The big Antenna S11 catalog

S11 is a parameter of how well the power you put into an antenna stays there and doesn’t reflect back into you radio. With “stays there” one includes the amount of power radiated out into the air, as well as losses of the energy inside the antenna. With losses I mean energy converted into heat. With the low power LoRa works at you won’t notice this heat. But if you would put 1000 Watt into the antenna and it has a loss of 50%, you will feel it.

Receiving a signal can be seen as the reverse of sending one, so for reception the S11 should be the same as for transmission.

But in short, you want your S11 parameter to be as low as possible at the frequency you are using your antenna at. This is the same as wanting your VSWR to be as close to 1:1 as possible. Lower S11 = VSWR closer to 1:1. S11 and VSWR are just two different ways of expressing the same thing. VSWR is a linear ration, while S11 is a ratio in dB.

Except for wanting a low S11, depending on your use case you want the minimum S11 point (in the frequency range) to be at the frequency you are going to use your antenna at. If it is a wide band antenna that is meant for GSM/3G/LTE, the S11 will be low over a wide range of frequencies.

For LoRa (EU) we want a low S11 at 868MHz plus minus a few MHz. We do not want a low S11 at any other frequencies as that will cause the antenna to receive signals strongly at those frequencies too. Remember that all signals that are received by the antenna is passed to the radio, and the radio need to detect the LoRa signal out of the noise caused by all the other signals.

So what we ideally want is an antenna that has a low S11 (below -10dB) around 868MHz, but a high (close to 0dB) at any other frequency.

I actually started this thread to document what the S11 looks like for a wide range of antennas. I’m using an AAI RF Vector Impedance Analyzer (DIY Antenna Analyzer) to do the measurements. We do not know how accurate this device is, but at least it gives a good idea.

Some other comments from readers:

It’s not too difficult to get the S11 tuned for a LoraWan external antenna. A good S11 response is important but it is only the first step, it doesn’t mean by any means that the antenna is radiating efficiently.

The measured radiation pattern of the antenna in an open field test site is the real indication of antenna performance, for most use cases you are looking to get good omni radiation out to the horizon.

5 Likes

Starting with the Kerlink Wirnet Station default antenna because I was really impressed by how well this one looks and it inspired me to start this thread.

03

47

We see an S11 of -24.3dB at 871MHz. Nice and sharp valley, so a narrow bandwidth around 868MHz. This is a really good example of what you want in a LoRa antenna and this is most likely the reason why Kerlink gateways perform so well.

Conclusion
I would definitely use this antenna on outdoor gateways - if I can find them for sale anywhere.

1 Like

The dipole whip antenna that comes standard with the Lorank8 gateway has a centre frequency at 850MHz, but still performs very well at 868MHz where it has an S11 of -20dB.

24

34

Conclusion
Ideetron found a really good antenna for their gateway. I would definetly use these antennas on indoor gateways and not-so-compact end devices.

Multitech IP67 Conduit

24

Antenna: R08063/21704NM, 806-960/1710-2170 MHz, 3-4dBi

17

S11 plot:

52

Values as 868MHz:

44

Conclusion
If I have a choice I would not use this antenna on a gateway. It is too wide band - receiving too many unwanted signals that increases the noise floor and therefore decreases the SNR.

Taoglas Barracuda OMB.8912.05F21

57

50

S11 plot:

00

Values at 868MHz:

53

(Apologies for the bad quality photos - the African sun isn’t for taking photos!)

Conclusion
This antenna is OK, but I have seen antennas that work better. Because this antenna resonates over a wide frequency range it will receive unwanted signals too, which will increase your noise floor. I doubt that the extra gain this antenna gives you will make up for the loss in SNR, because the unwanted signals also lie in the same reception area as your LoRa signals.

I recently bought an 868MHz whip antenna (rubber duckie antenna) from my local hobby shop. When I measured it using my VNA, I was shocked how bad some of these antennas are. The range I got on LoRa was way below what I expected (<100m where I normally get 5km).


07

Firstly we can see this antenna does not work at 868MHz. It as an SWR of 5! That’s a return loss (s11) of -3.6dB. This is means the antenna does not resonate at 868MHz at all. You’d be better off using no antenna than this one.
31
44

Zooming out a bit we can see this antenna resonates at 960MHz and 1800MHz, telling me that it is in fact a GSM antenna for the 900MHz and 1800MHz band. But a bad one at that as the S11 doesn’t go down low enough.
44
11

Conclusion
This whip can not be classified as an antenna on 868MHz. Do not buy it, and if you already did, throw it away.

1 Like

In my search for better whip antennas I ordered some from RS Components. The first one is:
https://za.rs-online.com/web/p/telemetry-antennas/0542563/

22

This antenna resonates at 885MHz, a little higher that promised. But it is also quite wideband, so even at 868MHz the S11 is very low. So this antenna performs good enough for an end device.

54

As an aside, this antenna also resonates at 1900MHz, saying that it might have been designed as a GSM antenna, but later tuned for 868MHz.

29

This SWR at 868MHz is 1.2, which is very good.

46

Conclusion
This antenna has acceptable performance for end devices. I will not use it on a gateway where I do not want to receive unwanted signals.

3 Likes

Another 868MHz whip antenna from RS Components:
https://za.rs-online.com/web/p/telemetry-antennas/7934363/

29

This antenna is slightly longer than the rest, looks exactly like the antenna on the Lorank8 gateway, and it is most likely a dipole antenna specifically designed for 868MHz.

01

This antenna resonates very well at 864MHz with a very deep trough. This trough/valley doesn’t move sideways when bringing your hands closer to the antenna, saying that the antenna is very stable at its resonant frequency.

Zooming out a bit one can see this antenna also resonates at 2.4GHz. Not as well as at 868MHz, but good enough to use this antenna for wifi or bluetooth too.

11

This antenna has an SWR of 1.12, which is excellent.

12

Conclusion
This is definitely the best whip antenna for LoRa I have seen. I’d use this antenna on not-so-compact end devices and indoor gateways.

4 Likes

5 posts were merged into an existing topic: The BIG and SMALL ANTENNA topic part 2

Glass Fibre 6dBi 915 MHz antenna from RAK Wireless
The first 915MHz antenna review

antenna

S11:
S11-1

925MHz_Antenna

This should work well on a Gateway. Good rejection in neighbouring channels where mobile phone networks reside which helps to reduce “noise” and therefore increase SNR for the wanted Lora signals.

Today I measured the Molex 105262 ISM 868/915 antenna. I did not read the datasheet or manual!

First measurement First I measure just on my wooden table:

afbeelding

afbeelding

Second measurement: The second attempt was while the antenna was held on a PVC housing with an acrylic plate:

afbeelding

afbeelding

Third measurement: This time i held the antenna in free space. Now it was in no way resonant on 868 or 915 MHz but much higher (near 1GHz)

Conclusion. From this I confirmed that the material on which an antenna is mounted influences the resonance frequency;
My advice: Glue this Molex antenna on PVC housing for best performance.

1 Like

Today I received the Arduino MKR WAN 1300, which will be used at The Things Summer Academy, in the mail. It includes a funky dipole antenna. First thing I did was to very the antenna. I’m not exactly impressed :unamused:

Antenna in free air resonates at 920MHz
20210811_153836

Antenna against wood resonates at 810MHz
20210811_153900

Antenna against a plastic box - which would be the normal mode of operation - resonates very well at 918MHz
20210811_155554

Problem 1

The antenna is very sensitive to its surroundings. For me this is a bad characteristic.

Edit:
Yes sure, wood is indeed a crazy example. I’ve tried different plastics, foams and objects close to the antenna. The resonant frequency shifts around a lot. Comparing this to the Aurel GP868 which stays resonant at 860MHz even if I wrap my fist around it. The important thing to remember here is that one can’t guarantee how an end user will install the sensor. There could be a metal pipe, or wooden beam running right next to the enclosure. An antenna that stays resonant no matter the surroundings is better than one that gets de-tuned easily.

Problem 2

This antenna operates in the 915MHz ISM band, and not the 868MHz ISM band which is used in Europe - where I bought this Arduino.

That wood be the case for most antennas, so what wood you suggest ?

2 Likes

Mount on wooden lolly ice stick and place assembly in plastic case… shave stick to tune to middle of two material properties… Wood that be a solution?

1 Like

@jpmeijers Have you tried to measure the antenna while “glued” to different surfaces?
I have observed multiple times that these antennas are tuned when installed on a specific surface.
One antenna got resonant on 868 when glued to ABS. So the material on which it is installed is of influence.

Saw exactly that with some patch antennas (used on early RAK 5205 builds) … then received some of these Protect your IoT device from ingress with the PnP sensor box – RAKwireless Store where LoRa and gps ants already ‘mounted’ to sidewall and found system improved vs leaving free floating inside. Not sure if they still ship with ants as have bought for a while… caveat emptor! Hence comment about lolly ice sticks above (yes was a joke but…) …I used one to hold ant in place for 1st device! But didn’t mount to side…

20210728_132223-scaled