw data file is 446 meg.

I get an empty output file and a file named “taps.txt”, which is
also empty.

I tried both the Gnuradio v3.1.2 (in Fedora 9) and the Gnuradio
from SVN.

I get a few output strings, like “>>> gr_fir_fff: using SSE” from
btl-fsd and “>>> gr_fir_fff: using SSE
Setting initial_freq: 3065000.000000” from fpll

Do you have any tips on finding the problem?

Thanks.

2

I got a few underruns during the “usrp_rx_cfile”.
About 10 in 10 seconds. The raw data file is 446 meg.

Underruns are bad.

Try capturing the data to a ram disk instead of your hard drive if you
do not have a faster hard drive setup available (i.e. RAID). I used the
ATSC demodulator a few months back and was never able to successfully
store that data down to my hard drive, so I allocated a 512 MB ramdisk
and used that which worked great.

Thanks,
Brandon

3

Wuest Brandon-WTVR47 wrote:

store that data down to my hard drive, so I allocated a 512 MB ramdisk
and used that which worked great.

I am having the same issue. When using a ramdisk I get no usrp overruns
but
still the output file is empty. When putting the output of each py file
in
its own file, I found btl-fsd.py outputs nothing. Are there any obvious
reasons why this would be happening? Is there anything I can do to
narrow
down the problem?

View this message in context:
http://www.nabble.com/ATSC-demod-tips-tp19446876p19611184.html
Sent from the GnuRadio mailing list archive at Nabble.com.

4

On Mon, Sep 22, 2008 at 09:05:56AM -0700, Ben Green wrote:

Try capturing the data to a ram disk instead of your hard drive if you
do not have a faster hard drive setup available (i.e. RAID). I used the
ATSC demodulator a few months back and was never able to successfully
store that data down to my hard drive, so I allocated a 512 MB ramdisk
and used that which worked great.

I am having the same issue. When using a ramdisk I get no usrp overruns but
still the output file is empty. When putting the output of each py file in
its own file, I found btl-fsd.py outputs nothing. Are there any obvious
reasons why this would be happening? Is there anything I can do to narrow
down the problem?

If you are trying to capture continuous data, there is no substitute
for a disk subsystem that can sustain the required throughput. In
general we’ve had better luck using ext2 file systems rather than ext3
file systems. (You can mount an ext3 as an ext2.) Posting the ext3
journal was causing a problem last time I looked at this in detail
(> 2 years ago).

Eric

5

On Mon, Sep 22, 2008 at 12:23:20PM -0700, Ben Green wrote:

For capturing short segments (<500 MB raw data) is a ram disk not
sufficient?

If that’s enough for your purposes, then sure, it’ll work.

Eric

6

Eric B. wrote:

If you are trying to capture continuous data, there is no substitute
for a disk subsystem that can sustain the required throughput.

For capturing short segments (<500 MB raw data) is a ram disk not
sufficient?

–
View this message in context:
http://www.nabble.com/ATSC-demod-tips-tp19446876p19614781.html
Sent from the GnuRadio mailing list archive at Nabble.com.