Order Tray | Contact Us | Home | SIG Lists

[vna] [Fwd: RE: calibration at a remote point?]

John Ackermann N8UR jra at febo.com
Thu Mar 11 01:06:10 UTC 2010

As I feared, N8ZM isn't subscribed to the vna at tapr.org list, so his very 
informative reply to my message bounced.  It's attached below.

-------- Original Message --------
Thanks, John...I think?

First, let's assume that the instrument has been calibrated to the point
where the bottom end of the coax connects. This will give us a very good
measurement of the coax/antenna combination impedance, and this can be swept
over whatever frequency range you desire, even well out of the 'band'. The
analyzer doesn't care about SWR, it just measures it, with one caveat: A VNA
is a 50 ohm centric instrument, which means that it is very accurate when
the impedance it sees is near 50 ohms. Just as on a Smith chart, the farther
you are away from the center, the more difficult it becomes to resolve the
measured value accurately.

Determining the antenna impedance at the far end of an unknown T-line can be
done, keeping in mind that the accuracy is not going to be as good as if the
coax was not included.

Certainly, if you know the electrical length and attenuation of the coax,
you can back these values out of the measurement to calculate the antenna
impedance. The trick is to come up with these values across the band. As an
aside, S-parameters make this pretty easy now that we have PC's to do the
number crunching.

Most VNA's have in their software the capability to do an Inverse Fourier
Transform, or IFFT, on the amplitude and phase data accumulated for each
swept frequency point. The resulting series of data values that come out of
the IFFT are in the time domain. If you are familiar with TDR, Time-Domain
Reflectometry, these data points will give you the same type of picture of
what is happening along the length of the T-line. For those unfamiliar with
TDR, if you lunch a very short pulse down a transmission line, every
impedance 'bump' along that line will cause a small reflection of energy
that will return to the source end at an elapsed time that can be converted
into the distance to that bump. Thus, a small dent at the midpoint of an
open coax will show a small bump of reflected energy at a time proportional
to the distance to that bump, and another, larger bump from the nearly total
reflection of the remaining energy that reached the open end, occurring at a
time proportional to the length of the cable.

At any frequency where the antenna is matched well to the T-line, there
should be a very small reflection. However, outside the band, the match is
usually just plain awful, so when we do a frequency sweep over a range much,
much wider than the bandwidth of the antenna, and then calculate the IFFT,
we will see a good sized bump at the feedpoint of the antenna. This will
give us the electrical length of theT-line.

The amplitude of the antenna's bump will give us a clue to the cable loss if
we assume that at frequencies that are way out of the band the antenna's
impedance produces a reflection coefficient of close to 1.0. If there is no
cable loss, then all of the energy sent down the line was reflected back,
and all of the reflected energy arrived at the VNA and was measured.
However, there was some energy lost, and we can pretty safely assume that
half of it was lost in each direction. So the amplitude of the bump compared
to 1.0 is the information needed to do that.

So, armed with the line length and loss, we can take our in-band impedance
measurements at the bottom of the coax and back out the effect of the cable
to get a decent picture of the antenna feedpoint impedance. Keep in mind
that the cable parameters that we calculated are based on the assumption of
a high out of band reflection coefficient at the antenna, and the limits of
the resolution ( number of data points swept and frequency span)) of the
IFFT. Also note that the cable losses reduce the dynamic range of the VNA,
which reduces the accuracy of both magnitude and phase measurement values
when they are far from 50 ohms (remember the Smith chart). And the
computer's math skills (or lack of same) may reduce the ultimate accuracy.

Nevertheless, with some reasonable thought put into it, you can get a useful
picture of the antenna feedpoint in this way.

Now, as to whether Tom McDermott has implemented this type of technique in
his software, I can't say.


Tom Holmes, N8ZM
Tipp City, OH

-----Original Message-----
From: John Ackermann N8UR [mailto:jra at febo.com]
Sent: Wednesday, March 10, 2010 5:20 PM
To: TAPR/Ten-Tec VNA Users List
Cc: Tom Holmes, N8ZM
Subject: Re: [vna] calibration at a remote point?

Hi Tom and Oliver --

I suspect that Oliver is asking about a capability to determine the SWR
at the antenna, without access to the far end of the feedline to do a
fixture calibration.

Some of the "antenna analyzer" products apparently claim the ability to
calculate feedline length, insertion loss, etc., with a one-way
measurement, and then apply that to effectively provide a fixture
calibration without requiring access to the far end of the cable.

My ham friends and I talked about this at our weekly "Lunch With the
Geeks" a couple of months ago, and we couldn't come to agreement on just
how you could do that without knowing a priori either the electrical
length of the cable, or its insertion loss.

I'm copying Tom, N8ZM, on this message as he's a network analyzer guru
and may be able to provide some insight.  Tom, can even a fancy Agilent
VNA look at the bottom end of a feedline plus antenna combination and
figure out the "true" antenna SWR (and for extra credit, the feedline
length and insertion loss) without knowing ahead of time at least some
info about the cable?


Tom McDermott wrote:
> Hi Oliver,
> Fixture calibration has been available in the TAPR software since the 
> beginning. There are some sources of error however when calibrating out 
> a feedline, a significant one being that a feedline is not exactly 50+j0 
> impedance. A two-port instrument can compensate only 4 error terms of 
> the 16 total.
> -- Tom, N5EG
> --- On *Wed, 3/10/10, O. Johns /<ojohns at metacosmos.org>/* wrote:
>     From: O. Johns <ojohns at metacosmos.org>
>     Subject: [vna] calibration at a remote point?
>     To: vna at tapr.org
>     Date: Wednesday, March 10, 2010, 9:19 PM
>     Hi,
>     Is this list still alive?  We'll see.
>     Awaiting arrival of a TenTec TAPR VNA.  I see in the user guide to
>     the Array Solutions VNA2180 a procedure for calibrating their VNA to
>     read the impedance, SWR, etc., AT THE ANTENNA, i.e. at the antenna
>     end of the feedline while the VNA is at the rig end of the feedline.
>     Is this possible with the TAPR unit?  Any tips?  Any reason it won't
>     work?
>     Tnx and 73,
>     Oliver
>     W6ODJ
>     _______________________________________________
>     vna mailing list
>     vna at tapr.org
>     https://www.tapr.org/cgi-bin/mailman/listinfo/vna
> ------------------------------------------------------------------------
> _______________________________________________
> vna mailing list
> vna at tapr.org
> https://www.tapr.org/cgi-bin/mailman/listinfo/vna

More information about the vna mailing list