[time-nuts] Square to sine wave symmetrical conversion (part 2)

Bob Camp kb8tq at n1k.org
Sun Jul 26 21:21:45 UTC 2015


Hi

> On Jul 26, 2015, at 2:48 PM, jerry shirᴀr <radio.n9xr at gmail.com> wrote:
> 
> Hi Bob,
> 
> Help me to understand your remarks.
> 
> A.  The oscillator stage has a limiter.  (Agreed.)
> B.  ​You can *easily* get <-160 dbc/Hz with an oscillator that has
> harmonics in the
> -10 to -20 range?

yes

> C.  But we are not talking about the signal from the oscillator stage
> itself?

yes

> D.  Tuned buffers can have these higher harmonics and all is well?

not if they are properly tuned. The tuning is bandpass in nature.

> E.  Okay.  We are now talking about downstream harmonics and not the actual
> oscillator stage harmonics?

nope.

> 
> Ideally the oscillator stage should have no tuned circuits so that the
> feedback of this oscillator stage utilizes only the tuned circuit of the
> high Q resonator.

Fine unless you have an overtone crystal. Roughly 99.99% of all low noise 
stuff runs on an overtone. That forces you to some sort of “zonal filter” in the 
feedback path. 

>  Okay.  What frequency components should the hi Q
> resonator exhibit?

The resonator has impedance properties. 

>  Should it be rich in harmonics or should it be pretty
> limited to one frequency represented by a sine wave?

All *real* resonators have “rich” impedance plots. An SC very much so. 

>  Are we talking about
> the harmonics at the output of the actual oscillator stage or after a few
> buffer stages?

No, we’re talking about the signals present in a typical oscillator stage it’s self 
*before* you decide to clean them up and *after* you make this or that choice 
about where to get a signal from. 

> 
> Help me out here.

Your oscillator *must* meat Barkenhausen’s criteria. It’s got to have a gain of
*exactly* one and a phase shift of *exactly* zero degrees (modulo 360 degrees). 
Anything else and it will not sustain oscillation. If the oscillator has a gain of one all
the time, it’s unlikely it will ever start oscillating. The only practical way to make an
oscillator work is to provide “excess gain” at starting. If that gain is in the 3 to 6 db range, 
the norm is to take care of it by limiting.

If you look at a very conventional circuit that runs a single transistor as the oscillator stage,
that stage is both a limiter and the sustaining stage for the oscillation. If you take a look at
the collector current, it’s far from continuous. Pick off from there without any bandpass tuning and shove the result into 
a phase noise analyzer. Phase noise is as good as any other point on the oscillator. Harmonics
*may* be down 6 db (more likely 10)  ….

Bob


> 
> Thanks,
> 
> Jerry N9XR
> 
> 
> On Sun, Jul 26, 2015 at 8:44 AM, Bob Camp <kb8tq at n1k.org> wrote:
> 
>> Hi
>> 
>> Ummm …. errr….
>> 
>> The oscillator loop has a limiter in it or it’s not going to work very
>> well.
>> ​​
>> You can *easily* get <-160 dbc/Hz with an oscillator that has harmonics in
>> the
>> -10 to -20 range. With some tweaking you can get into the 170’s.
>> 
>> It’s not the limiting by it’s self, it’s *how* the limiting is achieved.
>> 
>> In most cases the “oscillator output” you are looking at is not the signal
>> from
>> the oscillator stage it’s self. You are looking at a signal that has been
>> through
>> several (like tuned) buffers before you see it. The same narrow band tuned
>> stages
>> can be used to “clean up” harmonics on any signal. The old style rack mount
>> OCXO’s did a *lot* of this.
>> 
>> Bob
>> 
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.




More information about the Time-nuts_lists.febo.com mailing list