@Rob65
As @BoRRoZ mentionned, I put the LORIX one Antenna on my other Gateway with an SMA to N-Male Adapter (1 meter long). I got the same results than with LORIX hardware, so it’s definitively coming from the Antenna.
With this one gateway
with the picture antenna (+3dBi) from Farnell I’ve got -75dBi
If I put on this GW the LORIX Antenna, I’ve got -84dBi, but, My test device is an RN2483 without antenna.
So get out a Rocket Scream board (RFM95) with antenna and LMIC stack, keep original antenna on each gateway and results are much better
{ "gtw_id": "RPI IC880a",
"rssi": -35,
"snr": 9.5,
}, {
"gtw_id": "LORIX One",
"rssi": -33,
"snr": 9.8, }
Makes sense, but results are not always consistent (look below), same place, same device, same sofware
{ "gtw_id": "RPI IC880a",
"rssi": -41,
"snr": 7,
}, {
"gtw_id": "LORIX One",
"rssi": -24,
"snr": 7.8 }
{ "gtw_id": "RPI IC880a",
"rssi": -39,
"snr": 9.5,
}, {
"gtw_id": "LORIX One",
"rssi": -25,
"snr": 10 }
{ "gtw_id": "RPI IC880a",
"rssi": -43,
"snr": 10,
}, {
"gtw_id": "LORIX One",
"rssi": -29,
"snr": 11 }
@Rob65 The plant is a fake one, thanks to IKEA
@BoRRoZ, got an email from Wifx (LORIX One) about LoRaWAN certification, here is the transcript
For clarification, we have indeed passed the LoRa certification at Semtech in Grenoble with "very good"
Our data are identical to the design reference of Semtech whereas the PCB is much smaller.
which is perfect for me