Message-ID: <002901c27a04$47aa5bc0$cd2a58db@xtr743187>
From: "Vernall" <vernall@xtra.co.nz>
To: rsgb_lf_group@blacksheep.org
References: <3DB0C203.A90691A6@verizon.net> <000801c279ad$b339adc0$f530f7c2@a7j7r2> <3DB55E79.50901@usa.net>
Date: Wed, 23 Oct 2002 08:49:39 +1300
MIME-Version: 1.0
Subject: LF: Type of IF filter
Content-Type: text/plain; charset=iso-8859-1; format=flowed
Content-Transfer-Encoding: 8bit
Sender: <owner-rsgb_lf_group@blacksheep.org>
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org

Hi all,

Alberto commented:

> IMHO, the main reason for the more pleasant sound is the IF
> filter...supposing the '60s receiver
> uses a crystal filter, while the Hallicrafters not.
> Crystal filters are wonderful in the frequency domain, but terrible
> performer in the time domain,
> while LC filter just the opposite...

My opinion as well.  The issue is in "peaks" of group delay for frequencies
just inside the edges of the steep roll-off of a filter.  Crystal filters
have a "squarer" shape factor, hence more peakier group delay (differential
phase shift) characteristic.  Some professional receivers have "equalising
networks" to compensate the group delay over the desired passband.

It is a difficult ask to measure group delay of filters, and to explore why
the subjective result (by ear) of some filters is less pleasing than from
others.  It also has implications for data quality of digital modes, and how
"equalisation" may be possible.  Last weekend I discussed this with a friend
who is familiar with PC sound card software, and Windoze operation, and he
is confident he can develop an audio frequency response and group delay test
set using a full duplex sound card.  The basic line-up to investigate is:
audio sweep is generated on an output, and the received version is input to
the same sound card.  The desire is to display graphs of frequency response
and group delay, one above the other.  This would allow for tweaking of
filter tuning in real time, and saving files of results, or hardcopy
capture.  Such a test set could obviously test audio filters, including DSP
filters.  It could also deal with radio IF filters by using an SSB
transmitter to convert the sweep to RF, and the test receiver in a BFO mode
to downconvert to audio, near the original audio frequency.  Applying a
steady tone at setup may be sufficient for the sound card to find and
correct for a frequency offset, as well as find the delay at that frequency
(as a reference for sweeping other frequencies).  Anyway, that is an outline
of a potentially useful tool using a PC sound card.  If and when software is
developed and proven, I will advise where it could be downloaded from.  The
above description may also be a challenge for another developer to have a
go.

73, Bob ZL2CA