[time-nuts] oscillators

[Jim Lux]

On 8/27/12 10:45 PM, Chris Albertson wrote:
> On Mon, Aug 27, 2012 at 7:33 PM, Jim Lux <jimlux at earthlink.net> wrote:
>
>> On 8/27/12 4:15 PM, Rick Karlquist wrote:
>>
>>> Several decades ago, the concept of the "smart clock" arose
>>> at what was then HP.  The idea was as discussed here to
>>> "characterize" past aging, "predict" future aging, and
>>> then "correct" the aging.
>>>
>>
> We know what a OCXO is and a TCXO is.  I was at a presentation at work a
> whike back and they called what you describe a "MPCXO"  or MicroProcessor
> Compensated XO.    They said the characteristics were between the OCXO and
> TCXO
>
> Doesn't the thunderbolt do this.  I think it watches the aging rate of the
> OCXO and adjusts during hold over.
>


Not really.  The MCXO has a one time calibration for frequency vs 
temperature that's programmed into it (and some use *very* clever ways 
to measure the temperature).  The disciplining algorithms in a GPSDO are 
a bit smarter; some explicitly develop a model for the f vs T and apply 
it, in others it's essentially embedded in a higher order filter which 
takes the measured T, along with other parameters, into the filter.

I don't know if the GPSDOs try to do a time series fit/model (at least 
for low order terms) to deal with things like diurnal variation.  They 
could.

I should note that the MCXO approach, popularized as a "better TCXO" in 
a small low power package, is also used in some software defined radios, 
except that there's no separate microcontroller.  The frequency vs 
temperature characteristic is just embedded in the other algorithms in 
the radio's host processor.