S-Parameter plotting software

[joeqsmith]

AppCAD is a simple RF calculator which includes a Touchstone plotting feature.  Sadly the software appeared to change hands from Agilent to Avago Technologies (now Broadcom) and has not been updated since 2012. 

https://www.broadcom.com/appcad

It has a few limitations and I wonder if there are better tools available.   Google search hasn't helped yet:

https://www.rfpage.com/free-tools-to-analyze-and-plot-s-parameters-file/

Is anyone aware of a Windows 10 compatible viewer that will allow you to overlay 5 or more Touchstone files with different ranges, sample sizes and variable frequency steps?   

[rf-messkopf]

You may want to take a look at the METAS VNA Tools by the Swiss national metrological institute:

https://www.metas.ch/metas/en/home/fabe/hochfrequenz/vna-tools.html.

You have to register if you want the full version, but there is also a reduced one (VNA Data Explorer) free to download, which only contains the data plotter. The full version adds a lot of functions like offline calibration and S-parameter uncertainty evaluation according to the EURAMET calibration guide.

As a data plotter the VNA Tools are very convenient and hard to beat in terms of speed and versatility. They are ideal if you quickly want to look at different Touchstone files and make comparisons, etc. You can pot linear/logarithmic magnitude, phase, convert to Z-parameters, plot in one or several diagrams, take differences and quotients, do time domain analysis, calculate passivity and switch terms, etc. The user interface is a bit unusual, but actually very powerful, especially with the integrated file manager. The only lacking feature is a Smith chart plotting facility. You can however plot \$S_{nn}\$-data as a polar plot with rectilinear coordinates, but not with the curved Smith coordinate lines. For me this is not a real shortcoming as long as I only want to look at S-parameter data.

As was already mentioned, the most flexible option would be a math package like MATLAB, Octave, Python (scikit-rf) with good plotting capabilities. Then you can do arbitrary math on the data, the plots are infinitely customizable, and can be output in publication quality in any format you want. Takes longer to set up, though.

A third option is to use a simulation package with S-parameter simulation capability. I often use QucsStudio, which is free and very powerful: http://qucsstudio.de/de/start/. Besides plotting this allows to add circuit elements to your S-parameter data. Also takes longer to set up and is not too convenient for browsing files real quick.

[joeqsmith]

Thank you both very much.  I would like to avoid creating a viewer.         

I have downloaded the free METAS viewer and installed it.   I have also filled out the registration but doubt that will happen today.  It's much nicer than the AppCAD viewer.  Just selecting the files to load was a real pain in AppCAD. 

It works great with the basic files.  Very fast and user friendly.   The first thing I noticed is that the screen will flicker at random every 1-10 seconds.  It will do this with no files selected.  It appears it is rescaning the drive and refreshing the file list.  Is this normal?

[rf-messkopf]

It works great with the basic files.  Very fast and user friendly.   The first thing I noticed is that the screen will flicker at random every 1-10 seconds.  It will do this with no files selected.  It appears it is rescaning the drive and refreshing the file list.  Is this normal?

No such flickering on my machine with the full version.

They have set up a user forum for METAS VNA Tools at https://groups.google.com/g/vnatools. Perhaps it is a good idea to post there and report that problem.

[joeqsmith]

The flicker will continue even if the software looses focus (not currently selected).   

Now that I know it works well with typical good files I'll try something more difficult. 

In this first test, I have used my software to sweep the NanoVNA (nothing attached) from 10kHz to 1MHz with a 10kHz segments.   There's a bit or a rounding error which throws AppCAD into a tizzy, but METAS seems to handle it just fine.

****
To be clear, when I say rounding error, the frequencies I use are off a slight amount making the spacing between samples uneven. 

[joeqsmith]

Using what I call my quasi log mode to sweep the data, forcing the frequency spacing between samples to be very non-linear, AppCAD will not even attempt to load the file.

METAS will load the first segment and abort.  So close..

What's odd to me is the original Touchstone spec provides an example with uneven spacing.    I have attached one of the test files.

[joeqsmith]

It works great with the basic files.  Very fast and user friendly.   The first thing I noticed is that the screen will flicker at random every 1-10 seconds.  It will do this with no files selected.  It appears it is rescaning the drive and refreshing the file list.  Is this normal?

No such flickering on my machine with the full version.

They have set up a user forum for METAS VNA Tools at https://groups.google.com/g/vnatools. Perhaps it is a good idea to post there and report that problem.

Interesting if I use their built-in navigator to go to the directory where the files reside, the flickering stops.   I can't take a screen shot of the navigator but I select the C: pull down menu at the top left of the screen.   

[joeqsmith]

If I look at where it hung, it does not seem to like replicate frequency entries.  I am going to feel really stupid if this was the problem all along.   

[rf-messkopf]

If I look at where it hung, it does not seem to like replicate frequency entries.  I am going to feel really stupid if this was the problem all along.   

Yes, I can reproduce that. I have never run into that problem, but surely that is not intentional. Manually setting the plotting range to a higher frequency or switching off "Auto" in the "Setup"menu does not help either. That should be reported to them as well.

I can't get it to flicker, though.

Nevertheless, it is a very powerful software with a lot of useful features hidden in all sorts of context menus, and a with lot of science behind them (I'm referring to the full version). Takes some time to get used to.

[joeqsmith]

I r an enjinear.    

I removed the overlap for both the linear and log segmented sweep functions.  Both AppCAD and METAS are now able to read them correctly.       It was just bad coding on my part but I'm surprised that the software had an issue with it.  I noticed it had a problem early on and just assumed it was from the uneven samples.   

I really like that METAS program.  Even though AppCAD's viewer will now work, I doubt I will use it again. 

Thanks again. 

Attached ZIP is with no overlap.

[joeqsmith]

I wonder if it is something with my PCs causing the flicker.  When I tried running the software supplied with the original NanoVNA, it would flicker.   I had tried it on two different PCs with two different OSs and had the same results.   This flickering caused me to toss it out and roll my own.  I had people write me that did not see the flickering.   

[joeqsmith]

From the Touchstone spec.

All data sets must be arranged in increasing order of frequency.

I would say equals is not increasing and that both programs met the standards while mine did not.  AppCAD has a similar table format but with it not even loading the files, I had no idea what the problem was.  Thinking back, I'm surprised that the MicroCAP SPICE simulator didn't have a problem with it.  After all, that was the whole point of coming up with that segmented sweeps in the first place.   

[Marsupilami]

I r an enjinear.    

Ain't we all?

I write touchstone processing software on a daily basis and it never occurred to me that this might not be explicitly specified. I guess the intention is to look at the frequency values as unique keys but then the ascending order is somewhat redundant requirement.
Anyway if you're still looking for software (METAS VNA Tools is hard to beat) you should check out RSPlot. I think it can do up to 5 files at once and reasonably useful.
My personal choice was SPView from ARK RF Systems. The big downside that even the lite version costs if I remember correctly something around $300-$400 however while it has some issues it's a really good viewer and it has a build in linear simulator with schematic editor. That is really great for answering questions like ok I have the s3p for this splitter, let's see quickly how would this look if I cascaded two and shorted the unused ports.

http://www.rsmicro.com/rsplot-simulator
http://www.arkrfsystems.com/SPview.htm

[joeqsmith]

I installed the full version onto a second PC and interestingly enough, still see the flicker.  I have been in contact with them and they confirmed my suspicions of how the software was written.  As for the replicate frequency being used: "Frequency points may only occur once and must be ascending."     For the flicker:  "This is indeed an issue and got improved in release V2.4.5". 

I've been reading their materials for the software.  This is a really nice program. 

[rfclown]

I know that you are a LabVIEW programmer. I wrote vis to read and write Touchstone files and use LabVIEW to compare files.

[joeqsmith]

I know that you are a LabVIEW programmer. I wrote vis to read and write Touchstone files and use LabVIEW to compare files.

I had included support for Touchstone in the software I wrote for my HP8754A 20 some odd years ago.  Reading/writing text files isn't a major task in any language.  If you are asking why I never just rolled my own S-parameter potting software,  mostly it was lack of interest and lazy.   METAS is fairly nice but for some strange reason, they left out Smith and Polar graphs.  So I still find myself using the old AppCAD.   

[rfclown]

I haven't rolled my own. I just use the Smith Plot functions that come with LabVIEW. They do have bugs (with drawing arcs between points near the edge of the chart) but they work well enough that I've not needed to do anything else.

[joeqsmith]

My current software supports importing a single port Touchstone file and displays it.  That's the extent of it.   Making a full blown reader that used the standard graphs to display it seems easy enough.   I assume this is what you have done. 

The graphs in my software are something I created with a CAD program.  The overlays are raster images mapped to an XY graph.   The front panel is always showing the same XY graph, just with different overlays depending on the measurement selected. 

When I wrote the software for the 8754A, I don't think LabView had support for Smith charts.  If it did, it was too primitive for my tastes. 

[rfclown]

I haven't used your solftware, but from the screen shots I've seen in this forum your Smith Charts look much cleaner than what LabVIEW does with their functions. The LabVIEW functions use the picture structure, which I haven't used from scratch. I think the Smith Plot functions have been around since the 1990's, but they haven't changed (same bugs, same look). But there is a Smith Multi Plot which make it as easy to put multiple plots on a picture. I've tried drilling into their functions to fix the plot bug, but never got it right.

[joeqsmith]

I still have 6.1 from 2001 and it indeed has support.   Labview 5.0 from 1998 does not appear to support it from the examples, but looking in the examples directory, I found it.   Going back to 1996 with LabView 4.0, again it resides in the pictures directory under examples.    Going all the way back to 1994 LabView 3.0, again I find it in the pictures directory.   All support the multiplot function. 

That's as far back as I can go.  Back then there was no Undo.  lol.   Looking at the older versions, the LLB is the same size so I would expect no changes.   So, it was certainly there when I rolled the software for my first VNA.

[Marsupilami]

FYI
LabVIEW Network Analysis Library: https://forums.ni.com/t5/Network-Analysis/High-Frequency-Network-Analysis-Library/td-p/3516257?profile.language=en
S-Parameter De-embedding tool and Viewer: https://forums.ni.com/t5/Network-Analysis/Free-S-parameter-Embedding-and-De-embedding-Tool/td-p/3518899?profile.language=en
Both of these should work with the free Community Edition of LabVIEW. (The tool has binary too, only needs the LV runtimes)

The library support reading, writing files, conversion between different network parameters (e.g. S to T) and has some plotting functions too. I haven't used the latter so I don't know, but the rest is solid.

[rfclown]

....That's as far back as I can go.  Back then there was no Undo.  lol.   

They had undo before that, but took it away. I first used 2.0 on a Mac. Then 3.0 (I think) on a Sun workstation. At some point, maybe when they went multiplatform, the undo went away. Forget when it came back. Maybe 4.0?  I stayed with 6.1 for a long time even though we usually had the latest at work. I would run on the most feeble machine and did everything I wanted. I bought a used 6.1 copy for home use, and stayed with 6.1 also at work. At some point I switched to 7.1 and stayed with that until just recently. I haven't liked anything beyond 7.1, but it's a pain being behind when wanting to use other's stuff... like the stuff Marsupilami just shared. I'll probably download it at work were I have 2018 and see if I can save vis back to 8.0 for home use.

[joeqsmith]

They had undo before that, but took it away.

   The first time I had seen NI's software, it was for DOS.  I think it was called LabWindows but don't quote me on that one.  I think we were using Borland's Turbo Basic at that time to control the equipment.   Years later I saw LabView 3 running a test and I started to look into it.  That lack of an undo was enough to keep me from even trying to use it.  I heard some good horror stories. 

4 added the undo but it seems like the app builder was corrupting the projects.   Still, it was good enough I decided to start learning how to use it.   LV5 was good enough that I bought the base license for home use.   I used 6.1 for many years.  It  had a major problem with my GPIB Ethernet interface but did everything I needed at that time.  Eventually I upgraded again and they botched the comm ports.  Now days I am stuck on 2011.  I talk directly to my GPIB Ethernet controller with their Ethernet library and only use the FTDI USB serial controllers.   

Even with NI offering the tools free of charge for hobby use, I have never installed them.   There are some features I would really like to have that never make it in and the features they do add are of little or no value to me at home.   

[szoftveres]

Here's a hand-written Smith Plot implementation for GNU Octave. Just rename from .txt to .m