I want to put a TTN Outdoor gateway on each side of a tunnel and see what kind of coverage I could get in a long straight tunnel.
After a bit of googling I have found that the Metro in Paris has used LoRaWan to provide coverage during construction
http://www.wyres.eu/case-studies/wyres-indoor-geolocation-rer-paris-lora/ . They used directional antennas in their setup and I was wondering if anyone had used directional antennas with the outdoor TTN gateway or any antenna recommendations?
Directional ants good if you know where you are trying to reach and also helping deminish signals coming from other directions therefore potentially improving your local SNR, however, they concentrate available energy in given direction so you also need to consider local regulatory TX EIRP limits and dial down any GW TX pwr to offset the increased gain over stadard 2.15dbi radiator (e.g. in EMEA typically 14dbm tx power coupled with std radiator = 16.15dbm limit), if you see connector and cable losses then increased ant gain can be used to compensate - just be aware of fringing/notching at higher gains…such dir ants can be helpful in picking up weak/remote nodes - just dont expect to be able to talk back to them in extremis without breaching regs 
Directional antennas are not often used, because it most often causes the Gateway to breach legal power limits when it transmits.
You can of course reduce the power output of the Gateway to match the improvement in antenna gain. This will help the Gateway receive packets from remote nodes, but a lot of the nodes will now no longer be able to receive packets from the Gateway, the whole system has become very unbalanced.
I was involved in a recent range experiment with a couple of guys from UK where we all had ultra long range connections, 150 - 220km on LoRaWAN using The Things Network.
One of the guys shared this design for a home made Yagi directional antenna which he has had a lot of success with. I have printed a couple since but have yet to document the range and success of these.
I hope that this helps
G
Hi I made the antenna too, and measured it. 
The antenna is not resonant at 868 MHz but on 787 MHz (5 dB return loss)
Also resonance is at 2100 MHz. The antenna is not designed for that frequency so the radiation pattern is a wild guess.
I expect that the frequency goes off the design frequency because you encapsulate the elements in plastic. You need to adapt the size and position of the elements to the plastic you use or do not encapsulate the antenne.
