[time-nuts] Abrupt changes in TI with HP 5370B's own timebase

[David Kirkby]

I've set up my HP 5370B using its own internal 10811-60111 oscillator, a 
BNC-BNC lead to connect the oscillator output to the start input and the 
COM, SEP switch to the START COM position. This is used as a performance 
test (see page 4-2 of the manual).

The readings should be 100ns +/-1 1ns, which mine are, but some rapid 
changes can be observed when data is collected over a long time (75 
hours). See:

http://www.g8wrb.org.uk/useful-stuff/time/data/abrupt-changes-in-TI/75hours.html

Data is read down the GPIB bus about once every 100ms, so there are a 
lot of data points.

I'm surprised by the changes seen around 12, 18, and 33-43 hours. Can 
anyone explain what they might be? The data is not as "clean" as I would 
have liked.

The jitter seen on multiple reads have an SD of about 30ps over short 
(let instrument computer SD of 100k readings), but clearly on very long 
times the SD is higher (101.258ps to be exact), due to these jumps.

The instrument has its covers on, but is not in a temperature controlled 
room. Quite the opposite, the temperature does vary quite a bit, as and 
when a window is opened, which is not too far from the HP 5370B.

But I would not think even any change in oscillator frequency should 
cause this.


PS,
should anyone want to look more closely, the original data (all 30MB of 
its), can be found in the zip file in this directory

http://www.g8wrb.org.uk/useful-stuff/time/data/abrupt-changes-in-TI/


-----
David Kirkby,
G8WRB

Please check out http://www.g8wrb.org/
of if you live in Essex http://www.southminster-branch-line.org.uk/