Comments embedded below... And yes, this is a very interesting topic. Of
course the real problem will probably turn out completely unrelated.
----- Original Message -----
From: "Ron" <w8ron@stratos.net>
To: "jeff millar" <wa1hco@adelphia.net>
Cc: <W8JI@contesting.com>; "'Amps'" <amps@contesting.com>
Sent: Tuesday, May 21, 2002 10:23 PM
Subject: Re: [Amps] Non-linear subharmonics was: Nonsense, AL-1500 parasitic
> Hi Jeff.
> Thanks for you mail.
> I think the topic is drifting where one fellow had a problem with a spur
on his
> amp at 1.5X the fundamental.
> The discussion the came to suggestions that the AMP was creating a
sub-harmonic
> by itself....in other words , the amp was the problem.
> Other suggestions were that the exciter had a spur and the amp was
re-inforcing
> it or mixing or tripling it.
I still think something in the amp or something in the
feedline/antenna/tower/guy system can cause the problem. In response to the
initial question, I asked about ferrites in the tank. But maybe a solid
state driver could produce the sub-harmonic. If the measurements are
correct, then the spurious is truely a sub-harmonic and it becomes
interesting to find the source.
>
> Then we drifted to sub-harmonics and my statement was that any
non-linearities
> would be on the fundamental waveform and therefore would be higher in
frequency
> and sub is not possible in that it would be a non-linear variation from
one
> cycle to the next to be lower in frequency. To be an odd fraction of the
> fundamental would mean that the non-linearity effects one cycle of the rf
> fundamental but not the next!
We need to distinguish between
non-linearities in resistance that produce harmonics (diodes, tubes
clipping, etc)
non-linearities in reactance that can produce parametric amplification
( varactors (C vs V) or ferrites (L vs I) )
Either type of nonlinearity produces harmonics on the signal. But a
non-linear reactance can produce _gain_. And gain in the right circuit
produces oscillation. Maybe we should consider a parmetric divider not a
frequency divider, but rather an injection locked oscillator running at f/2
or f/3.
Parametric dividers have a circuit that produces sub-harmonic oscillation,
which then mixes with the pump frequency to produce products at f/2, f,
3f/2, 2f, 5f/2, etc.
> In your example , you are introducing two signals and mixing and the
statement
> didn't realy mean to cover the mixing operation even though an amplifier
can and
> is used as a mixer. If you have a mixer and provide only one signal , the
mixing
> products do not appear.
To produce 3/2f, the non-linear reactance mixes the f/2 oscillation with the
fundamental to produce 3/2f.
> So I think we are talking about one specific part of the discussion that
is
> theoretical in nature. A non-linear amplifier does not choose which rf
cycle it
> it will amplify linearly or non-linearly.......can you agree on this
point?
Yes this is theoretical. I agree the amplifier tube doesn't do any of this
sub-harmonic stuff. The source of gain is the non-linear reactance. Tubes
don't have non-linear reactance, so I suggested looking at the ferrites.
After the reference to sub-harmonics in solid state amps, maybe the problem
lies in the driver stage?
> All of you examples include some type of mixing operation and the center
of the
> discussion ruled out mixing.
Any non-linear reactance will also act as a mixer. So, if a parametric
oscillation exists at a sub-harmonic, then all the mixed products will
appear as well.
> The noise issue suggests that the amp is conditionally stable.
I don't understand what you mean here.
>
> ---
> Ron
>
> jeff millar wrote:
>
> > The reference to gain at f/2 is there...near the top of the second
> > reference...
> >
> > For proper operation, the capacitance must be decreased
> > when the input voltage is maximum and increased when the
> > input voltage is minimum. In other words, the pump signal
> > frequency must be exactly double the frequency of the input signal.
> >
> > The statement makes the point that a circuit with non-linear reactance
can
> > have gain at f/2. The the author goes on to generalize
> >
> > A parametric amplifier that is not phase-sensitive, referred to as
> > a NONDEGENERATIVE PARAMETRIC AMPLIFIER,
> > uses a pump circuit with a frequency higher than twice the input
signal.
> > The higher-frequency pump signal mixes with the input signal and
> > produces additional frequencies that represent both the sum and
> > difference of the input signal and pump frequencies.
> >
> > This statement makes the point that a circuit with non-linear reactance
can
> > have gain at any frequencies lower than the pump frequency.
> >
> > Steve's point about noise helps to understand how the subharmonic gets
> > started. quoting from his email...
> >
> > As best I remember the article I read, the argument goes that
> > if anything (even noise, maybe) causes f/2 to appear in a circuit,
> > then IMD between f/2 and f produces 3f/2, then both f and 2f
> > with 3f/2 produce f/2 and so on, so it can become stable and self
> > sustaining if the circumstances are right. The same analysis can
> > apply to f/3 etc.
> >
> > This describes how any oscillation gets started. In other words, a
circuit
> > with gain starts out with zero signal and then a bit of noise occurs
that
> > gets amplified and filtered and feeds back....leading to a build up of
self
> > sustaining oscillation. In familiar oscillators, a filter ensures that
gain
> > and oscillation occurs at a single frequency.
> >
> > It's valid to call a parametric frequency divider a sub-harmonic
generator
> > because the frequency of oscillation isn't controlled in by a filter,
but by
> > the mixing of oscillation, harmonics of the oscillation and the pump
> > frequency.
> >
> > Michael Tope provided a reference
> > (http://smirc.stanford.edu/papers/JSSC99JUN-hamid.pdf) that shows a
> > subharmonic generater using a mixer and gain. That circuit looks a lot
more
> > like a classic oscillator than a parametric amplifier, but the same
> > principle applies.
> >
> > Also recently on this list, Peter Chadwick provided a reference to a
book
> > that covers the subject
> >
> > Parametric dividers can be made with a varactor: see
> > Manassewitsch, Frequency Synthesisers, Theory and Design,
> > second edn., p376 et seq.:
> > One would expect a power dependency in the effect.
> > Carrier storage devices can also be used to make a divider -
> > see Manassewitsch again.
> >
> > jeff, wa1hco
> >
> > ----- Original Message -----
> > From: "Ron" <w8ron@stratos.net>
> > To: "jeff millar" <wa1hco@adelphia.net>
> > Cc: <W8JI@contesting.com>; "'Amps'" <amps@contesting.com>
> > Sent: Monday, May 20, 2002 10:43 PM
> > Subject: Re: [Amps] Non-linear subharmonics was: Nonsense, AL-1500
parasitic
> >
> > > I read these pages and couldn't find a reference to the f/2
discussion.
> > > It seemed to discuss injecting energy into a resonant cavity using a
> > charge pump
> > > running at twice the cavity frequengy and also injecting it at other
> > frequencies
> > > to obtain mixing products.
> > >
> > > Second, Steve's comment about "any thing (even noise) getting into the
> > amplifier
> > > is an invalid argument in that any ligitmate signal that enters the
> > amplifier is
> > > a mixing product and we already discussed that. If it were noise
....well
> > noise
> > > is random and so the harmonics and mixing products would be noise and
> > randon as
> > > well.
> > > ---
> > > Ron
>
> _______________________________________________
> Amps mailing list
> Amps@contesting.com
> http://lists.contesting.com/mailman/listinfo/amps
|