开云体育

Hi Brian

Thank you , and great work , i am still not familiar to , it seems great and can be developped more , if any youtube video can help us step by step how to use it , it will be very helpfull . I need a real time plotting to adjust in real time the reisistor trimmer terminaison to focus the impedances on one graphical dot as small as possible on smith shart to measure characteristic impedances of coax's or baluns , wondering if your software can download S11 and S12 data's on real time from nanoVNA as Nano-App does ?

Congratulation for your good Work.
73's Nizar .

Alan, if I were to display several Q contours simultaneously, what values would be most useful? This would require no user input.

Brian

On Mon, May 12, 2025 at 12:38 PM, Brian Beezley wrote:

I'm trying to avoid keyboard input, such as specifying Q for one of these
curves

Routine I did inputs a single Q contour value. The Q=1 is a default value. No input required.
This value of unity at one time was used and provided with vna instruments as an overlay or etched into the screen.

On Tue, May 13, 2025 at 06:42 AM, Team-SIM SIM-Mode wrote:

it will be appreciated if we can renormalise also the port2 impedance independly of port1

Nizar, this program will do that:



Brian

Hi Jhon

A graphical X4 zoom option on the future firmware around the center of the Smith shart can help to show more easily the impedance focusing on the center plot of Smith diagram .

73's Nizar

Hi Jhon

In fact I redo the measurement of the characteristic impedance of the first balun with more care of bad contactes, i measure Zc=60 Ohm see screenshoot below , wich gives SWR = 1.2 with 50 Ohm coax and 1.3 with 78.5 Ohm Coax wich explaine the more losses with 78.5 Ohm renormalized coax , I think the renormalized impedance function on Dislord firmware + H4 is accuracy enought here to give a correct results , we assume here that coax has 78.5 Ohm and antenna has also 78.5 Ohm S11 = S21 , No problem right now about firmware modeling , butit will be appreciated if we can renormalise also the port2 impedance independly of port1 one on the future firmwares: it will be really wonderfull .

One important conclusion : if we want to fully optimize out ferrit balun we should measure it's characteristic impedance with a resistor trimmer terminaison and adjust it to much as near as possible the coax characteristic impedance to have less losses , NanoVNA H4 + DiSlord firmware should help a lot .

73's Nizar

Jim, I had the same thoughts myself, and earlier today I decided not to pursue graphical matching any further. My first impulse for a design problem of any complexity is to implement an optimizer. I do this all the time, even for rather straightforward problems. I have a standard local optimizer (Nelder-Mead) and a standard global optimizer (differential evolution). I've cycled through many possibilities for both over the years, mostly for antenna optimization, and I have settled on these two. They are fast, effective, and sure-footed. They are also blind. They find the best performance, according to your criteria of what's good, but they offer no insight on how they did it. The global optimizer can find a design in left field that you would never have thought to check. This is a virtue, but it can leave you wondering: how did it do that? I can now appreciate the attraction of graphical matching network design with a Smith chart since the process is so intuitive and suggestive. But I was reading a technical paper earlier today about constant-Q Smith matching where more sections bought you wider bandwidth and my first thought was: the choice of number of sections could be automatically optimized!

It has been a lot of fun learning about the Smith chart. All I ever really wanted to do was implement impedance renormalization, which just seems magical to me (intentionally misloading a circuit to allow measuring it without a matching network, but then unraveling the misloading later in analysis). The other thing I wanted to implement was the Y21 method, which magically suppresses stray shunt reactance when doing a series-through measurement. It's my idea of something for free. I've looked at a lot of commercial L and C .s2p files. The Y21 method exposes the labs with sloppy fixtures.

Now I've got to get back to coding. I've implemented circles with a precise pixel-by-pixel method instead of using the compiler circle function, which didn't produce high-quality output. Neither did my usual method of drawing line segments between points. How short a line segment is adequate? My homebrew circle generator produces noticeably cleaner constant-R Smith circles. Now I'm about to do it for the constant-X curves, which currently use line segments. I belatedly discovered today that those curves are really arcs of circles.

Brian

At some point, too, it's addressing a pretty niche need (essentially lumped component multisection filters). I would think that most people today designing a multi section filter use some sort of tool like ELSIE (for LCR, which is free) or one of the multitude of design tools like HFSS, ADS, QUCS, etc.. because they'll take into account things like component tolerances, standard values, loss in L and C, transformers. Not to mention things like coupled microstripline filters or cavities. Or, perhaps Scikit-RF - which doesn't do filter design, but sure is able to model quite complex circuits, and you could lash up an optimizer in Python around it.

Once you get beyond a certain complexity, using the computer design tools is probably a better way to go.

Perhaps as a way to gain insight, a Smith Chart design might be useful. But I think you're going to have a tough time doing something like trading off Butterworth vs Chebyshev vs Elliptical/Cauer on a Smith Chart (I think - never tried it - Done it on a pole zero, and on a Bode plot). Isn't the "have all the sections have similar Q" is really about getting passband ripple low and/or not having a really tight tolerance on some, and less on the others.

On Mon, May 12, 2025 at 11:31 AM, Donald S Brant Jr wrote:

The idea is to have each intermediate matching step wind up on/near the same Q
circle.

Don, this sound like high values for the constant-Q curves would be necessary. Is that right?

I'm trying to avoid keyboard input, such as specifying Q for one of these curves. So far the only keyboard entry needed is when renormalizing the reference impedance.

Brian

On Mon, May 12, 2025 at 01:47 PM, alan victor wrote:

The Q=2, 3, etc.... where the R,X ratio is 2,3, etc... and this facilitates
the constructs for matching over a desired band when a load model is imported
to the chart.

When making a graphical multi-section matching network solution, the constant-Q curves guide your component choices, to avoid unnecessarily constraining the bandwidth.
The idea is to have each intermediate matching step wind up on/near the same Q circle. Having any steps with higher Q than necessarily will constrict the bandwidth. Having steps of lower Q will necessitate more sections which will increase the loss. Or, if the bandwidth need is narrow you can use high-Q sections for some "free" bandpass filtering.
73, Don N2VGU

A phase vs frequency plot helps identify where the middle of the transition is. Sometimes that's easier than trying to find the -3dB points. A classic S curve from +90 to -90 degrees phase, and 0 is the center frequency, and +/- 45 degrees are easy to find.

Thanks for all the details, Alan. My program does have group delay, which is differential phase shift. It does extract Q from S11 or S21 when measuring inductance.

I've changed to displaying SWR circles for SWR = 1.5, 2, and 3 simultaneously. I may add another curve for Q contours of 1, 2, and 3. I would leave calculation and interpolation to the user for now.

Any other suggestions for specialized curves or features?

What chart program are you using?

Brian

The rectangular plot of the one port S data is insufficient to
determine Q. You will need the differential phase shift
of the reflection coefficient as well. Taken together with
the appropriate formula the Q can be determined. The chart
facilitates this and overlays that could be clipped onto
the network analyzer screen were available. Or the glass
was etched on screen. Doing this with the nanovna simply
requires the touchstone file and then import this into
the chart program I displayed. There are a other measurement
types available, S21, for example, where the Q is extracted
directly from the screen. Typically, this is a rectangular plot,
although a chart plot can also be used.

The Q is obtained from the reflection coefficient crossing at
zero degrees, 7.003 MHz divided by the 3 dB down power points
where the sweep crosses the Q=1 contour at 7.022 and 6.974 MHz.

The result is 146.

The common name is "constant Q contour" and the Q=1 is somewhat unique in that
it is a locus of points connecting all cases where R=X.

Other contours are available. The Q=2, 3, etc.... where the R,X ratio is 2,3, etc... and this facilitates
the constructs for matching over a desired band when a load model is imported to the chart.

On Mon, May 12, 2025 at 10:17 AM, Jim Lux wrote:

Your only question is "linear frequency scale or log".

Thanks for reminding me, Jim. I've been meaning to look into log scales, both x and y. Right now everything is linear only.

Brian

OK, I see you calculated Q from the marker frequency values. Interpolation would be easy and would improve accuracy. Maybe I could add interpolated Q calculation to the marker. Looks like I would need two markers per curve. Right now I provide just one for each curve.

Does this Q = 1 curve have a common name? Something analogous to circles of constant SWR being called SWR circles.

Brian

There's lots of different ways to plot the same data, and people develop their approaches based on the ones they're familiar with.
I guess, too, in the "bad old days" when plots were done by hand (or plotter) doing it all with one plot was more convenient.

Today, where you can flip among plots in the blink of an eye, yeah, a gain vs frequency plot would be useful. Your only question is "linear frequency scale or log".

Thanks for the illustration. How did you get Q = 145.8? Why does the rectangular plot needs phase slope added to discern? Is the dip it too broad and shallow? A rectangular phase plot is available. How would you use it?

Alan, if I handed you a device and asked you to find the series-resonant frequency and the bandwidth at that point with a NanoVNA, what would you do? If the VNA was equipped with the feature you're proposing, what would you do differently?

Brian

Here is an illustration... attached.

This is true for some measurements but not all...

Consider a one port measurement, only S11, for a shunt series resonator.

The jig is made using one section of the commercially available NanoVNA test board. Available on eBay, Amazon and from Chinese sellers. You can find discussions about this board on several posts in this group.