[time-nuts] Basic regenerative-divider questions
bruce.griffiths at xtra.co.nz
Sat Sep 29 19:16:50 EDT 2007
John Miles wrote:
>> Am I missing something here?
>> I always thought mixers were non linear by definition, and
>> relying on that
>> non linearity to function:-)
> Sure, a mixer is nonlinear with respect to the multiplicative function it
> applies to its two inputs to obtain the desired output. It should, however,
> behave linearly with respect to multiple frequency components that may be
> present at any *one* input. You don't want it to modify either input
> signal, just multiply them together.
> Think of a mixer with a perfect sine wave at its RF input and a square wave
> at its LO input. It's nonlinear with respect to the switching action caused
> by the LO signal, but it had better be linear with respect to how it handles
> the sine wave being switched on and off. If it distorts the sine wave
> input, it will generate harmonics that you probably didn't want. And if you
> apply two or more tones to the mixer's input at once, you want only those
> same tones coming out, with the usual +/- translation by the LO frequency.
> To the extent that the mixer allows the RF input tones to interact or
> multiply with each other, it's nonlinear.
> This wasn't such a big deal in the old days when your radio had a high-Q
> tuned circuit in front of its first mixer, but modern designs work by
> shovelling a large portion of the spectrum into the mixer at once.
> Nonlinearity is a bad thing in that case.
> -- john, KE5FX
At drive levels below saturation, the loss of a mixer depends on the LO
Consequently the feedback loop gain of a regenerative divider depends on
the input signal level.
Hence one would expect there to be a well defined threshold at which the
lop gain is sufficient for regeneration to occur.
More information about the time-nuts