nanoVNA Real Resistance Measurement Range

Tom VA7TA
 

Greetings All,

About a year ago I made up a series of minimized reactance, precision RF resistance loads (see photo attached) for testing an antenna analyzer that I built based on a design by Yury Kuchura, EU1KY (1). It is a single port analyzer specified for 500KHz to 150 MHz that has harmonic extended coverage to 450 MHz. Similar to the nanoVNA it uses the Si5351 as a signal source. Of particular interest is that it has firmware that offers some unique calibration features.

A photo of my diy precision reference loads which are built upon BNC male PCB connectors and can be directly connected to my EU1KY analyzer is attached. The resistance elements consist of four paralleled, 1%, 0805 SMD resistors mounted like wheel spokes on tiny PCB disks with the hub consisting of the BNC connector center pin. Values of 5, 10, 50, 100, 300, 500, 1000 and 2000 Ohms were built. I found that my 50 Ohm BNC load, which must be connected via a BNC/SMA adapter to the nanoVNA, was a close match to the supplied 50 Ohm SMA load with a full 900 MHz sweep of the spectrum.

In consideration that some day I may wish to use the nanoVNA for determining component values needed for matching various load impedances I decided to run a series of tests using my BNC loads to determine the accuracy of the nanoVNA over a 5 to 1000 Ohm range of load values. Additionally I decided I would make a comparison to my EU1KY antenna analyzer which uses a modified SOL calibration technique.

As the measurement of high Z values within the VHF and UHF bands is challenging to say the least it was not surprising to find that the accuracy falls off for high resistance loads for those higher frequencies. As I understand it this loss of accuracy is expected for all VNA instruments that use similar measurement techniques(2). Apparently the method used here is optimum for our most common low Z transmission line measurement needs thus it is the best choice for this application.

In preparation for this test I first calibrated my nanoVNA with the BNC/SMA adapter attached. For the OPEN step I attached an empty BNC PCB male connector and I used my BNC 0 Ohm reference for the SHORT. This was done to include compensation within the calibration results for the BNC adapter and connectors. Screen captures of the OSL calibration Smith charts and all the other load measurements made subsequently are attached along with the s1p Smith chart data files for those that may wish to take a closer look at the data.

I arbitrarily decided that 10% measurement accuracy of the real R component of a complex impedance would provide sufficiently accurate data for most practical antenna Z matching endeavors. I then noted the frequency threshold where the accuracy fell off to 10% for each precision load value. Using this data I software generated the attached horribly(!!!) crude and vague graph. The graph illustrates the approximate frequency/load resistance limit for 10% or better measurement accuracy. As long as the measured value is below the curve at the measurement frequency then the probable accuracy for the real component is roughly within 10%.

Load values below 50 Ohms were found to be roughly within 10% across the entire 900 MHz sweep. The accuracy for up to a 1000 Ohm load was within 10% up to about 125 MHz. For the 2M band the 10% accuracy limit was found to be roughly 500 Ohms. For 70cm the 10% accuracy limit was found to be in the neighborhood of 150 Ohms.

For measurements that land above the graph curve it was found that the accuracy quickly deteriorated. For example the 1K load measurement drops to a real R of 300-400 Ohms between 400 and 500 MHz. The 500 Ohm load measured around 250 Ohms for the real R. Thus for 70 cm the error was 2:1 or worse for both the 500 and 1000 Ohm hi-Z measurements.

In comparison when using my EU1KY antenna analyzer I found the 450MHz measurement of the 1000 Ohm load was well within 10% (930 to 1010 +/- j30). The EU1KY firmware supports the use of modified values for the SHORT and OPEN calibration steps which can be user selected. I have several different calibration OSL combinations saved. For this measurement I used the 300/5/50 Ohm OSL which is my default choice. It seems to provide me the best load resistance range accuracy. Those that have the necessary programming and math skills might find it interesting to take a look at the EU1KY antenna analyzer calibration techniques - I believe the firmware code remains open source.

I wish to emphasize that my intent for posting this information is mainly to raise awareness about the limitations and possible pitfalls associated with direct VHF/UHF high Z measurements. It is not my intent to cast a shadow on this fabulous nanoVNA innovation. I consider the nanoVNA to be a wonderful instrument that provides a useful and affordable tool for radio amateur's interested in antenna, feedline and RF matching measurements. I suspect the nanoVNA load range performance described here may very well be better than would be found with many expensive commercial antenna analyzers. Being a two port analyzer it has functionality that the EU1KY analyzer does not offer and costs much less. It may be possible to use 1/4 wave transmission line stubs for indirectly measuring high impedance's. I find the nanoVNA is a delightful little instrument to play with!

1)https://bitbucket.org/kuchura/eu1ky_aa_v3/wiki/Home
2)https://www.mwrf.com/test-measurement/make-accurate-impedance-measurements-using-vna-pdf-download


--
73 & Enjoy!
Tom, VA7TA

Join nanovna-users@groups.io to automatically receive all group messages.