开云体育

When I measure 50 Ohm coaxial impedance with the 1/8 wave method, I always get 50 Ohms. I tried some 75 Ohm cable and still get 50 Ohms!

When I use my old method of adjusting the termination it works fine and I get the expected 75 Ohms.

I think I got a reading at 1/8 wave of roughly 200 pF and 15 MHz, if memory serves. So either the method is flawed or I am doing something wrong. What I do is sweep the frequency to the point where the Smith chart trace is at the bottom of the chart, or one-eighth wave out (open end), and read the figures on the screen. Then I compute the capacitive reactance.

Bob,
On a 50 ohm Smith chart, the bottom (6 o'clock) location will always be -j50 ohms.

The trick to measuring the cable impedance is to adjust the CW test frequency so that an open cable measures as a short: The cable is 1/4 wave long at that frequency. The CW frequency is then cut in half: The cable is now 1/8 wave long. A 50 ohm cable will measure as -j50 ohms, right at the 6 o'clock point. A 75 ohm cable will measure -j75 ohms, about 5 o'clock.
This all assumes proper calibration at the reference plane where the cables are attached, of course.

If you want, you can just do the frequency adjustments by changing the Stop frequency. The display will then be an arc instead of a point, but the high frequency end of the arc is the relevant place to read the value with the marker.

--John Gord

On Sat, Jan 25, 2020 at 10:01 PM, Bob Albert wrote:

"... When I measure 50 Ohm coaxial impedance with the 1/8 wave method, I always get 50 Ohms. I tried some 75 Ohm cable and still get 50 Ohms!.."
=====================================================================
Bob,
I haven't tried the 1/8 wave method with my NanoVNA-H4 yet so your question gave me a reason to do so. See the attachment for details.

1. I connected a 2 mtr length of RG-59 (75 ohm) cable to CH0. My calibrated range was 50k-1G and I reduced the stop frequency until I had one full revolution of the smith chart.

2. I set <marker 1> to half-way around the smith chart (1/8 wave length) and read the frequency (24 MHz)

3. I set <marker 2> to half that frequency (~12 MHz) and read the impedance (~75 ohms).

I could not set my second marker to exactly 12 MHz because my original calibration points landed me on either side of that frequency. One side was 76.4 ohms and the other side was 73.2 ohms. If I had done a 101 point cal from 50k-50M I could have landed on 12 MHz (~75 ohms).

I've used this method since it was introduced to the group and its always worked for me.

- Herb

the 1/8 wave method

added to /g/nanovna-users/wiki/Application-Notes#Collected-wisdom

On Sun, Jan 26, 2020 at 11:46 AM, hwalker wrote:

2. I set <marker 1> to half-way around the smith chart (1/8 wave length) and
- Herb

Herb, it seems to me that a half-way around the smith chart is 1/4 wave length. Isn't it? )))

On Sun, Jan 26, 2020 at 04:51 AM, Andy UA3RAW wrote:

Herb, it seems to me that a half-way around the smith chart is 1/4 wave length. Isn't it? )))

=========================================================================

Andy,
You may be able to put it in better terms than I did, as I was only parroting back what I understood from the original poster. I have only used smith charts sparingly in my career. The key thing that stuck in my mind was the original poster saying the 1/8 method started from the same position as a short on the first revolution around the smith chart.

Since the actual method has always worked for me I didn't give it any deeper thought. If the terminology I used needs correcting please add to the discussion so I can correct my notes and not perpetuate incorrect information (i.e. if that point is 1/4 wave length why is it called the 1/8th method).

- Herb

On Sun, Jan 26, 2020 at 04:50 PM, hwalker wrote:

Since the actual method has always worked for me I didn't give it any
deeper thought. If the terminology I used needs correcting please add to the
discussion so I can correct my notes and not perpetuate incorrect information
(i.e. if that point is 1/4 wave length why is it called the 1/8th method).

- Herb

Ok Herb, I'll try.
As it was described by DJ7BA in #8655 message, you must find the lowest (=Lambda/4) resonance frequency. (Impedance there is about 0 Ohm.) One full revolution in the Smith diagram corresponds to a displacement along the line by a distance equal to half the wavelength in it.
So the half-way around, is equal to 1/4 wave length. This is a point of the lowest resonance and it is corresponding to the Marker 1 on your screenshot.
In order to find the characteristic wave resistance of a coax cable, you must find the point that is a half of 1/4 wave length, or 1/8 wave length, and read the reactance at this point. That's why it is called Lambda/8 method.

On Sun, Jan 26, 2020 at 07:07 AM, Andy UA3RAW wrote:

As it was described by DJ7BA in #8655 message, you must find the lowest (=Lambda/4) resonance frequency. (Impedance there is about 0 Ohm.) One full revolution in the Smith diagram corresponds to a displacement along the line by a distance equal to half the wavelength in it.
So the half-way around, is equal to 1/4 wave length. This is a point of the lowest resonance and it is corresponding to the Marker 1 on your screenshot.
In order to find the characteristic wave resistance of a coax cable, you must find the point that is a half of 1/4 wave length, or 1/8 wave length, and read the reactance at this point. That's why it is called Lambda/8 method.
==============================================================

Andy,
Thanks so much for the clarification. Its something I'll definitely be adding to my engineering notes. Electronics has been both my vocation and avocation for the past 40 years. I still enjoy learning new things, however; few subjects make my eyes glass over as the smith chart does.

At the beginning of my career the company I worked for won a government contract to RF immunity test ignitors to ensure they didn't unintentionally misfire. I had to design rf matching networks for use between our power amplifiers and the ignitors at each test frequency. We brought in the great Christopher Bowick to teach a course on using the smith chart to design matching networks. After that project was completed I never had another practical application to apply what I had learned and as they say "if you don't use it you lose it".

The Lambda/8 method is the first real world application, other than verifying the quality of my solt calibrations or measuring components, that I have used in a while. I wish there was a way of automating the process.

Again, thanks for taking the time to reply. Perfect example of the members helping members format.

- Herb

Bob Albert: When I measure 50 Ohm coaxial impedance with the 1/8 wave method, I always get 50 Ohms. I tried some 75 Ohm cable and still get 50 Ohms!

Have you tried the 1/8WL open/short method? Where the two Zs cross is the Z0?

Well there seems to be no concensus about this.? I have a couple of things more to try.
However, starting at the right edge, open circuit, and increasing frequency as the trace swings clockwise, at the 6 o'clock point there is one-eighth wave.? The reactance there, I thought, should be numerically the same as the characteristic impedance.? It's not; it seems to be 50 Ohms regardless (as one responder says).
Continuing clockwise the quarter wave point is reached at the left edge, where impedance becomes resistive and minimum.? So there is no information other than the line length (and attenuation).
Continuing clockwise we return to nearly open circuit at the right edge.? Again, it's resistive but now high impedance.? No information regarding Z.
Of course, terminating the line in a variable resistance shrinks the trace until it becomes a dot regardless of frequency; at that point it's easy enough to read the impedance.? This works (theoretically) for any length line at any frequency.? But it requires the user to adjust the termination.? I can do it that way but was hoping that the eighth wave method would work, eliminating a step that has some potential for error.
Bob
On Sunday, January 26, 2020, 08:50:13 AM PST, W5DXP <w5dxp@...> wrote:

> Bob Albert: When I measure 50 Ohm coaxial impedance with the 1/8 wave method, I always get 50 Ohms.? I tried some 75 Ohm cable and still get 50 Ohms!

Have you tried the 1/8WL open/short method? Where the two Zs cross is the Z0?

Use nanoVNA Partner v0.20 link:

Well there seems to be no consensus about this

Using 1-900MHz device calibration from weeks ago,
I followed Herb's steps /g/nanovna-users/message/10210
for three generic CATV jumpers with F-connector barrels via RCA to BNC to SMA adapters on my worse clone
which reported about 72 Ohms reactive and 1 Ohm resistive.

Halfway around the Smith chart isn't 1/8 wave; it's 1/4 wave.

On Sun, Jan 26, 2020 at 10:37 AM, Bob Albert wrote:

Halfway around the Smith chart isn't 1/8 wave; it's 1/4 wave.
=======================================================


Bob,
Read message #10234 in this topic. We cleared that up.

- Herb

On Sun, Jan 26, 2020 at 08:05 PM, Bob Albert wrote:

Well there seems to be no concensus about this.? I have a couple of things

Did you read the original message #8655?
Let's try to do it step by step. )))
Switch on NanoVNA and connect a piece of open ended cable to CH0. No matter what is the length and Vf.
In my example, NanoVNA was calibrated from 50 kHz to 900MHz. You need to switch one of the CH0 traces to REACTANCE measurements in DISPLAY>FORMAT>MORE>REACTANCE menu.
You will see something like on the first screenshot.
Move the marker to the first (lowest) resonance frequency approximately where the REACTANCE trace cross the zero axiss for the first time.
Read the frequency value. Then, in order to rise accuracy, set START and STOP frequency a little bit lower and higher this frequency. See next two screenshots. In my example the lowest resonance frequency is somewhere near 20MHz. That's why I set START=10MHz and STOP=30MHz in STIMULUS menu.
The result is on the next screenshot. Now you can move the marker to a resonance frequency, where impedance is about 0 Ohm, more accurately.
Read the value of this frequency. In my example it is 21.2 MHz. Divide it by two. This is the frequency on which the cable length is 1/8 of wave length. Remember it. The final step is to switch NanoVNA into CW FREQ mode in the STIMULUS menu and set the calculated frequency. In my example it is 10.6 MHz. See last screenshot. Read the REACTANCE value. It is the cable's wave impedance.
Thanks! I hope, that it helps somebody. :-)

That should be -j72 ohms. Proper and good use of a VNA involves respecting
the sign of the complex portion of the impedance measurement. -j implies
capacitive and the below the central horizontal line on the Smith Chart.
+j implies inductive and above the central horizontal line on the Smith
Chart. Only the central horizontal line is purely resistive, no complex
portion and, therefore, no +/- j. Without the sign, a pure resistance
would be understood.

The 'device' or concept I used long, long ago to keep which part of the
Smith Chart is capacitive or inductive follows.

INDUCTIVE: Think of a coil spring. When compressed, it pushes
upward. The coil spring 'looks like' an inductor.
Therefore, it belongs to the top of the Smith
Chart. This is +j.
CAPACITIVE: Many times we use a bypass capacitor to shunt high
frequency to a return. It pulls the HF energy
'downward' or to the reference plane.
Therefore, the lower half of the Smith Chart is capacitive. This
is -j.

Maybe this helps? I used it some 40+ years ago to keep things straight.
Of course, I no longer need this 'tickler'. Your mileage may vary.

Dave - W?LEV

--

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

Hi Dave -

That should be -j72 ohms.

Sorry. I should have said "worse clone reported about -72 Ohms reactive and 1 Ohm resistive"
See attached image for repeat of Herb's steps with slightly different results, but still no "j".

My point was, those steps work fine for roughly sorting characteristic impedances,
which firmware in my nanoVNA does not directly report

Using 1-900MHz device calibration from weeks ago,
I followed Herb's steps /g/nanovna-users/message/10210
for three generic CATV jumpers with F-connector barrels via RCA to BNC to
SMA adapters on my worse clone
which reported about 72 Ohms reactive and 1 Ohm resistive.

On Sun, Jan 26, 2020 at 12:34 PM, Oristo wrote:
Sorry. I should have said "worse clone reported about -72 Ohms reactive and 1 Ohm resistive"
See attached image for repeat of Herb's steps with slightly different results, but still no "j".

My point was, those steps work fine for roughly sorting characteristic impedances,
which firmware in my nanoVNA does not directly report
==============================================================
Oristo,
Would you substitute Andy's message # 10256 for mine in the wiki applications notes? His general procedure is the same but Andy does a much better job than I did of explaining the steps and not confusing the issue by using incorrect terminology.

Much appreciated.

- Herb

substitute Andy's message # 10256 for mine

OK, since you ask, but I find yours much simpler to follow

SInce the NANO gives you both inductance and capacitance at the measurement
frequency and we know the characteristic impedance is given by SQRT(L / C),
we can calculate the impedance, Zo, directly:

Zo = SQRT([626E-9] / [153E-12]) = 64 ohms

This should change very little with frequency.

Dave - W?LEV

--

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