i still read in the gnuradio documentation, that we can prevent
freezing graphical elements by inserting the throttle element before,
but how can i locate which element is freezed ?
Is there any hlp in or python modules which can be used for
tuning/debugging ?
I inserted for every element a throttle, but this doesn’t help me out.
Can somebody of the Gnuradio Team add to the documentation, that
the peak detector may freeze the application too.
The solution is to rise the <look_ahead> value, so that the
peak detector may not detect to much peaks.
Is there any hlp in or python modules which can be used for
respond, or freezing as in the waveform has stopped updating on the screen?
-Josh
Hi and thanks for your response,
i am using a usrp-source and a fft-sink and peakdetect+scope-sink.
only the screen with the wave form freezes, the buttons of the scope
and the fft sink are not freezing. This seems to appear mostly when
there are many peaks i think, but i can not attest this.
The only think i change before it freezes is the Signal frequency
of my signal generator, which sends data to the LF-RX db.
The only think i change before it freezes is the Signal frequency
of my signal generator, which sends data to the LF-RX db.
Regards Markus
That says that throttling is not your issue. You are probably correct to
think that the peak detect block stopped outputting samples. -Josh
This affects my scope-sink and my fft-sink. Not only the scope.
But the buttons of both elements are ok.
I also thought it could be refer to samples that the peak-detector
spits out.
Any idea how to avoid this ? I am in need of the peak-detector.
Have you removed ALL gr.throttle blocks from your flowgraph?
I tried some combinations without any success, so i think it’s a bug
and i have to wait for until it is solved. The throttle element
doesn’t change anything for me.
The signal is at 1020 kHz. Some frequencies results in a freezing
of my graphical elements.
Do you think that it might be possible for peak detector to output
always the same number at “some frequencies”?
Replace the peak detector with a source of constant 0 or a source of
constant 1, do the plots appear frozen?
It seems that if you give the graphical sinks a signal that never
changes, the plotted waveforms will also not change.
Further, if your scope is triggering on channel 1 (the blue) and blue is
constant, the scope cannot find a trigger point, and stop plotting new
samples until a trigger point is found.
Do you think that it might be possible for peak detector to output
always the same number at “some frequencies”?
I think not. The Peakdetector(with TH factor Rise and Fall) works and
the
Peakdetector2 works not. Maybe there are some bugs in the peakdetector2
(with only TH factor Rise).
Replace the peak detector with a source of constant 0 or a source of
constant 1, do the plots appear frozen?
Yes i did this and also other combinations, but only when i play with
Peakdetecor2, my application freezes.
It seems that if you give the graphical sinks a signal that never
changes, the plotted waveforms will also not change.
My application is really freezed. I changed the signal frequency, so i
can watch how the signal in the scope changes, but when i change to
special frequencies, then i get my application freezed.
Further, if your scope is triggering on channel 1 (the blue) and blue is
constant, the scope cannot find a trigger point, and stop plotting new
samples until a trigger point is found.
I tried this also, but this doesnt solve this problem.
As i said this problem doe not appear in Pekdetector but in
Peakdetector2.
Further on the Peakdetecor finds peaks whether i changed the TH factor
Rise
to 10e15. It would be nice if you give me some instructions
for the alpha value and look ahead. Because i think i missunderstood the
documentation of the peakdetector.
Will there be always an average value calculated or only when the
threshold
founds some possible peak ?
Can you please explpain in your own words the alpha value ?
I really believe that the gui is not freezing. What you see is a result
of the fact the the peak detect outputs a binary 1 or 0 signal.
Connect the peak detect to a char_to_float block, and then to a
numbersink. Setup the numbersink for a min of 0 and a max of 1. The
numbersink will clearly show what state the peak detector is in. Make
sure to include a throttle somewhere.
Take a look at the source code for peak detect, perhaps the comments, or
the implementation can help you to figure out what the parameters like
threshold or alpha mean. Also, add a few printfs to the work method, to
verify what the block is outputting, and if its really freezing/stopped
working.
I really believe that the gui is not freezing. What you see is a result
of the fact the the peak detect outputs a binary 1 or 0 signal.
Connect the peak detect to a char_to_float block, and then to a
numbersink. Setup the numbersink for a min of 0 and a max of 1. The
numbersink will clearly show what state the peak detector is in. Make
sure to include a throttle somewhere.
Still done. It doesnt matter which graphical sinnk i use.
Only the <gr_peak_detector_xb_0> works with all one.
I have tried about 20 combinations, with different sinks, like
number sink, file sink, scope sink, with throttle, without throttle,
different connections.
The <gr_peak_detector2_fb_0> doesnt work correctly.
Take a look at the source code for peak detect, perhaps the comments, or
the implementation can help you to figure out what the parameters like
threshold or alpha mean. Also, add a few printfs to the work method, to
verify what the block is outputting, and if its really freezing/stopped
working.
Ok i will do this.
Thanks for your answer and help. I still read the comments in the Header
and sources, but if i understand this correctly than it should not
produce peaks by setting TH factor Rise to 10e15 for example. So i think
there should be a second Bug or i misunderstood something.
Do you think a peak detector should find peaks if it gets a noise
signal and it is setted up with TH factor Rise by 10e15 ?
I add to files:
peakdetector_LF_RX_08.grc --> works not (gr_peak_detector2_fb_0)
peakdetector_LF_RX_13.grc --> works (gr_peak_detector_xb_0)
peak detector can output 1 or 0. What is it outputting?
I get 1 and 0 when it is not freezed.
I uploaded a screenshot, but i am not sure whether this works,
There you can see that i get 1 at a peak otherwise 0. The blue wave
in the scope shows the peaks and the red wave is the signal
i am searching peaks for.
The signal is at 1020 kHz. Some frequencies results in a freezing
of my graphical elements.
Do you mean it is better to make my own signal processing block to print
?
As i understand in, http://www.gnu.org/software/gnuradio/doc/howto-write-a-block.html
the only way to make a signal processing block is by create and edit
with C++ ? Is this correct ?
Is there a way to make a signal prcessing block in Python ?
As i said i still tried the number sink and the file sink, and both
stopped,
when the application freezes. The size of the saved file of the file
sink
doesnt grow up any more. And the number sink doesnt change between 1 and
0
anymore, even if i change the signal frequency. When i use
<gr_peak_detector_xb_0> it still works.
regards markus
This forum is not affiliated to the Ruby language, Ruby on Rails framework, nor any Ruby applications discussed here.