[time-nuts] Fury Interface Board: 5MHz needed?

[Bruce Griffiths]

SAIDJACK at aol.com wrote:
> Hi guys,
>  
> here is a thought: why not use EFC parasitic ground current loops to  cancel 
> out OCXO Tempco?
> ___________________
>  
> Most single oven heaters will have a significant tempco.
>  
> If that EFC tempco happens to be the inverse to the heater current, then it  
> should be possible to carefully select the heater-current-induced EFC-ground  
> voltage change to cancel out the OCXO tempco?!
>  
> In other words, if heater current increases (due to falling temperatures),  
> then EFC voltage will increase due to the parasitic ground loop.
>   
Not quite, the heater current induced ground system voltage drop
subtracts from the EFC voltage so the effect depends on the sign of the
frequency vs EFC voltage slope.
However one could sense the OCXO current and add (or subtract) a voltage
proportional to the current from the EFC voltage to compensate for
internal and external ground bounce.
Some allowance would perhaps have to be made for the ground wiring
resistance tempco (complex because the temperature at one end is the
oven temperature and at the other the PCB temperature.
Either a pot could be used or a microprocessor can perform the
calculation and adjust the EFC DAC ouput as appropriate.
>  
> If one is lucky, then the OCXO has a negative tempco that requires a higher  
> EFC voltage at lower temps.
>  
> Then, it should be possible to pass just the right amount of current  through 
> the EFC ground connection to get a good tempco cancellation :) This  current 
> could be scaled by selecting the length of the EFC ground cable for  example.
>  
> Dirty, but probably effective once calibrated.
>   
With some ingenuity and an opamp or 2, either sign can easily be be
accommodated.
Using a length of PCB track for the current sense resistor may provide a
degree of temperature compensation for the tempco of the internal ground
wiring resistance, provided that the internal ground wiring resistance 
tempco tracks that of copper. However the finite thermal gain of a
single oven OCXO will in most cases dominate the OCXO thermal drift so
an adjustable tempco sensing resistor may be more useful.
>  
> bye,
> Said
>   
Bruce