[time-nuts] A little tail about using the time-nuts tricks at work

[Magnus Danielson]

Fellow time-nuts,

Every now and then, a customer runs into trouble. You end up in these 
meetings where vendor and customers discuss troubles. I heard about this 
one and invited myself along. Turns out my friends at the customer was 
attending. Troubles involved dropouts and blips on a transmission. They 
have had one problem with a cyclic error, but it reduced after finding 
out they had forgot to enable a sync input. The remaining problem was 
erratic and showing up only for some locations and not correlating to 
anything obvious. One of the techs had tried out to see what an external 
sync-box was doing, and he had found it to do strange things, but it was 
not conclusive. We then realized that it might be that handling of the 
box could be different for each occasion they where running, because on 
side is a travelling setup, so it depends on the wiring and power-up 
order. So, we had a nice candidate. Our boxes have nice logging, but you 
loose much of the data if you do not pull it out before power-down, so 
we had no real detailed logs from when real hits occurred, so we where 
blinded. Not really a clear picture, but we had a good meeting. As we 
where over at their place, I asked if I could borrow one of those boxes 
as I can measure things better. They showed me what they did with an 
oscilloscope, and well, it did move around a little, but it wasn't very 
clear what was going on.

As I got back to my lab-bench at work, I hooked it up with sync signal 
and then measured the frequency, and sure enough, it showed frequency 
errors similar to what we had seen before. Then I turned over to TIC 
measurements, meaning a HP5335A, a GPS as PPS and 10 MHz source and a 
laptop with GPIB and TimeLab. We do have better counters, but I get to 
use the 5335A without people stealing it and besides I like it better 
than the 53132A. Just doing basic operations as turning it on, providing 
sync and watching the phase and frequency, I could see *much* better 
what this box is doing, and the hunch that my customer had was 
completely confirmed. A quick report on email caused them to book a 
visit, and they sat down and could see it for themselves. The 5335A and 
TimeLab with a GPS turned out to be a valuable tool for this exercise, 
clearing up the issue in a few hours of play-time. Flipping between 
phase and frequency view, zooming in etc. all helped to illustrate what 
was going on.

I will assist my customer in reporting on the issue to the sync-box 
vendor, providing measurements and explanations.

I hope this little tail will encourage you to use the tools, learn to 
investigate them and learn the various ways of displaying data, so it 
can help you to pin-point the problems you are having.

Oh, I had fun doing it! :-)

PS. I will help a friend to get his PM6654C up and running, and try 
using it with TimeLab. With it's 2 ns resolution, it would still be 
sufficient for this exercise.

Cheers,
Magnus