Re: Scope hangs

I tried Tommi Rouvali’s mod to audio_fft.py with the throttle added as
suggested by Johnathan C. but my CPU usage still went to 100% and
the fft window locked up. This is on a MingW/Windows XP system with a
1.8GHz Pentium M with 512MB RAM.

As a further experiment, I added an audio.sink instead of the throttle
and I got the tone as well as a working FFT display and CPU usage
dropped to 13-15%.

I then tried running fftsink.py since it has throttle statements in the
standalone test app at the end and once again the window locked up with
100% CPU usage after displaying the first of the two fft screens in that
app. After some experimenting, I changed both gr.sig_source statements
to audio.source and got rid of the throttles and everything worked fine
with CPU usage around 35-50%.

It appears the throttles aren’t working the way I think they are
supposed to. Any advice?

From: Johnathan C.
Subject: Re: [Discuss-gnuradio] Scope hangs
Date: Mon, 29 Oct 2007 11:20:38 -0700
User-agent: Thunderbird 2.0.0.6 (X11/20071022)




  self.src = gr.sig_source_f (sample_rate, gr.GR_SIN_WAVE, 650,

And again I run it with -S to get the scope view. But this time the
scopes user interface is hang. Buttons do not do anything, and even
when
trying to close window linux tells me ‘Program not responding’ and I
have to force quit this app. The sin wave is scope runs there just
fine.

Johnathan C.
Corgan Enterprises LLC
http://corganenterprises.com

Thanks,

Rob Walker
Crane Division, Naval Surface Warfare Center (NSWC Crane)
Harnessing the Power of Technology for the Warfighter
300 Highway 361
Crane, IN 47522-5001
Ph: (812) 854-3921 (DSN 482)
Cell: (812) 617-1217
Email: [email protected]

Hello,

You probably made right connections too?

Here is my working version:

http://www.rouvali.com/fileadmin/Download/audio_fft2.py

“Walker, Robert CIV NSWC Crane, WC NAE Staff”
wrote:[email protected]

standalone test app at the end and once again the window locked up with
100% CPU usage after displaying the first of the two fft screens in that
app. After some experimenting, I changed both gr.sig_source statements
to audio.source and got rid of the throttles and everything worked fine
with CPU usage around 35-50%.

It appears the throttles aren’t working the way I think they are
supposed to. Any advice?

What version of GNU Radio are you using? In version 3.0 (and earlier)
gr.throttle did not work on MinGW (see ticket 103 in gnuradio.org/trac).
It
should be ok in version 3.1 and in the latest trunk.

– Don W.

Don - Thanks for the info. I’m using 3.0.3 since I haven’t yet seen
instructions for installing 3.1.1 on MingW on the website.

Rob

Thanks Tommi - By the way, this version of audio_fft as well as the one
posted on the GNU Radio website seem to have some unused “subpanel”
code. It appears that this subpanel might have been developed for
usrp_fft and then wasn’t removed when audio_fft was created from it.

I believe the line “self.show_debug_info = True” and everything from
#self._build_subpanel(vbox)” to “vbox.Add(hbox, 0, wx.EXPAND)” is
unused. Correct?

Rob