Re: Vector Block Sample Rate

[Marcin Puchlik]

Hi Daren and Fabian,

Fabian: Daren is using cyclic transmission so starting from the moment when the Pluto hardware buffer is full it repeats the samples over and over again. I think that CPU load has nothing to do here since as I said, samples are not being calculated in a real time manner.

Marcin

niedz., 5 cze 2022 o 11:14 Fabian Schwartau <fabian@opencode.eu> napisał(a):

Hi Diren,

I am not exactly sure what your problem is.
What you should know: gnuradio does not inherently has any sample rate
(unlike Simulink for example). Each block receives samples, processes
them, and outputs the same number of samples as fast as it can (except
if the block is interpolating/decimating - which is not the case for you
LPF). Between the blocks are buffers, which buffer the samples until the
next block will process it. The sample rate you set in most blocks is
just to calculate stuff (like the filter response) based on other
frequency settings (like your cut-off frequency). One exception here is
for example your PlutoSDR Source/Sink - both will actually
provide/require the sample rate you specify.
So, as your LPF has an interpolation of 1, it does not interpolate and
will output the same number of samples it receives. Your vector source
will provide samples as fast as possible until some buffer is full and
it will then keep it full.
One problem might be that your CPU cannot keep up filtering the
61.44MSPS, which is quite a bit. Try increasing the transition width of
the filter, which will reduce the filter order and thus the CPU load.

Hope that helps,
Fabian

Am 04.06.22 um 23:09 schrieb DİREN ERDEM AYDIN via GNU Radio, the Free &
Open-Source Toolkit for Software Radio:
> Hello All,
>
> I'm working with Adalm Pluto SDR, trying to transmit and receive a
> custom burst signal given in the figure. To create the signal I'm using
> a vector block that creates a pulse and then due to a low pass filter I
> can extract the desired waveform. Flowgraph is also added.
>
> As far as I know, input/output sample rates of connected blocks should
> be matched. However, I'm providing 10k fewer samples to LPF because of
> the high sample rate of pluto, 61.44 M. When I tried to increase the
> samples in the vector source to Mega ranges frequency response of the
> signal drops drastically. I have also tried to use interpolation in
> between vector block and LPF, but it doesn't work either.
>
> My question is the LPF block normalizes the input sample rate to 61.44
> M? The whole system works OK but this thing stays unclear for me. Thank you.
>
> Regards,
> Diren