开云体育

I having fits trying to measure lossy "soft" ferrites against the manufacturers curves. Over the range 1 to 300 MHz, I'm not even getting close, looking at either S11 into a short or S21. I've tried a few home-brew fixtures, and they all yield a Z which is much lower than the published data, so I'm thinking it is my fixturing and calibration that are doing me in.

For those that have been successful:
- Do you perform you OSL cal at the input port to the fixture, or the far side?
- Is your preferred method S11 into a short, S11 into 50 ohms, or S21?

Just to set the level of conversation; I'm quite comfortable driving around a Smith Chart and have been involved in RF for a very long time, so feel free to pick me apart!

What ferrites and where's the reference for them.
Jim K9YC has done some fairly extensive choke testing.

Mike W9MDB

The parts I'm currently playing with are all Laird ( formerly Steward) 28 alloy, with the Z peaking in the 100-200 MHz region. The data I'm referencing is from their data sheets.
I'll read through Jim's paper. Thanks!

Brent AB1LF

On 11/4/22 8:04 AM, Brent DeWitt wrote:

The parts I'm currently playing with are all Laird ( formerly Steward) 28 alloy, with the Z peaking in the 100-200 MHz region. The data I'm referencing is from their data sheets.
I'll read through Jim's paper. Thanks!

What test fixture are you using? 1 or a few turns through the core?

Note that these things have about 20% tolerance with respect to the datasheet values, too.

Single turn. Tolerance is a good note, but my readings are consistently less than half the specified value. Unfortunately, Fair-Rite is the only company who has written anything about how they measure their product, and they made the comment that other companies may do it differently.

I still think my "errors" may come from the way I calibrate. For an S11 calibration into a fixture, should the fixture be included in the OSL cal, or performed at the input to the fixture. For an S21 cal, should the fixture be included at all, or only the connecting cables.

Hopefully, I'll stumble on to an answer that agrees with the vendor one of these days!

Thanks again for your input, Jim!

I have a Copper Mountain VNA in my lab at work, so I will probably try to get some generic advice from them next week, without mentioning that it's really for use with a nanoVNA at home.

Greetings, Just as an FYI: There is a great deal of downloadable generic reference material from?Rhode & Schwarz relating to all things VNA.

I always calibrate to the interface of the "vna/jig/cable combination" with the DUT because that interface is whatever circuitry the DUT is connected to will be seen as.
E.g. if you are testing an antenna mounted on a tower and are interested in the impedance of the feed point of the antenna then calibrate the vna and included feeder to the point the feeder connects to the antenna on the tower.
Similarly it is also important to carefully consider your reference point for calibration when sweeping/ tuning filters; in this case a set of cavities used in multiplexing.
On Sat, 5 Nov 2022 at 9:59, Brent DeWitt<bdewitt@...> wrote: Single turn. Tolerance is a good note, but my readings are consistently less than half the specified value.? Unfortunately, Fair-Rite is the only company who has written anything about how they measure their product, and they made the comment that other companies may do it differently.

I still think my "errors" may come from the way I calibrate.? For an S11 calibration into a fixture, should the fixture be included in the OSL cal, or performed at the input to the fixture.? For an S21 cal, should the fixture be included at all, or only the connecting cables.

Hopefully, I'll stumble on to an answer that agrees with the vendor one of these days!

Thanks again for your input, Jim!

On Thu, Nov 3, 2022 at 05:43 PM, Brent DeWitt wrote:

I having fits trying to measure lossy "soft" ferrites against the
manufacturers curves. Over the range 1 to 300 MHz, I'm not even getting
close, looking at either S11 into a short or S21. I've tried a few home-brew
fixtures, and they all yield a Z which is much lower than the published data,
so I'm thinking it is my fixturing and calibration that are doing me in.

You need a good test jig setup if you want to measure up to 300 MHz. Could you post some pictures of your test jig and the cal loads you use for calibration?
When you measure the impedance are you referring to the complex impedance R +/-jX or the impedance magnitude |Z|?

For those that have been successful:
- Do you perform you OSL cal at the input port to the fixture, or the far
side?

You always do the OSL cal at the point where you connect the Device-Under-Test (DUT). This establishes the Reference Plane for measurements. You should verify it is working OK by measuring some known resistors that are non-inductive

- Is your preferred method S11 into a short, S11 into 50 ohms, or S21?

The S11 method is the preferred method. S21 measurements using a NanoVNA are subject to error for several reasons.
- NanoVNA source and load impedances that are not 50 ohms
- Stray capacitance in the test jig significantly affects results
- NanoVNA does not perform 12 point calibration error correction

You can read more about S11 vs S21 measurements in these links.





Roger

Brent,
I use a small, simple jig for such measurements. I use just one port, and measure directly. No resistors nor any other stuff.
Of course calibration is done at the very spot where the DUT will be connected.

For reference, in case it's useful to you, I'm attaching my complex impedance measurements for several beads. There are small and large ones, and of several different materials.

I find the complex impedance curves far more useful than the simple impedance modulus data published by some manufacturers. Because in many cases it's VERY important whether the impedance at a given frequency is mostly resistive, mostly inductive, or even has a capacitive component!

Hello
With which software do you produce these pretty curves?
--
F1AMM (Fran?ois)

On Sun, Nov 6, 2022 at 08:03 AM, Fran?ois wrote:

Hello
With which software do you produce these pretty curves?
--

NanoVNA App

Roger

NanoVNA App
Roger

** Thanks.

It's pretty but I don't understand this software at all. It runs on my Windows Seven.

Is there a notice?

I'm like a chicken that just found a knife :).

I use nanaovana-saver which I think I know how to use.

73
--
F1AMM Fran?ois

-----Message d'origine-----

De la part de Roger Need via
Envoyé : lundi 7 novembre 2022 02:58

Fran?ois, you can easily learn to use NanoVNA App by the good old tried method of clicking on each icon and seeing what it does! The one thing you need to know is to right-click on the graph, to get all the graph options.

It's important to always watch that you are using a correct calibration. The top frame of NanoVNA App tells you which calibration is in use. You can do calibrations, store them with meaningful names, and load them, from the calibration screen, called up from the calibration icon at the top. It's convenient to do complete, detailed, averaged calibrations, with many points, over the whole frequency range, for each of your test jigs, test cables, etc. Then, when using the NanoVNA, you simply call up the correct calibration, instead of re-calibrating every time.

Practice leads to mastery!

Manfred

On Sun, Nov 6, 2022 at 09:30 PM, Fran?ois wrote:

NanoVNA App
Roger

** Thanks.

It's pretty but I don't understand this software at all. It runs on my Windows
Seven.

Attached is a description of what all the buttons are used for. You right click on a graph for more options.

This is a link to a previous discussion in this group about NanoVNA App

/g/nanovna-users/topic/91143866#28096

Roger

Yes. I agree, R&S has published some very good information. My current questions revolve, very specifically, around fixturing and calibration for measuring "soft" ferrites, of which there is much less information. I may be homing in on though!

Roger
Thanks for the annotated screen shot. Very helpful. I am a newbie, lying in the weeds observing what is going on. I have made a copy of your post for future reference.
Thanks again.
Don KC3MPO

Can someone provide the link to pre compiled version of NanoVNA-App for Windows - link provided here does not work, and github page only offers source so one can compile and mess it up with it if wanted :)

Google search did not reveal much more then that.

Thanks!

Can someone provide the link to pre compiled version of NanoVNA-App for Windows - link

:)
--
F1AMM Fran?ois

-----Message d'origine-----

De la part de Miro, N9LR via

Envoyé : mardi 8 novembre 2022 14:53

Brent, see this topic. It shows the jig I used to measure the ferrite beads:

/g/nanovna-users/topic/79953788#20089

But I calibrated it with a 50? SMD resistor, rather than the leaded resistor 50? load shown there, to get better accuracy into the VHF range.

See this too:

/g/nanovna-users/topic/85417716#24627

Manfred