I have both a NanoVNA H4 and a NanoVNA V2 Plus 5 (pro) and I am trying to measure the group delay of an RF power enclosure box, which is primarily a Bias Tee with some adapters. I noticed some wiggles in the signal, so I found the source of the wiggles to be a 2m SMA cable I used to connect to the box. Can someone explain where these wiggles come from (reflection?) and why are the wiggles greater in the NanoVNA H4 compared to the V2 and why are they at a higher frequency? I am using NanoVNA-Saver to calibrate and take the measurement. I'm doing 1MHz - 701MHz 1212pts 3 times averaging sweep.

The impedance of flexible cables is not 50.0000 ohms but around 47-53
ohms. The centre conductor is almost never exactly central, the
dielectric materials vary and the braid is subject to flexing which
opens /closes the braid. So the  C per unit length along the cable
varies a small amount in an irregular way. You might try another cable
or flexing the cable. But does a few picoseconds matter?


Regards,

Alan G8LCO

Have you used port extensions to "remove the cable"{ from the measurement

The characteristics of coax vary with the lot, temperature, installation and
aging. I know that hams use coax for phased arrays and I have seen it used in
CB and Marine arrays. Are there any commercial applications that use coax for
matching or phasing these days? I would assume commercial installations would
want something adjustable and rated over tempetature.



Mike N2MS

> On 06/15/2023 10:07 AM EDT Alan
<[g8lco1@gmail.com](mailto:g8lco1@gmail.com)> wrote:

>

>
>

>
>

> The impedance of flexible cables is not 50.0000 ohms but around 47-53

>

> ohms. The centre conductor is almost never exactly central, the

>

> dielectric materials vary and the braid is subject to flexing which

>

> opens /closes the braid. So the C per unit length along the cable

>

> varies a small amount in an irregular way. You might try another cable

>

> or flexing the cable. But does a few picoseconds matter?

>

>
>

>
>

> Regards,

>

>
>

> Alan G8LCO

>

>

_._,_._,_

* * *

On 6/15/23 7:07 AM, Alan wrote:
> The impedance of flexible cables is not 50.0000 ohms but around 47-53
> ohms. The centre conductor is almost never exactly central, the
> dielectric materials vary and the braid is subject to flexing which
> opens /closes the braid. So the  C per unit length along the cable
> varies a small amount in an irregular way. You might try another cable
> or flexing the cable. But does a few picoseconds matter?
>
>
I'm assuming it's more of an "interesting, is this significant" kind of
question.

Actual variation in coax tends not to be that wide (in my experience)
and is more subtle. The classic one is "one turn around the spool" when
it's been stored with the spool on edge, so you have a sort of repeating
pattern of deformation.

And, of course, there's the classic "someone drove over/stepped on it"

But an interesting thing is "what if the cable is 52 ohms (a standard
value - RG-58A is 52 ohms), and the source and terminations are 50
ohms"... Gamma = 0.02, which is -34 dB.

On 6/15/23 8:55 AM, N2MS wrote:
> The characteristics of coax vary with the lot, temperature, installation
> and aging. I know that hams use coax for phased arrays and I have seen
> it used in CB and Marine arrays. Are there any commercial applications
> that use coax for matching or phasing these days? I would assume
> commercial installations would want something adjustable and rated over
> tempetature.

Pretty much every broadcast station uses a phased array of radiators,
whether it's a stack of dipoles or bowties for VHF, FM, etc.; or a
directional array for AM.

But aside from that (where the phasing tolerances can be pretty loose -
your Tx pattern doesn't change much from small phasing errors):

Phase matched assemblies are quite common in direction finding or
multichannel systems. There's the notorious bump for PTFE/Teflon
dielectrics around 25C, so if you want good phase matching, you need
good temperature control (or matching) at least.

I built an experimental system at work which measured the received phase
from 4 X-band (7.15 GHz) antennas several meters apart to control the
transmitted phase from 4 Ka-band (32 GHz) antennas. The idea was to be
able to compensate for pointing errors and do limited "retrodirective"
beam steering. When you have a the equivalent of a 6 meter diameter
antenna (0.1 degree beamwidth), you need to be able to point back to
earth very accurately. We were looking at tenths of a degree
variations. Just measuring and calibrating the system was non-trivial.

Multichannel test systems (both VNAs and things to test multi port
devices) use phase matched (and stable, repeatable) coax assemblies.
There's a reason those test port cables for a good VNA cost $5000 each.

So anything where you are measuring the phase between two signals -
monopulse feeds for radar, for instance - you want good phase behavior.

Near field scanners for antenna ranges rely on the precise measurement
of phase as received/transmitted at the probe, because the phase and
amplitude is transformed into the far field radiation pattern.

Any precision measurement is probably going to care about the cable -
probably more about stability in general, and repeatable behavior over
temperature. You can calibrate out the phase shift vs frequency, but if
it changes, then your cal is all fouled up

Am 15.06.2023 um 17:55 schrieb N2MS:
> The characteristics of coax vary with the lot, temperature,
> installation and aging. I know that hams use coax for phased arrays
> and I have seen it used in CB and Marine arrays. Are there any
> commercial applications that use coax for matching or phasing these
> days? I would assume commercial installations would want something
> adjustable and rated over tempetature.
> Mike N2MS
>> On 06/15/2023 10:07 AM EDT Alan <g8lco1@gmail.com> wrote:
>> The impedance of flexible cables is not 50.0000 ohms but around 47-53
>> ohms. The centre conductor is almost never exactly central, the
>> dielectric materials vary and the braid is subject to flexing which
>> opens /closes the braid. So the  C per unit length along the cable
>> varies a small amount in an irregular way. You might try another cable
>> or flexing the cable. But does a few picoseconds matter?
>> Regards,
>> Alan G8LCO
>

Mike,

when reading the posts I remind, that high priced and phase stable
coaxial cables are avaliable for high priced commercial VNA´s. The price
for such a cable can be in the 1000 US$ range!

Also cables with PFTE (teflon) do exibit in a specific temperature range
a high phase variation.  For instance look at the whitepapers available
at the German company elspec:

Whitepaper | el-spec GmbH (elspecgroup.de)
<https://www.elspecgroup.de/whitepaper>

 Regards

Henning Weddig

This is a very sensible practical answer. I actually have to measure and
verify 50-ohm coax is close enough to 50 ohms to be suitable for bench
measurements. I have found similar ranges...not because of how a cable
was stored (that seems to have no bearing unless it was overheated or
moisture contaminated). The problem seems more about manufacturing errors.

I have found some name brand 50-ohm nominal impedance cables to be as
high as 55 ohms and as low as the upper 40 ohm area. I've also found
variations along a cable's length.

A problem very notable is the connection from an aluminum or metalized
foil inner shield inner shield to the outer braid at connectors. Lenz's
Law dictates the differential mode current concentrates on the inner
wall of the inner shield. At connectors the RF current has to move to
the inner wall of the connector, which is usually connected outside the
outer braid.

I avoid foil-shielded cables in any critical application.  They can make
some spectacular bumps at VHF and upper UHF, as well as noise and IMD.

73 Tom

On 6/15/2023 10:07 AM, Alan wrote:
> The impedance of flexible cables is not 50.0000 ohms but around 47-53
> ohms. The centre conductor is almost never exactly central, the
> dielectric materials vary and the braid is subject to flexing which
> opens /closes the braid. So the  C per unit length along the cable
> varies a small amount in an irregular way. You might try another cable
> or flexing the cable. But does a few picoseconds matter?
>
>
> Regards,
>
> Alan G8LCO
>
>
>
>
>
>
>

--
This email has been checked for viruses by AVG antivirus software.
www.avg.com

What exactly are you all trying to match?