K3 - A case for wider roofing filters?

[Joe Planisky]

... or There's No Such Thing as a Free Lunch.

Until recently, I only had the 13 kHz and 2.7 kHz roofing filters  
installed in my K3.  I mostly use CW and usually found the DSP  
filtering perfectly adequate for my narrow bandwidth needs.  But after  
having a few QSO's disrupted by nearby strong QRM, I decided to try a  
narrower roofing filter.  So I ordered the 500 Hz 5-pole filter from  
Elecraft.

It arrived, I installed it, and configured it (including the frequency  
offset).  To set the gain, I used the birdie at 49.380 MHz and  
Spectrogram to make sure that the audio signal amplitude with the 500  
Hz filter active was the same as with the 2.7 kHz filter active.  This  
required 6 dB of gain on the 500 Hz filter.  However, I noticed that  
boosting the filter gain by 6 dB also boosted the noise floor by 6  
dB.  In other words, the S/N ratio at 500 Hz bandwidth was 6 dB worse  
than at 550 Hz bandwidth.  After thinking about it for a moment, I  
realized it made perfect sense.  I checked with Gary at Elecraft and  
he said that in general, the narrower roofing filters will have  
greater loss than the wider ones, and nothing he said made me think  
that 6 dB of loss for a 500 Hz filter was unusual.

I tuned around and found some strong signals and did some A/B  
comparisons with the 500 Hz filter on and off.  The filter definitely  
works as advertised for dealing with strong signals on a nearby  
frequency.

I then sought out some weak stations, some of those 'RST 219' guys.  
These stations virtually disappeared into the noise at a DSP BW=0.20  
with the 500 Hz filter enabled, but were somewhat copyable with the  
2.7 kHz roofing filter at the same DSP BW.  In the absence of strong  
QRM, the wider roofing filter gives me better performance because of  
its lower loss.

But does it make any practical difference?  I encounter weak signal  
conditions (usually due to QSB) more often than I encounter s9+30  
signals 500 Hz away.  So given the current situation, I'll probably  
keep the 500 Hz filter turned off most of the time. I wish there was  
an easy way (i.e. one button press) to enable or disable a roofing  
filter so that I could engage it only when really needed.

I've heard that the 8-pole filters tend to have a little less loss  
than the 5-pole ones, but I'm undecided as to whether to try that  
route.   Any voices of experience?

73
--
Joe KB8AP

_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.):
 http://mailman.qth.net/mailman/listinfo/elecraft   

Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com

[W7GJ, Lance]

Joe Planisky wrote:

> ... or There's No Such Thing as a Free Lunch.
>
> Until recently, I only had the 13 kHz and 2.7 kHz roofing filters
> installed in my K3.  I mostly use CW and usually found the DSP filtering
> perfectly adequate for my narrow bandwidth needs.  But after having a
> few QSO's disrupted by nearby strong QRM, I decided to try a narrower
> roofing filter.  So I ordered the 500 Hz 5-pole filter from Elecraft.
>
> It arrived, I installed it, and configured it (including the frequency
> offset).  To set the gain, I used the birdie at 49.380 MHz and
> Spectrogram to make sure that the audio signal amplitude with the 500 Hz
> filter active was the same as with the 2.7 kHz filter active.  This
> required 6 dB of gain on the 500 Hz filter.  However, I noticed that
> boosting the filter gain by 6 dB also boosted the noise floor by 6 dB.  
> In other words, the S/N ratio at 500 Hz bandwidth was 6 dB worse than at
> 550 Hz bandwidth.  After thinking about it for a moment, I realized it
> made perfect sense.  I checked with Gary at Elecraft and he said that in
> general, the narrower roofing filters will have greater loss than the
> wider ones, and nothing he said made me think that 6 dB of loss for a
> 500 Hz filter was unusual.
>
> I tuned around and found some strong signals and did some A/B
> comparisons with the 500 Hz filter on and off.  The filter definitely
> works as advertised for dealing with strong signals on a nearby frequency.
>
> I then sought out some weak stations, some of those 'RST 219' guys.  
> These stations virtually disappeared into the noise at a DSP BW=0.20
> with the 500 Hz filter enabled, but were somewhat copyable with the 2.7
> kHz roofing filter at the same DSP BW.  In the absence of strong QRM,
> the wider roofing filter gives me better performance because of its
> lower loss.
>
> But does it make any practical difference?  I encounter weak signal
> conditions (usually due to QSB) more often than I encounter s9+30
> signals 500 Hz away.  So given the current situation, I'll probably keep
> the 500 Hz filter turned off most of the time. I wish there was an easy
> way (i.e. one button press) to enable or disable a roofing filter so
> that I could engage it only when really needed.
>
> I've heard that the 8-pole filters tend to have a little less loss than
> the 5-pole ones, but I'm undecided as to whether to try that route.  
> Any voices of experience?
>
> 73
> --
> Joe KB8AP
>
> _______________________________________________
> Elecraft mailing list
> Post to: [hidden email]
> You must be a subscriber to post to the list.
> Subscriber Info (Addr. Change, sub, unsub etc.):
> http://mailman.qth.net/mailman/listinfo/elecraft 
> Help: http://mailman.qth.net/subscribers.htm
> Elecraft web page: http://www.elecraft.com
>

Hello Joe,

You raise an interesting viewpoint.  Most HF operators seem concerned with
filtering out strong nearby signals, which is where the narrower roofing filters
really come into their own.  However, if you look at the charts on the Elecraft
web page, you can see the attenuation introduced through the use of these filters.

That is the main reason why I ordered the 6 KHz wide filter with my K3.  I use
that wide filter setting for all my EME contacts, to make sure I get the most out
of the antenna.  And when I am looking for that weakest CW signal on multi-hop Es,
I try to use a wide filter setting and on the radio, and use a narrow audio filter
(such as the one in Spectran) with my headphones.  In fact, that was one of the
great things that came with the K3 - the ability to very finely tune the incoming
CW signal to move it right into very narrow passband of the Spectran audio filter.

GL and VY 73, Lance



--
Lance Collister, W7GJ (ex: WN3GPL, WA3GPL, WA1JXN, WA1JXN/C6A, ZF2OC/ZF8)
P.O. Box 73
Frenchtown, MT  59834  USA
QTH: DN27UB
TEL: [phone]   URL: http://www.bigskyspaces.com/w7gj
2m DXCC #11, 6m DXCC #815


_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.):
 http://mailman.qth.net/mailman/listinfo/elecraft   

Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com

[Barry N1EU]

Joe Planisky wrote

This required 6 dB of gain on the 500 Hz filter.  However, I noticed that  
boosting the filter gain by 6 dB also boosted the noise floor by 6  
dB.  In other words, the S/N ratio at 500 Hz bandwidth was 6 dB worse  
than at 550 Hz bandwidth.  After thinking about it for a moment, I  
realized it made perfect sense.  

Not sure if it makes sense.  My experience is that the signal and noise are both attenuated or boosted together and the S/N remains the same.

73,
Barry N1EU