On Wed, 28 Jan 2009 20:12:15 +0000, Ian Jackson
wrote:

Thanks for the advice, however......

Why? A half volt signal level

(that's surely one hell of an RF signal?)

I can see you don't know what environment you are living in. That, or
you live out in a pastoral setting and the house has no significant RF
contribution. Most of the world is urbanized and folks live within
short distances of large transmitters.

Varicap diodes will always suffer from having their capacitance
modulated by an RF signal impressed upon them. However, they do seem to
work well enough - even when the RF level is pretty high (which must be
the case especially with varicap-tuned oscillators - ie most VCOs and
PLL systems).

You don't put the Varicap into the high level part of the circuit.
This is obvious from the outset.

Presumably, the effect of this modulation will be to generate
intermodulation products. In VCOs, this will simply appear as harmonics
of the oscillator signal (which you would get anyway - even with a
conventional tuning capacitor).

A conventional cap (and a conventional inductor, much less) will not
contribute harmonics because it is linear. Harmonics comes from
nonlinear components added by the circuit (partial conduction of an
active component that aids in the oscillation).

Where multiple-frequency signals are
present (like you have with the receiving loop), the most apparent
effect of the modulation of the diode capacity will appear as
crossmodulation and other nasties on the other signals in the passband.
However, as varicaps ARE used for the tuning of the input of receiver RF
stages, how do they 'get away with it'?

They don't if you are in such an environment. Again, if you don't
know your environment, then it's all a crap shoot. There is no
"getting away with it." The alternative is that the designer of a
product fully anticipated these issues and purposely chose a design
that minimized the effect of accidental contributions. Not all
designers are up to speed on the topic. The web is full of reported
failed projects that do not take bias Z and filtering into
consideration.

This swing is also a problem for PIN diodes used as voltage variable
attenuators.

My understanding of things is that the effectiveness of PIN diodes
relies on them having a very poor performance at RF (especially at VHF
and UHF). In attenuator circuits, they are forward biassed, and the DC
current passing through them varies their RF resistance. However, the
charges flowing through the junction are so 'sluggish' that they don't
react to the rapidly-changing RF voltages. There is therefore negligible
modulation of the RF resistance by the RF signals passing though them.

Poor performance when a PIN diode is first and foremost a diode for RF
and microwave applications? Only if you are using it for the wrong
reason (like using a 1N23 for 120V 60HZ power rectification).

PIN diodes are used as resistors and switches, not rectifiers.
Further, don't confuse the switch application as meaning fast turn-off
in the pico-to-subnanoseconds. Speed is relative to the application
of signal path steering where 100s of nanoseconds is more than
adequate.

Consult:
http://www.ieee.li/pdf/pin_diode_handbook.pdf

73's
Richard Clark, KB7QHC