I’m using a TTIG gateway, with custom boards. They all used to work fine, sending/receiving data through the gateway. However, now nothing seems to go through the gateway to TTN. The gateway shows connected in my account, solid green LED. However, no data is ever uploaded. Firmware hasn’t changed on the devices sending to the gateway.
Using the LMIC library, shows EV_TXSTART, then EV_JOIN_TXCOMPLETE: no JoinAccept. Nothing shows up in the live data for any device tested.
Has something changed on TTN in the past year? If not I’ll get a new gateway, and try it out.
What do the GW console logs show if you look at live traffic? (Whilst power cycling the devices)
Is it seeing the Join Request?
Is there then a Join Accept to send back?
If you just stopped using the devices the LNS may still be retaining old session info. Reset session for devices in the console then trigger a re-join, you never know
If you decommissioned the devices were antenna removed? Have they been put back correctly? Did you power up without ant in place? (Dont - may cause damage!)
No, there’s no data making it through the gateway at all. I’m an EE, and design PCBs etc - these boards all have LoRa modules, but primarily use Satellite and GSM currently for main comms. They’re not decommissioned - LoRa is still used for local area communication with some other devices, but LoRaWan has not been needed in about a year due to lack of coverage where the devices are installed.
I only noticed this while doing some testing on a unit that was in for service, and found it could not send data via TTN any more. So I registered a new device with TTN in the same way as usual (I hadn’t been bothering to register new devices for quite some time) and the new unit would not send either. Same with all other boards tested. There have been no firmware changes related to LoRa in quite some time, so I also decided to update LMIC libs and retry - still nothing. Gateway shows zero activity apart from connecting, realtime device data shows nothing at all either. The devices always have LoRa antennas, as they’re used locally (proprietary protocol, not LoRaWAN) and have no problems. The indoor gateway I use for testing is only around 2 metres from the devices under test, so not a range issue.
It could possibly be the TTIG has failed I guess, but it’s just been sitting on a shelf for around a year till now. I might just buy another, and try it out - it’s not really urgent, but I wanted to see why this has failed.
Actually it is! Potentially far to close - search forum TL:DR look for 5-10m seperation pref with a wall or window in the way as signals will blled channels/distort etc. Testing with RSSI in range ~ -65 to -105dbm is kind of the goldilocks range to avaoid having to debug RF induced problems Often leading to CRC/MIC errors and hence dropped packets or simply not seen correctly on right channels. If nothing at all in GW from any devices then would likely agree:
If devices still send/receive using LoRa + proprietary protocol then unlikely to be a LoRa RF fail on the nodes. Best test by GW substitution?(*)
Though I assume you do see the GW status update ~every 30sec?
(*) Where are you in UK? Perhaps test against a local TTN Community GW if you can get close? If not far away can test against one of mine or I may be able to loan a unit
As you can see, the gateway only updates status after I plug it in. After that there’s nothing. In the screenshot, I unplugged it to move it to another socket about 10m away at around 17:03. Screenshot - EDIT: After this I turned on verbose logging, and it DOES show gateway statistics updated every minute.
I don’t think the range is the issue. Previously units were tested with the gateway in the exact same place. When the units communicate locally via LoRa the gateway is not used - it’s similar to a mesh network, but not quite - not really relevant here I’m in Norfolk - there are a few other TTN gateways relatively close by (though not in range from here) - I can just order another here through the company anyway though, just in case it’s this TTIG that’s failed.
Was about to say do that! All looks good then, so maybe the RF FE has failed - a quick drive by of GW’s with you nodes may tell you more if they work fine but sure looks like a TTIG problem then. Good luck!