开云体育

I am about to measure a 1/4 wave of 450 ohm windowed twinlead for the 2m band using my NanoVNA. My question is, since I will be making an unbalanced to balanced connection, should I use a common mode choke, balun or add ferrites to the coax side to make the connection, or does it really matter at 2m frequencies? The coax lead from my VNA to the twinlead will be about 6" to 12" long. I will probably terminate the coax in two short wires to connect to the twinlead.

You should be able to measure transmission line as if it was inductor at electrical 1/8 wavelength to get Zo.

Short out far end and other end terminals looks like an near infinite impedance at 1/4 lamda. You need to sweep freq to find true electrical 1/4 lamda. Cut in half for electrical 1/8 lamda and should get +j450 impedance.

If there is plastic type fill with cut windows, the cut windows are to get the characteristic impedance up. It also increases Vp because net average dielectric is closer to air. Velocity factor is likely around 0.85 range so use that as a sanity check on 1/4 lamda measurement.

450 ohms is starting to get up there on resolution accuracy on 50 ohm VNA. Just plot a j450 ohm point on a 50 ohm ref. Smith chart to see how much a little error in Rho length means to 450 ohm inductance reading accuracy.

Keep it away from environmental items, particular metals and high dielectric surface. Hanging in free space with couple of feet from anything should suffice. Any choke or balun will iikely just add more errors to measurements. Look for freq at true real impedance horizonal line and it actual physical length gives velocity factor, again likely in 0.85 range.

You also might try TDR mode but Nano likely not too accurate for Zo measurement at that high iimpedance.

In my last fw i add measure Cable Option and Port Z normalization

You can try use this feature for measure cable Z0 and lenght (use Marker->Measure->Cable S11), and see how it look if measue on 450 Om port (use Port Z: 50 -> 450 in DISPLAY menu)

On Wed, Jul 28, 2021 at 09:55 PM, <kk7xo@...> wrote:

I am about to measure a 1/4 wave of 450 ohm windowed twinlead for the 2m band
using my NanoVNA. My question is, since I will be making an unbalanced to
balanced connection, should I use a common mode choke, balun or add ferrites
to the coax side to make the connection, or does it really matter at 2m
frequencies? The coax lead from my VNA to the twinlead will be about 6" to
12" long. I will probably terminate the coax in two short wires to connect to
the twinlead.

Are you trying to verify the impedance of the 450 ohm twinlead or do you need a method of measuring how long it needs to be for 1/4 wavelength at 2M (144 MHz.)?

Roger

I'm not exactly sure what you want to measure. I assume you'd like to measure the velocity factor and the characteristic impedance.

Might I suggest the following:
Build a test fixture to connect the two leads of your 450 ohm line to BOTH ports of the nanoVNA. This could be just a couple of coax cables whose shields are tied together and the conductors are 1" (or so) apart.
Calibrate this arrangement.
Connect your piece of 450 ohm line to the center conductors.
Connect a 200 ohm (or something about that) load to the other end.
Measure the resulting impedance across a range of frequencies well below to well above the desired operating point (See below for 'well below' and 'well above').
Compute the resulting impedance. (There are two ways to do that, one using S11 and one using S21. I would recommend the S21 approach)
From that sweep you can determine:
1) The length of the line: The frequency at which the impedance is 200 ohms tells you the frequency at which the line is 1/2 wave and therefore the electrical length of the line.
2) The impedance of the line: At twice that frequency the line is 1/4 wave. The 200 ohms will be converted to something like 1k. The characteristic impedance is just Sqrt(Zmea * 200). (Again, this would be better measured using the S21 because 1k is a bit high.)

If all you are interested in is the correct length for 1/4 wave, you can make life a bit easier... no math necessary. Leave the line 'open' and simply measure S21 and S11. The frequency at which the line is 1/4 wave will cause the 1/4 wave to be a 'short'. At that frequency, S11 will be '0' and S21 will be '1'.

If someone knows of any pitfalls to this approach, please let us know!!!

Interesting questions. How about the same scenario for the 300ohm balanced
windowed line. How would you measure for the 1/2WL of that cable

Fred - N4CLA

On Thu, Jul 29, 2021 at 1:22 PM ward harriman <ward.harriman@...>
wrote:

I'm not exactly sure what you want to measure. I assume you'd like to
measure the velocity factor and the characteristic impedance.

Might I suggest the following:
Build a test fixture to connect the two leads of your 450 ohm line to
BOTH ports of the nanoVNA. This could be just a couple of coax cables whose
shields are tied together and the conductors are 1" (or so) apart.
Calibrate this arrangement.
Connect your piece of 450 ohm line to the center conductors.
Connect a 200 ohm (or something about that) load to the other end.
Measure the resulting impedance across a range of frequencies well
below to well above the desired operating point (See below for 'well
below' and 'well above').
Compute the resulting impedance. (There are two ways to do that, one
using S11 and one using S21. I would recommend the S21 approach)
From that sweep you can determine:
1) The length of the line: The frequency at which the impedance
is 200 ohms tells you the frequency at which the line is 1/2 wave and
therefore the electrical length of the line.
2) The impedance of the line: At twice that frequency the line is
1/4 wave. The 200 ohms will be converted to something like 1k. The
characteristic impedance is just Sqrt(Zmea * 200). (Again, this would be
better measured using the S21 because 1k is a bit high.)

If all you are interested in is the correct length for 1/4 wave, you
can make life a bit easier... no math necessary. Leave the line 'open' and
simply measure S21 and S11. The frequency at which the line is 1/4 wave
will cause the 1/4 wave to be a 'short'. At that frequency, S11 will be
'0' and S21 will be '1'.

If someone knows of any pitfalls to this approach, please let us know!!!

Window line is great at HF.? But remember that it consists of tandem sections of transmission lines that have two different characteristic impedances.? When the wavelength is so long that a large number of such sections in tandem will approximate a continuous line, the line will work well.

Another example of such a structure is at voice frequencies.? A telephone cable pair's voice frequency transmission characteristics can be improved significantly by adding series inductances, usually in the old Bell System 88 mH every 6000 feet.? Up to about 3300 Hz, the transmission characteristics are made more uniform and the attenuation is reduced.? But beyond 3300 Hz (approximately) the structure looks like a low pass filter with parallel capacitance and series inductance and the cutoff is suprisingly sharp and abrupt.

Although I've never used window line except at HF, I suspect that there will be some frequency above which the window line will begin looking like a series of different lines of two different characteristic impedances with all sorts of interesting reflections.? Maybe this won't be important below UHF, but it might be something to keep in mind.

73,

Maynard
W6PAP

There were a lot of interesting answers to other questions, but the question I had was whether to use or not to use a balun, which roncraig1 did address. Thank you roncraig1.

On Fri, Jul 30, 2021 at 08:14 AM, <kk7xo@...> wrote:

There were a lot of interesting answers to other questions, but the question I
had was whether to use or not to use a balun, which roncraig1 did address.
Thank you roncraig1.

The answer to your question is that you need to do several things in order to accurately determine the length of a 1/4 wave length of balanced 450 ohm ladderline for operation at 2M (144 MHz.).

1. You need to keep the ladderline well away from any other objects
2. You need a proper test jig in order to establish a "reference plane" for the measurements. I use SMA connectors with a pin jack and SMD loads. Photo attached.
3. The NanoVNA must be running on battery power. The USB lead and PC will have a significant effect on measurements.
4. A balun and/or common mode choke (on the coax cable to the test jig) is required when testing a balance line.

Any common mode current on the coax cable will affect the measurement. If measuring at HF which has a longer wavelength you can probably get away without a balun or choke but at VHF this is not the case. A 1/4 wavelength of generic 450 ohm ladderline (VF=0.91) is 18.6 inches (472.5 mm) long at 144.3 MHz. If the length is 19 inches (482.6 mm) it is 141.3 MHz. So a 0.4 inch (10 mm) change in length results in a 3 MHz. difference.

A discussion of how to make a suitable balun or RF choke for VHF use is a detailed topic..

Once 1 to 4 above are in place I would make the measurement by connecting a 50 ohm load and the ladder line in parallel at the jig. The far end of the ladder line is short circuited. At 1/4 wavelength this results in a very high impedance in parallel with the 50 ohms resistor. By measuring Return Loss an accurate measurement of the 1/4 wavelength frequency can be made.

Roger

To measure 1/2 wave you could simply short the far end.
Sweep the frequency.
When s11 is -1 (impedance is zero) you're at the 1/2 wavelength frequency (or a multiple thereof, start the sweep very low).

You can learn all sorts of things by playing with a transmission line simulator. LTSpice is good.
(Unapologetic pitch for my stuff) SimSmith is better.

www.ae6ty.com/Smith_Charts.html

ward

On 7/30/21 10:40 AM, ward harriman wrote:

To measure 1/2 wave you could simply short the far end.
Sweep the frequency.
When s11 is -1 (impedance is zero) you're at the 1/2 wavelength frequency (or a multiple thereof, start the sweep very low).

or open.? Either way, you get mag s11 = 1 when it's a 1/2 wavelength long.

However, the OP wanted 1/4 wavelength. So you could use the fact that 1/4 wave transforms the impedance at the far end. So it also will be either a short (if far end open) or a open (far end shorted), both of which show mag S11 = 1 (or maximum... there *is* some loss)

On Thu, Jul 29, 2021 at 10:21 AM, ward harriman wrote:

If all you are interested in is the correct length for 1/4 wave, you can
make life a bit easier... no math necessary. Leave the line 'open' and simply
measure S21 and S11. The frequency at which the line is 1/4 wave will cause
the 1/4 wave to be a 'short'. At that frequency, S11 will be '0' and S21 will
be '1'.

If someone knows of any pitfalls to this approach, please let us know!!!

Ward,

What you say is true for theoretical transmission lines with no attenuation. A physical TL has attenuation and you don't get exactly 0 ohms at the end of a open 1/4 wavelength transmission line or infinite impedance with a shorted line.

I don't know if you read my latest post where I pointed out that common mode results in incorrect S11 and S21 measurements when measuring a balanced TL with an unbalanced VNA. You need to reduce it with a suitable balun or RF chokes on the coaxial cable especially if you are operating at VHF or higher like the OP posed in his question.

Even with a proper balun then there are some issues with accuracy. You might find the attached article by Rohde and Schwartz interesting.

I absolutely agree I was talking about ideal lines... and ideal VNAs.
One must always take these things into account.

Thanks for keeping me honest....

It occurs to me that it might be easier to characterize a longer piece of line (which allows one to work at lower frequencies and have fewer capacitance problems) and then cut the desired length…….

Just thinking’ outloud……

Ward

On Fri, Jul 30, 2021 at 02:38 PM, ward harriman wrote:

It occurs to me that it might be easier to characterize a longer piece of line
(which allows one to work at lower frequencies and have fewer capacitance
problems) and then cut the desired length…….

Just thinking’ outloud……

I agree with you. I have been doing some tests on a 38.75 foot piece of 300 ohm twin lead with the intent of creating a 1/4 wavelength stub. I used a signal generator with a fast pulse and a digital scope to perform TDR. The round trip delay was 103.2 nanoseconds. This works out to a velocity factor of .763. Next I used the Transform mode on a NanoVNA-H4 and got 102.0 nanoseconds which is a VF of .772 which is pretty close considering the NanVNA is using a IFFT method and not an actual pulse.

Once we know the VF to some degree of precision a 1/4 wavelength piece can be cut . A 20 inch piece with a VF of .763 is 1/4 wavelength at 112.49 MHz. With the NanoVNA calculated VF of .772 that same 20 inch piece is 1/4 wavelength at 113.81 MHz. So a difference of 1.32 MHz.

Next step was to cut a 20 inch piece and see what the NanoVNA measures on the Smith Chart using a short or an open on the end. A photo of the test setup showing the jig and the ferrites on the short coax cable is attached. Without the ferrites the common mode was really affecting the measurements as shown in the attached screenshots. This was particularly the case when the termination was shorted and high impedance was being measured by the NanoVNA. 1/4 wavelength frequency was measured with reactance as close to zero as possible.

A comparison of the results (20" 300 ohm twinlead) 1/4 wavelength frequency..

Using TDR method (calculated) 112.49 MHz.
Using NanoVNA Transform method (calculated) 113.81 MHz.

Open termination - NanoVNA - without ferrites 112.02 MHz.
Open termination - NanoVNA - with ferrites 112.02 MHz.
Short termination - NanoVNA - without ferrites 80.529 MHz. ** see screenshot
Short termination - NanoVNA - with ferrites 108.02 MHz.

Conclusions
-----------------

1. Common mode is an issue and adding ferrites gave better results. I used 3 mix 31 after trying 1 and 2 with poorer results. Mix 43 have higher permeability and would be better but I didn't have any snap-on ones.

2. Results are closer to the TDR method when an open termination is used. The impedance being measured is low and in a better range for the NanoVNA. With a low impedance there is less common mode.

3. A proper balun was not investigated and takes more effort to construct and test.

Roger

This may be of interest for those testing balanced lines with a VNA for HF applications. From the 2019 ARRL Handbook

Roger

I suspected a ballun was necessary. I have a drawer full of those ferrites that I use to make 2m coax bead balluns. And I've made a jig like the one in your picture also, except without the ferrite. I used it to measure the 1/4 wave and had good results. Now I will make a jig with the ferrite and compare.

Thank you. An excellent article. Goes right to the heart of my question.

A while back I built an antenna for NOAA 137 MHz satellites, and I did a detailed search for what's the best coax feedline choke for 2 meters, and determined from the manufacturer's graphs (Fair-Rite) that (drum-roll):
--- Type 31, 2 turns ---
provides the best common-mode resistance (CMR) at 2 meters. This was a surprise, as I have always associated type 31 with the lower HF range -- but that's true for RF transformer application, not choking.
A single pass-thru provides about 400 Ohms CMR, with a gentle peak up towards 1 GHz.
2 passes provides about 1250 Ohms, peaking at ~100 MHz; at 144, it's ~1100 Ohms CMR. Two passes thru two beads would double that, etc.
Do not use 3 turns, as it will move the peak CMR (~2400 Ohms) way down around 40-50 MHz, and rapidly falling above that.

So I ordered a dozen Fair-Rite 2631540002, which is a thick medium-sized bead, about 1/4" ID by 5/8" OD by 1" long. Provides about 340 Ohms CMR at 144 MHz, with 2 passes providing about 1100. This thick & long type-31 bead has the highest CMR at VHF that I could find in all the datasheets.

Here's the Fair-Rite 2631540002 datasheet (see the 2 graphs therein) -
RG-316 will easily pass through it twice. Put it as close to the antenna end as possible. This will significantly de-couple the outside of the shield (and the nanovna) from the measurement. One on each end is even better. Small alligator clips at the end right as it comes out of the bead; calibrate open, short, and load (2x100 Ohm in parallel) at the alligator clips. Well, ok I hear someone saying "not consistent spacing! capacitive coupling changes!" OK, use a li'l ol' 2-terminal block instead of alligators.

Maybe this will help somebody.
73, --kv5r

On Sat, Jul 31, 2021 at 11:40 AM, KV5R wrote:

A while back I built an antenna for NOAA 137 MHz satellites, and I did a
detailed search for what's the best coax feedline choke for 2 meters, and
determined from the manufacturer's graphs (Fair-Rite) that (drum-roll):
--- Type 31, 2 turns ---
provides the best common-mode resistance (CMR) at 2 meters. This was a
surprise, as I have always associated type 31 with the lower HF range -- but
that's true for RF transformer application, not choking.

Thanks for the detailed info on Mix 31 and your winding tips. That is what I have in the parts box and I was pleasantly surprised to see it was better than Mix 43 and 61.

Roger