Ok I have a NanoVNA-H4

I have three coils I want to measure their inductance.
I tried what I saw a few you tube videos of where, I simply calibrate it and put the coil across.

well yeah it works maybe.

Two of the coils of known value both I tested at 3 to 4 Mhz freq wise.

The small one was supposed to be 8.5uh but measures as only 3.7

the bigger one i supposd to be 200uh but measures at 170uh.

Now BOTH of them on the smith trace onlu make  very small segment of a circle up in the capacitor land like around 1 and 2 o clock.
Looking for suggestions.

Joe

On 2022-04-15 17:36:-0500, you wrote:

Now BOTH of them on the smith trace onlu make very small segment of a circle up in the capacitor land like around 1 and 2 o clock.
Looking for suggestions.

The top half of the Smith chart is inductive and the bottom half capacitive. Can you send an image or a s2p file?

~R~

Hi Joe
Well, you don't mention exactly what your method is, so I'll drop 2 that I have used.
1. S11. Put the coil in series with a cap of known value, and find the phase change from +180 to -180. You can use that freq to find the coil value.
2. S11. Cal the VNA with a coax with clips of some sort. Do SOL, using the clips. Put the coil in the clips. Find a freq band that covers the top of the Smith with some good detail. For example, you might cal from 50k to 5MHz. With Phase and Smith, you should see an arc along the upper outer ring of the Smith. Move the Smith marker until it is about 12:00 high, and the phase is ~90 degrees. The Smith should be fairly close to the coil value.
I am not a GHz kinda guy, so I don't have any experience at that freq, but these methods should both work w the coil values you listed.
Please include an s1p file if possible, and tell us how you are doing it.
~R~

On Fri, Apr 15, 2022 at 03:36 PM, Joe WB9SBD wrote:

I have three coils I want to measure their inductance.
I tried what I saw a few you tube videos of where, I simply calibrate it
and put the coil across.

well yeah it works maybe.

Two of the coils of known value both I tested at 3 to 4 Mhz freq wise.

The small one was supposed to be 8.5uh but measures as only 3.7

the bigger one i supposd to be 200uh but measures at 170uh.

More information is required in order to understand your measurements.

1. What type of coils (air-wound, iron powder, ferrite) are you trying to measure?
2. Did you calculate 8.5 uH and 200 uH for the coils or were they marked as these values?
3. Are you using a test jig to do the measurements or some alligator clips at the end of a coax cable?

Roger

Hi Joe
This is one method I have used.
After setting up the circuit and calibrating (50k-5M) for the leads I am using,
I moved the markers (they all move together, so choose one) so that
1. The reactance is purely inductive (Smith @ 12:00)
2. The phase is ~90�
3. The reactance is ~50 ohms

For a pure inductor, and this is not, though it is SMT, with short traces, so is close,
all these things should be true at the same time.
I can't say why it is not closer . I'm pretty sure I grabbed a 4.7, but might be a 5.6.

~R~

Thank you this is exactly what I was looking for!

Thanks!

Joe WB9SBD

On Fri, Apr 15, 2022 at 05:57 PM, Rich NE1EE wrote:

I moved the markers (they all move together, so choose one) so that
1. The reactance is purely inductive (Smith @ 12:00)
2. The phase is ~90�
3. The reactance is ~50 ohms

It is not necessary to measure at a S11 phase angle of 90 degrees, a reactance of ~50 ohms or at the top of the Smith chart in order to get good results.

I imported your s1p file into NanoVNA app and plotted inductance, resistance, reactance and phase angle. The results are attached.

You can see that the measured inductance over the entire frequency range was between 5.31 uH and 5.22 uH. The phase angle ranged from 175 to 35 degrees and the reactance from close to 0 to 250 ohms. Measurements were made at several positions on the Smith Chart.

If you search this group for "measure inductance", "measure capacitance" or "pitfalls" you will find many posts by group members discussing how to measure components accurately.

Roger

This illustrated the power of measuring inductance at different
frequencies, or the frequency at which the inductor will be utilized.

Using only the values read from the three markers on the Smith Chart, I
calculated the resulting inductance for each of the three frequencies using

+jX = 2 X pi X F X jX

*FREQ (MHz)* *+jX (Ohms)* *INDUCTANCE (µH)*

0.533 17.6 5.89

1.52 50.0 4.78

4.0 132 3.32

So, what is occurring here? For an inductor this large, nominally about 6
µH, there is a reasonably large accumulated series capacitance due to
winding-to-winding capacitance. This capacitance cancels a portion of the
inductance. Therefore, as frequency increases, the amount of available
accumulated capacitance cancels more and more of the intended inductance.
If the inductance is critical as in a tuned circuit, this is why measuring
the complex reactance at the intended design frequency is important.

If you had run your sweep well above 4.0 MHz you would eventually reach a
point where the inductive reactance cancels the capacitive reactance. That
would be where your inductor exhibits self resonance. Nothing is left of
the impedance other than pure resistance of the windings, themselves.
Above that frequency, the inductor would measure as a capacitive reactance.

Dave - WØLEV



On Sat, Apr 16, 2022 at 5:00 PM Roger Need via groups.io <sailtamarack=
yahoo.ca@groups.io> wrote:

On Fri, Apr 15, 2022 at 05:57 PM, Rich NE1EE wrote:

I moved the markers (they all move together, so choose one) so that
1. The reactance is purely inductive (Smith @ 12:00)
2. The phase is ~90�
3. The reactance is ~50 ohms

It is not necessary to measure at a S11 phase angle of 90 degrees, a
reactance of ~50 ohms or at the top of the Smith chart in order to get
good results.

I imported your s1p file into NanoVNA app and plotted inductance,
resistance, reactance and phase angle. The results are attached.

You can see that the measured inductance over the entire frequency range
was between 5.31 uH and 5.22 uH. The phase angle ranged from 175 to 35
degrees and the reactance from close to 0 to 250 ohms. Measurements were
made at several positions on the Smith Chart.

If you search this group for "measure inductance", "measure capacitance"
or "pitfalls" you will find many posts by group members discussing how to
measure components accurately.

Roger

--
*Dave - WØLEV*
*Just Let Darwin Work*

On Sat, Apr 16, 2022 at 11:35 AM, W0LEV wrote:

This illustrated the power of measuring inductance at different
frequencies, or the frequency at which the inductor will be utilized.

Using only the values read from the three markers on the Smith Chart, I
calculated the resulting inductance for each of the three frequencies using

+jX = 2 X pi X F X jX

*FREQ (MHz)* *+jX (Ohms)* *INDUCTANCE (µH)*

0.533 17.6 5.89

1.52 50.0 4.78

4.0 132 3.32

Dave,

In my post the inductance graph shows that there is very little change from 0 to 5 MHz - 5.31 to 5.2 uH range. The data was taken from a measurement made by Rich NE1EE and posted earlier as a s1p file.

Unfortunately you made a mistake in your calculations.....

Reactance = +jX = 2*pi*Frequency*Inductance so Inductance = +jX/(2*pi*Frequency)

*FREQ (MHz)* *+jX (Ohms)* (2*pi*Freq) *INDUCTANCE (µH)*

0.533 17.6 3.35 E06 17.6/(3.35 E06) = 5.25

1.52 50.0 9.55 E06 50/(9.55 E06) = 5.24

4.0 132 25.13 E06 132/(25.13 E06) = 5.25

I agree with you that the inductor should be measured at the frequency where it will be utilized. That was the intent of my initial post where I stated "It is not necessary to measure at a S11 phase angle of 90 degrees, a reactance of ~50 ohms or at the top of the Smith chart in order to get good results."

If the inductor Rich used was measured beyond 5 MHz. the "apparent inductance" would have increased until it reached the self resonant frequency. I have attached two plots I made using a 330 uH inductor. One explains why the reactance and apparent inductance increase and the second shows what happens at resonance when the component now acts as a capacitor.

Follow up to last post. It does not take much self capacitance in order to have an inductor quickly reach self resonance. I had a 100 uH inductor in my junk box that had 1.4 pF of self-capacitance (not very much at all). I measured it with my NanoVNA in S11 mode and then saved the results as an s1p file. This file was imported into a spreadsheet and the reactance and resistance plotted on the same graph (attached). The measured reactance and resistance both rise quickly after 10 MHz. until they reach the self-resonant frequency (SRF) of 13.57 MHz. After the SRF the component acts as a capacitor.

The effect of such a small capacitance dictates that a suitable "test jig" must be used to avoid stray capacitance. Alligator clips at the end of a piece of coax are only suitable up to a few MHz. if accurate results are desired when measuring inductors.

Roger

Hi Joe

Enjoy :-)

https://youtu.be/iJ1qKE5O0bY

Regards
--
VE6WGM

Be sure to read the comments below the video. They explain why the measurements are taken at 90 degrees on the smith chart :-)

--
VE6WGM

On 2022-04-16 18:10:-0700, you wrote:

Be sure to read the comments below the video. They explain why the measurements are taken at 90 degrees on the smith chart :-)
--
VE6WGM

Thanks for the link and the comments. Very interesting vid. The 2nd comment on that page is one of the reasons I posted for Joe...I wanted some simple set of guidelines that he could apply. I was aware that the same coil value would be display over a wider freq range. We still need to get in the ballpark w freq, but that is simply part of the process. Once there, the comments at this link apply, IMO.

~R~

On Sun, Apr 17, 2022 at 04:17 AM, Rich NE1EE wrote:

Thanks for the link and the comments. Very interesting vid. The 2nd comment on
that page is one of the reasons I posted for Joe...I wanted some simple set of
guidelines that he could apply. I was aware that the same coil value would be
display over a wider freq range. We still need to get in the ballpark w freq,
but that is simply part of the process. Once there, the comments at this link
apply, IMO.

Rich,
The references under discussion are technical papers by the staff of Copper Mountain Technologies, a commercial manufacturer of VNA products. Links to journal papers are below:

https://www.signalintegrityjournal.com/blogs/8-for-good-measure/post/1344-using-a-vna-for-power-plane-impedance-analysis

https://www.clarke.com.au/pdf/CMT_Accurate_Measurements_VNA.pdf

Copper Mountain Technologies also wrote an application note with more detail on this subject.

https://coppermountaintech.com/wp-content/uploads/2019/03/Measurement-of-Electronic-Component-Impedance-Using-a-Vector-Network-Analyzer.pdf

The essence of these publications is that the uncertainty of the S parameter measurements made by a VNA will affect the accuracy of components measured using these methods: shunt measurement, shunt-thru measurement and series measurement. Their plots showing measurement accuracy are all based on the measurement limitations of specific products they manufacture and are similar, but not the same, to what we would get with a NanoVNA

The method we are all discussing in this post is the S11 shunt measurement technique and how it can be used to measure components like an inductor. With the shunt method the uncertainty of the component measurement will get worse as the impedance Z of the component under test gets further from the system impedance Zo of 50 + j0. The reason is that the uncertainty of the magnitude of the S11 reflection coefficient gets progressively worse as it increases from 0 to 1. For ideal inductors or capacitors the magnitude of the reflection coefficient is 1 and for practical components slightly less than 1. This can be seen as measurements that lie close to the circumference of the Smith chart and traverse this boundary as the frequency is changed. Even with these limitations reasonable estimates of the value of an inductor (or capacitor) can be made using a NanoVNA in S11 shunt mode and this has been shown previously and in others made by several members of this group.

Nowhere in the references above does it indicate that measurements are made at a 90 degree reflection coefficient angle which is the 12 o'clock position of the Smith chart. In fact you can see from my graphs (that used your s1p file) that good results are obtained at reflection coefficient angles over a wide range.

I think part of this "90 degree belief" is that the S11 refection coefficient phase is sometimes confused with the phase of the impedance Z. Somehow this leads to the belief that measurements of inductors have to be made at 90 degrees. This subject has been discussed several times in this group. I just noticed that Owen Duffy (who writes technically thorough articles on the NanoVNA and other electronic subjects) just wrote an analysis of the s1p file you posted and it is interesting reading.

https://owenduffy.net/blog/?p=24727

Roger

On 2022-04-17 10:49:-0700, you wrote:

I think part of this "90 degree belief" is that the S11 refection coefficient phase is sometimes confused with the phase of the impedance Z.

Thanks for your comments...I have only recently come to realize that I wanted to understand phase better...so this discussion is timely.

~R~
72/73 de Rich NE1EE
The Dusty Key
On the banks of the Piscataqua

Rich… I went on this learning journey regarding phase on the nanoVNA not too long ago also. Here’s what I learned ;-)

https://youtu.be/-gGJHhJ2lUs

Regards,
--
VE6WGM

The reason why people make measurements at +/- 90 degrees on the smith chart is because the measurement accuracy using the shunt configuration when trying to measure the nominal value of an inductor or capacitor is highest at 0+j50 ohms.

"Nowhere in the references above does it indicate that measurements are made at a 90 degree reflection coefficient angle which is the 12 o'clock position of the Smith chart. In fact you can see from my graphs (that used your s1p file) that good results are obtained at reflection coefficient angles over a wide range.

I think part of this "90 degree belief" is that the S11 refection coefficient phase is sometimes confused with the phase of the impedance Z. Somehow this leads to the belief that measurements of inductors have to be made at 90 degrees. This subject has been discussed several times in this group. I just noticed that Owen Duffy (who writes technically thorough articles on the NanoVNA and other electronic subjects) just wrote an analysis of the s1p file you posted and it is interesting reading."

--
VE6WGM

OK, I'll add a small correction to that last post

...measurement accuracy...the nominal value of an inductor or capacitor is highest at 0+j50 ohms or 0-j50 ohms, respectively.

Just so people don't get their knickers in a twist. ;-)

Regards,
--
VE6WGM

Hello everyone,



Sorry to hijack this thread, however I also see some nanovna users recomending to add a 50 ohm resistor to one of the legs of the inductor/capacitor to be measured, from the S11 port output.

Does this makes sense?

Cheers!

"...measurement accuracy...the nominal value of an inductor or capacitor is highest at 0+j50 ohms or 0-j50 ohms, respectively.
"
And when you change the frequency, you'll see how much frequency matters. And of course you might find the resonance frequency too, when the sign changes. It is a good idea to vary the frequency, if your results are good in one frequency only, you have problems.