everything you always wanted to know or ask about :
different types of antennas
diy antennas
cable influence (loss / cable types / quality )
best location to place them (indoor or outdoors and where)
lightning protection
measurement equipment
lots of pictures of your setups
*
welcome in The BIG ANTENNA topic part 1 
Hello 
Very good Topic!
Have a nice day
Jens
Hello
Good timing BoRRoZ and a first post from me, I have plenty of newbie questions rattling around in my head.
So I have a metal pole above the roof line of my house, about 10m over ground level. I live on the east coast of Ireland and would like to do some learning and testing with some nodes and a āstandard specificationā multi channel gateway. Currently I plan to put the following sort of equipment together.
Mounted on the pole:
Commercially available 10dBi antenna
1metre N to SMA cable to a IP rated enclosure mounted on the pole.
IMST iC880A inside the above enclosure
Interface board like this one by Hallard
Raspberry Pi Zero with usb to Ethernet adaptor
I have power and space in the attic area of the house and a network switch already there also. I plan to inject POE via a network cable approx 10m to the concentrator using a DC-DC buck converter on the interface board for power inside the IP rated box on the pole etc.
My questions would be
I am happy to get involved in discussion so welcome any feedback positive or negative on these parts and compatibility. Thanks in advance.
Garry
I have tested several China Export antennaās, including the exact one that you are using. While it isnāt exactly 10dBi gain, it still performs quite well and is my antenna of choice for my next project.
The rest of my setup is pretty much the same as yours. You will need a u.fl-to-N pigtail, not a SMA-to-N. I bought this one. My coax cable was from a A-brand, but even a lower quality brand shouldnāt matter that much if itās only 1 metre. My iC880A is mounted in a IP65 enclosure with the RasPi and connected with jumper wires.
For your setup I wouldnāt worry too much about lightning protection. You can only do it decent or donāt do it at all. Doing it decent involves running a dedicated 35mmĀ² copper wire up to the antenna from grounding rods youāve put in to the ground. Because itās on top of your house, youāll probably be alright.
Very reassuring, thanks.
I have tested between 10 to 20 Chinese 868 MHz / 915 MHz antennas, and my findings are that they are mostly crap. It also completely depends on what youāre going to mount it on. Most antennas in China are measured on an old benchtop VNA, and they check the return-loss to see if it works or not. However, they usually screw the antenna directly on the VNA, so the ground area around the antenna is completely different than mounting the antenna on a PCB for example.
If you want to get antennas from China, I would advise you to buy 5 different antennas from different manufacturers, and checking the RSSI on the gateway (if you donāt have access to a VNA).
Connecting the antenna through a U.FL to SMA pigtail usually makes the return-loss even worse, so for those situations a dipole antenna would be more suitable, but those generally cost more.
which do you recommend ?
did you ever tested this type of outdoor node antenna ?
By far the best antenna Iāve used (and probably the āgeneral bestā 868MHz antenna available) is the Procom CXL 900-3LW-NB/868MHz, but at ā¬165 itās not particular realistic for the average tinkerer.
Unfortunately Iāve never tested those kind of outdoor antennas, usually we try to get the most performance out of small and cheap antennas! I am a big fan of the molex antennas, but you need to mount them to plastic in order to tune them properly. Iām still in the process of finding good āwhipā antennas that have a SMA connector!
yes Iāve bought also several different āchinaā types 
We did an 868 MHz antenna test recently:
https://www.coredump.ch/2017/04/13/lorawan-868mhz-antenna-test-part-1/
https://www.coredump.ch/2017/04/30/lorawan-868mhz-antenna-test-part-2/
Interesting how far some antennas diverge from their specification.
I have a question related to dipole antennas: A single element dipole cannot have more than 2.15 dBi gain, correct? So all those dipole 5dBi gain whip antennas available on Ali Express are by definition impossible to have that much gain, correct?
like this one, less then one euro in quantities and shipped with all kind of lora products
Whip antennas are by definition monopoles, but when placed on a conducting surface act (somewhat) like dipoles. Some whips have grounding radials in the base of the antenna, but that does not make it a real dipole though. Most whip antennas increase gain by adding resonators or loading coils, but I would take those gain figures with a big grain of salt.
A new and very interesting outdoor antenna option: Sirio GP 868 C (dedicated for the European 868 ISM-band) is now available for ordering: http://www.k-po.com/SIRIO-GP-868-C.2.html
Did someone do actual VSWR & gain test for those 868MHz fiberglass antennas on Aliexpress?
Iāve seen that @GryKyo and @Epyon use similar antennas but I find the technical characteristics announced by the seller to be, ahem, extremely dubious. Still, even if you āonlyā get 6dBi out of their 10dBi antenna, itād be cheaper than the Procom.
I did some quick testing of the popular Sunhans 13dBi antenna and the A&C 14dBi compared to a regular ground plane antenna (2.19 dBi) by analysing the RSSI of a fixed node around 500m LOS away. While both antennaās functioned better than the ground plane, they certainly donāt reach the advertised gains. I would guess something between 3-6dBi at most.
While Iām at it, I should add my contribution:
A few months back, we bought cheap antennas on eBay, advertised as ā7dBi 868MHz antennasā.
They were dirt cheap, so while we knew the 7dBi figure was utter nonsense we decided see if they couldnāt replace our plastic indoor antennas.
Well, no. We had them tested in a RF chamber at Orange Labs, and they expert opinion was that they were āunusableā. Something about a 15dB lossā¦
Hi, is anyone using this collinear coaxial antenne solution for lorawan?
http://jeroen.steeman.org/Antenna/collinear-coax
http://www.rason.org/Projects/collant/collant.htm
The solution seems very popular on 1090MHz ADS-B;
Mostly, antenna gain is misunderstood.
For example, if you have a mobile antenna youād better off with a 0dBi antenna rather than a 10dBi antenna. The more gain the antenna has, the more energy will be bundled in one direction. Think of putting a reflector behind a light bulb. All the light hitting the reflector will be reflected, no light will go through.
For a mobile antenna you donāt want āblack spotsā. Effectively, you want your sensitivity/gain be the same in each direction (0dBi). This is called isotropic. Any misalignment in the pattern will yield gain in some direction. A dipole is not fully isotropic (radiating less in line of the dipole and radiating more perpendicular to the dipole), think of a radiation pattern like a squeezed ball.
Stationary (gateway) antennas will have more likely gain in horizontal direction and less gain vertical direction. This can be achieved by stacking dipoles for example.
An antenna can not output more energy than goes in. On the other hand, an antenna can ālooseā energy. When thereās a mismatch in the antenna (system), part of the energy getās reflected and will not reach itās destination. This is measured in VSWR (or in S parameters). A VSWR of 1 is most ideal: no energy is reflected.
