Common mode current at VHF/UHF and varying SWR with transmission line length

[Rack201]

Hi All

In testing some UHF & VHF/UHF dual band mobile antennas, I'm trying to understand why I see significantly different SWR / return loss figures based on transmission cable (coax) length, tested with a calibrated nano-VNA.

I assumed that changing SWR with cable length indicated the presence of common mode current on the outside of the coax, caused by an imperfect impedance match at the feedpoint of the antenna.  One of the antennas is marketed as a 'raised feed' antenna for 477Mhz - essentially a sleeved dipole, another is a colinear within a fiberglass radome (also 477Mhz) and the 2m/70cm dual band antenna is a typical wire phased element setup with a loading coil in the base.  They all exhibit the same behavior when measured, so I assumed they just were not perfectly matched to 50 ohms at the test frequencies and CMC occurred as a result.

To validate this, I modeled and built a simple 1/4 wave antenna with 4 radials at 45 degrees at the base cut/tuned up for 477Mhz.  The VNA returned an SWR of 1.0x:1 with an impedence of ~48 Ohms -jxx milli Ohms - all pretty good.  Then as an example. inserting an additional 3ft of coax into the feedline shifts the SWR to ~1.4:1 / ~70 Ohms-j8 Ohms.  Different cable lengths have differing effects.

What could I be seeing here?  The coax lengths test fine for return loss.  It's possible the connectors I'm using are junk and not at 50 ohms?

I also note that for the 477Mhz retail antennas, they are typically sold with a pre-terminated length of coax (and manufacturers advise not to cut them).  Maybe this suggests that they are not exactly matched to 50 ohms and the included length ensures the radio is connected at a point where it will see SWR around 1.5:1 as they advertise.

Any thoughts appreciated!

[profdc9]

A coax cable has two sets of modes.  The internal TEM guided mode, which is the desired mode, which is conducted between the outside surface of the center conductor and the inner surface of shield  The internal TEM mode has equal and opposite currents on the conductors.  The second mode is a surface wave mode (Sommerfeld or Goubau mode) that is loosely bound to the outside surface of the shield, and is usually bound there due to the dielectric layer on the outside.

When an antenna is attached to the end of a coaxial cable, the currents at the inside and outside surfaces of the shield are joined together.  Kirchoff's current law, where the sum of all currents into a closed surface must be zero, tells us what happens.  If there were only two currents into a node  (the inner conductor and the inner surface of the shield) these must be equal and opposite.  These must always be equal and opposite because the coax shield prevents any net electric charge inside the shield.  If the outer conductor carries current, it is due to an imbalance between the opposite currents on the inner conductor and inner surface of the outer conductor.

So how does one get imbalance?  If a cable is attached to the coax, the antenna can capacitively couple to the outer surface of the shield.  This capacitive current path unbalances the two currents at the inner conductor and the inside surface of the shield.  If there is a high voltage node at the end of an antenna (like a short dipole), even a small capacitance (on the order of pF) can conduct a high current.  For example, the end of a radial is a high voltage node which becomes transformed to a low impedance through a 1/4 wave propagation.

You can build an RF ammeter using a small ferrite toroid, a coil of wire, a Schottky diode (like a 1N5711) and an analog meter, and that will let you see if there's a standing wave on the cable.  You can use a small ferrite toroid as a choke to try and eliminate this current path and see if the sensitivity to the cable length is reduced.

[radiolistener]

In testing some UHF & VHF/UHF dual band mobile antennas, I'm trying to understand why I see significantly different SWR / return loss figures based on transmission cable (coax) length, tested with a calibrated nano-VNA.

This is an indication, that your antenna is bad matched with coax cable.
As result your coax cable works a part of antenna and this is why it's length affects VSWR.

You're needs proper match between coax cable and antenna with balun transformer to isolate your cable from antenna and use RF chokes on the cable to reject induced interferences on the cable. With proper matching and RF chokes it helps you to significantly reduce noise interference from nearby electronic equipment and VSWR will not depends on cable length.

If you're using balun transformer tuned for 50Ω coax, with RF chokes and still had VSWR dependency from cable length, most of all your coax cable is not 50Ω, so you're needs tune matching circuit installed between antenna and coax cable.

It's possible the connectors I'm using are junk and not at 50 ohms?

Yes, bad connectors can cause some reflection, once I had 1 meter cable with UHF connectors which leads to VSWR about 1.3 at 28 MHz... These UHF connectors are crap. If your power is within 100W, I recommend to use SMA. For higher power use N connectors.

[Rack201]

Thank you for the replies.

If what I'm seeing is correct, I might have some poor quality coax.  The ~3ft section I was adding in seems to return almost a ~1.5:1 SWR reading on its own, when terminated with a 50ohm resistive load.  And/or it could be the selection of adapters I'm using.  Don't know why I didn't think to test it on its own first.

[A.Z.]

for VHF/UHF add a Maxwell (W2DU) type choke on the coax placing it near the antenna end of the coax, that should solve/relieve the issue



in your case you may just place the beads directly over the coax feeder

[radiolistener]

it can be due to different cable impedance, or due to bad cable quality (for example braid made from cheap/thin conductor). In the first case you can re-tune matching circuit between antenna and coax to match impedance of your coax. In the second case there is no solution, just replace cable with a better quality one.

VSWR=1.5 looks like your cable is 75Ω. These 75Ω cables works better on VHF than Chinese 50Ω RG58, they have less loss, but it needs to use a little different matching between cable and antenna.

[Solder_Junkie]

It is easy to make a common mode current meter. I built this one and while it is a little insensitive on 2m and 70cm, it works well.

Details: http://www.ifwtech.co.uk/g3sek/clamp-on/clamp-on.htm

[Rack201]

VSWR=1.5 looks like your cable is 75Ω. These 75Ω cables works better on VHF than Chinese 50Ω RG58, they have less loss, but it needs to use a little different matching between cable and antenna.

The cable is Times Microwave LMR-240 Ultraflex 50 Ohm.  I'm suspect on the e-bay UHF barrel connectors I used to join it to the antenna section...

[Rack201]

for VHF/UHF add a Maxwell (W2DU) type choke on the coax placing it near the antenna end of the coax, that should solve/relieve the issue

Thanks for that - I've been looking for some mix 61 snap-on ferrites to try - don't seem readily available here in aus.  Will put an order in.

[A.Z.]

for VHF/UHF add a Maxwell (W2DU) type choke on the coax placing it near the antenna end of the coax, that should solve/relieve the issue

Thanks for that - I've been looking for some mix 61 snap-on ferrites to try - don't seem readily available here in aus.  Will put an order in.

you're welcome, and in case you're curious about the birth of the W2DU, have a look here

https://www.nonstopsystems.com/radio/pdf-ant/antennas-article-W2DU-origin.pdf

[A.Z.]

as for placing the choke, an RF ammeter would be useful, slide the "clamp" along the coax until you find a current minimum, then place the ferrites so that the center of your W2DU will be near/over that spot

[CaptainBucko]

I am reading this with interest, because I built a coaxial co-linear and had exactly the same problem.

[A.Z.]

I am reading this with interest, because I built a coaxial co-linear and had exactly the same problem.

A "flowerpot", by chance ?

https://vk2zoi.com/articles/double-five-eighth-flower-pot/

if so it's strange, the bottom air wound choke usually suffices

[mtwieg]

Bad cables/connectors/adaptors could definitely explain why SWR changes with cable length. But even if that's the reason, that would not lead to CM currents, so adding ferrites/chokes would not help. profdc9 already gave a good explanation for why the common mode gets excited, it has everything to do with the antenna itself and its nonconservative electric fields (i.e. parasitic capacitance). So I would make sure I'm chasing after the right problem first. If touching the cable shield at various points causes measured SWR to change significantly, then you have CM behavior in play.

[A.Z.]

so, an OCF dipole should be used w/o a good choke ? Come on. And then, the same old wives legend about "touching the cable"; use an RF ammeter, not some folklore, measure ALONG the line, and then make some educated guess.

And, since we are at it

https://www.eznec.com/Amateur/Articles/Baluns.pdf

[mtwieg]

so, an OCF dipole should be used w/o a good choke ? Come on.

Never made such a claim.

And then, the same old wives legend about "touching the cable"; use an RF ammeter, not some folklore, measure ALONG the line, and then make some educated guess.

Obviously measuring the common mode excitation with a voltage/current probe is ideal. I only suggest the "touch" method to get a rough idea of whether the common mode excitation exists. Especially from my experience with high-channel account phased arrays, where the touch test takes a minute as opposed to adding and then removing current transformers for each channel which would take an hour...

[A.Z.]

the touch method is unreliable, would you go to a shaman or to a doctor ?

[A.Z.]

It is easy to make a common mode current meter. I built this one and while it is a little insensitive on 2m and 70cm, it works well.

Details: http://www.ifwtech.co.uk/g3sek/clamp-on/clamp-on.htm

good suggestion, and NO, the issue has little to nothing to do with impedance mismatch or losses, it's due to common mode currents due to an inbalance between the transmission line (coax) and the load (antenna) which makes the outer surface of the braid an active part of the antenna, hence, changing the TL (coax) lenght affects the antenna match and radiation

Note: a badly installed or off-specs connector may cause CMC, but that will be the cause, the effect will be insurgence of CMC and those need to be cured by both checking the connectors installation and using choke(s)

[CaptainBucko]

A "flowerpot", by chance ?

No, I don't think so - it was a long time ago, I remember calculating the coaxial co-linear lengths and cutting them with accuracy, but I found it only provided a good return loss when connected with specific lengths of coaxial cable feeder. So I am pretty sure this is exactly the same problem I was having.

[A.Z.]

A "flowerpot", by chance ?

No, I don't think so - it was a long time ago, I remember calculating the coaxial co-linear lengths and cutting them with accuracy, but I found it only provided a good return loss when connected with specific lengths of coaxial cable feeder. So I am pretty sure this is exactly the same problem I was having.

I see, it was probably something like this one

https://www.rmham.org/wp-content/uploads/2020/04/k2ad-r.pdf#page=14

now, that antenna needs a choke (a W2DU will do) to "separate" the radiator from the feedline, in theory the choke won't be needed, in practice, small size differences may induce CMC along the feedline and cause problems, so, since a W2DU at V/UHF isn't as big and heavy as one used at lower frequencies (e.g. HF) and introduces almost no loss, I think it would always be a good idea using it, at least it won't harm

[Geoff-AU]

The cable is Times Microwave LMR-240 Ultraflex 50 Ohm. 

Nice.

I'm suspect on the e-bay UHF barrel connectors I used to join it to the antenna section...

Ew yuck.  1) ebay  2) UHF connector.

UHF connectors are poorly named, they are far from ideal above VHF.  The impedance is no longer 50 ohms.

I am reading this with interest, because I built a coaxial co-linear and had exactly the same problem.

They are very difficult to get right, usually if you can't verify the radiation pattern you're better off sticking with simpler antennas.  Also be aware that a lot of the benefit might be coming from having the top of your antenna elevated higher, not the "extra gain", as this page describes about 1/4 wave, 1/2 wave, and 5/8 wave if the tips are set at the same height:

https://practicalantennas.com/designs/verticals/5eights/

A "flowerpot", by chance ?

https://vk2zoi.com/articles/double-five-eighth-flower-pot/

if so it's strange, the bottom air wound choke usually suffices

Air wound choke is an important part of the design, ferrite choke won't have a high enough Q to choke it at that location

[A.Z.]

I'm suspect on the e-bay UHF barrel connectors I used to join it to the antenna section...

Ew yuck.  1) ebay  2) UHF connector.

UHF connectors are poorly named, they are far from ideal above VHF.  The impedance is no longer 50 ohms.

Aw, I missed that, the so-called "UHF connectors" are usually good up to around 50MHz, better going for BNC connector and, possibly, some decent ones

A "flowerpot", by chance ?

https://vk2zoi.com/articles/double-five-eighth-flower-pot/

if so it's strange, the bottom air wound choke usually suffices

Air wound choke is an important part of the design, ferrite choke won't have a high enough Q to choke it at that location

The bottom, air wound choke is usually ok, but in some cases it won't suffice, so placing a W2DU choke about 1/4 wave below the airwound will usually cure any issues, not just that... in my experience the coax below the "flowerpot" carries some RF and has a bit of impact on the radiation pattern, by placing the second choke as described will have the side effect of lowering a bit the takeoff angle which isn't bad imHo

note: the above was tested with the "regular" 1/2 wave flowerpot antenna

https://vk2zoi.com/articles/half-wave-flower-pot/

not with the other variants

[Geoff-AU]

placing a W2DU choke about 1/4 wave below the airwound

That's a good idea 

[T3sl4co1l]

This is an indication, that your antenna is bad matched with coax cable.
As result your coax cable works a part of antenna and this is why it's length affects VSWR.

You're needs proper match between coax cable and antenna with balun transformer to isolate your cable from antenna and use RF chokes on the cable to reject induced interferences on the cable. With proper matching and RF chokes it helps you to significantly reduce noise interference from nearby electronic equipment and VSWR will not depends on cable length.

No, that is not correct -- it is perfectly normal for mismatch to occur within a cable, and have no radiating or CM consequence.  As profdc9 noted, outer shield current is only due to imbalance: asymmetry in the antenna itself, proximity to nearby structures, a radiation pattern that includes (or gets reflected back onto) the feedline, etc.

Note that mismatch should be evident on the VNA: impedance will lie somewhere on a circle centered on the origin, the angle varying with feedline length.  This is on top of all the circles (across broad frequency ranges) due to the antenna's resonance and whatever, so keep in mind how this needs to be measured: plot a point at resonance, change the cable, plot another point at resonance, etc.

Further counter-example: consider driving a 100Ω twisted pair with one terminal grounded.  The common mode voltage is Vin/2 by definition.  This launches an outer-shield or CM current into whatever the surrounding impedance is (I = (Vin/2) / Z)), where "outer shield" in this case refers directly to the balance of currents in the wires, but applies equally well in the case of coax (where the one-terminal-grounded configuration just happens to correspond to the low-CM case).  The far end of that twisted pair can be perfectly terminated with a 100Ω resistor, SWR = 1 anywhere on the line, and still the line acts as an antenna, because you've coupled 6dB into the common mode at the feedpoint.  You might (rightfully) complain that this is a contrived and erroneous situation, because obviously you use a balun on a twisted pair TL -- but that's just the thing, your criteria didn't require a balun, and a properly balanced antenna feedpoint won't do that anyway.

So, just to clear that up, as no one else seems to have dug into this particular point so far.

As for mitigation, note that series chokes can only reduce, never eliminate it.  Much higher attenuation can be had by pairing the series choke with some shunting impedance: tie feedline shield to ground every once in a while, to make an equivalent ground circuit of a ladder divider, with the series spans being either blank line (still has some CM impedance) or chokes (whether coiled up or ferrite loaded, the ferrite of course introduces loss that dampens the otherwise resulting resonant modes), and the shunt elements being ground clips (cut off the outer jacket, add a spring clamp), inline bulkhead connectors (gets a ground connection, maintains environmental protection), etc., to whatever nearby large metal structure (poles, beams?), tower, (metal) wall, or stake into earth, is available.  What counts as "ground" should be as wide as possible: a tower is better than a pole; a grounding grid attached to a tower base is better than single stakes into the ground; etc.

Grounding the feedline has the same effect as ground stitching vias on a two-layer PCB, keeping a good ground around a microstrip trace -- or even more particular, suppose it was not just a microstrip trace but one with a signal conductor on top.

You can never eliminate the support/feedline/tower from the antenna, of course, whether in terms of reflection off the nearby ground plane (earth itself, or a large structure the antenna's installed on), or that the width of the line/pole/tower has some (even if slight) effect on radiation pattern, balance, etc.  But these general guidelines can make it good enough not to matter, and for the rest, standardizing on a support dimension (width and height, and assuming a good ground beyond it) at least fixes the remaining variables.

Tim

[iMo]

the touch method is unreliable, would you go to a shaman or to a doctor ?

People who spent some time with practical tuning the antennas and measurements in the field know this method (or better "a quick check") is pretty reliable provided you are looking at a nanovna connected to the coax, watching the SWR graph on the display and grabbing the coax with your hand (and I've been using it as well).

For VHF/UHF a single ferrite bead at the antenna feedpoint usually helps. While messing with the antenna (especially the VHF/UHF one) you (or any objects) have to be at least one-two lambdas off the antenna, otherwise you will couple the antenna via your body to the ground (thus you imbalance the antenna).

Anyhow, none VHF/UHF antenna will be perfectly balanced, none coax perfectly 50ohm, no connector perfectly 50ohm - therefore you will always see some SWR and common mode mess.

Therefore I advice people to have the final length of the coax cut, equipped with connectors, etc. before they start with the final tuning of the antenna's elements. The changes in the coax length after the tuning the antenna usually lead to changes in the SWR (because the coax repeats the antenna's feedpoint impedance in a "specific pattern" with the coax lengths which are not exactly even multiplies of lambda/4 for the specific cable).