[time-nuts] OCXO Power Study

[Bob kb8tq]

Hi

There’s a lot more to an OCXO than how the power changes. Power changing quickly *might* mean that the 
temperature at the crystal is more stable (heat is applied as soon as the temperature changes a very small 
amount). The crystal in one uint might be tightly coupled to the oven and in another design, more isolated (buffered)
from the oven enclosure. There are lots of variables. 

Insulating an OCXO will change its thermal gain. Thermal gain can quickly be described as how many degrees
C the crystal changes for a given power change. It turns out that there is an optimum thermal gain for an OCXO.
To little gain and the crystal changes a lot. Increase up to a point and things get better. Go past that point and your
temperature curve inverts (= gain goes negative). Add more gain still and the stability gets worse. 

While it might seem best to have an enormous gain right at the crystal, the rest of the circuit *does* get into the act. 
A modern device is “tuned” to balance the crystal against the TC of the rest of the circuit. The net result is a fairly
precise combination of gain and set point that gives the minimum temperature impact. 

If you want to get into all the “why” of this stuff - check out Rick’s papers from back in the 1990’s.

Ultimately there is no substitute for actually measuring the performance of the device by directly monitoring frequency. 
For temperature performance you have to run the device over temperature in a controlled manner. 

Bob

> On Feb 27, 2019, at 8:51 PM, AC0XU (Jim) <James.Schatzman at ac0xu.com> wrote:
> 
> Time Nuts-  You may find the attached OCXO power study interesting. I will be adding additional OCXO models and also direct time stability measurements to the data...
> 
> One thing I think that I conclude is that I would not recommended this particular Trimble OCXO if you are looking for ultra stability. Also, if a device containing an OCXO permits, adding additional insulation around the OCXO case may improve stability.
> 
> Jim<OCXO Power Consumption1.pdf>_______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.