2021 ~ We are being tricked --That Nano VNA you have may be Unreliable!

 The Terrible Truth about the Nano VNA!

The Nano VNA H4 equivalent from 1953. Don't recognize it? It is a 1953 Hudson Hornet owned by Steve McQueen. Not many Hornets around today so will the H4 still be around in 5 years?


One of the reasons you bought that Nano VNA aside from trying to be associated with rubbing elbows with the professed illuminati is to measure stuff that has inductors in the mix. This is another "we tricked you!"

Unquestionably most solid state homebrew equipment has ferrite and toroid inductors somewhere in the innards and that means winding stuff. But you have to know stuff to wind stuff (inductors). Basic to the winding is first knowing how to calculate the number of  turns to wind on the inductor. 

There are two separate formulas that are used for the calculations, with one for Ferrite and the other for Iron Powdered Cores. The form of the formulas is the same only that the inductance value for iron powdered cores must be in microhenries and the multiplier is 100 whereas the Ferrite the unit is millihenries and the multiplier is 1000. Interesting is that one millihenry is 1000 microhenries. 

Another factor is what is called the Al value  of the core material and the Al values are different for various sized cores. For instance a type 6 (yellow) powered iron core has a different value for a T-50-6 versus a T-37-6. For the anal retentive types the Al values are units of henry's per so many turns. Ferrites tend to have values in hundreds versus Iron Powdered cores that are typically less than one hundred.

The calculation of number of turns involves the value of the inductance divided by the Al value and the square root of this result times the multiplication factor is the number of turns. Let us give an example: the ever popular 2.5 millihenry choke! We have a box of FT-37-43 cores and the CURRENT Al value is 350. So the square root of 2.5/350 = 0.08451542547 [I am using this many digits as the guys with the H4 Nano VNA love to read to 5 places]. Now take that times 1000 and you need 84.51542 turns of wire on the FT-37-43 core to yield 2.5 Millihenry.

A couple of (tricked) problems here and that is that the Al value has shifted in recent years for the FT-37-43 cores. Several years ago it was 420 and then it got changed to 380 and today it is 350. That translates for a 1 millihenry choke from 49 (old) to 53 (new)  turns. 

Also a problem is that which is shown on the Amidon website is wrong! I looked up the Al value for the T-50-43 core and as of yesterday the webpage shows the Al Value as 440 Per 100 turns -- not 1000. I talked to the Chief Engineer and it seems the website is being done by an outside 3rd party (probably in Wuhan China) and they likely thought that there was no difference in Ferrite and Iron Powdered Cores. The Chief Engineer loved my TKT on how to mount Ferrite and Iron Powdered Cores to a chassis. If you want to know send me an email. 

But lets go back to TURNS and an example of why your inductor may be wrong and you will get erroneous readings on your beloved H4 Nano VNA. I have a T-50-6 core and it has 20 Turns of wire on the core --so what is the inductance? Remember the Iron Powdered formula has the inductance in microhenries and the multiplication factor is 100.  The Al value for a T-50-6 is 40. The result is 1.6 microhenry. [ I will do the math as I likely I lost someone 20/100 = (x/40)^.5 so that (20/100)^2 = x/40 so that x = 40*(20/100)^2 = 1.6]

So ok now what wire size do we use and this is the trick trap. This chart (below) is from Amidon and it lists the core sizes and wire size that will fit on that core. For our 20 Turn winding I would pick wire size #22 as the max number of turns is 28. While wire size 20 would work -- the wire size is stiffer and harder to work. But with the size 22 after winding the core leave a bit of a gap between the start and end of the winding and evenly space the turns around the core. 

What you don't want to do is wind 20 Turns of #34 -- although you are winding 20 turns --that is only 1/6th of the possible number of turns and with an almost certainty your 20 turns of #34 will not be close to 1.6 microhenry. This is why the BPF you built and measured with your beloved H4 is no where near centered over the 20 Meter band! You have been tricked into thinking (wire) size doesn't matter --It does!

More to think about with a 2.5 millihenry choke on a FT-37-43 and that is the #34 wire is current carrying capability and the actual wire resistance. I think the continuous current capability is like less than 100 ma. 

Now if you move up to wire size #30 -- much less resistance for the same length of wire and about 3X the constant current capability. So If you went to the FT-50-43 core (Al=440)  then 2.5 millihenry requires less turns like 75.37783 Turns. (Does that match your H4?)  

Those number of turns will fit nicely on the FT-50-43 using #30 wire. So this is a better choice --  same inductance, less wire resistance and more current carrying capability. QED. [That which was to be proved!]


It is like wearing Big Boy pants and to be one of the cool guys you bought a Nano VNA. Now by association you run with the illuminati. When you watch the Nano VNA you tube video's you have this warm feeling in your shorts --because you have one in your shack.

But firstly something needs to be clear as often numbers are thrown around without a cogent understanding of the basis or a clarifications of the units. This leads to untruths -- right out of the gate. The frequent measurement of choice is the dBm or the power gain (or loss) as referenced to 1 milliwatt (that is the "m" in dBm). Measurements made with a DSO (O Scope for the OT's) can be selected in some menu but typically a Peak To Peak Value is displayed. 

However Peak to Peak values cannot be used directly to derive true power. The industry standard is to use the Root Mean Square (Vrms) in power calculations so that Pout = Vrms*Vrms/R. Since we typically use 50 Ohms as the load, the R = 50. 

In going from PTP to RMS -- two actions must take place and the 1st action is to use 1/2 the PTP value as we are only looking at 1/2 the cycle. Next the 1/2 PTP must be conditioned by the 1/2 of the Square Root of 2, or .5 * 1.414 = .707. So Google it and you can graphically see the derivation of 0.707. Now that we have the RMS value which is .707*0.5*PTP, we can square that and divide by 50 (50 Ohms) and we have power. BUT wait that answer is in Watts. So we now must multiply that value by 1000 to get milliwatts. 

The dBm values requires dividing Pout in Milliwatts by 1 milliwatt (our reference). The dBm is a log function so that now we have the expression log(Pout mw/1mw) and since it is a power measurement we must multiply the log value by 10.

Recently I was looking at some exciting work being done by my good friend N2CQR on a project called the Myth Buster. See  Soldersmoke.blogspot.com. On a posted drawing he had a note 0.453 volts across 47 Ohms is close to  7dBm. So that started me on the derivation process. 

The spreadsheet below does the heavy lifting in going from Peak to Peak to RMS to dBm at tenth volt increments to 3 volts. Of note is that Bills 0.453 volts is an RMS value and that is akin to 1.3 Volts Peak to peak and thus just slightly less than 7 dBm.

Now here is a TRUTH revealed. You can take your scope's (O or DSO) Peak to Peak Value and simply square that value. With that result multiply by 2.5 (got it) next take the log and multiply by 10. So OK my Scope (DSO) reads 1.414 Volts. Squaring that we get 2 and then multiplying that by 2.5 = 5, we get the value  (drumroll) 10*log(5) = 6.989 dBm rounded to 7 dBm. Lets try 2 volts PTP and that yields 10*log(10) = 10 dBm.

So for just 50 Ohms --you can get dBm with a few calculations using PTP. The 2.5 factor was not the result of a hemp mist party so prevalent in SoCal; but is an accounting of milliwatts, 50 Ohms and the square root of 2 somewhere in there. I know I will get an email asking how. So

Some Important News to those who are always calculating. You absolutely have to be measuring 44.72 Volts PTP across 50 ohms to proclaim I have 5 watts out. 
44.72^2 = 1999.878 and now X 2.5 = 4999.696 and dividing that by 1000 to get watts we have 4999.696/1000 ~ 5 watts. But we are not done yet so if you take 10*log(5000) = 
36.989 dBm or 37 dBm. 

So you use WSPR in WSJTX and must set your power level to announce to the world what you are transmitting -- for 5 watts it is 37dBm -- now you know how that got there.


My story with the untruth of the Nano VNA begins a year ago when a dear friend gifted me a Nano VNA. It was like now I could roam with the Cool Guys who literally carried around the H4 in their shirt pocket always at the ready to make measurements and espouse how good a BPF it was or that their 4 pole filter had straight skirts. 

Well the first untruth -- a guy like me with FFS (Fat Finger Syndrome) can no way operate the unit as a standalone. Did I also mention that Navigation with a stylus is like a hunting for Nemo with a Treasure Map. A wrong turn and you have to go back to square one.

Thus the connection to a computer. The Nano VNA Saver software appeared as a solution to FFS  -- BUT my unit would not reliably connect the host computer. When you tried to connect you would get an error message that Microshaft was being alerted and then the program would close. Needless to say that was a huge lack of a confidence builder. I did get it to work once or twice but then said: "Not ready for Prime Time"! It now could put on that special shelf where you place questionable devices.

Besides there was life on earth before the Nano VNA and I had optional methods for measuring BPF's both in hardware and software simulation -- it was not the end of the world!

Then just a week ago another offshore ham sent me some boards he made to test crystals for use in a crystal filter. So I thought maybe I would give the Nano VNA another chance at redemption. 

Now I have a very nice Windows 10 machine in the shop  that is the size of two CD case stacked and so I loaded Nano VNA Saver on that jewel and when I connected, it did not crash as before. But when you would run the sweep only one out of 10 tries would actually produce a plot -- typically you get an error message "too many tries". Finally I once got a plot of a 6 pole filter that looked like this. (But could never repeat it.)

That is a plot of a 6 pole filter (below) I built using the Dishal method.

The filter Cf is near 11.5 MHz and was intended for use in a Dual Conversion SSB transceiver similar to the architecture I used in my KWM-4 transceiver. [See my QRZ page if you want to be dazzled by a brilliant piece of engineering -- the KWM-4]

But unable to get the Nano VNA to reliably sweep, I bought another H4 unit. Here is where the prices range from $49 to $200 for the same H4 -- I picked something Mid-range with excellent reviews.

Well it arrived yesterday and that same filter above resulted in plot that looks like this which is similar but different! This one can be repeated and you get data.

The BW (-3dB) is 2.15 kHz (Like a Collins 455 kHz Mechanical Filter) and the -6 dB BW is 2.37 kHz and calculated from the data the  -60 dB band width is 9.7 kHz so I think they used to measure the shape factor as a ratio of the -6dB to -60dB so that gets you 9.7/2.37 = 1:4 --not sure of this; but if that  is the construct then it is better than a 1:6 .

I also have a 4 Pole filter I built using the Desperate (Dishal) Software with the same series of crystals, 11.5 MHz as the devices of choice. This makes clear why you want something better than 4 crystals.

Two obvious points (well maybe three) and that is the  (-3dB) BW is more narrow and two the skirts are flared at a lesser dB level. The 6 pole has better than 10 dB of skirt before flaring so that  translates to better rejection of close in signals. Go with 6 poles or more and forget that 4 pole BS. The FLEX Radio 40M SDR Police will give you a star on the spectrum, because that same rejection works on transmit too!

But then I took a look at the plot and it was sort of like Maggie's Drawers (A term used on Military firing Ranges when you miss the target-- I don't understand that connection but when I was in the US Navy --lots of things did not make sense.) I opened up the measurement window to look at more of the Spectrum and you get a different answer.

Now we can see that the -3dB bandwidth is 2.5 kHz and he -6dB is almost 2.8 kHz which is telling that this filter likely is not as good at the 6 pole. So my comment again go for the 6 Pole. 

Thus my conclusion is that both filters will work in a rig. But the 6 pole offers advantages for out of band pass signal rejection. The lesson learned --look at enough of the spectrum before making a suitability decision.

[Note this four pole filter is installed in a homebrew radio which by the way is dubbed Maggie May as she has a lot of various crystal filters pass through her. But the old girl does well as you can hear in the video at the front of this posting. Using the data from the Nano VNA it was really easy to simply poke in (Maggie May's ears just perked up) the BFO frequencies into the sketch.]

I next moved the second H4 to a 40M BPF that I know works and its design is almost (I said almost) like the Bitx40 only I use a combination of fixed and trimmer caps to tune the BPF. 

The 2nd H4 is obviously operating because I could tweak a trimmer and instantly see a plot of that result. My that was refreshing, although still bothered by how accurate was it. Mind you I did nothing about terminations nor run any of the calibrations -- I just was anxious to see if it connected and could I sweep plots. With the trimmers using  a non-magnetic adjusting tool I could make a small change and run a plot in a matter of seconds. 

What is really cool is that you can run the analysis on a BPF and it will tell you the bandwidth and the 3 dB points. This filter will need to have its Cf shifted from 7.068 to 7.150 MHz and then it would be centered over the band -- But once all is calibrated and terminated properly I can do that. What also is evident --this BPF is too "peaky" and more work is needed to flatten the Band Pass. But now with a properly working Nano VNA (an assumption)-- I can fine tune the response.

But that may be a second untruth --that you believe your Nano VNA. I see a close similarity between the 1st and 2nd measuring the 6 Pole filter; but the 1st is suspect as I can't repeat the test. 

A long time ago (reaching back to my Aerospace roots) the truth was: In God we Trust all others Bring Proof! That is why we had Metrology Labs with laboratory grade standards to evaluate the Calibration of the test equipment. Relying on an internal calibration is questionable and puts us in the league of: Trust me your Tesla can drive from San Diego to Eureka CA (764 Miles) without a recharge! That is simply untrue. 

Could the Nano VNA be trusted? Well my personal bent is to be towing a small trailer behind that Tesla with a spare battery and a transfer switch to make that 764 mile trip. There is a chasm of difference between absolutely accurate to very useful. I found the Nano VNA to be useful in making circuit adjustments to put the BPF in the right lane. BUT keep in mind mathematical answers to 5 or 6 decimal places -- is only as accurate as the data acquisition ahead of it. 

What I have seen is that the Nano VNA is touted as perhaps a far greater discovery than what happened in St Louis (the invention of the bread slicer). What I have not seen (it may exist but I haven't seen it) is the accuracy of the data acquisition to give those 5 decimal place answers. A 10% error at 7 MHz -- is 700 kHz. 

So look at the hard truth -- the Nano VNA is a slick tool; but do not put it on a pedestal. My 1st H4, it is quite apparent is not functioning properly but gave me a plot like I wanted to see. But with a high degree of certainty is likely inaccurate! Keep in mind the 2nd H4 Connects and Sweeps reliably -- so which one is true?

There is another sad truth -- a proliferation of knock offs that are being sold cheap and simply don't work. Caveat Emptor. What you have in your wallet may not be a Capital One Card!

The Terrible Truth: Your Nano VNA may not be working properly. Trust but Verify! So it is the wise man who looks not at the slick advertising or the plethora of you tube videos or that you believe as Gospel that presentation at the local ham club. Trust this --always look at what is too good to be true --literally always isn't. 

Should you buy a Nano VNA? That is a personal decision; but likely the $39 special may have issues. Look at the specifications in detail and look at the ratings of actual user's. Insure that your test setup is "kosher" with a calibration and proper termination and only then think about the YL who says she is safe -- likely she is not! (The "safe" point is not about Covid19). 

Now on to watching at least 1/2 dozen videos on how to calibrate the Nano VNA using the Nano VNA Saver software. Then again that pesky problem -- not a really independent calibration.

Pete, N6QW

Popular posts from this blog

New Technology for 2020 ~ The Hermes Lite 2.0 SDR Transceiver

21st Century Homebrew SDR SSB Transceiver Project

New Technology for 2020 ~ A Direct Conversion Receiver You Can Build