[time-nuts] Spacecraft Timekeeping

[SAIDJACK at aol.com]

Hi Tom, Kevin,
 
the JLT DROR GPSDO mentioned by Kevin in the original  email may be 
overkill for this application, as it is designed to provide  very low phase noise 
under extreme vibration conditions, and it  is state-of-the-art technology 
and priced accordingly. It is designed to  interface directly with a vehicle 
power supply, so that may help in simplifying  the installation.
 
Phase Noise doesn't seem to be a requirement here, and we do have a  
"standard" version that does not have the phase-noise-under-vibration cleanup  
loop installed to save substantial cost. It should work quite well in the  
application, as it has various extremely low-g-sensitivity oscillators on board  
that will keep drift to very small values even under very large 
acceleration and  vibration, and temperature changes. It was designed to be essentially 
 indestructible.
 
If cost is an issue, we have ultra-low-g versions of our FireFly-IIA  
products available (our g-force product line) that are designed and tested to  
provide a guaranteed holdover under the conditions that Tom mentions below. 
They  are designed to guarantee drift in "backpack" type applications without 
GPS, but  with lot's of thermal changes and a very wide temp range, loads of 
acceleration  (shock, vibration, and tilt), and fast warmup. To give an 
example, we can  achieve better than 0.1ppb per g per axis acceleration 
sensitivity on the  worst-case axis on our g-force products.
 
So for example the FireFly-IIA g-force would have the following drift in a  
hypothetical high-g application:
 
  g-force on average over entire trip - 3g (this extreme acceleration  
would never happen in reality of course)
  average thermal change - 40 Degrees C
  allowed warmup time with GPS - 1 hour
 
Drift over 5 hour mission:
 
   3g x 0.1ppb/g + 40C * 0.005ppb/C + 0.1ppb initial error =  0.6ppb 
overall average error over 5 hours
 
So in this extreme scenario we have 0.6ppb * 5 * 3600 = 10.8 microsecond  
drift over 5 hours.
 
Aging and retrace is not taken into consideration here, but that should be  
more than compensated by the unrealistic constant 3g acceleration error we 
added  to the calculation.
 
The cost of our g-force products is only about 2x to 2.5x the price  of our 
standard DOCXO products.
 
bye,
Said
 
 
 
 
In a message dated 3/9/2011 12:11:49 Pacific Standard Time,  
tvb at LeapSecond.com writes:

Some  GPSDO "learn" their hold-over parameters over time,
sometime days, in which  case you may be in for a surprise.

Is there any way you can simulate  the temperature, vibration,
and g-force (or lack of) trauma that the GPSDO  is going to
experience during this hour of launch-to-orbit?

If  not,  I suggest a simple test: just throw your GPSDO and
battery in a  backpack and go jogging. Twirl the backpack in
the air every now and then.  After an hour I'd be very curious
by how many microseconds it's  off.

/tvb