Testing antennas with nanoVNA v2

Hi, I bought a nanoVNA v2_2 to check the quality of the default antennas as well as some that might be better suited for a rooftop relay. I am using 433MHz devices, and the antennas I bought are supposed to match that frequency as well (but they are just a cheap experiment after all…)

Since I am quite a noob at all of this, I would like to share what I did and hope for some hints on what I might have done better :wink:

  1. I calibrated the VNA
  2. I attached the antenna
  3. I attached it to my computer and started the NanoVNA Saver app
  4. I selected the range from 400Mhz to 500MHz


My assumption would be, that the antenna doesn’t do too bad with -5.4dB return loss at 433MHz, but it would work even better at frequencies with a local minimum like 416MHz and 448MHz, where the return loss i even higher, up to -6.9dB.

Is that about right or do I need to consider other values as well, if I want to judge the antennas effectiveness at 433MHz?


Here is the result of one of the T-beams original antennas: I guess with this one, I can’t expect great range results at 433MHz (but it would work well at 470MHz):

1 Like

Here is another result of one of the T-beams original antennas: I guess this one works much better at 433MHz (-7.8dB return loss), but it would work even better at the minimums frequency (441MHz with -13.8dB return loss)

And this one (original t-beam) is… broken? Looks like a flatline on the ECG…

but when I extended the sweep from 100MHz to 1.000MHz I found a minimum around 838MHz, so I sweeped from 800MHz to 900MHz and got this result:

So maybe this was accidentally a 868MHz antenna that is a bit off. But still, -31.1dB return loss at 838MHz doesn’t seem reasonable, does it?

Hello! It’s good that someone else here is busy measuring antennas! :slight_smile:

Your reasoning is generally correct. The higher the return loss, the better the antenna works. In general, it makes sense to focus on the following numbers: readings 0 …- 10 dB - this is a poorly working antenna; -10 …- 20 dB - average quality of the antenna. -20 …- 40dB - excellent working antenna. If you wish, you can convert these numbers into the SWR parameter.


For example, here are my antenna parameters from the TTGO Lora 32 unit 433MHz.


For example, here are my antenna parameters from the TTGO Lora 32 unit 868MHz.

1 Like

Hi, thanks, glad to know that I am not completely off :smiley:

I found a source claiming

Although more return loss is better here, there is little benefit above 10 dB return loss, since more that 90% of available power is already being delivered to the antenna. Return losses above 10 dB have little practical benefit.

That gave me some hope I have at least some antennas that perform okay at “my” frequency of 433MHz - but that seems to collide with your statement, doesn’t it?

Is rp1 SWR and rp2 return loss? And could you see here that the best performance of this antenna would be around 550MHz, where the minimum is?

Still looks impressive, what kind of antenna do you have to get this performance?

0 …- 10 dB is poor performance of the antenna in general, conditionally. Then, probably, correct, therefore 0 …- 6 dB - the antenna does not work at all.
-6 …- 10dB - does not work very well.

1 Like

These are the stock antennas for TTGO Lora 32 for 433MHz and 868MHz

1 Like

Wow, you had those results with the stock antennas… I must have had back luck with the ones I received or I did something wrong while measuring…

1 Like

I don’t want to spoil the party, but return loss only tells you how well antenna and radio are matched, not how well the antenna performs in term of radiating power/receiving signals.

A case in point would be a near-ideal 50 Ohm resistor such as the “dummy” load part of your VNA’s calibration kit: return loss might be in the region of -60 dB, but that would make it a both a great load and a really bad antenna!

A finite-elements simulation and a field-test will give you a more definite answer regarding the real performance of the antenna.

That’s correct. That’s one more reason not to be too concerned about high return loss, which still desirable, just not over real-world performance.


Thanks for the clarification, that is just what I was hoping to learn. So this is just one factor to judge the antenna, but could you conclude that there is no point in “real world testing” if the return loss value is already “bad” for the frequency to be used, so the VNA results could be used to have some sort of per selection to filter out the worst and find the most promising antenna from my selection?

1 Like

I wouldn’t say so, particularly at HF, but at VHF and above it’s easier to make a well-performing antenna that also exhibits a high return loss.

Nobody in their right mind would use a Windom or a Beverage antenna at UHF, though they are solid performers in the low HFs and they have a low return loss.

Hm, not sure I get that… if my antennas are supposed to be designed to work at 433MHz, that would be above VHF. So you are saying they might perform well and they might have a high return loss, but those things don’t need to be related?

Do you think there is anything the VNA can tell me rather quickly about the quality of my antennas, e.g. if one value is high and another is low, that is a good or bad thing? Even if the device wasn’t that expensive, I bought it hoping I could check the quality of the antennas…

Do you think there is anything the VNA can tell me rather quickly about the quality of my antennas, e.g. if one value is high and another is low, that is a good or bad thing?

Nanovna is an excellent and inexpensive device for this!

1 Like

Yes, I apologize I didn’t explain myself.
The higher the frequency, the smaller the antenna, due to the decreasing wavelength. This means that it’s easier, from a mechanical point of view, making a high-gain antenna at 433 MHz than, say, at 4,3 MHz because the latter would be 100 times smaller while having the same gain.

Now, a half-wave vertical antenna (i.e. a vertical dipole) for 4,3 MHz would be 34,9 metres long, while only 34,6 centimetres at 433 MHz. Both antennas could be made to exhibit a high return loss and of course will have the same gain (same configuration, same radiation pattern), but one might be tempted to make the 433 MHz antenna much longer by “stacking” so-called “collinear elements”. Why so? Because more elements, if “fed” correctly, will add “gain” to the antenna, by radiating more in the desired direction (usually more towards the horizon than the sky).
If I’m not mistaken, at 433 MHz, a 1,7 m collinear antenna will “gain” 9 dB over a standard dipole, an increase in power close to 8 times! And that will happen with a manageable antenna length, as opposed to doing the same at 4,3 MHz.

That’s why return loss alone will not tell much about the real performance of an antenna.

Well, it’s an extremely useful tool to validate and troubleshoot antennas and transmission lines. You could check whether an antenna working frequency is as advertised (or designed, if you’re building one). It’s something every radio tinkerer should have!


Just a quick note in reference to the original post. You must set the frequency range FIRST and THEN calibrate. Anytime you change the frequency upper or lower, you have to recalibrate.


True! Also, anytime you change connectors, cables, etc.