[time-nuts] 5070B once more....

[John Ackermann N8UR]

Tom Van Baak said the following on 05/30/2009 03:44 PM:

>> ... as long as you remember to use TI+/- mode.
> 
> For stuff like this I use plain old TI mode and just set the
> UUT 1pps to lag the reference 1pps by a fraction of a
> second, or maybe even just a couple of microseconds.
> 
> That way you get a small positive time interval reading
> every second.

For what it's worth, I evolved a system where one of my GPS units and 
one of my Z3801A run as close to UTC(GPS) as I can get, taking into 
account cable delays, etc.

I run another Z3801A at about 500uS early relative to GPS.  Then, I use 
that 3801A and an HP 5359A time synthesizer (which can generate delays 
of up to 180ms) as the reference when I sync the dividers in the clocks 
that I am monitoring.

I set up a hierarchy -- with the nominally highest drift rate clock set 
maybe 250uS early, the next unit 150uS early, and the final unit 50uS 
early.  Now, I can compare #1 against #2, #2 against #3, and any of them 
against GPS using positive TI.  (This assumes the convention that the 
DUT goes on the START input, and the reference on the STOP input.)  I 
can usually take data this way for quite a while before any of the units 
cross over.

John