[time-nuts] Tbolt instability

WarrenS warrensjmail-one at yahoo.com
Fri Oct 9 01:25:33 UTC 2009


----- Original Message ----- 
From: WB6BNQ 

WarrenS wrote: 
8) Do not put any pressure on the case top or bottom, & put some rubber feet under it in the 4 corners.
>What do you mean by pressure ?  Can you explain the pressure issue ? 
If you squeeze the case or set something on top if it on set it so that it applies some pressure to it's  bottom cover
 all can have small effects on its Freq, much the same as tilting it.
  9) Do not let it Tilt, move or vibrate it. (I put mine on its zero G axes)
>Which axis is that and how did you determine it ? 
This is an overkill unless you are moving the unit around some and can not be sure it is placed back at the exact same level position.
The Null may be any axis and it will very unlikely be on a flat axes.
Just tilt it around and see there the freq change effect nulls and reverses its direction.
FINE check it by tilting it in any angle and the first 10 deg or so should do nothing and then all tilt directions will cause the same freq anount and direction of change.
Using Lady Heather

12)  Make sure  the Dac gain is correct, mine is -3.4, not the 5Hz/ volt default
>How did you determine this value ? 
  One way is with an external counter and the unit in disable mode.
 You can output various small value changes to the Dac and measure their effect on the freq.
Using Lady Heather?  Or it may take 'trimble monitor' for that because LH may have a S/W bug that effects the Dac out command?
The three things Dac Gain effects is the accuracy of the Damping, TC and the accuracy of the OSC Freq plot.
If you do not care about them, then not a big lost to leave it uncalibrated.  Watch the sign, It's neg!

13) The default TC is way too slow, I've found 500 sec with a damping of 1.5 NOW works good for me. 
Without the TLC modes, It is safest to leave it at its default of 100 sec and 1.2.

>Ditto on this one as well. 
If you mean how or why, 'Tom Van Baak's  leapsecond.com is the place search for more info.
He is the expert on how to do that using ADEV data.
BUT, Simplest is set TC to 330 sec and leave everthing else alone. 

Assuming you like to tinker, and that you have already taken reasonable steps so the unit does not see large or fast temp changes or any other movement changes, One way is:
Start at 500 seconds and 1.5 Damping and if the phase error stays within YOUR  acceptable limits over time, 
change the TC to 1000 second. 
If the phase error goes outside your acceptable limits decrease the TC to 250 second.
One of those 3 values of TC should be good enough for just about everyone. 
And what is the BEST "acceptable phase error" to use?  
Around 15- 25 ns  peak over a 1/2 day or so is a good place to start. 
It depends on the changing environment that the units sees. 
Another way to set it a little coser it is to keep making the TC longer say by x1.4 until the Dac voltage plot noise reduces to a min  and then a bit more so the Phase error plot starts to increase a little.  
all Using Lady Heather plots


More information about the time-nuts mailing list