Google Fourier Gibbs phenomenon.
73
Bill wa4lav
________________________________________
From: amps-bounces@contesting.com [amps-bounces@contesting.com] On Behalf Of
Gary Schafer [garyschafer@comcast.net]
Sent: Monday, January 02, 2012 10:38 PM
To: 'Dan Mills'
Cc: amps@contesting.com
Subject: Re: [Amps] IMD
> -----Original Message-----
> From: Dan Mills [mailto:dmills@exponent.myzen.co.uk]
> Sent: Monday, January 02, 2012 8:22 PM
> To: garyschafer@comcast.net
> Cc: amps@contesting.com
> Subject: RE: [Amps] IMD
>
> On Mon, 2012-01-02 at 19:57 -0500, Gary Schafer wrote:
> > Hi Dan,
> >
> > I don't know about a DSP radio but I do know that if you modulate a
> regular
> > old fashion SSB rig with a square wave you will not get anything out
> of it
> > at RF that resembles a square wave. Looking at it in the time domain.
>
> Bear in mind that a communications SSB rig is only a decade wide (300Hz
> to 3K give or take), and a square wave is odd harmonics, so at most you
> will fit a 300Hz fundamental, plus 4 harmonics, which is not enough to
> give anything that remotely looks like a square wave (Especially as the
> radio is unlikely have particularly constant group delay)....
>
> Inputting at less then 300Hz does not help because you loose the
> fundamental, and increasing the frequency just means that you get ever
> fewer harmonics, above about 1Khz you just get a sine wave as all the
> harmonics are cut away.
>
> an input at much less then 300Hz will indeed give a series of spikes,
> these being the components that fit into the radios passband.
>
> This is hard to demonstrate as most radios use one of the IF filters to
> implement the bandwidth control, so you cannot easily see that the
> envelope really does follow the input to the modulator as long as the
> input results in RF that lies within the passband of the downstream
> filters.
>
> This is why you must band limit the clipper output, and ensure that the
> band limiting does not reintroduce overshoot, tedious but it can be
> done.
>
> 73, Dan.
>
It is not that simple. The leading and trailing edge portions of the square
wave stack up and create an infinite amplitude at each. If the bandwidth was
not limited there would be infinite amplitude at each with no signal in
between.
If you modulate the same transmitter in the AM (double side band) mode it
will reproduce a square wave fairly well.
This is easily seen with a phasing type of transmitter where there is no RF
filter to limit bandwidth. Transmitting both side bands will give a fair
looking square wave provided the audio frequency response is not severely
limited. Switching to SSB the square wave goes away and you get very high
amplitude spikes where the leading and trailing edge of the square wave is
located and little output in the middle.
This is why when you first clip the audio where the sin waves more resemble
a square wave then feed it to the SSB transmitter you will get high spikes
and low average power. Just the opposite of what you are trying to
eliminate.
Of course some moderate clipping and filtering does help a little to
increase average power on SSB but it is much more effective with a double
side band transmitter.
Somewhat effective audio has been done with multiband audio clippers.
73
Gary K4FMX
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|