[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Linrad] Re: [perseus_SDR] Re: Auto Notch Filter Demo Movies
- Subject: [Linrad] Re: [perseus_SDR] Re: Auto Notch Filter Demo Movies
- From: Leif Asbrink <sm5bsz.com; leif@xxxxxxxxxxxxxxxx>
- Date: Fri, 6 Mar 2009 21:41:41 +0100
> The file I have uploaded contains a small segment from 17MHz broadcast
> band tonight. At this time of night the band is closing so only a few
> stations are remaining. You can see the PLT interference very well and
> tails off when we get near 18 MHz ham band where there is a conveniant
> notch. Enjoy the file.
Thank you Mark:-)
This file is very interesting. In the time domain (the oscilloscope
view) the interference looks like bursts of white noise that come at
a rate of about 670 Hz and that have a duty of about 20%.
The spectrum shows that the noise is not at all white noise......
There is a lot of structure in it (no surprise)
What can be done with Linrad today is to gate out the 20% of the
time when the interference occurs. That means apply 100% AM modulation
to the desired station. (Other stations are not AM modulated because
Linrad can be set to consider the carriers as strong signals to
route outside the gate.)
When the desired signal is AM modulated, false modulation sidebands
occur at +/- 670 Hz, +/- 1340 Hz etc. Those sidebands can be removed
by applying notch filters (max 9 notches in Linrad) and in the end
a modest improvement of copy is possible.
I have received a file from Mauritz, SM2BYC, with another powerline
QRM. It seems to be a corona discharge from one of the three phases
on a 400 kV line. Presumably there is a sharp tip somewhere on which the
field strength becomes high enough. The corona QRM is similar in nature
to the PLT interference in that it looks like noise bursts in the time
domain. They repeat at 100 Hz and have a duty of 50%. Another difference
is that the envelope is flat on the PLT but it looks like sine squared
or perhaps Gaussian in the corona discharge signal. The corona noise
does not resolve into pulses even at 2MHz sampling rate and I do not think
it would even at much higher bandwidth. The corona discharge is probably
emitting true white noise - although modulated by the discharge power
that is zero behind some threshold.
In all, it will be posible to do better than the current blanker plus
notch filters that is available in Linrad-03.03. I am not yet convinced it
would be worth the effort because in the recordings I have so far, the
possible improvement is limited by the noise floor between the interference
pulses. It might become quite different if someone would make a recording
available where the interference really degrades reception severely.
Particularly the PLT which would require loss of 20% of the signal only
might be possible to remove with a small degradation of reception
(compared to a situation where the QRM would not be present.)
Leif / SM5BSZ
> --- In perseus_SDR@xxxxxxxxxxxxxxxx Leif Asbrink <leif@xxx> wrote:
> > Hi Mark,
> > > Interesting reading Leif. I wonder how the new filter would cope with
> > > PLT or BPL interference. I have PLT interference here in the UK and is
> > > quite bad on upper HF freqs. Though the ham bands are notched quite
> > > well some broadcast bands are wiped out.
> > I have never seen what a limited slice of BPL spectrum would look like.
> > Presumably it is OFDM of some kind. If you can record 30 seconds
> > of it ay the highest possible bandwidth I would be interested to
> > have a look (and to verify for myself that nothing can be done.)
> > On the other hand, with enough bandwidth and knowledge of the protocol
> > it should be possible to decode the transmission and to use the
> > decoded information to reconstruct the RF signal. Then one could
> > subtract the interference:-)
> > The simple solution is to use a two channel radio with two
> > antennas so one can combine the signals and cancel the interference.
> > Can be done with analog circuitry, combining two antennas for
> > a specific signal to be canceled, but the computer would be
> > much faster and more accurate. The risk that the desired signal
> > will also be canceled is small but not zero.
> > 73
> > Leif / SM5BSZ
You received this message because you are subscribed to the Google Groups "Linrad" group.
To post to this group, send email to linrad@xxxxxxxxxxxxxxxx
To unsubscribe from this group, send email to linrad+unsubscribe@xxxxxxxxxxxxxxxx
For more options, visit this group at http://groups.google.com/group/linrad?hl=en