[time-nuts] 1 PPS from Z3801A Does it need conditioning.

ed breya eb at telight.com
Wed Sep 21 21:17:31 UTC 2011

If you have a definite ground-loop problem, you can reduce the 
sensitivity to direct triggering on it with the counter settings, but 
that doesn't necessarily eliminate the interfering signal or its 
possible effects - if it happens just at the right time on one of the 
desired edges it will still add errors to the information. It may be 
worth first fixing the ground loop by maximizing sensitivity to it, 
in order to see and track down the source, and find an effective solution.

First try securely tying the chassis grounds of all the equipment, 
and make sure they all feed from the same line circuit - as close 
together as possible. Each piece of modern equipment typically has a 
line filter with caps from the lines to chassis. Any line transients, 
regardless of the source, will pass a signal through the caps that 
ends up being distributed through the ground system, so keeping the 
chassis grounds tightly linked will minimize what ends up on the 
signal cables. The lead inductance will limit the effectiveness, but 
it's a start.

Once this is done, the next level is to try common-mode chokes in 
various forms - ferrite "prayer beads," toroids,  or split cores - 
first on the signal lines, then on power cords, to steer the 
undesired currents away from the signal interconnections. It can take 
a lot of experimenting.

If that is all not enough, then consider going fully-differential, 
and shielded, using impedance matched drivers and cable to best 
preserve the edges. BTW, if you have any old analog scope carcasses 
around, the delay lines inside make excellent differential 
transmission cables which are very stable because they are fairly 
stiff, using solid dielectric (usually polyethylene or PTFE).  They 
are not good for general purpose use, but great for temporary 
experimental setups - once formed up and connected, they will barely 
move. The Z is typically around 100 ohms differential, as I recall. 
The typical DL is perhaps 50 feet long, so you can make a lot of 
one-time-use cables.

PS: one piece of equipment that I have found to be very noisy is the 
common Weller type soldering iron station. They are always around in 
the electronics lab, and typically running during experimenting and 
building things. When the temperature control cycles off it makes 
very large line transients. If you ever find a seemingly random line 
surge noise problem in the lab, check your soldering stations first.


More information about the time-nuts mailing list