[time-nuts] Fine frequency adjustment in 5370B
Tom Van Baak
tvb at leapsecond.com
Sun May 15 17:43:04 EDT 2005
> That's what I was intending doing, although the adjustment as far
> courser than ideal on a counter with 11 digits.
Many of us use HP 5370A/B's as *time interval*
counters; typically to precisely compare two 1 PPS
In this case the 25 ps rms single-shot resolution of
the counter is the key feature, not the 11-digit display,
nor even the quality of the timebase. You're 10811
OCXO can be off by huge amounts and you can
still get perfectly accurate 1 PPS TI measurements.
As a example, roughly speaking, if you tune your
10811 today to be accurate to 11 digits, it will be
down to 10 digits by the day after tomorrow, to 9
digits after a week, only 8 digits by the end of the
month, and 7 digits by the end of a year. There is
a good chance it will still be accurate to 6 digits
(1 ppm) after 10 years.
Still, even when off by as much as 1 ppm (e.g., ten
years since calibration), if the 1 PPS time intervals
you are measuring are down in the microseconds --
as would typically be the case when comparing Rb,
Cs, or GPS references -- your tired old 5370B will
continue to give you perfect readings. 6 digits of
accuracy is more than enough when measuring
time intervals with 25 ps of resolution and under
10 microseconds in range.
More information about the time-nuts