[time-nuts] Fury Interface Board: How about TI OPA277?

kevin-usenet at horizon.com kevin-usenet at horizon.com
Sat Nov 3 13:14:53 EDT 2007


Yeah, it gets a bit heated.  Since you have the data sheets and all,
could you spare us a bit of work and summarize, for each OCXO:
- What EFC and power voltages it needs.
  (Using an input voltage lower than your own supply rail seems odd.)
- What EFC voltage stability (uV per day or whatever) is needed to match
  the oscillator stability (ppb per day).
- And what RFC voltage tempco (uV/C) is required to match the oscillator
  tempco (ppb/C).

Obviously, we want to do better than just *matching* the OCXO specs,
because that would degrade performance noticably, but it's a good
number for comparison.  And doing more than 10x better is probably a
waste.

On particular, if the +/-10V EFC oscillator is an older model, we might
not need the same specs on the level-shifter as on the 0V-based
amplifiers.


I'd also be curious what the bandwidth of the Fury discipline loop
is, approximately.  A simple passive low-pass filter can cut off
higher frequencies without causing problems, but phase shift at
too low a frequency can cause oscillation.

That determines the bandwidth over which the buffer amplifiers have to
be quiet.  If the frequencies are low enough, it might be worth using a
chopper-stabilized "zero-drift" amplifier.  They have some high frequency
hash in their output, but it can be filtered out.



More information about the time-nuts mailing list