What does a VNA tell us about this antenna ?

[tautech]

In this thread I want to demonstrate the use of SVA1015X for analysis of a number of antennas I have hoarded over the years.

In another thread I built and tuned a 315 MHz J-Pole antenna to get 500m range with a key-fob remote and in doing so learned getting maximum antenna emissivity at the fundamental frequency was paramount to getting good results.
https://www.eevblog.com/forum/rf-microwave/antenna-project-log/

Recent demonstrations of SVA1015X to customers doing various wireless projects have shown the 'off the shelf' commercial antennas they have selected couldn't be considered as optimal for their intended frequencies and their wireless range expectations are therefore generally not met.
As antennas have the useful property of reciprocity we can sweep them with a signal and find their transmission properties which are also very close to their receptive properties. We can get basic information with a spectrum analyser and a reflection bridge but not the degree of info that can be acquired from a VNA.


2.4 GHz SMA WiFi antenna analysis
While I can't fully sweep a 2.4 GHz antenna with a 1.5 GHz VNA let's look at a full 1.5 GHz sweep and see what info arises.
The subject:

It should be noted that this antenna connected directly to the S11 port has its performance impacted by adjustments of the VNA due to proximity effects and it need be connected more remotely to reduce them for accurate repeatable measurements.

Emissivity, aka Log Mag
Where we look for the largest dip in the trace that represents the greatest absorption of energy from the feedline/transmitter and the frequencies this might occur at.

Interestingly here it's at ~1/2 of 2.4 GHz that this antenna is designed for.


Well could this antenna be useful at 1.28 GHz ?
We need to check the SWR to look for efficiencies:


Results of 1.6 and 1.2:1 indicate it might useful however how will they match the optimal 50 Ohm impedance we always seek ?

A Smith chart will give us the info of impedance and inductive or capacitive match/mismatch.


Overlay all these results using a trace for each measurement and we can see the full picture.


We need remember all these results are with a full 100 Khz-1.5 GHz sweep which is not what we would normally do.
Instead we would narrow the sweep considerably and set the center frequency to the frequency of interest.


Next up, some antennas under 1 GHz.

[T3sl4co1l]

But S11, SWR, Z/Zo and the need for tuning, are all equivalent (give or take the first two being shown in scalar form)... 

The real question is: is it absorptive or transmissive at those frequencies?  It could very well be a better antenna say in that 400-800MHz range, despite the poor SWR, if it's more transmissy than lossy.  (In that case you'd apply a matching or tuning network to use it there, at 50 ohms.)

Most antennas with reasonable construction and no intentionally introduced losses, do in fact couple nicely, so it's probably that the markers show the best.  But in general, you never know

Tim

[hagster]

As commented, the Return Loss and SWR are the same, but on a different scale.

You can often get an idea of how lossy an antenna is likely to be by inspection. If the antenna is basically just metal and air, then there really isn't much mechanism available to create loss.

You also need to check that the power is radiated from the antenna and not the cable. Normally you can get a good idea by moving your hand over the cable and watching for large changes on the plot. A good antenna should not have currents flowing on the outside of the coax and hence should not be sensitive to changes in cable length of proximity of other objects. For small low efficiency antennas it can be a real challenge to separate the cable effects from the performance of the antenna.

[tautech]

Next subject:
Magnetic base whip antenna ~360 mm long.



What is it as there's no markings and just a ~3m long coax feedline BNC terminated ? ? ?

Full 1.5 GHz Log Mag span to find frequencies of resonance.

Marker 3 indicates 876 MHz best Log Mag

Let's narrow the Span to 500 MHz centered on 876 MHz


And a SWR check at the same settings


Smith chart with just 100 MHz span on 876 MHz


Finally Smith, SWR and Log Mag with markers all on 876 MHz and a 100 MHz span.


It should be noted to get selected marker measurements displayed it is necessary to select the correct trace from the trace menu. Once selected the diamond marker indicator is filled with color and not a hollow shape.

Antenna labelled and stored.
Next up a little Yagi...........

[borjam]

For an antenna you need to check a couple of things:

- First, coupling. When checking S11, as a rule of thumb better than -8 dB (less than 2.5 VSWR) will be good.

- Second a resistor would give perfect matching but it would be useless. So you would need to check how well it radiates, which is expensive.

The best thing you can do is to buy antennas from reputable sellers who provide a datasheet with the correct measurements.

Antenna looks can be deceiving, sometimes there is a bit of trickery hidden inside a plain looking one that helps it work at more than one frequency.

And unreputable antennas... I remember a batch of super cheap Chinese antennas. I tested 30 of them with a VNA and all of them were different.

[tautech]

For an antenna you need to check a couple of things:

- First, coupling. When checking S11, as a rule of thumb better than -8 dB (less than 2.5 VSWR) will be good.

Even for mW ?
Nah, when you have a VNA why not get them way down under 1.1:1 ?
With the one hendrog and I conjured up we got it down to 1.0137:1 !

The best thing you can do is to buy antennas from reputable sellers who provide a datasheet with the correct measurements.

Antenna looks can be deceiving, sometimes there is a bit of trickery hidden inside a plain looking one that helps it work at more than one frequency.

And unreputable antennas... I remember a batch of super cheap Chinese antennas. I tested 30 of them with a VNA and all of them were different.

Yep all that and that's one of the reasons I started this thread especially after sweeping a customers '925' MHz antenna only to find it worked best at 870 MHz ! No wonder they had range issues !

The little Yagi up next is labelled and I can find some info on it so it will be interesting to see how well it meets its spec. After that I have a couple more at least to write up here and one is a monster sized whip so that will need a fine day, something that's rare here ATM.

[borjam]

For an antenna you need to check a couple of things:

- First, coupling. When checking S11, as a rule of thumb better than -8 dB (less than 2.5 VSWR) will be good.

Even for mW ?
Nah, when you have a VNA why not get them way down under 1.1:1 ?
With the one hendrog and I conjured up we got it down to 1.0137:1 !

Yes, even with mW level powers. Otherwise efficiency would be too low and at least according to manufacturers you can damage a chip if the mismatch is too large. And I forgot, even if you don't break anything a large mismatch will likely cause distortion. So the problems will mount

Always read the datasheets

Regarding such awesome VSWR it will depend on the application. On HF you can get an outstanding match for a bandwidth of about 100 KHz. But if you are using it for WiFi (mininum channel size is 20 MHz) or cellular (several multi MHz bands on different frequencies) it will be difficult.

Yep all that and that's one of the reasons I started this thread especially after sweeping a customers '925' MHz antenna only to find it worked best at 870 MHz ! No wonder they had range issues !

You would be amazed I had an issue at work with that. Some think that as long as it has a connector and it looks like an antenna it will work

The little Yagi up next is labelled and I can find some info on it so it will be interesting to see how well it meets its spec. After that I have a couple more at least to write up here and one is a monster sized whip so that will need a fine day, something that's rare here ATM.

Also, remember that antenna position matters, especially if you have stuff within the near field.

[OwO]

For transmit I would say any VSWR < 1:3 is "good enough" because that's -6dB return loss which means 75% of the power is radiated. Another way to think of it is in terms of "mismatch loss", which at 1:3 VSWR is -1.25dB. It might be unintuitive to most but 1-2dB is completely negligible in a typical wireless system. I've done propagation experiments on various ISM bands and you can easily get 10-20dB variations in received signal just walking around.

For receive it's usually even less important because the ambient noise is going to be at least around 10dB above thermal noise floor so any antenna with less than 10dB mismatch loss (VSWR 1:37!) is likely not noticeable. That is the reason most broadcast receivers get away with very crappy antennas that are completely untuned. The exception to this is the higher bands (think >5GHz) or satellite receivers where you can have a noise floor below thermal noise (because the antenna is pointing at the sky).

[TheUnnamedNewbie]

What S11 is good enough also really depends what your antenna is connected to and what it's application is. If you are dealing with some kind of radar, a poor S11 could result in your receiver being saturated more than you want due to the transmitted pulse.

And mismatch can hamper the performance of the amplifier - the amplifier's matching network will be designed for maximum power (usually) and a poor S11 can impact that significantly.

I have always frowned at people assuming good S11 = good antenna, because of two reasons: S11 can be changed with a matching network, and just because it is not reflected back doesn't mean it is going anywhere (losses).
Yeah, at the few GHz frequencies you might be able to ignore any dielectric/conductor losses in an antenna, but that stops once you get past a few GHz or want high accuracy of your antenna (eg for imaging or sensing)

[T3sl4co1l]

Heh, the S11 notch is probably the antenna's tuning, but the SWR around the notch isn't bad either -- suspect cable loss, is that just ordinary shitty RG-58 or 174?

So we see that good SWR may not necessarily give good antenna performance; obviously, a transmission test would be required to prove that.

Tim

[borjam]

Heh, the S11 notch is probably the antenna's tuning, but the SWR around the notch isn't bad either -- suspect cable loss, is that just ordinary shitty RG-58 or 174?

So we see that good SWR may not necessarily give good antenna performance; obviously, a transmission test would be required to prove that.

Indeed, new product idea for Siglent: The PAC1000T (Portable Anechoic Chamber 1000 Tati Anniversary Edition)

Outside view:


Inside view:

(Operator not included)

[TheSteve]

Tautech you missed out the on the best part - now that you found a resonant frequency do a little math on it and determine how the antenna works and see if it all makes sense.
Likely the lower section of that antenna before the coil will be roughly 1/4 wavelength long at your resonant frequency. The coil itself is a matching network(many people think it is just there to look cool). The upper section should also work out to be mathematically related to the resonant frequency. From there you will get a picture of the gain it will ideally have and the ideal radiation pattern.

[coppercone2]

is that like a russian hippy van?

[tautech]

Tautech you missed out the on the best part - now that you found a resonant frequency do a little math on it and determine how the antenna works and see if it all makes sense.
Likely the lower section of that antenna before the coil will be roughly 1/4 wavelength long at your resonant frequency. The coil itself is a matching network(many people think it is just there to look cool). The upper section should also work out to be mathematically related to the resonant frequency. From there you will get a picture of the gain it will ideally have and the ideal radiation pattern.

Yeah, maths is my weak point.

Anyways, from base to start of coil = 90mm
Coil = 3.5t over 20mm @ 15mm dia.
Coil to tip (top section) 200mm

[TheSteve]

Tautech you missed out the on the best part - now that you found a resonant frequency do a little math on it and determine how the antenna works and see if it all makes sense.
Likely the lower section of that antenna before the coil will be roughly 1/4 wavelength long at your resonant frequency. The coil itself is a matching network(many people think it is just there to look cool). The upper section should also work out to be mathematically related to the resonant frequency. From there you will get a picture of the gain it will ideally have and the ideal radiation pattern.

Yeah, maths is my weak point.

Anyways, from base to start of coil = 90mm
Coil = 3.5t over 20mm @ 15mm dia.
Coil to tip (top section) 200mm

Let's see - speed of light in meters per second divided by the frequency in Hz. 300000000/876000000 = 0.342 meters - divide by 4 for a quarter wave and we get 85.6 mm for the lower section.
The upper section is likely a 5/8 wave which would be 214 mm.
The numbers line up reasonably well.
That would make the antenna a collinear, the 5/8 over 1/4 design typically gives around 3.2 dB of gain.
To work well and keep the RF radiating in a nice pattern(vs wanting to go up) the antenna should have a ground plane under it - such as the roof of a car. This would likely alter the resonant frequency a little as well but it should work fine for any use between 800 and 900 MHz.

[borjam]

But it can get much more complicated at high frequencies, where its shape, the actual material used for the antenna, its thickness, the radome if present, etc, can have a significant impact.

Some apparently simple antennas, for example, can work on several LTE/UMTS/GSM/WiFi bands.

[radiolistener]

any piece of wire can work as antenna. The problem here is to put RF energy into this piece of wire

Not all RF energy which is put into a wire will be radiated as EM, just a small part. Some part will be spent on heating loss. And the rest of energy will be returned back to the transmitter, where it can be added with a new transmitting energy and the sum may damage your transmitter

In order to avoid that, it's better to use piece of wire with resonant length for desired frequency. When wire has resonant frequency it will not return energy back, it will accumulate it. So, RF energy will oscillating within wire and don't come back to transmitter.

Each oscillation in the wire will leads to heat loss and radiation loss. Since we using resonant length, the energy will be accumulated in the wire and amplitude will be increased for each oscillation. Since amplitude is increased, heat loss and radiation loss also will be increased. At some point incoming energy will be balanced with energy loss (heat and radiation).

The oscillation cycle count needed in order to radiate applied energy is named as antenna Q factor. Usual half wavelength dipole has Q factor about 12. Higher Q factor leads to higher amplitude in the wire, so it will have higher thermal and radiation loss. But at the same time higher Q factor leads to smaller antenna bandwidth.

With vector analyzer you can analyze if your antenna with matching network and with feeding cable returns energy back to transmitter or not. But you cannot say if this energy is spent on heat loss or it is spent on EM radiation.

[tautech]

any piece of wire can work as antenna.

Yep, this:
Inside a tin shed !


~250mm of stainless wire dropped into a BNC socket worked sporadically for us over 500m for years until I made a J-Pole that solved all our problems.
It need be mentioned good design and proven measurement can vastly improve reception quality as that's often all you have to work with when transmission antennas are set in stone by the manufacturers of say a short range keyfob device.

[tautech]

Next subject:
Yagi
HG909Y-NF

Datasheet attached.

Coupled to SVA1015X with Siglent 6 GHz rated 700mm N-N cable.
Full sweep Log Mag


 
This is looking better so lets narrow the span to 100 MHz centered on 930 MHz.
Log Mag


SWR


Smith chart


Phase


All 4



And for giggles we'll look at a DTF of the feedline and antenna with default DTF settings.......set marker to peak and out with the tape measure and just 50mm difference in results. 

[borjam]

I will post later the S11 graph of an antenna with a crazy high Q: a tuned loop antenna for HF.

It has an extremely small bandwidth, like 4 KHz or so.

[borjam]

So this is it

It's a MFJ1788 loop antenna. It's in a bad placement (balcony) so the SWR is higher than it should be.

Tuned to 7080 KHz offering a VSWR of 1.6:1, climbs to 2.5:1 with an offset of just 3.3 KHz.

By the way this is an example showing the VNA function working below 10 MHz.

Now, a really interesting feature would be a "continuous valley" monitoring and linked markers for the desired performance.

For example, let's see I am adjusting this antenna. While I adjust the tuning controls the resonance frequency will change. Unfortunately the VNA function doesn't adjust the "valley" peak marker when it changes. So I must keep pressing peak->valley while I tune the antenna if I want the marker to be updated. It should "follow" the valley.

Linked markers: It would be great to specify something like: I'd like markers set to a certain RL or VSWR value. Again, updated if the antenna behavior changes because I am making adjustments.

These two features would be really awesome because thanks to the web server you can be monitoring the VNA while you are adjusting the antenna, even if you are high on a tower.

This would be a good use case for a ham operator.

[TheSteve]

Does it not have an auto-tracking feature when a max or minimum peak has been selected?

[borjam]

Does it not have an auto-tracking feature when a max or minimum peak has been selected?

Do you mean the VNA?

If you press the peak buttona and select peak or valley it will place a marker on the maximum or minium value. But it won't follow the maximum or minimum value if you change something in the DUT (like for example tuning the antenna).

[TheSteve]

Yes, that is what I was referring to. It is a feature that should be requested.

[tautech]

Yes, that is what I was referring to. It is a feature that should be requested.

Done !

Thanks borjam and Steve.

Useful indeed and a feature that certainly would've helped when I did final tuning of the J-Pole I made to resolve the proximity effects once it was installed.