NanoVNA Custom Software

[purpose]

Another option, now that the data is within reason between your pull fixture, RLC meter and Nano, is try to measure your bag of parts and build up the filter you are after.   

I could do that. I got most of the way through the 50 using the old hardware, so I might just finish them off with the same. The fixture is new, but using the components from the old. I'll compare some results to see if I have to rebuild the old.

[purpose]

First attempt at a 5 pole jobbie. Needs some fyne choo ning (correct capacitors) , but joeqsmith's software, coupled with the Dishal software certainly gets you right in the ballpark.

[joeqsmith]

Sounds like you have made some good progress with your filter. 

After trying to measure an RF capacitors ESR, I was attempting to use the shunt through method to read a 100uOhm resistor.   After designing a custom common mode transformer, I was able to get something sort of close.   
https://www.eevblog.com/forum/rf-microwave/impedance-measurement-with-vna-using-series-shuntseries-through-methods/msg3546156/#msg3546156

The transformer led me down another path...   The last time I played around with a breadboard, we were making an oscillator for the fun of it.  This one is simulating a very bad power plane and bypass capacitors.      

[purpose]

That lot is above my pay grade, but here is a pic of my shunt/series component fixture.
Flick a switch between the two.

[joeqsmith]

The log sweep mode is getting a workout.    Shown is a 1nF silver mica in series with a 10 ohm on the breadboard so simulate a fairly high ESR.    All the spurs above 80MHz are present with nothing plugged into the breadboard.   

[joeqsmith]

Using the default 101 data point, linear sweep but viewed as a log.   Shown with a 100uH in series with the 10 ohm.  The inductor's DCR is about 5 ohms.  Seems like a poor choice for a decoupling network, even for a breadboard.   

[joeqsmith]

Sweeping 20KHz to 100M, still no transformer and unit is not calibrated.    Plot A, starting with a tant for the bulk. B, adding a 0.1.  C adding a 0.001uF.

[joeqsmith]

Tried bunching up the caps onto one small section of the breadboard then trimmed the leads and jammed them on a bit of FR4.  Tames some of the peaks, sort of.

[purpose]

Still way above my pay grade, yet appears useful to those earning a goodly chunk.

[joeqsmith]

I'm not sure how useful this particular setup actually is.  Obviously you can measure components this way.  From my software, you can guess that normally I will use the shunt method.  You need to use the setup that best fits what you are trying to measure.   I had thought about changing my software to have a way to change the measurement type for the readouts.   It may help clean up some of the front panel.   For the low frequency work, that common mode transformer will be important but I doubt the people buying the Nano would have the $800 or so to buy one.   I want to roll on for about the same price as the Nano that the RF/digital hobbyist can build.   If you read through that other thread, the one I built has really good performance at the lower frequencies but it falls apart before you get to a GHz. 

There are posts about repairing GPUs and wanting to look at the power planes.  Basically, how to look at a PDN with such a low impedance.   Something like Nano may aid in the troubleshooting.   

When I first reviewed the original Nano, I mentioned that a friend had bought these and given me one.   That video starts out with me reading a paragraph about high speed digital design.   My friends whole interest in the Nano was to strengthen their understanding of PDNs.   When I first introduced them to the VNA, we were looking at part of their PDN design as an example.   

[joeqsmith]

Parts are taking longer to get than I expected.  The new cores are due next week.   The H4 is still a few weeks out.   

The software was diverging a fair amount from what was last released.   My plan is to release the software without all the PDN functions and just try and address the the incompatibility with the H4 and what ever flavor or the week firmware is out there.  Also I uncovered another bug that happened during the port that will be corrected.   

Later, if I get anything that seems useful working on the Nano, I will roll the major revision and release new software for both the old hardware and the V2 Plus.  The main menu was getting too tight.   I want to keep the size small enough to fit on my laptop.   In the attached picture, upper left, you can select the type of measurement and the cursors will change.   It makes the readouts more generic.   All those Transmission readouts are gone, freeing up a bunch of space.   Most likely I will add more buttons to automate some of the measurements I tend to make.  This would also require a major update to the manual.   

[purpose]

All sounds good to me.

[joeqsmith]

Checked the status and it's still on the boat. 

I have been giving some thought to your being able to change the number of data points independent from the software.   With the software you are using, it will detect the number of data points during the connect and anytime you change the frequency settings.   You could however fool it easy enough.  The problem with that is you could potentially waste a bunch of time collecting bad data.   

The performance hit comes when you want to change the frequency a lot (segmented sweeps, zooms).    Assuming that your H4/firmware combo allows you to save the number of points such that on power up it is automatically set,  another option would be to have the software detect it during connection only.   

One problem is that I store the number of data points that were used when creating a calibration file.   When you load a calibration, it will properly setup the VNA.    The V2Plus has provisions allowing the software to change the number of points remotely.  The original Nano was always hard coded to 101 so it posed no problem.  Now we have a case where the cal file could be set to 401 and the Nano is set to 101.  It doesn't seem like this was a well thought out idea.   You could contact the firmware group and see if they support remotely changing the number of data points.   
***
It seems Dislord has posted in the V2 group.  I have reached out to them to see if we can sort some of this out.   

[joeqsmith]

Talking with Dislord and a few others, its possible to select the number of points remotely.  They also have a few other features that may help improve the performance (update rate).   I will just need to play around with it.  The first question I have is if it will give us reliable data or does it send random garbage like the other new releases.   After seeing the problem first hand in your early posts,  I assume you have ran some long term tests with it and if it was a problem you would have brought it to my attention.  After all, that would be a show stopper. 

[purpose]

When you say 'long term', are we talking something like overnight data gathering?
I've taken 2000 sample sweeps that appear to be glitch free, which of course I can't find now.

[joeqsmith]

How long does it take to sort a bag of crystals?  I would expect it to run over night with no trouble at all. 

[purpose]

How long does it take to sort a bag of crystals?
I would do 300ish sample sweeps, so not an aeon.

I would expect it to run over night with no trouble at all.
Of course you would. Pretty sure you won't be disappointed.

[coppercone2]

what are people doing with crystals that makes them so interesting?

Is it crystal filters?

[joeqsmith]

It should be here this week so hopefully we can start looking at the code again next weekend.

[purpose]

what are people doing with crystals that makes them so interesting?

Is it crystal filters?

In my case, yes.

[purpose]

It should be here this week so hopefully we can start looking at the code again next weekend.

Marvellous.

[joeqsmith]

There are countless posts talking about what people have bought.   It's rare to see people like the two of you who are actually experimenting and building something.  Good to see it.     

[coppercone2]

well at least the gear will end up on ebay eventually (partially) restored, even if people are just collecting you still get some saves because people don't turn on ancient capacitors, replace swiss filter technology, catch corrosion, save it from the landfill and sometimes offer pictures and information much more useful then old catalogs and manuals

if no one is buying it then its very possible to get it thrown into a e-waste dumpster and reclaimed for scrap

theoretically, even not looking at specifications, there is usually a chinese PCB you can get on ebay that theoretically does the work. practically, its unlikely to make it into even a plastic box with connectors on it to turn it into a usable tool (half of the stuff people claim makes things obsolete is kind of like attaching a saw to a motor body without a gearbox and calling it a useful carpenters tool (it does not even have a handle). but on paper both spin?

also, I expect there is at least some suppression from hardware developers.. those guys selling DC/DC converter minipcbs on ebay.. they don't want you to make your own thing, they wanna take over because its more sales. when they have cheap parts and a big inventory of course the trend will be to discourage people from doing otherwise.

effort required to use it as a test system in acquired state: obscene (who the hell actually can stand using a bunch of circuit boards strewn over a table connected by screw terminals and solder as a 'useful' laboratory. ?)
specification: facilitates blinking of light if you manage to connect it all without breaking something

so that gear has alot of use. its a step up above a shoe box with a bunch of cards. some people can get places with that though. imo its nuts to work like that. I think thats why people are going for it, but then nothing is developed because it 'feels' like you can just buy every simple thing on ebay, so I think people get stuck. feels like a bit of a youtube popularity contest (how many people are getting stuff to try to get followers? how many percent of followers actually know anything?). if you get started like that and then realize your projects are all seemingly out there, the equipment gets frozen. then in that case, just playing antiques road house will get more popularity then doing something with the equipment, because 'i can just buy that board and hook it up look it has these connections on it', the equipment body is kind of like art.. very few people actually understand the utility of the specification


try showing someone a thought out project, the end many times is a graph or number on something, often written on a note pad. then show someone glowy tubes.. whats going to get you props? the social media / marketing aspect that is common in this hobby completely drives some people crazy, but it reduces costs and increases availability. so frustration is inevitable?

plus it helps younger people in electronics careers because sometimes you can get through a non technical hiring manager because they saw something silly about what you are doing. any large company is going to have that problem. if there is some boss somewhere in the building that just happens to collect cool looking standard resistors to put into a display case near the model cars, you are kind of in luck. I once saw a gigantic radio transmitter tube in a hardware store, for no reason. i bet if you knew what it was you can get hired easier if you need money.

but it does clog the forum.. and sometimes stuff you need gets locked up in art prison..


and the electronics industry is just completely nuts because of things like Chinas mail/shipping subsidies (which some say is a part of some kind of economic war). for small and cheap things, the effects of shipping on purchasing can be described as chaotic. i have a feeling small changes in postage price, can already make things weird, and large ones are just ludicrous (not sure if its good or bad). and things like amazon and counterfeits, they reported 2 million listings that were removed just right now..

[joeqsmith]

It should be here this week so hopefully we can start looking at the code again next weekend.

Marvellous.

They missed their long range projected date and have added another week.   

[purpose]

How nice of them.