This is very interesting, because where I live there are sometimes strong noise sources perhaps from OTH radar or other applications which take out big chunks of a band. The technique described is suboptimal because you sacrifice the roofing filter function in order to narrow the blanker's passband. Is there some way to do this without interfering with the filters? My guess is that it would require a hardware modification. But maybe it could be a plugin replacement for the noise blanker board? Noise is a huge problem for many of us and it will only get worse in the future. Any improvement to anti-noise features would be a big advantage over the competition.
Vic 4X6GP/K2VCO > On 4 Feb 2016, at 5:03 AM, Wayne Burdick <n...@elecraft.com> wrote: > > Hi all, > > If you have... > > - really ugly noise sources that neither of the K3/K3S noise blankers > completely clean up, and > > - a 6-kHz crystal filter, and > > - a narrow crystal filter (200-1000 Hz) > > ...then you may want to try an experimental technique I've been using the > past couple of days. In many cases it produces dramatically improved > blanking, at least in narrow-band modes (CW, PSK, FSK). I've been able to > hear many weak signals that I simply couldn't hear before. > > It may also work for SSB signals in conjunction with a 15-kHz crystal filter, > but I haven't tried that yet. > > The kind of noise I'm talking about is often quite unstable, with a buzzy > sound, possibly drifting around a bit in frequency and amplitude. Light > dimmers, switching power supplies, and various other devices create such > noise. The noise may be narrowband: as you tune the VFO, you may find there's > a "hump" of noise that's anywhere from 2 kHz to 50 kHz wide. It may also have > very complex waveform with multiple noise pulses back-to-back in a burst. > > These types of noise are difficult to deal with. The IF blanker's signal path > may be too wide (0.2 to 2 MHz), resulting in too little energy in-band to > trigger the gating signal. The DSP blanker's RF signal path may be too > narrow, making it hard for the DSP to distinguish noise from desired signal. > > * * * > > Setup: > > 1. Connect the radio to a computer running K3 Utility. Go into the > Configuration / Configure Crystal Filter setup screen. > > 2. Find your 6-kHz filter (probably FL1 or FL2). Now the fun part: fake out > the firmware by entering a bandwidth for this filter that's just 50 Hz wider > than your narrow CW filter (ideally 250-500 Hz). *Do not* change the filter > offset. But *do* make sure that the 6-kHz filter's CW and DATA enable boxes > are checked. > > 3. Click "OK" to save this experimental crystal filter configuration setup. > > 4. You will now find that when the WIDTH control is rotated from, say, 0.40 > to 0.45, the XFIL selection will jump from something like FL4 directly to FL1 > or FL2 (your 6-kHz filter). That, hopefully, is the boundary where magic may > occur, below. > > * * * > > The Experiment: > > 1. Find one of your most offensive local noise sources. I have them on most > low bands. The stronger the amplitude the better. Narrowband sources may > provide the most dramatic results. > > 2. Back down the AF gain control, then *turn off AGC*. You may need to use > the RF gain to keep the signal from clipping. > > NOTE: The reason for doing this test without AGC is to make sure you can hear > the full effect of applied noise reduction. AGC flattens out the receiver's > audio response, making it hard to compare different settings. (If you find > that the noise-remediation trick works, you can later turn AGC back on, and > while the effect won't be as obvious, any benefit in signal-to-noise ratio > will still apply.) > > 3. Select CW mode and adjust the WIDTH control for your narrow filter's > bandwidth (example: "BW 0.40"). > > 4. Turn on the noise blanker (tap NB) and hold NB (LEVEL) to access the > blanker parameters. > > 5. Set the IF blanker to OFF (VFO B). Then experiment with the DSP blanker > settings (VFO A) to obtain the best possible reduction in signal. > > 6. While still the LEVEL parameters are still displayed, adjust the WIDTH > control to the next step up (example: "BW 0.45"). This should kick in the > 6-kHz filter, *but the DSP bandwidth and filter graphic will still show a > narrow passband*. In other words, you're widening out the crystal filter but > making very little change in the DSP's internal filter bandwidth (15 kHz IF, > and AF). > > 7. Now re-optimize the DSP noise blanker settings for the 6-kHz filter case. > Did the noise drop? (If you have a signal generator, e.g. an Elecraft XG3, > you might put an antenna on it and generate a weak signal right in the middle > of the noise to get more definitive results.) > > 8. Try it on other noise sources. It may help on some but not others, due to > the wide variance in noise signals. > > Please log your results and report them to the list, at least until Eric > shuts down the thread :) > > * * * > > IMPORTANT: > > As you can imagine, opening up the crystal filter bandwidth much wider than > the DSP bandwidth will make the receiver more susceptible to in-band > interference. If necessary, use RF GAIN, preeamp, and attenuator settings to > reduce all interfering signals to a manageable level. > > I find there are many occasions on which better blanking is really critical, > even if gain must be reduced in order to take advantage of it. > > * * * > > If we get enough positive responses from this experiment, we'll provide a > simply, intuitive way of selecting the 6-kHz filter for noise blanking > purposes. And maybe the 15 kHz filter for SSB use, if applicable. For > example, we might add more selections to the DSP blanker parameter (presently > t1-1 to 3-7). Suggestions welcome. > > 73, > Wayne > N6KR > > > > > > > > ______________________________________________________________ > Elecraft mailing list > Home: http://mailman.qth.net/mailman/listinfo/elecraft > Help: http://mailman.qth.net/mmfaq.htm > Post: mailto:Elecraft@mailman.qth.net > > This list hosted by: http://www.qsl.net > Please help support this email list: http://www.qsl.net/donate.html > Message delivered to k2vco....@gmail.com ______________________________________________________________ Elecraft mailing list Home: http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/mmfaq.htm Post: mailto:Elecraft@mailman.qth.net This list hosted by: http://www.qsl.net Please help support this email list: http://www.qsl.net/donate.html Message delivered to arch...@mail-archive.com
