开云体育

I'm wondering if the correct way to measure the length of window line
with a NanoVNA is fundamentally different than that of measuring coax.

I'm following the method described by Arie
(/g/nanovna-users/message/26832). I have 5.175m of "JSC
1320 300 Ohm Ladder Line 300 Ohm 20 AWG / 7 Strands Bare Copper". That's
a quarter wavelength of 20.7m, or a frequency of 14.5MHz.

The first step is to sweep it to determine the velocity factor. Yet,
when I sweep from 12-17MHz, I get the Smith chart attached. There's no
point when the impedance is close to zero.

Am I doing something wrong (most probable), or is measuring the length
of window line with an NanoVNA fundamentally different than measuring coax?

Thanks for your advice and suggestions.

-Kevin
KC3KZ

I've measured windowline many times using the NANO - something I can not
easily accomplish with the HP 8753C. Be sure the NANO is not connected to
any conducting cabling, chargers, USB cables, or sitting on or near
anything conducting. At HF frequencies the NANO is small enough to not
influence measurements. However, if it is connected to anything
conducting, it is no longer "small" as a function of frequency /
wavelength.

The attachment indicates you have connected the NANO to a PC. The NANO
effectively becomes much larger at RF frequencies as its connected to: 1)
the USB cable, 2) the PC, and 3) the house wiring, and finally, 4) the
power grid. This will greatly influence correct readings of windowline.
Measurements must be made with the NANO alone.

Dave - W?LEV



On Tue, Feb 15, 2022 at 4:40 PM Kevin Zembower via groups.io <kevin=
[email protected]> wrote:

I'm wondering if the correct way to measure the length of window line
with a NanoVNA is fundamentally different than that of measuring coax.

I'm following the method described by Arie
(/g/nanovna-users/message/26832). I have 5.175m of "JSC
1320 300 Ohm Ladder Line 300 Ohm 20 AWG / 7 Strands Bare Copper". That's
a quarter wavelength of 20.7m, or a frequency of 14.5MHz.

The first step is to sweep it to determine the velocity factor. Yet,
when I sweep from 12-17MHz, I get the Smith chart attached. There's no
point when the impedance is close to zero.

Am I doing something wrong (most probable), or is measuring the length
of window line with an NanoVNA fundamentally different than measuring coax?

Thanks for your advice and suggestions.

-Kevin
KC3KZ

--
*Dave - W?LEV*
*Just Let Darwin Work*

Has anyone wound a high common mode impedance 1:1 balun with 50 ohm coax and calibrate at the output end of the balun?

I made a balun years ago for use with GR Z bridges (GR 1606 and 916) for this purpose with good results.

I didn't wind the coax on a ferrite toroid at the time. I wound the coax on a length of PVC pipe and with a variable cap across the coax (shield to shield) I tuned it to resonance for the frequency of the test. The common mode Z is very high and higher than a ferrite-cored balun.

Obviously this would not work for a swept frequency test but should work with a broadband ferrite-cored balun.

I'm wondering if the correct way to measure the length of window line
with a NanoVNA is fundamentally different than that of measuring coax.

I'm following the method described by Arie
(/g/nanovna-users/message/26832). I have 5.175m of "JSC
1320 300 Ohm Ladder Line 300 Ohm 20 AWG / 7 Strands Bare Copper". That's
a quarter wavelength of 20.7m, or a frequency of 14.5MHz.


First things first, a FULL ELECTRICAL wavelength at 14.5MHz is 300/14.5 ~= 20.7m. A quarter wave ELECTRICAL length is 5.172m.
To find the PHYSICAL length for a quarter wave stub at 14.5MHz, MULTIPLY the ELECTRICAL length by the cable VF.
The Spec Sheet for 1320 TV Ladderline is 0.82. Therefore the PHYSICAL length required is 5.172*0.82 = 4.241m.

If you now want to prove this, you can use the NanoVNA, BUT it is best if you use a broadband matching transformer and measure the 300 Ohm ladderline in balanced mode.
If you try measuring the ladderline directly on the 50 Ohm unbalanced S11 input of the NanaVNA, you will get unreliable results as you have already discovered. Possible reasons are mentioned in other posts in this thread.

To match 50 Ohm to 300 Ohm, the impedance ratio is 300/50 = 6. The turns ratio for the transformer is the square root of the impedence ratio, sqrt(6) = 2.45.
So let us use 8 turns on the 50 Ohm Primary and (8*2.45) = 19.5 turns on the 300 Ohm Secondary.
The half turn is convenient as we end up with the primary at one end and the secondary at the other end of the ferrite core.

To make the matching transformer, use a binocular "balun" ferrite core. I have used a 14mm (9/16") long, 13.2mm wide ferrite with two 3.8mm (5/32") holes.
It is best if the primary and secondary windings are bifilar wound.
It is even better if the smaller 8 turn primary is wound in the centre of the 19.5 turn secondary.

As the power levels used here are quite low, the thickness of the wire is not critical. The main thing is that you can fit 28 "turns" of wire down each hole.
You will need about 1200-1500mm (4-5ft) of fine enamelled copper wire for a 14mm (9/16") binocular balun core.
Take your wire, fold it over, mark one leg at each end so you can identify it later and twist the legs together lightly. About 1 twist per cm, 2 twists per inch is fine.

It is difficult to describe what to do next, but what we want is to start with 5.5 turns of the secondary wire only, then 8 turns of the double bifilar wire, and finally 5.5 turns of the secondary wire only to finish up.
Unwind about 250-300mm (10"-12") of the unmarked leg so you can pass 5 full turns of the marked leg only through both holes of the ferrite and then pass it through the hole you started at to give 5.5 turns.
Now twist the unmarked leg around the marked leg to start your bifilar combined winding.
Make 8 complete turns with the twin bifilar wire and finish at the same end you started the 8 bifilar turns.
Separate the two legs and wind the marked leg ONLY through the ferrite another 5.5 times. You should finish at the same end you started with the single wire.

The "single" wire end is your 300 Ohm secondary, the bifilar 8 turn end is your 50 Ohm primary.

Mount the finished transformer on a piece of pcb or perf board. Attach a female SMA connector to the 8 turn winding and some terminals at the 300 Ohm end to attach your ladderline.
I am not being too prescriptive here. Use your imagination and your ingenuity to make a neat, relatively robust piece of test adaptor for you VNA test kit.

FINALLY, connect your SMA test lead to the transformer SMA connector and terminate the 300 Ohm terminals with a NON-INDUCTIVE 300 Ohm SMT or metal film resistor/s.
Now do a Short/Open/Load Calibration over the frequency range of interest and save to one of your Calibration memories.

You can now run your test on your ladderline. It should be relatively accurate as the Calibration Plane is at the 300 Ohm Terminals of the Transformer.

I used a similar transformer design when I was a design engineer working on early VDSL Broadband Network deployment. This used multiple carriers from 2MHz-35MHz on 100 Ohm Cat 5 cable.
I used a 10 turn primary and 14 turn secondary for the 50/100 transformer. The system was very sensitive to cable balance and the transformer worked well.

73...Bob VK2ZRE

Bob, thank you so much for your very detailed answer. It'll take me awhile to get the ferrite and build the test jig, but I'll let the group know of my results. Thank you, again, for so generously answering my question.

-Kevin
KC3KZ

Agree with Kevin. Very good explanation.

John, wa3jrs

We're measuring impedance here. No matching transformer is needed. Just a 1:1 high common mode Z balun. From the Z's measured you determine the characteristics of the balanced transmission line OR and that includes its electrical length. Let's not over complicate things here.

On 2/16/22 11:28 PM, Bob Ecclestone VK2ZRE wrote:

I'm wondering if the correct way to measure the length of window line
with a NanoVNA is fundamentally different than that of measuring coax.

I'm following the method described by Arie
(/g/nanovna-users/message/26832). I have 5.175m of "JSC
1320 300 Ohm Ladder Line 300 Ohm 20 AWG / 7 Strands Bare Copper". That's
a quarter wavelength of 20.7m, or a frequency of 14.5MHz.


First things first, a FULL ELECTRICAL wavelength at 14.5MHz is 300/14.5 ~= 20.7m. A quarter wave ELECTRICAL length is 5.172m.
To find the PHYSICAL length for a quarter wave stub at 14.5MHz, MULTIPLY the ELECTRICAL length by the cable VF.
The Spec Sheet for 1320 TV Ladderline is 0.82. Therefore the PHYSICAL length required is 5.172*0.82 = 4.241m.

If you now want to prove this, you can use the NanoVNA, BUT it is best if you use a broadband matching transformer and measure the 300 Ohm ladderline in balanced mode.
If you try measuring the ladderline directly on the 50 Ohm unbalanced S11 input of the NanaVNA, you will get unreliable results as you have already discovered. Possible reasons are mentioned in other posts in this thread.

It is unclear why you'd get unreliable results for measuring *electrical length* without a balun - given that the NanoVNA is tiny, and isolated from other things. That is, the "unbalanced-ness" of the VNA is irrelevant when looking at a piece of balanced transmission line, unless there's some weird layout that has the transmission line laying next to the VNA.

The 300 ohms (or 100-120 ohms looking at twisted pair) means that one can't directly use the S11 or R+jX values, but the *distance* measurements should be just fine. The impedance of a balanced line with the other end open goes through the same cycle (at the same frequencies) of apparent short (at 1/4 wavelength), to apparent open (at 1/2 wavelength), to apparent short (at 3/4 wavelength), etc.

What wouldn't be necessarily valid is loss measurements (which is a handy feature of a VNA measurement, not only do you wind up with the length of the cable, but you get the loss at various frequencies)

On 2/17/22 5:52 AM, WB2UAQ wrote:

We're measuring impedance here. No matching transformer is needed. Just a 1:1 high common mode Z balun. From the Z's measured you determine the characteristics of the balanced transmission line OR and that includes its electrical length. Let's not over complicate things here.

Considering that the NanoVNA is smaller than a lot of HF baluns for balanced lines, I don't think you need the balun, either (until you get up to where the NanoVNA is a significant fraction of a wavelength - maybe don't use it for UHF TV twinlead.

Just make sure, as others have pointed out, that you don’t have any other wires (power supply, USB to computer,, etc) connected to the nanoVNA.

73, Willie N1JBJ

On 2/17/22 7:52 AM, William Smith wrote:

Just make sure, as others have pointed out, that you don’t have any other wires (power supply, USB to computer,, etc) connected to the nanoVNA.

I'm not sure a USB cable would make a difference if the feedline is properly "isolated" from the environment. (i.e. the USB cable isn't coiled up with the transmission line).
I can't think of a physical mechanism for that to cause a problem.

One could easily test it with the NanoVNA by itself, and then with the cables hooked up. If there no difference, then knock yourself out using a PC app to control the VNA.

73, Willie N1JBJ

On Feb 17, 2022, at 10:10 AM, Jim Lux <jim@...> wrote:

On 2/17/22 5:52 AM, WB2UAQ wrote:

We're measuring impedance here. No matching transformer is needed. Just a 1:1 high common mode Z balun. From the Z's measured you determine the characteristics of the balanced transmission line OR and that includes its electrical length. Let's not over complicate things here.

Considering that the NanoVNA is smaller than a lot of HF baluns for balanced lines, I don't think you need the balun, either (until you get up to where the NanoVNA is a significant fraction of a wavelength - maybe don't use it for UHF TV twinlead.

As happens so often in these threads, we need to go back to the original post and read the question.

Kevin was puzzled as to why his S11 results did not show a distinct "short" when he tested his 300 Ohm ladderline connected directly to his NanoVNA unbalanced 50 Ohm S11 port.
He posted the sweep results in that original post..
Some folks offered possible reasons, others possible work arounds. But no one explained to him why he got the results he did. No one ANSWERED his original question.

I am not going to dispute any of the subsequent comments made regarding direct connections, common mode chokes or anything else posted here.

I offered one possible solution using a broadband impedance matching transformer because it is quite possible that Kevin is actually building an antenna being fed with 300 Ohm ladderline and that by building the transformer, he could verify his stub and then go on to test the completed antenna. As the system is now impedance matched back to 50 Ohms, his S11 results on the complete antenna system would now better reflect (pun intended) the actual system performance.

If folk are not comfortable winding their own transformers, there are bare transformers and packaged units readily available on the web.
If you are comfortable winding your own, it is almost as easy to wind a 300/50 matching transformer as it is a CM choke.
The difference is, you now have a piece of test kit specific to the task at hand.
The task in this case centres around 300 Ohm ladderline.

I do not know why Kevin got the results he did originally, but I am confident that if he uses a matching transformer, the results he gets next time should be closer to what he expects.

And as a final comment, the ladderline stub should be tested and trimmed well clear of any thing as even moisture on a non conducting pole will be enough to add distributed capacitance across the line and affect the apparent Velocity Factor. Metal, including other cables, in close proximity is a definite no no.

Which brings me to offer a possible REASON for Kevin's results.
He states he had 5.175m of "JSC 1320" cable. A quick check of the specs for 1320 ladderline gives a VF of 0.82
Therefore the ELECTRICAL wavelength of the line is 4(5.175/0.82) = 25.24m. This gives a frequency of 300/25.24 = 11.89MHz
Since Kevin only started his sweep at 12MHz, he did not sweep the actual stub frequency of 11.89MHz.
If the ladderline was lying on the ground or close to any metal, the "apparent" VF would be even lower and the electrical length and thus quarterwave frequency would be even lower again.
Looking at the Smith Chart, we can see it is approaching a zero Z but does not quite make it. Why? Because he should have started the sweep at a lower frequency.
That's my theory anyway.
So it pays to go back and read the entire original post.

73...Bob VK2ZRE

Great post Bob, patient, thoughtful, and to the POINT. i.e. people do your homework before jumping into the 'fray'. There are lots of 'us' guys out there willing to help, but it is your responsibility to do your 'own' homework FIRST. That advice being, READ the original question and then go from there.

Mike C.

Do I understand this correctly from this discussion? For every Zo a balun or transformer is required for that specific Zo? One for 100 ohm cat 5, one for 300 ohm TV twin lead, another for 600 ohm ladder line and another for 450 ohm window line, etc.?

On 2/20/22 6:43 AM, WB2UAQ wrote:

Do I understand this correctly from this discussion? For every Zo a balun or transformer is required for that specific Zo? One for 100 ohm cat 5, one for 300 ohm TV twin lead, another for 600 ohm ladder line and another for 450 ohm window line, etc.?

Not really - there's two aspects to this:

1) Different characteristic impedance - You could calibrate with 300 ohm load, short, and open and directly measure 300 ohm lines (or 75 ohm lines, etc.)? The S parameters will be correct, but anything which turns that into Z won't necessarily be right (because the formula needs the calibration impedance.
- Or, you can mathematically transform measurements made assuming 50 ohms into measurements using some other impedance. The NanoVNA won't do this, but other software can.

2) Balanced to unbalanced transformation - This is more about "how accurate do you want to be?" The NanoVNA is a unbalanced system, but it's also tiny. It's also (in usual situations) not "connected to ground" so the unbalance shouldn't affect much (i.e. it's a two terminal device, the balanced transmission line is a two terminal device, there's no way to "know" that it's not balanced)

On Sun, Feb 20, 2022 at 07:01 AM, Jim Lux wrote:

- Or, you can mathematically transform measurements made assuming 50
ohms into measurements using some other impedance. The NanoVNA won't do
this, but other software can.

Jim - DiSlord has added the capability to change the system impedance for measurements in his recent firmware. It is under the Display menu..

Roger

On 2/20/22 10:20 AM, Roger Need via groups.io wrote:

On Sun, Feb 20, 2022 at 07:01 AM, Jim Lux wrote:

- Or, you can mathematically transform measurements made assuming 50
ohms into measurements using some other impedance. The NanoVNA won't do
this, but other software can.

Jim - DiSlord has added the capability to change the system impedance for measurements in his recent firmware. It is under the Display menu..

Roger

Then hook that twinlead up directly and go for it. That's pretty cool.

I thought I had some 300 ohm twinlead in the garage, and I was going to do some tests with transformer and not, etc.? But I didn't, so it will have to wait for the next trip the to store.

Good luck finding 300 ohm twin lead in this day and age. When I go to Lowe's and Home Depot or other stores that sold TV hardware, I don't see it any longer. There's plenty of RG-6 and OTA TV antenna hardware.

To measure the characteristics of balanced line I first find the freq where the line under test is a quarter wave long. I do this by connecting the line under test, with its far end shorted, across a 50 ohm load at the output of the balun (the terminals that were calibrated). I then sweep from the lowest frequency looking for the freq where the return loss is the highest. The return loss is often higher than 20 dB at this freq. The return loss is the highest at this freq because the shorted transmission line under test looks like a high impedance (parallel resonance) leaving the 50 ohm load hardly disturbed and having a pretty high RL. Next I remove the 50 ohm load from the test terminals and attach it to the far end of the transmission line under test. The impedance measured (Zin) by the Nano at the freq where the line is a quarter wave long should be resistive. Using the old Q section formula where Zo = (Zin x Zload)^0.5. Zload is the resistance placed at the far end of the transmission line under test.

For 300 ohm line, with Zload being a 50 ohm resistance, Zin should be a resistance of about 1800 ohms. Check: 300 = (50 x 1800)^0.5. Note that everything is a resistance which makes the math pretty easy not requiring complex algebra. ( on the Smith chart the resistances are 9 o'clock and 3 o'clock centered on Zo)

I have JSC twin lead myself. It is not the part number mentioned above. I think it is JSC 1018, spec'd to be 300 ohms. It looks like TV twin lead but it has 18 gauge stranded conductors. Years ago (like 15 to 18 yrs), using the above Q section formula the Z is not the nominal 300 ohms and came in at about 230 ohms. I wondered why this could be so I asked the mfr and never received a reply. I believe the usual TV 300 ohm line uses 22 gauge wire instead of 18 so the JSC 1018, made with the 18 didn't get its spacing adjusted wider to get Zo to be 300 ohms.

On Sun, Feb 20, 2022 at 02:11 PM, WB2UAQ wrote:

Good luck finding 300 ohm twin lead in this day and age.

I needed some for a project about a year ago, and wound up buying a roll of "vintage" Radio Shack twin lead on Ebay. Wasn't expecting that.

On Sun, Feb 20, 2022 at 5:12 PM WB2UAQ <pschuch@...> wrote:

Good luck finding 300 ohm twin lead in this day and age.

Wireman has 300 ohm twin lead. I bought some about a year ago.