here is how to "get a feel" for a roofing filter

[DOUGLAS ZWIEBEL]

Hi all:

I knew this would be a tough one for many (most).  Try this on for size....

Imagine that you walk into a very large office space (the band),
filled with desks and a hundred people working at those desks.  You
sit down at a desk in the middle of the room.

At first, nobody is talking (everyone is tuning the band).  You sit
down at a desk in the middle of the room and you strike up a
conversation with the person next to you (one qso on the band).  You
can easily hear them.

Now assume that five other "deskmates" (hams) within the room (band)
also start talking and that they are randomly distributed about the
room (the band).  Your conversation will likely continue without
issue.  But maybe, ONE of the other conversations if taking place only
5 desks (kc) away form yours.  It's a bit annoying.  What to do?
Well, you would ask your conversation-mate to talk louder (run more
power), move to another desk (qsy), or you could put up some walls
around the four desks adjacent to you (make a big cubicle).  This
helps to shut out some of the adjacent conversation (you narrow a
conventional IF filter).

If another conversation strikes up only 3 desks away from you, you
might compensate by making the cubicle even smaller - say around just
your two desks (narrow the conventional IF filter even more).

Now assume that everyone in the room is communicating verbally with
another person.  A few people are whispering to each other (qrpers),
some are at normal conversation levels, many are shouting (KW), and
some are screaming at the top of their lungs (you know who they are).
You have already build the smallest cubicle possible, but the other 99
conversations are still audible.  You can not actually understand the
other conversations, but the "noise" (IMD) is clearly there and making
it hard or impossible for you to hear the person you are talking with.
 What to do?

One solution is to make a smaller room with the big room (add a
roofing filter).  You might build a sound-insulated, concrete room
that is one-tenth of the original size, but housed within the big
room.  This must mean that there are fewer other conversations within
ear-shot because there are fewer desks in the "new" sub-room.

So...lets suppose that you now have your smaller room built within the
larger room.  You now discover that there are NO other conversations
that are so loud that you can hear them.  In fact, it is so quiet now,
that you can even take down your cubicle walls (open up - widen - the
conventional IF filter) and still hear your own conversation without
any other noise (IMD).

[By the way, noise here = IMD squeaks, pops, bleeps, boops, etc., not
static crashes or ignition noise, etc]

Let's go back to the beginning.  If you are having your solitary
conversation in the original GIANT room and then build a walled-off
section (sub-room), you WON'T notice any difference because it was
quiet to start with and making your private room (smaller part of the
big room) had no impact.  It is still just as quiet after the new
walls are put up.

In other words, you can NOT hear the difference that a roofing filter
makes unless there are lots of other LOUD signals within the first
(typically broad) IF.  I can guarantee you that Wayne and Eric will
get "complaints" that the roofing doesn't work because folks just
don't understand it's purpose.  They will say (and I'll take bets on
this), "I don't hear any difference whether I use the wide or narrow
roofing filter."  And they'll be correct.

Hope this analogy makes some sense for you.

de Doug KR2Q
_______________________________________________
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

[Bill VanAlstyne W5WVO]

REALLY nicely done, Doug. An excellent analogy. And I would wager heavily on
the correctness of your prediction.  ;-)

Bill / W5WVO


DOUGLAS ZWIEBEL wrote:

> Hi all:
>
> I knew this would be a tough one for many (most).  Try this on for
> size....
> Imagine that you walk into a very large office space (the band),
> filled with desks and a hundred people working at those desks.  You
> sit down at a desk in the middle of the room.
>
> At first, nobody is talking (everyone is tuning the band).  You sit
> down at a desk in the middle of the room and you strike up a
> conversation with the person next to you (one qso on the band).  You
> can easily hear them.
>
> Now assume that five other "deskmates" (hams) within the room (band)
> also start talking and that they are randomly distributed about the
> room (the band).  Your conversation will likely continue without
> issue.  But maybe, ONE of the other conversations if taking place only
> 5 desks (kc) away form yours.  It's a bit annoying.  What to do?
> Well, you would ask your conversation-mate to talk louder (run more
> power), move to another desk (qsy), or you could put up some walls
> around the four desks adjacent to you (make a big cubicle).  This
> helps to shut out some of the adjacent conversation (you narrow a
> conventional IF filter).
>
> If another conversation strikes up only 3 desks away from you, you
> might compensate by making the cubicle even smaller - say around just
> your two desks (narrow the conventional IF filter even more).
>
> Now assume that everyone in the room is communicating verbally with
> another person.  A few people are whispering to each other (qrpers),
> some are at normal conversation levels, many are shouting (KW), and
> some are screaming at the top of their lungs (you know who they are).
> You have already build the smallest cubicle possible, but the other 99
> conversations are still audible.  You can not actually understand the
> other conversations, but the "noise" (IMD) is clearly there and making
> it hard or impossible for you to hear the person you are talking with.
> What to do?
>
> One solution is to make a smaller room with the big room (add a
> roofing filter).  You might build a sound-insulated, concrete room
> that is one-tenth of the original size, but housed within the big
> room.  This must mean that there are fewer other conversations within
> ear-shot because there are fewer desks in the "new" sub-room.
>
> So...lets suppose that you now have your smaller room built within the
> larger room.  You now discover that there are NO other conversations
> that are so loud that you can hear them.  In fact, it is so quiet now,
> that you can even take down your cubicle walls (open up - widen - the
> conventional IF filter) and still hear your own conversation without
> any other noise (IMD).
>
> [By the way, noise here = IMD squeaks, pops, bleeps, boops, etc., not
> static crashes or ignition noise, etc]
>
> Let's go back to the beginning.  If you are having your solitary
> conversation in the original GIANT room and then build a walled-off
> section (sub-room), you WON'T notice any difference because it was
> quiet to start with and making your private room (smaller part of the
> big room) had no impact.  It is still just as quiet after the new
> walls are put up.
>
> In other words, you can NOT hear the difference that a roofing filter
> makes unless there are lots of other LOUD signals within the first
> (typically broad) IF.  I can guarantee you that Wayne and Eric will
> get "complaints" that the roofing doesn't work because folks just
> don't understand it's purpose.  They will say (and I'll take bets on
> this), "I don't hear any difference whether I use the wide or narrow
> roofing filter."  And they'll be correct.
>
> Hope this analogy makes some sense for you.
>
> de Doug KR2Q
> _______________________________________________
> 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 

_______________________________________________
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

[Eric Swartz - WA6HHQ]

The confusion here may be primarily one of semantics. The K3's main
narrow IF filters -are- the first filters seen after the first mixer,
just like those in the K2. Our first IF at 8.215 MHz acts as our primary
narrow IF filter stage and can also be considered our roofing filter
stage, since it is followed by the DSP IF and its additional filtering.
We had so many questions about what our 'roofing' filters were on the K2
and K3,  that we decided to refer to the 8.215 MHz narrow filters as
roofing filters. Since they are the first IF filter this is also
technically correct.  :-)

In our case, since the roofing filters ARE the main, narrow, IF filters
(just like the 4.915 1st IF filter is in the K2),  dynamic range is
measured for signals OUTSIDE the filter b/w, not inside. The K3's 1st IF
(8.215 MHz) can easily use a crystal filter down to 200 Hz. Higher
frequency first IF stages in other radios are limited to about 3 kHz
minimum and typically use filters in the 5 kHz to 10 kHz width range.
They only have one filter at their 1st IF (their roofing filter) so it
has to be at least wide enough to handle SSB. Plus it is very difficult
to make a filter narrower than 3 kHz economically at an IF of 40-70 MHz.

On traditional up-converting IF radios with wide (>=3 kHz) roofing
filters at their first IF (40 MHz to 70 MHz) the interfering signal can
easily be inside their first roofing filter but outside their next stage
narrow (500 Hz etc) IF filters. This causes havoc with the stages
between their roofing filter and their narrower IF filters.  (The ARRL
test sweeps to 1 kHz away with a 500 Hz main IF filter.)

73, Eric   WA6HHQ
Elecraft
_______________________________________________
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

[Brett gazdzinski-2]

Is there no way to put dsp at/as the first filter?
What are the limitations of the dsp, frequency, filter quality?
Is there a problem getting dsp to work at say 455 khz?

What would be the result of a real good variable filter
at the 8.215 Mhz IF just after the mixer and NO dsp
for filtering? (K2?)


The 2nd IF of 15 khz is audio, is it not?
At what point does the audio spectrum end and
RF begin?
I notice that the K3 info says the dsp is in the 2nd IF,
but the 2nd IF is high pitched audio?

This modern stuff sure spins my head, the homebrew receiver
is a single conversion to 455khz, no rf amp, good filter
just after the mixer, wide IF chain, detector.
It seems to work real well, and is basically the old tube table top
radio (all American 5?) with a good filter, s meter, and a digital
frequency readout.

Brett
N2DTS

> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of Eric
> Swartz - WA6HHQ, Elecraft
> Sent: Wednesday, May 02, 2007 11:12 AM
> To: [hidden email]
> Subject: Re: [Elecraft] here is how to "get a feel" for a
> roofing filter
>
> The confusion here may be primarily one of semantics. The K3's main
> narrow IF filters -are- the first filters seen after the first mixer,
> just like those in the K2. Our first IF at 8.215 MHz acts as
> our primary
> narrow IF filter stage and can also be considered our roofing filter
> stage, since it is followed by the DSP IF and its additional
> filtering.
> We had so many questions about what our 'roofing' filters
> were on the K2
> and K3,  that we decided to refer to the 8.215 MHz narrow filters as
> roofing filters. Since they are the first IF filter this is also
> technically correct.  :-)
>
> In our case, since the roofing filters ARE the main, narrow,
> IF filters
> (just like the 4.915 1st IF filter is in the K2),  dynamic range is
> measured for signals OUTSIDE the filter b/w, not inside. The
> K3's 1st IF
> (8.215 MHz) can easily use a crystal filter down to 200 Hz. Higher
> frequency first IF stages in other radios are limited to about 3 kHz
> minimum and typically use filters in the 5 kHz to 10 kHz width range.
> They only have one filter at their 1st IF (their roofing
> filter) so it
> has to be at least wide enough to handle SSB. Plus it is very
> difficult
> to make a filter narrower than 3 kHz economically at an IF of
> 40-70 MHz.
>
> On traditional up-converting IF radios with wide (>=3 kHz) roofing
> filters at their first IF (40 MHz to 70 MHz) the interfering
> signal can
> easily be inside their first roofing filter but outside their
> next stage
> narrow (500 Hz etc) IF filters. This causes havoc with the stages
> between their roofing filter and their narrower IF filters.  
> (The ARRL
> test sweeps to 1 kHz away with a 500 Hz main IF filter.)
>
> 73, Eric   WA6HHQ
> Elecraft
> _______________________________________________
> 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
>

_______________________________________________
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

[KK7P]

Hello Brett!

> Is there no way to put dsp at/as the first filter?
> What are the limitations of the dsp, frequency, filter quality?
> Is there a problem getting dsp to work at say 455 khz?

To answer these questions would take a textbook!  Or at least a very
long essay.

The short version is: there are many ways to design a radio, and to
partition the analog and digital circuitry.  We chose a design
architecture that provides world-class performance at a realistic cost.

> What would be the result of a real good variable filter
> at the 8.215 Mhz IF just after the mixer and NO dsp
> for filtering? (K2?)

It would be a K2 without the KDSP2 :-)

> The 2nd IF of 15 khz is audio, is it not?

The use of a final IF in the 12 kHz to 40 kHz range is used by virtually
every manufacturer of DSP-equipped radios.  There are solid technical
and economic reasons for this.

73,

Lyle KK7P

_______________________________________________
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