Introduction

Vector Network Analyzers are low cost nowadays – they can be found integrated inside spectrum analyzers so that a single tool can be used for a range of requirements that an engineer would have, such as discovering and solving EMC-related issues with the spectrum analyzer portion, and creating better-performing wireless receivers and transmitters with the aid of its VNA function. Standalone VNAs are available at decent price points, starting from under $100.

This blog post shows how to get started with a VNA, through to the point of performing very useful work with it. No RF knowledge is required to follow the blog post, and there are no equations to use beyond the level of Ohm’s Law.

When reading this blog, any text in italics is background information that can be ignored if desired. Everything else is in simple, plain language.

By the end of the blog post, it will hopefully be clear what core functions a VNA performs and how to use a VNA for successful, well-performing RF projects. I’ll demonstrate with an FPC1500 VNA (see the FPC1500 review here), but all the information applies to any VNA; I won’t use any instrument-specific features.

Briefly, what is a VNA? And what is a Smith Chart?

When delivering maximum power from a DC source battery with internal resistance (let’s call it R_s) to a destination load resistance R_L, you’ll notice, if you sit with pen and paper and use Ohm’s Law and Power=Voltage x Current, that the destination receives maximum power when R_s is equal to R_L. The source and destination resistances must match for maximum power transfer to the destination. And if you didn’t know what the resistance was, the tool you’d use would be an ohmmeter, which measures resistance using the formula R = V/I.

The same thing applies to AC signals, except that we have to match impedances because sources and loads are not purely resistive. An impedance has two components: the DC resistance as before, and the AC resistance (known as a reactance to distinguish it from resistance) caused by inductance or capacitance. The sine-wave AC voltage and AC current are in-phase for resistances but not for capacitance or inductance; therefore, the impedance formula has to use vectors for V/I rather than the resistance formula, which uses scalar V/I, i.e., the familiar Ohm’s Law.

The rightmost diagram below shows an arrow or vector with a short horizontal component, which is the resistance, and a vertical negative component, which represents capacitive reactance (inductive reactance would be indicated if the arrow was pointing above the horizon).

Even if a circuit doesn’t use resistors or capacitors, at high frequencies, even a short length of wire can have significant reactance; therefore, the inductances and capacitances within the circuit might not be visible as physical components soldered to the circuit board!

In any case, it doesn’t matter how complex the circuitry is and how invisible the inductances and capacitances are because if the circuit is treated as a black box with a two-pin connector at the end (for instance, a coax connector), we can call that connection a port, and instead of just measuring its resistance to DC with an ohmmeter, we can measure its impedance with a VNA.

Just as a side note, when people say an “impedance is 50 ohms,” they mean that the impedance is 50 ohms resistance and there are zero ohms reactance. If you ever see something written as (say) 40 + j20 ohms, then that’s an impedance with 40-ohm resistance and 20-ohm inductive reactance. If the imaginary component was negative, such as 40 - j20 ohms, then that’s 20 ohms of capacitive reactance. Ordinarily, you’d also need to mention the frequency at which the impedance value is relevant since it is frequency-dependent. For background information, the reactance X in ohms equals 1/(2*pi*f*C) or 2*pi*f*L. You don’t need to know this to follow the blog.

Impedance measurement is important for communications circuits because it’s all about getting the energy from the transmitter circuit into the antenna and into free space and, at the other end, getting the energy from the receiving antenna into the circuit for amplification and decoding. To maximize power delivery, you want the source and destination impedances to match between different stages closely.

The block diagram below of a typical radio receiver shows many circuit building blocks. If they all had a different impedance to each other, there would be a lot of loss through the system. All the green connections carry RF signals, and the blocks usually have a specific impedance, usually 50 ohms, but different values are possible for some blocks. At the antenna end (shown with an orange line), the impedance may be (say) 38 ohms, 75 ohms, or something completely different, and a VNA is used to determine its impedance so it can be matched to the radio receiver.

Since impedance varies with frequency, you will want to be able to measure it at various frequencies of interest and chart the results. The most popular chart is a Smith chart, which represents the DC component of the impedance, i.e. resistance, on a horizontal line, and reactance (positive for inductance, negative for capacitance) either north or south of the horizontal line, following curves.

The precise values do not often matter; sometimes, all one cares about is that the measurements are approximately in the center.

For background information, on the chart above, the numbers are normalized to 50 ohms, which just means that any point on it needs to be multiplied by 50. The blue center dot is at the horizontal 1 position, which means 1*50 = 50 ohms. The yellow dot represents 0.2*50 = 10 ohms with 0.5*50 = 25 ohms inductive reactance. You could write it as an impedance of 10+j25 ohms. The green dot is left as an exercise for the reader.

There are more detailed descriptions of the Smith chart elsewhere. All that’s important for now is to know that any impedance measurement gets represented by a point on the chart, and the center of the chart represents a pure 50 ohm DC resistance. Normally, RF circuits operate with 50-ohm inputs and outputs, so when checking such circuits with a VNA, you want to see your measured value at your frequency of operation, looking like a dot sitting in the center of the Smith chart. In real life, circuits are not that well matched to 50 ohms, and you may see the point away from the center. Since a VNA can measure the impedance for a range of frequencies (it will sweep across a user-defined range), instead of a dot, you will most likely see a line on the Smith chart, with one end of the line representing the impedance at the starting frequency, and the other end of the line representing the impedance at the ending frequency. There's no scale on the line, but all VNAs support markers (they behave like oscilloscope cursors), which can be added to the Smith chart to show frequency values.

The photo here shows some examples of RF circuits or building blocks and the impedance you might expect (at the frequency or frequencies of operation) when measuring them. The ceramic filter in the photo is a 3-pin device where the two outer pins form the input and output ports, with the center pin being used as the ground reference for them.

Ordinarily, the circuit is unpowered for measurements (otherwise, it will severely damage the VNA). In some circumstances you may need to power up the circuit (for example, where the circuit may switch things at the port you’re trying to measure, but generally you need to be extremely careful about that).

To summarize, a VNA will allow you to disconnect and probe into circuits, treating them as black boxes, to obtain a scan of impedance measurements across a range of frequencies. The information is often displayed in Smith Chart form.

Solving a Practical Problem: Impedance Matching

From the position of points on the Smith chart, a VNA user can recognize what one needs to do to better match the desired value (usually 50 ohms, but it can be other values). A neat trick is to provide a better match by simply adding an inductor and/or a capacitance to the port. It has the effect of shifting the point on the VNA, at a given frequency, toward the desired impedance value. No energy is wasted in such a match because inductors and capacitors do not consume energy; they store and release it during the AC cycle, i.e., average energy consumption is zero for ideal (theoretical) components. Sometimes a match is required across a range of frequencies; that requires different types of circuits; transformers could be used.

The photo and schematic below show what impedance matching can look like on a real board.

Here’s a photo of the antenna on a Raspberry Pi 3B. It looks like there is no matching circuit nearby..

.. however, it’s on the other side of the board. You’ll also see two large copper areas where the VNA was connected to perform the antenna matching. The connector that would have been soldered there has its copper trace directed toward the antenna-matching components, using either a zero-ohm resistor or a sufficiently large capacitance (to keep the reactance low at the frequencies of interest). The large white object is most likely a filter.

Now that it’s hopefully reasonably clear what a VNA can do, you may wish to see it in action, solving a real-world problem. In the video, the problem was to take a component, a ceramic filter with 330-ohm impedance (according to its datasheet), and match it to 50 ohms. The procedure to perform a match using a single inductor and capacitor is shown in the video.

If you watch the video, you’ll notice that before the VNA is used for the impedance measurement, I did a seemingly weird thing where I ran a calibration step with either nothing connected to the end of a cable, or shorting the cable, or attaching a 50 ohm resistor to the end of it. That calibration step is critical to perform with a VNA prior to an impedance measurement. The reason for it is discussed next.

Edit media

Dimensions x Small Medium Large Custom

Subject (required) Brief Description Tags (separated by comma) Video visibility in search results Visible Hidden

Parent content

FPC1500 Spectrum Analyzer - Review Vector Network Analyzer (VNA): A Practical Getting-Started Guide for the General Reader Comment on NanoVNA - Antenna testing Comment on RoadTester Poll: What Are You Purchasing or Planning to Purchase in the Near Future?

Poster

Upload Preview

Ok

Cancel

Open, Short, Load (OSL) Calibration

A VNA is a highly sensitive and accurate instrument! So you may wonder why it needs calibration every time it is used! Calibration is a misleading term. What you’re really doing is ‘calibrating out’ connected leads (usually coax cable) and connectors. To perform an impedance measurement, internally, the VNA generates a signal at different frequencies and sees what gets reflected back by the device-under-test (note: this is just one measurement that a VNA can make! It is the only one that I’ll discuss since it is perhaps the most used, but some VNAs can also make measurements using more than one port).

For more about signal reflection, you’ll be interested to read up on how signals travel over cables such as coax and twisted pair or even an RF signal trace on a PCB; these are all known as transmission lines, and in a nutshell, they consist of distributed inductance and capacitance, which acts like an (often) 50-ohm resistance, but only until the signal reaches the end of the line, at which point it will reflect unless the line is connected to the same resistance. If you’re wondering about the origin of the strange name ‘transmission line,’ it came about because vector measurements were initially required for AC power systems, where electricity is distributed at a low frequency; they were mains electricity transmission lines!

The purpose of the calibration is to allow the VNA to send signals (at the frequency range of interest) to the connected cables, and the VNA can then record the reflected signal phase and then associate that with a specific point on the Smith chart; the specific points are usually infinite ohms, zero ohms, and 50 ohms (known as an OSL or Open, Short, Load calibration). Once you have done that, you have set your ‘reference plane’ of measurements. If you later shorten or extend the cable length, you’ll need to re-do the calibration.

Calibration Standards

Electrical signals travel fast! At a rate of several tens of picoseconds just across the length of a connector. Since a VNA makes phase measurements internally, it is important that the reference plane position is known and does not shift, even by a fraction of a millimeter, because that will impact the measurement accuracy, especially once you’re in the GHz range.

A VNA doesn’t know what connectors you are using, and the 'end of the connector' as a reference plane is a vague term because there is an amount of capacitance at the disconnected end. If you short a connector, it has an amount of inductance at the end of it! So, to solve this problem, you can buy things called calibration standards. They look like RF connectors (such as type N or SMA) supplied with a specific open, short, or load on the end, with a known length. That length is supplied by the manufacturer, because they will have measured it very precisely (to a thousandth of a millimeter). In practice, it’s not a physical length but an electrical length (electricity travels at around 2/3 of the speed of light through typical RF connectors, although that's not the case for high-end calibration standards). If you program those electrical length values into the VNA (there will be a menu for it, or you can upload it as a file), then the VNA can more precisely know where the dots need to go on the Smith chart during the calibration phase. When you do the 'Short' calibration step, for instance, the VNA will know that the dot is slightly north of the zero-ohm position of the Smith chart because it knows from the electrical length differences that the short calibration standard is slightly longer than the open standard you’re using.

Please note that the above is a layman’s explanation! I’m no expert in calibration standards. The maths that the VNA does is pretty complex, because the manufacturer doesn’t just supply the length values, but also several coefficients for the VNA to better understand the capacitive and inductive reactances present on the open and short standards respectively.

Here's a problem; decent calibration standards (i.e. those which are repeatable and supplied with the calibration data file) cost a lot (into the thousands of $ for a set of three, and you’ll need as many sets as the types of connectors you use, and then multiply that further by two for male and female connector variants).

It is possible to buy lower-cost calibration standards, at about $70 for a set of three, but nevertheless as a design engineer, there will always come a time when you need to perform a calibration using your own DIY calibration method, and in that case you do need to accept that there will be an element of error in your measurement (which obviously you will try to reduce through as best a calibration as you can manage, and verify it if possible using known impedances).

Note: one thing definitely not worth doing, is to buy an OSL calibration kit from AliExpress! Some are available there for $40, and others for less, and all are useless, because it is unlikely they are supplied with any measured values at all, or any values you can trust and use! If you’re buying an OSL kit, it really does need to be from a trusted source.

The next part of this blog post covers ultra-cheap DIY SMA calibration standard construction. I believe it would be reasonable to use for at least a few hundred MHz, and perhaps even approaching a GHz. I have used a very short-cut, unorthodox method to come up with the length values. If you know of better DIY techniques, please share them!

Building DIY OSL Calibration Standards

The first step is to determine which connector you need to use! And then buy at least three of the same ones. Purchase reasonable quality connectors; if the internals slide around, the results will not be repeatable, and the measurements will have more errors than desired. Depending on the connector type, you might need to try several different connector manufacturers, because some are easier to modify than others.

A popular connector choice is SMA; it looks as shown below-left. BNC connectors (shown for scale below-right) are best avoided for use with a VNA.

Identifying the Reference Plane

For this blog post, I used female SMA connectors. Once you know which connector you need to use, Google is used to find the relevant standards document, which will contain a mechanical drawing of the critical dimensions, and in particular the location of the reference plane. From the diagram below, it can be seen that the plane is 1.93mm inside the connector.

Your Connector Measurements

Examine the connector and see how it can be modified. You are looking to create a surface that hopefully will still maintain the 50 ohms impedance. So, that will entail soldering direct to a surface and not on long component legs that will deviate away from 50 ohms.

I cut off the legs and tried to ensure that the cut end was as flush as possible with the connector body.

Either consult the datasheet or (preferably) measure the connector end-to-end to that flush surface. In my case, the length was 9.5mm (i.e., 13.5 minus 4, according to the datasheet drawing below).

From that drawing, one can work out the distance from the reference plane to the back surface of the connector; it is 9.5-1.93 = 7.57mm. I added a guesstimated 0.1mm because things are not fully flush on the back of the connector. That makes a length of 7.67mm.

Now, using the formula length / 0.3, we can see that the time for light to travel that distance would be 25.6 picoseconds. However, since the speed through the connector may be approximately around 2/3 of the speed of light, we need to multiply by 3/2, i.e. 25.6*1.5 = 38.3ps.

That value just calculated is called the Offset Delay. Some VNAs will want it in length format, so we need to multiply by 0.3. That results in an Offset Length of 11.49mm.

Assembly

Using the leg-trimmed connectors, for the 'Open' standard, nothing further needs to be done apart from protecting the cut center pin with a layer of epoxy glue so it doesn’t move around. The glue will have a slight impact, but it is what it is; something needs to be done to fix it in place and protect it.

For the 'Short' standard, you could use a small piece of thin copper sheet. I put some solder paste on the connector surface and heated the copper with a soldering iron from the other side. It is a bit fiddly. Afterward, the copper sheet can be trimmed more neatly to the size of the connector with scissors.

For the Load standard, I used two 100 ohm resistors in parallel, arranged at 180-degree spacing. Actually, I used about 50 resistors. All were measured until any two were found to equal 50.0 ohms in parallel. The optimal resistor size for the SMA connector is 0603; it fits just right from the center pin to the outer shell. Two resistors are recommended to reduce inductance. Afterward, epoxy glue was used to secure the resistors and the center pin in place.

Finally, for use as a verification device, I took a fourth SMA connector and soldered two 200-ohm resistors in parallel, again with 180-degree spacing. This would give me a 100-ohm reference to check that the calibration was sane.

Configuring the VNA

You'll need to explore the VNA or read its user manual to see how to actually turn on the VNA mode, select the single-port operation (known more accurately as an S11 measurement mode), and select the frequency sweep range. I won't cover that because it is fairly self-explanatory.

Next, each VNA will be different, but all will have the capability to configure the calibration standards. On the FPC1500, the fields were configured as shown in the screenshot below. All the coefficients were set to zero (which is reasonable for the 'short', although ordinarily, the 'open' does have a capacitance coefficient).

Using the Calibration Standards

The VNA is a precision instrument that needs decent cables for tests, even when operating at just a few hundred MHz. Avoid using regular coax if you can. It is far better to use either rigid or semi-rigid cables, although, in the earlier video using a breadboard, I broke all the guidelines. For general use, RG-402 is a suitable choice. It is about half the thickness of RG-58 and is easy to bend into a position, and then you’ll want to keep it as unmoved as possible for both the calibration and the actual measurement. Clamps, stands and so on may be needed to move the device under test (DUT) into position to align it with the cable.

RG-402 off-cuts can be obtained at a low cost from eBay (it might not even matter what connector is on the end; it can be cut off, and SMA connectors are easy to solder onto the end of it, provided you’ve got a large soldering bit for the outer conductor. The photo below shows what RG-402 and its associated SMA connector look like.

If you can, it is useful to put ferrite cores around the cable (as many as you can fit!). These act like a common-mode choke, and allow you to better isolate the cable from the device under test. You can use any ferrite cores intended for use as chokes.

The photo below shows just such an example with RG-402 cable and a load of ferrites (the photo happens to show N connectors being used, and a ready-made OSL calibration standard that comes in the form of a T piece). 

For more intricate work, RG-402 is still way too thick. I’ll address that in a possibly later blog showing a different calibration standards technique since it’s almost time to terminate this blog at a suitable point! One last thing however; how does one confirm that the DIY calibration standard is any good?! That is discussed next.

De-Embedding VNA Measurements

If you recall from the earlier section, the purpose of the calibration step is to end up with a position close to the end of the cable/connector, known as the reference place, from which the VNA can make accurate impedance measurements. To get to that, open/short/load calibration standards are temporarily attached to the cable, and the VNA is made aware of the length of the open and short by uploading the electrical length or delay information into the VNA configuration screen.

The trouble is that it’s impossible to tell from the open/short/load calibration standards if the actual calibration is any good. You could have attached garbage standards, and the VNA will blindly assume you know what you are doing and assume them to still represent close to infinite, zero, and 50 ohm impedance. One way to get a bit of confidence is to attach a completely different, known impedance after you have done the calibration, and see if what gets displayed makes any sense.

I used a 100 ohm impedance, assembled in the same way as the 50 ohm load, but with two 200-ohm resistors in parallel. When you attach it, you’ll see something like this:

That curve of measurements happens to cover 10MHz to 1GHz (that is the range I set the VNA to), and it looks nothing like a nice dot at the 100-ohm resistance position. Instead it is spread, with the impedance at high frequencies having an element of capacitive reactance. It looks awful.

However, if you think about it, the VNA was calibrated to a reference plane which is close to the end of the connector (well, about 3.4mm inside the connector actually – that's the approximate depth of the reference plane off the end of a SMA male connector. However, the 100-ohm verification standard isn’t at that position. It is an electrical length of 11.49mm away from the reference plane. No wonder the impedance doesn’t look right, because the VNA isn’t measuring precisely at the position of the resistance. What’s required is to calibrate out the extra length. There are several techniques that can be used. When this blog was first written, a method called de-embedding was used, however subsequently it was found that electrical length compensation worked better for this situation.

Not all VNAs support these methods. However, it is possible to perform them using a PC. The general idea is to save the VNA measurement to a file format that the PC can read, and then find an algorithm that can fix it and display a corrected Smith chart.

Most VNAs will allow you to save a screenshot of the Smith chart display, but they will also allow you to save in a special format called Touchstone format, also known as S1P format (the Touchstone file suffix is .s1p). Note: If you’re curious, the format is actually a simple text file, containing a list of frequencies, and the measurements, one per line).

Once you’ve done that, you can transfer it to your PC using a USB cable or memory stick for instance.

You will also need to remove the 100-ohm verification standard, and re-attach the 'Open' standard, and grab another .s1p file. Then, if you're performing de-embedding, repeat with the 'Short' standard. For electrical length compensation, you don't need to repeat with the 'Short' standard. You now have either two or three files. PC software will be used to do the maths to convert all that into a corrected .s1p file for the 100-ohm verification measurement. If you're performing de-embedding, then continue reading this section. Otherwise, you can now skip to the next section titled Electrical Length Compensation.

Here’s the Python code to perform the de-embedding. Change the filenames section in the code to suit your needs. I saved it as a file called de_embed.py

# de_embed.py - VNA Measurement De-Embedding
# rev 1 - shabaz - January 2024
# This code reads three .s1p files, (for S11 Open, Short, and for the DUT).
# Then, the DUT is de-embedded using the Open and Short files.
# The corrected DUT is saved to a new file, and is also displayed on a Smith chart.

import numpy as np
import matplotlib.pyplot as plt
import skrf as rf
from scipy.optimize import minimize
from skrf.calibration import OpenShort
import sys

# set the plot style
rf.stylely()

# set the filenames
short_fname = 'short-meas.s1p'
open_fname = 'open-meas.s1p'
dut_fname = '100-meas.s1p'

# output filename for corrected results
corrected_dut_fname = 'corrected.s1p'

# read the three files and convert to RF Network objects
short_ntwk = rf.Network(short_fname)
open_ntwk = rf.Network(open_fname)
dut_ntwk = rf.Network(dut_fname)

# perform the de-embedding using the skrf calibration module's OpenShort class
dm = OpenShort(dummy_short=short_ntwk, dummy_open=open_ntwk)
corrected_dut_ntwk = dm.deembed(dut_ntwk)

# save the corrected DUT network as a .s1p file
corrected_dut_ntwk.write_touchstone(corrected_dut_fname)

# plot the results on a Smith chart
dut_ntwk.plot_s_smith(lw=2, label='Uncorrected')
corrected_dut_ntwk.plot_s_smith(lw=2, label='Corrected')

plt.title('VNA Raw and Corrected Measurements')
plt.legend() # re-generate legend
plt.show()

print('Done, exiting..')
sys.exit()

The software relies on, amongst other things, a Python library called scikit-rf. To run the software, type:

pip install scikit-rf
pip install matplotlib
pip install scypi
pip install numpy
python de_embed.py

When the program is run, it will generate a file called corrected.s1p, and it will also plot the output to the screen:

The above is now an acceptable result. Across the 10MHz-1GHz range, the verification standard can be seen to be very close to 100 ohms. This provides a bit more confidence.

Electrical Length Compensation

You can follow the same steps as described in the De-embedding VNA Measurements section, but once you have the .s1p files for the open and the device-under-test, different code will be used.

Some VNAs will support electrical length compensation capability from a menu option, but we are going to use the following code called elec_comp.py:

# elec_comp.py - Electrical length compensation
# rev 1 - shabaz - January 2024 
# based on code by Keisuke Kawahara
# This code reads two .s1p files, (for S11 Open and DUT).
# Then, the electrical length is calculated from the Open file,
# and applied to the DUT file. The corrected DUT file is saved to a new file,
# and is also displayed on a Smith chart.

import skrf as rf
import matplotlib.pyplot as plt

# set the filenames
open_fname = 'open-meas.s1p'
dut_fname = '100-meas.s1p'

# output filename for the corrected results
corrected_dut_fname = 'corrected.s1p'

# read the two files
dut_ntwk = rf.Network(dut_fname)
open_ntwk = rf.Network(open_fname)

# electrical length compensation
lossless_tline = rf.media.DefinedGammaZ0(frequency=dut_ntwk.frequency, z0_port=None, z0=50, Z0=None, gamma=1j)
electrical_length = open_ntwk.s_deg[:,0,0] / 2
corrected_dut_ntwk = lossless_tline.line(electrical_length, unit='deg') ** dut_ntwk

# save the corrected DUT network
corrected_dut_ntwk.write_touchstone(corrected_dut_fname)

# plot the results on a Smith chart
dut_ntwk.plot_s_smith(lw=2, label='Uncorrected')
corrected_dut_ntwk.plot_s_smith(lw=2, label='Corrected')

plt.title('VNA Raw and Corrected Measurements')
plt.legend() # re-generate legend
plt.show()

print('Done, exiting..')
sys.exit()

The software relies on several Python libraries, particularly one called scikit-rf. To run the software, type:

pip install scikit-rf
pip install matplotlib
pip install scypi
pip install numpy
python elec_comp.py

When the program is run, it will generate a file called corrected.s1p, and it will also plot the output to the screen:

As you can see, the result is excellent. For even more confidence, you can extend the cable and try a completely different Open and DUT.

In the the result below, I attached a short extension cable from the reference plane, and then attached a different Open, and a DUT that happened to be a 150 ohm resistor (which, to be more accurate, actually measured at 148 ohms with a multimeter). Again from 10MHz to 1GHz, the result was very consistent. The Smith chart shows the corrected result to be a resistance of about 146 ohm at 10MHz, and 160 ohm at 1 GHz, which is not bad. Based on that result, I would feel confident about using the created OSL calibration standards, to at least 1 GHz.

A couple of extra things to note: firstly, the electrical length compensation, and de-embedding techniques, are actually very powerful, and the same code can be used not just for verification but also whenever you need to measure very close to an impedance, closer than the reference plane is.

The other point is the fact that one needs to get used to using software features, either within the instrument or on a PC, for a lot of RF-related tasks. This is a benefit. In the past, a lot of knowledge was required to make good use of spectrum analyzers and VNAs, and it was very easy to make mistakes. Nowadays, manufacturers have distilled upward of a hundred years of experience into the software. Just the calibration alone involves dozens of complex maths calculations internally within the VNA. For hobby use, it may not matter, but for work-related use, it is worth using known brand VNAs to trust that there are no bugs within the algorithms. It will make the difference between a product with just 'OK' or even maybe mediocre RF performance, and one that successfully benefits users with its range.

Summary

A VNA is one of the most useful instruments for engineers working with RF circuits or antennas, performing the important task of measuring impedance (i.e., resistive and reactive components) across a range of frequencies. To get going, all that’s required is any VNA (even a $100 one will work), plus an RF cable or two that will connect to the equipment you wish to measure (for instance antennas, filter modules, and so on), plus a handful of the same type of bare connectors, that can be DIY’d into calibration standards, which are used to set a reference plane position for making impedance measurements.

Procedure known as electrical length compensation, or alternatively de-embedding, are used to move the reference plane, so that additional cables and connectors can be eliminated from impedance measurements.

Once the impedance is known, a capacitor and/or inductor can be used for impedance-matching, to ensure maximum power gets from source to destination. The procedure to impedance-match was shown in a video demonstration.

Hopefully, with all of the above, you should now have enough information to get going with VNAs.

If you want to read more, check out  Vector Network Analyzers for the General Reader, Part 2: Inductance Measurement!  

Thanks for reading!