Nano VNA Reads Low SWR

[radiolistener]

Does this dinky gizmo ( i do not have one ) has settings to specify calibration standards coefficients or electrical delay? Maybe that numbers are incorrectly entered or corrupted which may lead to the problem OP is experiencing.

Capacitance is hardcoded to 50 fF. But you can setup delay through menu DISPLAY=> SCALE => ELECTRICAL DELAY.
By default delay should be 0 picoseconds

[joeqsmith]

Have you looked into making the settings programmable and using a common format?  With the V2, it would seem they must have this by now.   

[edtyler]

@joeqsmith and others

Thanks for the ideas and input. I have ordered a V2 from the HCXQS group.  I'll stockpile the failed V1 for disassembly and recycling of the LCD and various parts. As you say, repair is hardly worth it. I have worked with RF stuff for a few years and it is possible that the device was connected to some bench top setup while a legal limit transmitter was in operation a few meters away. I would hope these devices were not that fragile, but who knows. For $50, it is hard to complain about this item.

As you mention, I have not upgraded the firmware after the initial upgrade 2 years ago. In general, I never change the firmware/software on test equipment that is working unless I am experiencing a problem that new software is known to fix and I have a way to go back to the old software if the upgrade causes other problems. In this case, the unit was working properly - open SWR was infinity - and now it is not with no change to the firmware.

Again - thanks everybody for your help and ideas. My conclusion is that this NanoVNA has a hardware failure.

 

[0culus]

They are not well shielded at all, so damage from a high power transmitter nearby is plausible. They also create wideband splatter for a non-trivial distance when sweeping...best turn it off when you want to work the radio!

[joeqsmith]

@joeqsmith and others

Thanks for the ideas and input. I have ordered a V2 from the HCXQS group.  I'll stockpile the failed V1 for disassembly and recycling of the LCD and various parts. As you say, repair is hardly worth it. I have worked with RF stuff for a few years and it is possible that the device was connected to some bench top setup while a legal limit transmitter was in operation a few meters away. I would hope these devices were not that fragile, but who knows. For $50, it is hard to complain about this item.

As you mention, I have not upgraded the firmware after the initial upgrade 2 years ago. In general, I never change the firmware/software on test equipment that is working unless I am experiencing a problem that new software is known to fix and I have a way to go back to the old software if the upgrade causes other problems. In this case, the unit was working properly - open SWR was infinity - and now it is not with no change to the firmware.

Again - thanks everybody for your help and ideas. My conclusion is that this NanoVNA has a hardware failure.

It's a really odd condition as I would expect even if the unit had a a large amount of leakage, it would still show an open.  The 100ish ohms makes no sense. 

This guy demonstrates how the calculations really don't care about the absolute values of the standards.  With your unit being repeatable and throwing up good numbers for a short and load, it makes no sense why it would compress towards the open.   
https://www.youtube.com/channel/UCMznY--PyHE4yLc8jfSxmYQ

Dang, 50% down votes for that demo.       I'm interested in knowing what the problem is.   

[edtyler]

@joeqsmith

Interesting video. As you say, a VNA's calibration transformation equations do not attempt to rationalize the input characteristics. What this implies is that somehow the NanoVNA 'saw' something different during calibration than when not in calibration. To me, this seems like a very significant problem with the way the VNA firmware manages the display during calibration. As you say an open that is not really infinite impedance should consider that same (non-infinite)  impedance as an open. However, there must be some value for open, below which the equations give bad results.

You asked about my term 'stable Smith chart'. What I was trying to communicate is that the chart had a non-changing appearance. If there was some jitter in the display, that might indicate a poor connection at the SMA port.

[edtyler]

@0culus The transmitter is only 1500W maximum and the antenna is not very near the test bench. I get a higher power localized RF field from the cell phone sitting on the bench. I do take care to not have the SDR hooked up to any external antenna when other external antennas are transmitting.

The wideband sweep is pretty obvious on SDRs. My 8753 does the same thing to both my Flex 6600 and IC-7300 and other SDRs. First time I saw it, I wondered what is that crazy pattern of noise, then realized it was the VNA.

[joeqsmith]

@joeqsmith

Interesting video. As you say, a VNA's calibration transformation equations do not attempt to rationalize the input characteristics. What this implies is that somehow the NanoVNA 'saw' something different during calibration than when not in calibration. To me, this seems like a very significant problem with the way the VNA firmware manages the display during calibration. As you say an open that is not really infinite impedance should consider that same (non-infinite)  impedance as an open. However, there must be some value for open, below which the equations give bad results.

You asked about my term 'stable Smith chart'. What I was trying to communicate is that the chart had a non-changing appearance. If there was some jitter in the display, that might indicate a poor connection at the SMA port.

Maybe it's an ordering  problem.  Anytime I have used the firmware to cal the Nano, I go in the same order as the menu, top down.  Still, I would expect if you had deviated from your 2 year old procedure for the Nano, you would have mentioned it.  So, I really wonder what the actual problem is as from the information you have supplied, it makes no sense to me.  Then again, not being there.....

I was looking at YT the other day and noticed the countless videos on the Nano.   I had a few running for background noise and came across this gem:
https://youtu.be/F0oMpq_Jz5E?t=794

This isn't something I would have ever expected to see and stresses the fact that without seeing every detail of what people are doing, its hard to say.   I can imagine this person,  "I went through my cal and verified it with the same standards that I have been using.   I don't understand why my results don't match up"  "Yes, I verified these standards"   

I did like the double 100 ohm axial parts.   Dammit, now I want to test that....

[joeqsmith]

In a somewhat attempt to verify what was shown in the amateur radio video, I used two 100 ohm Metal Film 1/2W 1%,  two 100 ohm MF 1/4 1% and one 49.9 ohm MF 1/4 1% axial resistors.  I trimmed the leads as short as possible to minimize the lead parasitics.

It does appear there are some gains to be had.  Maybe hand selecting a couple of 1/4W parts wouldn't be too bad.   

Now that short     

[edtyler]

@joeqsmith

I did go in sequence through each step as I later described, top to bottom, in oder, no skips, including ISO and THROUGH. I always use the same procedure and a similar procedure on the 8753.

It is kind of hard to tell from the colors in your graph, but I would expect the HP 85033E cal kit to be very good but it looked like it still was not as flat as I would expect, assuming it is the top graph line. That said, it would also depend on what cal kit was used to calibrate your Nano VNA.

I've made a few cal standards using 2pc 0402 100 ohm 0.1% non-inductive resistors, soldered to a AMP SMA Male connector. The short was 2 pieces of wire. In  order to keep the reference plane as close to the same as possible for all three standards, I mounted the AMP connectors on my mill, and cut the center connector and outside mounting posts to be 0.020" +/- 0.001"above the plane of the connector shell. These actually are pretty accurate, almost as good as the R&S SMA standards. To be complete, the cal kit should have male and female versions of the SMA connectors, but just male cal standards are probably good enough, especially for the NanoVNA.

The V2 should be here tomorrow and I'll post preliminary results on that unit.

Thanks again for your ideas. It would be nice to get to the bottom of this mystery, but I think spending a lot of time probing the NanoVNA is not really worth the effort. I won;t scrap it though, in case I have time later. I only have a 1GHz bandwidth scope & probes, but maybe that would be good enough.

I liked the YouTube video you linked and they talked about some of the issues and features that make the NanoVNA interesting.. The only thing I did not really agree with was the statement that the unit was usable above 300MHz. Perhaps usable has a different meaning for quick SWR measurements. Their comments about variable quality address widely held concerns about buying anything, expecially test equipment from an unaccountable supplier.

[joeqsmith]

@joeqsmith

I did go in sequence through each step as I later described, top to bottom, in oder, no skips, including ISO and THROUGH. I always use the same procedure and a similar procedure on the 8753.

It is kind of hard to tell from the colors in your graph, but I would expect the HP 85033E cal kit to be very good but it looked like it still was not as flat as I would expect, assuming it is the top graph line. That said, it would also depend on what cal kit was used to calibrate your Nano VNA.

For the problem you describe, you would only use the SOL.   Isolation and thru are not used. 

Your next comment is a bit concerning.  We are looking at S11.  If the loads were perfect, nothing would be reflected back.   If the loads were a short or open, all of the signal would be reflected and we would read 0dBm.   

In the attached graph, you can see how the 2X100 ohm MF 1/4W parts compare with the 85033e and the Minicircuits ANNE.   The Nano wasn't used to collect this data and we are sweeping to 3GHz.   That particular ANNE was one of six that I sorted to find the one with the best performance.    Notice how the axial parts behave as we start to go up in frequency. 

I've made a few cal standards using 2pc 0402 100 ohm 0.1% non-inductive resistors, soldered to a AMP SMA Male connector. The short was 2 pieces of wire. In  order to keep the reference plane as close to the same as possible for all three standards, I mounted the AMP connectors on my mill, and cut the center connector and outside mounting posts to be 0.020" +/- 0.001"above the plane of the connector shell. These actually are pretty accurate, almost as good as the R&S SMA standards. To be complete, the cal kit should have male and female versions of the SMA connectors, but just male cal standards are probably good enough, especially for the NanoVNA.

I just use the PCB standards I've shown for lower frequencies.  They work fine for this.   Above a GHz, home made loads were out of the question.  The shorts and opens were a challenge for me to make.  Wire isn't going to cut it for a short.   

The V2 should be here tomorrow and I'll post preliminary results on that unit.

Thanks again for your ideas. It would be nice to get to the bottom of this mystery, but I think spending a lot of time probing the NanoVNA is not really worth the effort. I won;t scrap it though, in case I have time later. I only have a 1GHz bandwidth scope & probes, but maybe that would be good enough.

I liked the YouTube video you linked and they talked about some of the issues and features that make the NanoVNA interesting.. The only thing I did not really agree with was the statement that the unit was usable above 300MHz. Perhaps usable has a different meaning for quick SWR measurements. Their comments about variable quality address widely held concerns about buying anything, expecially test equipment from an unaccountable supplier.

I've shown using my Nano, which is about a year newer than yours, running a few experiments above 1GHz.   

Maybe someone in this forum, or the Nanovna groups would have a look at it for you.   Normally, when I use the Nano I use the calibration that is built into my software rather than using the Nano's.   I would guess that the open source software would also have this ability.    I wonder if you clear your Nano's cal, then cal it using their software, would it still show the same low numbers.   

Actually, with it being 2 years old, it would be interesting just to see detailed pictures of it.  I wonder if it has changed at all.

[edtyler]

@joeqsmith

Although I could do a one port calibration on the NanoVNA, doing a "full" 2 port is not much more work. But, as you say, my problem is just with S11.

Of course, home made SOL, even made with high quality parts and precise construction are not usable at higher frequencies. However, at frequencies below 500MHz, they work well enough for less critical measurements. For critical measurements, much greater care in calibration is needed and more than just the impedance of each standard is important, as you know.

This is why that less expensive tools are good only for less critical activities. Your comparisons of the NanoVNA V2 to a HP 8753E are interesting and illustrate the "good enough" nature of the NanoVNA V2. For measuring the SWR or complex impedance of an antenna up to 450MHz, it is fine and much better than the original NanoVNA. Another problem with Tools like the NanoVNA is the wide variation in quality of items labeled as NanoVNA or NanVNA V2. A friend bought one of the NanoVNAs and it was a very poor example with sloppy parts placement, cold solder joints and questionable components.

It would be interesting if someone was local to me in Southern California that would like to look at the apparently defective unit with me. Not really interested in sending it anywhere. If it is disassembled, I'll take some pictures and post them here.

I did receive my NanoVNA V2. it does fairly well and does not have the incorrect open impedance measurement issue. However, it measures the 150 ohm leaded resistor as 140 ohms at 146 MHz. Other tooling, including the 8753 measures 150 +/- 1 ohm. Not saying this is an issue, just interesting. My use for the V2 is the same as the V1, casual measurements of antennas and feed lines. Also, measurements of swap meet items like attenuators, coax switches, etc.  Portability and ease of use are key for my uses.

Need to get a more precise understanding of the NanoVNA V2 limits, but it is good enough for now,

[joeqsmith]

@joeqsmith

Although I could do a one port calibration on the NanoVNA, doing a "full" 2 port is not much more work. But, as you say, my problem is just with S11.

Of course, home made SOL, even made with high quality parts and precise construction are not usable at higher frequencies. However, at frequencies below 500MHz, they work well enough for less critical measurements. For critical measurements, much greater care in calibration is needed and more than just the impedance of each standard is important, as you know.

This is why that less expensive tools are good only for less critical activities. Your comparisons of the NanoVNA V2 to a HP 8753E are interesting and illustrate the "good enough" nature of the NanoVNA V2. For measuring the SWR or complex impedance of an antenna up to 450MHz, it is fine and much better than the original NanoVNA. Another problem with Tools like the NanoVNA is the wide variation in quality of items labeled as NanoVNA or NanVNA V2. A friend bought one of the NanoVNAs and it was a very poor example with sloppy parts placement, cold solder joints and questionable components.

It would be interesting if someone was local to me in Southern California that would like to look at the apparently defective unit with me. Not really interested in sending it anywhere. If it is disassembled, I'll take some pictures and post them here.

I did receive my NanoVNA V2. it does fairly well and does not have the incorrect open impedance measurement issue. However, it measures the 150 ohm leaded resistor as 140 ohms at 146 MHz. Other tooling, including the 8753 measures 150 +/- 1 ohm. Not saying this is an issue, just interesting. My use for the V2 is the same as the V1, casual measurements of antennas and feed lines. Also, measurements of swap meet items like attenuators, coax switches, etc.  Portability and ease of use are key for my uses.

Need to get a more precise understanding of the NanoVNA V2 limits, but it is good enough for now,

I don't have the two V2 yet.  OWO and their team keep improving the performance and I am looking forward to seeing some of these newer versions they have.

I have the original NanoVNA that my friend sent me that I used in my reviews.  Of course that old 70's HP8754A, the HP3589A and now a PNA. 

I just put together some software for the old 3589A.   
https://www.eevblog.com/forum/projects/hp3589a-spectrumnetwork-analyzer/msg3203052/#msg3203052

The 8754A with the doubler is good for about 2.7GHz but with my home made standards and what not, about 1.5GHz was it.  I had constructed an up/down converter for it to run experiments in the 5-7GHz range.   Interesting to play with but the results were sketchy at best. 

My PNA is far more advanced than my other network analyzers, it's still an antique and comes with it's own problems.  The hardware is nice with its four receivers but the software looks like a group of fresh outs wrote it that were managed by a fresh out.  Progress in writing my own software to control it has been slow and it still needs some work.   

When I eventually get one of OWO's new units, I may make some sort of review comparing it against my old VNAs.  I am not sure what critical activities would be but I've been impressed with the Nano since getting it.  When I made that demo, I had no idea how it was going to turn out.  IMO, it's one of the best products I have seen in a long time as it allows pretty much anyone access to the tools needed to learn some of the basics.     

I no longer have an amateur radio license.  The only time I have used a VNA to look at an antenna was when I made a demo with the Nano showing how it could be done.   It's certainly possible to use it to make these measurements (and many many others) but judging on the comments I received, its beyond the typical ham.   I hope that anyone interested in the subject (not just hams, but the EEs and beginner electronic hobbyists) would use the opportunity to learn more about it.   

Hopefully someone in your area will step up and give you a hand with your original Nano.  Enjoy the new V2.
 

***
Rough reading....

[joeqsmith]

Just an FYI on the Nano's ability to be used for critical measurements.  Attached is a recent link where I use the Nano to look at an old active probe from Tektronix.  The specified 2.0pF tip capacitance appears correct, until I get above 300 MHz.   S21 errors also increase above 300 MHz.  It's still useful, but you just need to be aware of the limitations.   

https://www.eevblog.com/forum/testgear/hi-z-probe-for-50-ohm-spectrum-analyzer/msg3241778/#msg3241778

[hector.pascal]

Damage to these VNA's from ESD, or mixer overload from nearby RF fields, usually shows up as a very noisy S21 trace.  The "open" sandwich clone versions seem more susceptible to this type of damage than the metal cased ones.  ESD protection is not included in any of the designs (as far as I can see).  This particular problem sounds like different hardware fault to me.  If a firmware reload/update doesn't fix it, sadly it's probably toast