On 08/03/2023 15:43, DİREN ERDEM AYDIN via GNU Radio, the Free &
Open-Source Toolkit for Software Radio wrote:
Dear All,
Changing only FFT size of the freq sink block during simulation drops
signal power drastically, screenshots are given in the attachments. I
am providing the pulse burst from the vector source as 6144 points (6
ones and the rest of the points are zero). Since FFT sink blocks show
the average power of the points on freq domain, increasing FFT size
would increase the number of zeros in the buffer so power is reduced.
What do you think about this approach, is it ok? and there are
fluctuations in the 32k example, that's why it is thicker than 4k
plot, what is the reason for this?
Moreover, no unit is written in vector or signal sources amplitude
sections. Are units assumed as volts?
Regards,
dea
------------------------------------------------------------------------
/Yasal Uyarı: MEF ÜNİVERSİTESİ bu mesajın içeriği ile ilgili hiçbir
yasal sorumluluk taşımaz./
/Legal Disclaimer: MEF UNIVERSITY does not accept any legal liability
or responsibility for the content of this message./
Units are *UTTERLY UNSCALED*, but Gnu Radio normally operates within
{-1.0,+1.0} for floating-point values.
Most hardware drivers honor this system and scale their (complex,
usually) samples appropriately--so those samples will
be *linearly related* to the instantaneous *voltage* as seen at the
antenna, but turning that into real-world values
is up to the user, typically. Gnu Radio itself has no idea what
these samples *mean* in terms of the real world.
Any actual hardware device will perform considerable analog (and then
digital) signal processing on the antenna signals.
So all that you know when you get those samples into Gnu Radio is
that they're (mostly) linearly-related to an antenna voltage.