NR Revisited (trying again in plain text mode)

[john lawrence-3]

Over the past two months I've been enjoying weak signal low band DXing on good antennas with my K3 using roofing filters for CW (500Hz) and SSB (2.1 KHz) . My K3 has the latest DSP board and the firmware is up to date with the thirty-two(?) NR choices.  What I'm finding is that the NR Firmware likes to see a wide bandwidth from which the weakest signals can be dug out of the in the noise on SSB or CW. If I narrow up the bandwidth with the optional filters I can't copy the signals as well. i.e. with NR enabled together a 2.1 or 500 Hz filter.  I hear best with NR together with the 2.7 KHz filter. Interesting!
 
This means that QRM must be considered as a secondary matter within the bandpass if the NR feature is to work at its best for me. I find different NR settings depending on band conditions and the ultimate setting varies with incoming signals. I've been working lots of JAs on 80 meters where the NR truly helps. The narrow CW/SSB filters are great but I don't use NR with them. Weak signal reception comes down to using the narrow roofing filters without NR.  OR, just the 2.7 KHz wide filter with NR enabled. Yes, over and over, the 2.7 KHz bandwidth together with NR brings them out from the depths of ESP level to readable signals.  
 
So, why have the optional filters? I use them to get signals so I can copy them through QRM when the band is busy.
.
Happy Holidays

73,

John, W1QS
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Guy, K2AV]

The DSP is behind the roofing filters.  I set my 400 Hz filter to cut
in at 450 Hz width.  If I set width to 500, the K3 uses the 1.8 kHz
roofing filter.  The information diversity of the noise vs CW signal
as a ratio is much larger there and the algorithm seems to take
advantage of that.  To my ears there is a lot more "bite" by the NR
when the roofing is wide.  This wide roofing won't do for contests,
but it's quite nice for listening and regular QSO's.

In your case I would compare NR with width at 550 and 500 for CW in
uncrowded non-contest situations. That should tell you a lot.

In any event, at some level of signal going down past merely weak, NR
no longer "digs into" the noise and IMHO begins to obfuscate the
signal.  At this point only narrowing bandwidth helps. I <never> use
the NR in contests for this reason. NR seems to kill about a couple dB
range where I can dig it out with the dB between the ears if the NR is
off. In a contest, those are the very contacts that if made will put
distance between one and the competition.

And lest anyone think that a dig on the K3, I have exactly the same
observation/experience on the Orions.

73, Guy

On Tue, Dec 29, 2009 at 12:53 PM, John Lawrence <[hidden email]> wrote:

> Over the past two months I've been enjoying weak signal low band DXing on good antennas with my K3 using roofing filters for CW (500Hz) and SSB (2.1 KHz) . My K3 has the latest DSP board and the firmware is up to date with the thirty-two(?) NR choices.  What I'm finding is that the NR Firmware likes to see a wide bandwidth from which the weakest signals can be dug out of the in the noise on SSB or CW. If I narrow up the bandwidth with the optional filters I can't copy the signals as well. i.e. with NR enabled together a 2.1 or 500 Hz filter.  I hear best with NR together with the 2.7 KHz filter. Interesting!
>
> This means that QRM must be considered as a secondary matter within the bandpass if the NR feature is to work at its best for me. I find different NR settings depending on band conditions and the ultimate setting varies with incoming signals. I've been working lots of JAs on 80 meters where the NR truly helps. The narrow CW/SSB filters are great but I don't use NR with them. Weak signal reception comes down to using the narrow roofing filters without NR.  OR, just the 2.7 KHz wide filter with NR enabled. Yes, over and over, the 2.7 KHz bandwidth together with NR brings them out from the depths of ESP level to readable signals.
>
> So, why have the optional filters? I use them to get signals so I can copy them through QRM when the band is busy.
> .
> Happy Holidays
>
> 73,
>
> John, W1QS
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Joe Planisky]

Hi John,

Exactly so.  Lyle Johnson explained how NR works in the K3 a while ago  
and he said exactly what you've observed: NR works best with wider  
bandwidths.  You can see his explanation at

http://www.zerobeat.net/mediawiki/index.php/K3_DSP

73
--
Joe KB8AP


On Dec 29, 2009, at 9:53 AM, John Lawrence wrote:

> What I'm finding is that the NR Firmware likes to see a wide  
> bandwidth from which the weakest signals can be dug out of the in  
> the noise on SSB or CW. If I narrow up the bandwidth with the  
> optional filters I can't copy the signals as well. i.e. with NR  
> enabled together a 2.1 or 500 Hz filter.  I hear best with NR  
> together with the 2.7 KHz filter. Interesting!

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html