NanoVNA (rev3.4): Correct 2-Port Calibration

This video presents an overview of how to correctly perform a 2-port calibration of the NanoVNA rev3.4 using the included cables. Common pitfalls created by the limitations of the stock NanoVNA kit and common hobbyist misconceptions about RF measurements are addressed to perform a technically correct SOLT+Isoln calibration.
#NanoVNA #RFConnectors #MeasurementPlane
Thumbnail Credit: Francis Fletcher

[@w2aew]

Three other tips to mention: 1) It's a good idea to tap on RESET on the CAL menu, after setting the measurement range and before starting the CALIBRATE process. 2) When attaching adapters, standards, etc., it is good practice to ensure that the male/female parts do NOT rotate when you are attaching them, you should only spin the nut. For example, when connecting the female SMA through, do not rotate the through adapter, only rotate the SMA nut. Spinning the male pin inside the female socket can cause damage to the pin/socket over time. 3) When you can't avoid adding/removing an adapter compared to your DUT connection, you can usually get away with applying a Port Extension (called ELECTRICAL DELAY under DISPLAY>SCALE) to adjust the calibration plane to account for the added/removed adapter.

[@RobertResearchRadios]

Thanks for the constructive feedback. If the user intends to calibrate to a non-default frequency range, they should definitely make sure to set it before pressing the reset and beginning the cal as you say. Thanks for pointing that out to other viewers. About point 2, this video was filmed before I was educated about the danger of turning female connectors inside SMA connectors. I have since found in my experience that this is 100% an important part of ensuring the longevity of my male SMA devices. The complaints about cheap radios using female SMA antenna connectors also appear to be well founded. I make a point of following this advice in my later videos after this point was raised on the first NanoVNA V2 video. With regards to point 3, Do you have some resources for exactly how much electrical delay to apply when accounting for measurement plane shifting as a result of adding additional connectors? For example, can I find velocity factor data and effective physical length when connected to a male SMA connector for SMA female to female adapters? I've always wondered how to deal with connectors of different and sometimes weird types in the calculation of the electrical length of a line. Understanding ideal transmission lines is one thing, but the practical application is difficult when connector length matters. Microstrip lines are easy enough since the line geometry doesn't really change at the junctions, but this was an especially challenging issue with the phasing harness for my 137MHz cross yagi where I used ring lugs as the connectors between my coax stubs. I wonder if for standard connectors there is more of a science to it even when coax connectors are used.

[@w2aew]

@@RobertResearchRadios It is usually easiest to determine the proper electrical length experimentally. Try this... Calibrate the unit over your intended frequency range, then add an adapter or short length of coax. Leave the far end open, and take a look at the Smith Chart. You'll see that at the high end of the frequency range that the trace is not sitting at the 3 o'clock position, but has rotated clockwise. This is due to the extra length added after the calibration plane. So, you can then go into the Electrical Length setting and start adding small values until the arc collapses back down to a single point at 3 o'clock. Now you extended the calibration plane to the end of your added adapter/coax. Of course, this is only an approximation because it assumes a perfect 50ohm impedance all the way to the end. Another way to do this is to look at the Phase of CH0. It will be a flat line after calibration, but will show a slope vs frequency if you add an adapter. Again, adjust the Electrical Length to make it flat again. You could do the same by looking at the Delay of CH0, but it will just show a up/down shift in the flat line - harder to see when you have it properly adjusted.

[@barrykery1175]

Good video. I do mine a tad different. First I replace the 50 ohm load that came with my Nano, that measured 51.7 ohms with a load that was as close as I could get to 50 ohms which was 49.9 ohms. I have two of them. As for the cables, I used one barrel connector and connected both of the cables together. Now when I did my calibration, I added a 50 ohm SMA inline 10 DB attenuator to the cables. I placed it where the one cable connects to the S21 port. Now when I did the calibration I put my loads on the end of the attenuator. I had to use a barrel connector at that point as well. When I normalize my spectrum analyzer I put a a 10 db pad on both my SA port as well as the tracking generator port. I was going to do the same to the Nano but figured one attenuator, properly placed, is enough. My DUT goes where the two cables connect. Barry

[@RobertResearchRadios]

Seems fine by me. Using the attenuator is smart and adds nice protection for measuring active devices too. The NanoVNA does lack a very large dynamic range though, so it can pose a problem for very lossy networks, but I hardly need more than 40dB of dynamic range for hobby passive networks. The standards that come with the NanoVNA are indeed hot garbage, but for most of my use cases the diff between a 49 and 51 ohm normalization isn't so important to me.

[@barrykery1175]

@@RobertResearchRadios The loads that come with the NanoVNA are good enough for the average ham. Surely good enough for HF work checking antennas. It's a great tool and I'm glad I have one. I find more uses for it everyday. Barry

[@WaschyNumber1]

My came with a exact 50 Ohm load, but it's difficult to find other replacements, I have bought a 5pcs pack and they vary between 50.5 Ohm to over 51.xx Ohm 🤔 Maybe the replacement complete calibration Kits are worth to being tested, if they are better quality. 🤔

[@johnwest7993]

I sometimes need to tap on one of the cal 'buttons' more than once. But I've found that the system is actually slow enough that it takes a couple of seconds to complete and indicate a cal function. So I've learned to be patient and wait for a second or 2 after I tap on one of the cal functions to see if it highlights it.

[@ivansavitsky449]

Thanks. Please don't forget to specify the frequency range before you do the calibration as the device resolution is just 101 points across the whole range.

[@RobertResearchRadios]

Yep; this is very important. Thanks for pointing this out for other viewers. If the user wants a more precise sweep, they need to change it BEFORE calibration, not after. If you do it after you have calibrated out the effects of your transmission lines for the original sweep points your calibration will not be valid for the new sweep points. While we like to think that transmission lines will behave linearly and we can just take one point and calibrate out all the effects of the line for all points of interest, they don't always act this way in real life, and it's important to calibrate for each specific frequency point you will measure. Notice that if you adjust the sweep settings after calibration the indicator text on the left hand side of the screen about the save/recall info and calibration status will disappear, showing that you have not recalled or made a valid calibration. This is a courtesy to the user in the firmware of the nanoVNA.

[@scharkalvin]

If you'r running the nano vna at UHF frequencies, then you're right about those adapters. However, If you're using them below 150mhz, the difference in that extra cm of cable/adapter isn't worth worrying about. In my case, I'm using it below 30mhz, so it's really "close enough for government work".

[@RobertResearchRadios]

Definitely agree on the "close enough for government work" at HF and probs most VHF work haha. I've done some work where it mattered at VHF though, but tuning your typical ham radio antenna with a one element match where you're only looking at return loss wasn't it. To figure when it may matter consider 1/400 is basically a gradian with VF=1 for a 1cm long connector at 2m freqs. For VF=.7 it's close to 1 or 2 degrees electrical length. If that matters to your application it matters; if not, whatever works. More power to yah. At HF we can get away with almost anything tehe.

[@RealLifeLifeHacks]

like the SWR of an antenna i have tested is like 1.0 at 144mhz.. but if i hold the VNA the SWR changes to 1.1 at 142.800mhz.. so whats is the real SWR???

[@RobertResearchRadios]

input impedance of a anything other than an ideal resistive termination varies with frequency, so the SWR will change over the frequency bandwidth of any real antenna. The "relevant" SWR is one considered w.r.t your used center frequency. All the measurements are "real".

[@RealLifeLifeHacks]

one big question.. coz the SWR values change if you hold the vna compared to when it just sits on the table.. so what is the proper way to measure SWR of an antenna? you hold the nanovna or not?

[@RobertResearchRadios]

You want to be as far from the antenna as possible and not to be in a space with many reflections off walls and such. Ideally, this is called an anechoic RF chamber. Less ideal would be like a large field. And in hobby work at home, I'd say just get the antenna away from ppl and metal objects in its near field and it's going to be an ok enough measurement for most Ham things.