[time-nuts] HP 5370B vs SR620

Pete peterawson at earthlink.net
Tue May 19 14:27:37 UTC 2009


Roy,

The 5370s have a few quirks that you should know about to avoid
possible surprises. These aren't newly discovered & have been the
subject of various postings a few years ago.

1) Both instruments have some front end coupling with the data path
    & the timebase. This was mostly corrected in the 5370B. 

2) Both instruments have a low level issue with non-linearity of the
    interpolator which provides the ~20ps resolution.

The result of these imperfections is jitter readings taken on the 
reference output signal (this is used in performance verification)
will display the lowest readings you'll ever see. Real world signals
locked to this source behave the same way. These readings are
satisfying, but not real. When H-P spec'd the jitter floor at 30-
35ps, that's the true performance limit. Still quite good by any
measure.

Another candidate counter you should check out is the Fluke/
Pendulum PM6681. These aren't common, but sometimes are
on *bay for reasonable ($500-600) prices. This is a much 
more versatile instrument than the 5370s with lower front end
noise, thus somewhat better jitter floor.

Regards,
Pete Rawson



More information about the time-nuts mailing list