Message-ID: <011901c48adc$6554a1e0$49540150@captbrian>
From: "captbrian" <captbrian@ukonline.co.uk>
To: rsgb_lf_group@blacksheep.org
References: <DC9AE469E761D611BD420002A5EC855C77271E@cerberus.herts.ac.uk>
Date: Wed, 25 Aug 2004 20:47:45 +0100
MIME-Version: 1.0
Subject: LF: Re: RE: Key Clicks
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

Do I assume that "Gaussian" is being used loosely here to indicate a gradual
slope increase at first and a gradual slope decrease at the end of the rise
followed by a similar  shape on the decline. ? I cannot think why there
should be a connection between a curve of statistical  random distribution
and a modulation curve for least harmonic production...but then Maths was
never a natural gift...I found it hard work...

Bryan G3GVB
----- Original Message -----
From: James Moritz <J.R.Moritz@herts.ac.uk>
To: <rsgb_lf_group@blacksheep.org>
Sent: Wednesday, August 25, 2004 6:36 PM
Subject: LF: RE: Key Clicks


-----Original Message-----
From: owner-rsgb_lf_group@blacksheep.org
[mailto:owner-rsgb_lf_group@blacksheep.org] On Behalf Of Hugh M0WYE
Sent: 25 August 2004 10:33
To: rsgb_lf_group@blacksheep.org
Subject: LF: Key Clicks

>My question is ...
>Is there a circuit that will produce a nice "gaussian wavform" that I could
>feed into an AM modulator to produce really clean CW - or is there no such
>thing ?

>73
>Hugh M0WYE

Dear Hugh, LF Group,

As far as filtering waveforms with analogue circuits goes, the pedantic
answer is probably "no", but in practice it is possible to produce good
approximations to Gaussian responses by using Bessel or linear phase or
other similar types of low pass filters. To get something that really looks
like a Gaussian response, you need a fairly high-order filter, but for
practical purposes, a second-order filter is quite good enough, for example
the one in the attachment clickfilter_circuit.gif. The simulated response of
this circuit to a 100ms "dot" is shown in the other attachment
click_response.gif (the pale green trace), with a simple RC filter response
for comparison (the upper, blue trace). Comparing the two, the Bessel filter
has a more rounded start to the transition compared to the sharp "corner" in
the RC response, and reaches the final level more quickly, while the RC
filter has a longer "tail" before it settles. A higher order Bessel filter
(the purple trace is 5th order) produces a more symmetrically shaped
response, with the start and finish of the transitions being nearly mirror
images of each other - but in practice for CW keying this would make little
detectable difference compared to the simpler circuit, while requiring a
more complicated circuit with more critical component values.

The 2nd order active filter circuit shown is quite practical - it is
designed to give transitions lasting about 20ms. The time can be increased
or decreased by increasing or decreasing the resistor values in proportion.
If you are using a single supply as shown, a single-supply op-amp is needed,
such as an LM324 or one of the CMOS types - something like a 741 requires an
additional -ve supply rail. There are many other filter configurations that
can be used, which can be found in handbooks. There is no reason why you
should not design an LC filter (the 5th order response was one), except that
the component values are not very practical, unless you have some hefty iron
cored chokes handy. I have used a filter similar to the one in the
attachment for my LF TX - it produces a keying envelope that looks very like
the one shown in the simulation.

Cheers, Jim Moritz
73 de M0BMU