[time-nuts] My Racal-Dana 1992

[Charles P. Steinmetz]

Don wrote:

>In that case, what and where is the reference for the heater servo?

It will be a voltage or current, suitably regulated and 
conditioned.  The temperature sensor's signal will be (or will be 
converted to) a voltage or current that is compared to the reference 
voltage or current.  Zeroing and scaling are done in the servo 
amplifier and/or the reference conditioning.

In principle, you could do the same thing with frequency -- feed a 
voltage-controlled oscillator into a frequency-to-voltage converter 
and servo the FVC output to a reference voltage.  But the physical 
signal we can generate directly most accurately is frequency (using 
quartz crystals or atomic resonators).  Think about it -- most time 
nuts have one or more frequency standards that are not only stable, 
but accurate to E-9 or better (sometimes quite a lot better), and can 
scale that accuracy to a wide range of frequencies.  The volt nut, on 
the other hand, struggles to maintain repeatability -- much less 
accuracy -- to parts in E-7 at just one voltage.

Best regards,

Charles