[time-nuts] Safely getting the electrical length of a connected antenna feedline

Dr. David Kirkby (Kirkby Microwave Ltd) drkirkby at kirkbymicrowave.co.uk
Tue Aug 9 06:34:38 EDT 2016

On 8 Aug 2016 21:23, "Bob Camp" <kb8tq at n1k.org> wrote:
> Hi
> An even more significant question:
> Is it worth doing?

> Your antenna and module could easily have delays
> in the 40 ns range. It has no impact on a “frequency” GPSDO. It is one of
a number of static offsets in a time transfer system.
> Even the NIST level outfits seem to have issues coming up with a purely
mathematical answer to “what is the offset”.

I thought i might be possible to do this with a VNA, and a test antenna
located a known distance apart. So I asked in the Keysight forum on 8th Feb


There's a replay by Dr_Joel, who is a VNA guru. Dr. Joel Dunsmore,
suggesting the use of the group delay function. In one sentence he wrote

"What is the level of delay accuracy you are looking to achieve.  With this
method, 1 nsec is reasonable, but if you need 100 psec or 10 psec, then we
will have to be much more careful."

The thread never got a complete solution, but it might help if other look
at that, and perhaps start another thread on a similar topic, as that one
is very old. However, if you don't have a VNA, I would not bother asking,
as you are not going to get a response (no pun intended), to a non-VNA
question on a VNA forum.

But IF it was possible to determine the delay through the antenna/filter
within a ns, it would make measurements of coax length useful, whereas as
Bob said, it is pointless unless you know the characteristics of the active

I don't know if there would be a way of generating a pulse and feeding that
into two antenna

1) Active one.

2) Passive one at the same distance, and same length of cable.

The pulse should arrive at the same time if the two antennas had equal
delay. But the signal from the active one will arrive later due to the
delay. That might be possible to see on a scope.

I think assuming that the delay in the active antenna can't be measured is
maybe an assumption that is untrue. You could perhaps do better if you
built your own antenna, and characterized the SAW filter separately.


More information about the time-nuts mailing list