There's also a bunch of '50' ohm loads around from old 10base2 networks where things were less critical, I've seen quite a few that actually registered as 47 ohms
Just a word of advice. Never "spin" on a connector. As a example when you installed the first SMA-M to BNC-F adapter to your VNA you initially tried to hold the BNC end while you tightened the nut. Had some difficulty and just spun it on. This destroys the center pins in the connector. Not something you would ever notice at HF /VHF/UHF frequencies but get into the microwave range and just doing that a couple of times will make a noticeable and measurable difference. The easiest way to destroy connectors is to allow the center pin to rotate while tightening the nut. Get caught doing that in a metrology lab and they would probably fire you on the spot for destroying their $3K cal standard. Had a friend who works at a metrology cal lab catch me doing just that one day at my shop when he was visiting and I got the lecture. I don't work on equipment at frequencies over 10GHz (honestly anything above the 23cm amateur band is usually just my microwave amateur radio setup) but he showed me some literature that demonstrates that at very high frequencies even not properly torqueing connectors make a significant difference. Mike KC3OSD
"Never spin-on a connector..." Yeah yeah yeah... But ! If you look at the NanoVNA and the little tiny SMA 'Open', 'Short', and 'Load' that come with them, you'll soon discover that there's no practical way to connect the 'Short' without just "spinning" it on. Good ahead, prove me wrong. Same with the little tiny 50-ohm 'Load' provided, the entire SMA 50-ohm load is too small to be practical to connect without spinning it on (What? You seriously want me to use needle nose pliers, Geesh!! Get real.). The Open doesn't matter since it lacks a center pin. So yes, you are correct. But the obvious rebuttal is staring us in the face and it must be acknowledged. Your advice is simply not practical with small SMA devices that lack the hand-hold of (for example) a cable. So, what I did was to purchase a pair of SMA-male to SMA-female 'adapters', also known as connector savers. Now my under $100 NanoVNA-H4 is protected with the addition of ~$10 set of connector savers. Even I'm not sure that it was actually a good investment. Spending $10 to protect a gadget that only costs $100, and I could even fix quite easily if I ever wore out the connectors? Hmmm... Dubious investment. So, from now on, you can temper your advice with this staring-us-in-the-face exception. Cheers!!
It never occurred to me that there could be such a large difference in quality between terminators. I'm going to try this on the terminators I have. Great video.
I especially appreciated the use of the delay format as a way of knowing what electrical delay to use. I built some 30db PI attenuators as suggested by w2aew but they didn't look as good as I expected. The smith chart seemed to indicate that I needed an electrical delay so I played around with that but never found a good value. Now I have a new trick, so I'll go at it again. Thanks.
Important point that must be explicitly mentioned more often: People with NanoVNAs spend their waking hours repeatedly calibrating them with the 'Open', 'Short', and 'Load'. Very nice. But for testing and adjusting most simple antennas, the frequency calibration depends only on the internal TCXO, and I don't see anyone calibrating the TXCO frequency 127 times a day. (Of course the frequency is more than precise enough; yeah got it. I'm trying to make the point clear here.) Point being, if all you're doing is calibrating the length of your 80m dipole, then the only parameter worth looking at is where the impedance dip is located in the frequency spectrum. You can perform this measurement using a NanoVNA even if the amplitude and phase calibration are a bit off. You're only looking for the dip, and trimming the antenna so the dip is where you want it in the band. (This presumes that you've not set the Impedance References to something odd like 300 ohms.) Me thinks people are bringing precision amplitude and phase calibration into the mix far too often. Likely not required in many use cases. Something to consider.
That was a great video. Full of useful information. I liked how you zeroed out the smith chart with the delay. I learned something. I also found that all termination loads are not equal. I mostly focused on only SMA figured that would have least chance of having any stray reactance because of their size. I used a spectrum analyzer with an RF sampler to make my tests and none were as advertised. Not all were bad, just not up to spec. Thanks for sharing. I really enjoyed your video. Barry, KU3X
PS...... forgot. Isn't you Nano only as accurate as the calibrated load use to set up the VNA? In other words, if the load that came with the unit was used and it would be a not so good load, would that not make all testing done after that no accurate? Barry
Have you done a video showing duplexer and/or multicoupler tuning, the type used in LMR repeaters? I’d REALLY like to see how well these little nanos do for that purpose.
This is great. What is your calibration procedure? Do you calibrate directly onto the vna without the adapter or with it on? I'm assuming the delay is turned off during calibration. Also, how can you display a large frequency range at once? My understanding is nanovna has a low dynamic range since it only sweeps 101 points. Thanks for the help.
Yes delay off while cal, normal cal. added delay for the SMA/BNC adapter since it adds electrical length. The 101 point does not affect dynamic range. It is just point to point accuracy but since loads are very broad band you don't need a fine detail cal. Just a note: if you put the Dislord firmware in your NanoVNA it extends the calibration number to 401.
My recently-purchased NanoVNA-H4 came with the 401 points option, so I leave it set to that all the time. If it slows the plots down, then I haven't noticed or been bothered in the slightest. It's an amazing device, and I find it better than the $10,000+ R&S FSH-3 that we've used at work before, a device that is relatively awkward to use and (in transform modes) shows MORE artifacts (digitization artifacts). Every ham radio operator should own one, if not several.
great video! Incredible how they can manufacture a 50 ohms resistance that has no significant reactive component up to 1GHz!?!?!? We live in the future!
@@vettonator You almost certainly do not need a "better" set of 'Open', 'Short', and 'Load', but you should probably order up a spare set as they so tiny that it's fairly likely that you'll eventually lose one. Cheap insurance. Was it $6 I think ?
Except 'Primary' standards. E.g. the "Prototype" kg lump in Paris by definition had (for many decades) precisely zero tolerance. I love to mention the exceptions, as it shakes up the assumptions, and that's a good thing.
50ohm should be 50ohm regardless but goes to show tolerances, heat and even 50.0001ohm can change further up the spectrum. not all 50ohm is actually 50ohm.
There's also a bunch of '50' ohm loads around from old 10base2 networks where things were less critical, I've seen quite a few that actually registered as 47 ohms
Just a word of advice. Never "spin" on a connector. As a example when you installed the first SMA-M to BNC-F adapter to your VNA you initially tried to hold the BNC end while you tightened the nut. Had some difficulty and just spun it on. This destroys the center pins in the connector. Not something you would ever notice at HF /VHF/UHF frequencies but get into the microwave range and just doing that a couple of times will make a noticeable and measurable difference. The easiest way to destroy connectors is to allow the center pin to rotate while tightening the nut. Get caught doing that in a metrology lab and they would probably fire you on the spot for destroying their $3K cal standard. Had a friend who works at a metrology cal lab catch me doing just that one day at my shop when he was visiting and I got the lecture. I don't work on equipment at frequencies over 10GHz (honestly anything above the 23cm amateur band is usually just my microwave amateur radio setup) but he showed me some literature that demonstrates that at very high frequencies even not properly torqueing connectors make a significant difference. Mike KC3OSD
"Never spin-on a connector..." Yeah yeah yeah... But !
If you look at the NanoVNA and the little tiny SMA 'Open', 'Short', and 'Load' that come with them, you'll soon discover that there's no practical way to connect the 'Short' without just "spinning" it on. Good ahead, prove me wrong. Same with the little tiny 50-ohm 'Load' provided, the entire SMA 50-ohm load is too small to be practical to connect without spinning it on (What? You seriously want me to use needle nose pliers, Geesh!! Get real.). The Open doesn't matter since it lacks a center pin. So yes, you are correct. But the obvious rebuttal is staring us in the face and it must be acknowledged. Your advice is simply not practical with small SMA devices that lack the hand-hold of (for example) a cable.
So, what I did was to purchase a pair of SMA-male to SMA-female 'adapters', also known as connector savers. Now my under $100 NanoVNA-H4 is protected with the addition of ~$10 set of connector savers. Even I'm not sure that it was actually a good investment. Spending $10 to protect a gadget that only costs $100, and I could even fix quite easily if I ever wore out the connectors? Hmmm... Dubious investment.
So, from now on, you can temper your advice with this staring-us-in-the-face exception. Cheers!!
Also, some small devices don't even include a slip surface between the hex flats and the center pin. They inherently must spin on. Worth mentioning.
It never occurred to me that there could be such a large difference in quality between terminators. I'm going to try this on the terminators I have. Great video.
The impedance of feed through terminations will be dominated by the instrument they are connected to. I have a video that shows that.
This is a simply-fantastic video. Thank you.
When you put a 50 ohm through on a scope there is the input capacitance on the scope that will impact the termination.
Nice Video!!
I especially appreciated the use of the delay format as a way of knowing what electrical delay to use. I built some 30db PI attenuators as suggested by w2aew but they didn't look as good as I expected. The smith chart seemed to indicate that I needed an electrical delay so I played around with that but never found a good value. Now I have a new trick, so I'll go at it again. Thanks.
Great advice, super instrumentation tutorial.
Important point that must be explicitly mentioned more often:
People with NanoVNAs spend their waking hours repeatedly calibrating them with the 'Open', 'Short', and 'Load'. Very nice. But for testing and adjusting most simple antennas, the frequency calibration depends only on the internal TCXO, and I don't see anyone calibrating the TXCO frequency 127 times a day. (Of course the frequency is more than precise enough; yeah got it. I'm trying to make the point clear here.)
Point being, if all you're doing is calibrating the length of your 80m dipole, then the only parameter worth looking at is where the impedance dip is located in the frequency spectrum. You can perform this measurement using a NanoVNA even if the amplitude and phase calibration are a bit off. You're only looking for the dip, and trimming the antenna so the dip is where you want it in the band. (This presumes that you've not set the Impedance References to something odd like 300 ohms.)
Me thinks people are bringing precision amplitude and phase calibration into the mix far too often. Likely not required in many use cases.
Something to consider.
That was a great video. Full of useful information. I liked how you zeroed out the smith chart with the delay. I learned something. I also found that all termination loads are not equal. I mostly focused on only SMA figured that would have least chance of having any stray reactance because of their size. I used a spectrum analyzer with an RF sampler to make my tests and none were as advertised. Not all were bad, just not up to spec.
Thanks for sharing. I really enjoyed your video.
Barry, KU3X
PS...... forgot. Isn't you Nano only as accurate as the calibrated load use to set up the VNA? In other words, if the load that came with the unit was used and it would be a not so good load, would that not make all testing done after that no accurate?
Barry
Yes that is right. That is why cal standards are $4000 or more
Have you done a video showing duplexer and/or multicoupler tuning, the type used in LMR repeaters? I’d REALLY like to see how well these little nanos do for that purpose.
I don't have a duplexer. Alan W2AEW did a good video and showed the nanovna is not good at duplexer tuning
This is great. What is your calibration procedure? Do you calibrate directly onto the vna without the adapter or with it on? I'm assuming the delay is turned off during calibration. Also, how can you display a large frequency range at once? My understanding is nanovna has a low dynamic range since it only sweeps 101 points. Thanks for the help.
Yes delay off while cal, normal cal. added delay for the SMA/BNC adapter since it adds electrical length. The 101 point does not affect dynamic range. It is just point to point accuracy but since loads are very broad band you don't need a fine detail cal.
Just a note: if you put the Dislord firmware in your NanoVNA it extends the calibration number to 401.
My recently-purchased NanoVNA-H4 came with the 401 points option, so I leave it set to that all the time. If it slows the plots down, then I haven't noticed or been bothered in the slightest.
It's an amazing device, and I find it better than the $10,000+ R&S FSH-3 that we've used at work before, a device that is relatively awkward to use and (in transform modes) shows MORE artifacts (digitization artifacts). Every ham radio operator should own one, if not several.
great video! Incredible how they can manufacture a 50 ohms resistance that has no significant reactive component up to 1GHz!?!?!? We live in the future!
much much more than 1GHz. the SMA load I use for calibration is rated at 18GHz. You need better connectors to go higher, 100's of GHz
It supposes you have calibrated the VNA with the best load to compare with the others !
yes. I check against a very good load that I've put on a 20GHz VNA.
how bad was the one that came with your nanovna?
quite good. I compared it to an expensive one.
@@IMSAIGuy thanks. I'm thinking about getting a nanovna and wondered if I needed a better calibration load.
@@vettonator nope
you can measure it with a multimeter to see if yours is 50 ohm
@@vettonator You almost certainly do not need a "better" set of 'Open', 'Short', and 'Load', but you should probably order up a spare set as they so tiny that it's fairly likely that you'll eventually lose one. Cheap insurance. Was it $6 I think ?
any manufactured product has a tolerance.
yes, all the big names of course. they will usually specify VSWR. there will be an actual datasheet. they range around 1.05:1 TO 1:3:1
Except 'Primary' standards. E.g. the "Prototype" kg lump in Paris by definition had (for many decades) precisely zero tolerance.
I love to mention the exceptions, as it shakes up the assumptions, and that's a good thing.
👍🍟
50ohm should be 50ohm regardless but goes to show tolerances, heat and even 50.0001ohm can change further up the spectrum.
not all 50ohm is actually 50ohm.