|
You are right. The crystal oscillator (channel element) noise is
multiplied up in the GE multiplier exciter. But the same thing applies to
a GE PLL exciter. The phase noise of reference oscillator (channel
element) is also multiplied up, it's just not done in the same
fashion.
Noise is a form of instability. The best way to think of it is,
instead of 100% of all of the energy being at exactly one frequency, it is
not. The oscillator is jittering slightly from center frequency, thus
producing sideband noise. The better the oscillator, the less the jitter
(or instability) the less the noise will be. This is why some oscillators
have more "hiss" than others.
As far as actually quantifying the phase noise of any oscillator or
exciter, it done in terms of its spectral power density, which is the power
contained in a 1-Hz bandwidth at any particular
frequency.
Additional noise contributors that unique to a PLL
circuit, include VCO noise (which can be far worse than just the noise
of a crystal oscillator multiplied up due to it's lower Q), divider noise,
active loop filter noise, phase detector flicker noise, among
others.
Don't get me wrong, I'm not trying to argue a "multiplier" design is a
better way to go. I'm just trying to shed some light on how it works,
what actually causes "hiss" and make the statement that just because one
attempt at a particular approach did not yield very good results,
doesn't mean that other approaches are automatically superior. With
today's improved PLL techniques and integrated devices, PLL is an
excellent approach for new designs.
Chris Hudgins - N5IUF
Unless these multipliers without "issues" have very high Q tuned circuits, Yahoo! Groups Links
|
