Re: FFT on RF Partial Discharge signals

On Mar 18, 7:57 am, Su...@xxxxxxxxxxxxx wrote:
On Mar 18, 9:59 am, Rick Lyons <R.Lyons@xxxxxxxxxxxxxxx> wrote:





On Mon, 17 Mar 2008 15:25:22 -0700 (PDT), Su...@xxxxxxxxxxxxx wrote:
On Mar 17, 9:51 pm, Mark <makol...@xxxxxxxxx> wrote:
On Mar 17, 3:17 pm, Su...@xxxxxxxxxxxxx wrote:

  (snipped by Lyons)

I would think you DON't want to window these signals since they
transient events.  If your FFT starts and ends during the "silent"
portion of your time data, then there should be  no need to window the
data.

Imagine the limit case where the time data is a single impulse.

Mark

It is part of my thesis to window these signals as a "high-resolution
spectrum" can be achieved and where zero-padding will provide a "high-
density spectrum" as well.  I have a jpeg of the signals but I am
unsure how I can put it on this thread for you to see what they look
like.  What I am trying to achieve is to obtain the best possible
frequency spectrums.  Applying windows to the whole signal (including
the "silent" bits) and then applying an FFT will give me the
corresponding spectrum with respect to window applied, be it hamming,
bartlett, etc.  But if I hone into the actual main signal then I
should be able to get an even better spectrum.  the problem I am
having is how can I do this.
The full time-domain signal starts at about -2.5x10^-7 secs and ends
at about 2.5x10^-7 secs.  The burst signal occurs at 0 secs and
eventually dies out at about 2.5x10^-7 secs.  In some signals, there
is still a little signal left at the end but does not die out
completely.  I'm assuming that discontinuities can occur here where
zero padding will be able to stop this remembering that FFTs repeats
itself.
I would say that the "silent" bits aren't exactly silent since there
are other noise signals from operating machinery within this
substation which can cause a bit of distortion but not too much pre 0
seconds (I call this the pre-buffered stage).

Hi Sub-K,
  I agree with Mark.  Perhaps you should not
be applying a window (e.g., Hamming) to your signal.

If I were you, I'd take my non-zero-valued time samples,
zero-pad to the next larger FFT size, and then
perform an FFT.

Sub-K, you said you want to obtain "the best possible
frequency spectrums".  Of course none of us here knows
what that means.  What does the word "best" mean?

Good Luck,
[-Rick-]- Hide quoted text -

- Show quoted text -

Hi Rick,
What I meant by the "best possible frequency spectrum" is that when
you window a signal and apply an FFT, this provides a good spectrum
due to the fact that discontinuties have been sorted out by the
windowing techniques.  I have obtained a frequency spectrum from my
original data but I can do one better and concentrate only on the
actual pulse and disregard any data leading up to the pulse and just
at the end of the pulse too.  Therefore, the data observed can be
repeated via the FFT computations without any discontinuties.  I
understand that zero paddings add no info to the actual data but can
provide a "high-density spectrum" thus showing enough of the spectrum
which, without it, I will be unable to see it in the actual frequency
spectrum.  The windowing is important as this will provide a "high-
resolution spectrum".  With both these issues addressed I will be able
to obtain "best possible frequency spectrum".  The problem I am having
is how can I limit myself to the actual pulse.

Thanks- Hide quoted text -

- Show quoted text -

Another way to "sort out" the discontinuities is to naturaly window
your data by including samples from both before and after the event. I
imagine you are taking your samples and running them through a big
buffer and there is a monitoring circuit that detects the discharge
and then pulls the data out of the buffer. By adjusting the buffer
size and the delay of the discharge detect you should be able to
include the front of discharge in your data.

IHTH,
Clay

.

Relevant Pages

  • Re: Best way to measure precise harmonics?
    ... to a spectrum analyzer, but it just doesn't seem to be precise enough ... ostensibly due to a limited-sample FFT. ... But any other signals, especially ... spectrum analyzer, and have the added benefit of being able to observe ...
    (sci.electronics.design)
  • Re: Best way to measure precise harmonics?
    ... to a spectrum analyzer, but it just doesn't seem to be precise enough ... ostensibly due to a limited-sample FFT. ... But any other signals, especially ... spectrum analyzer, and have the added benefit of being able to observe ...
    (sci.electronics.design)
  • Re: Spectral analysis
    ... convolved with the spectrum of the signal. ... spectrum lines become at least as wide as the main lobe. ... that with a rectangular window we can easier distinguish between frequency ... amplitude signals mutiple bins from each other because they will be burried ...
    (comp.dsp)
  • Re: Cheap Spectral Analyzer
    ... The spectral analyzer FerFT is cheap as from $29.85. ... draw FFT components and amplitude waveforms with this Java-based ... realtime spectrum analyzer for XMMS ... "a software tool designed for the analysis of signals and time series ...
    (sci.electronics.design)
  • Re: FFT on RF Partial Discharge signals
    ... density spectrum" as well. ...  I have a jpeg of the signals but I am ... the "silent" bits) and then applying an FFT will give me the ... corresponding spectrum with respect to window applied, be it hamming, ...
    (comp.dsp)