[time-nuts] 5370B problem

[SAIDJACK at aol.com]

Hi,
 
on the nanosecond-scale jumps I see, I attached an actual measurement in  
which you can clearly see the full-nanosecond jumps with only single-digit  
picosecond noise. The output rate is about 8 measurements per second:
 

TI = 1.38230000000E-08
TI = 1.38290000000E-08
TI =  1.38190000000E-08
TI = 1.28270000000E-08
TI = 1.18240000000E-08
TI =  1.28210000000E-08
TI = 1.38280000000E-08
TI = 1.08360000000E-08
TI =  1.28240000000E-08
TI = 8.82600000000E-09
TI = 1.08300000000E-08
TI =  1.38240000000E-08
TI = 1.38180000000E-08
TI = 1.28220000000E-08
 
You can even see the unit switch scale correctly going from 1.2824E-08 to  
8.826E-09. Note that the picosecond range only changed by two picoseconds, but  
the nanosecond range jumped a full 4 nanoseconds.
 
And this is a "good" sequence in comparison. Note that only the single  
picosecond digits are changing due to noise(!):
 

TI = 1.28240000000E-08
TI = 1.28240000000E-08
TI =  1.28260000000E-08
TI = 1.28250000000E-08
TI = 1.28260000000E-08
TI =  1.28210000000E-08
TI = 1.28230000000E-08
TI = 1.28260000000E-08
TI =  1.28230000000E-08
TI = 1.28220000000E-08
TI = 1.28280000000E-08
TI =  1.28290000000E-08
TI = 1.28240000000E-08
TI = 1.28250000000E-08
 
Side-note: I do really like the single-digits RMS noise at 8  
samples-per-second  for it means better than 1E-012 per  second noise-floor :) I can only 
achieve this single-digit picosecond noise  with my FTS-4050 against the Fury, 
the PRS-10 is much more noisy.
 
I just wish the unit wouldn't jump around in the nanosecond digits.
 
Again this is with 100 averages, +-TI mode, FTS-4050 Cesium 10MHz against  
Fury GPSDO 10MHz.
 
Does anyone have any ideas how to fix the issue?
 
thanks,
bye,
Said



************************************** See what's free at http://www.aol.com.