Modbus Class C sensors behaving like Class A and disconnecting (Dragino gateway + ChirpStack)

ChirpStack Application Server
1

Hi everyone,

I’m working with some Modbus Class C sensors through a Dragino gateway, but I’m seeing strange behavior:

  • The sensors connect, but they don’t behave like Class C devices — they act more like Class A.
  • As a workaround, I set up a device profile with Class A, which worked fine at first and I was receiving data for a couple of days.
  • Then suddenly the devices disconnected I receive the Dev-Nonce error so i flush the devNonce and now they just keep doing continuous rejoins. I see the join requests and join accepts, but no uplinks afterwards.

At this point I’m not sure where to look:

  • Could this be related to the MQTT bridge?
  • Or maybe the versions of ChirpStack, Redis, or the gateway firmware?

If anyone has seen similar behavior or knows of a related topic, I’d really appreciate a pointer.

Thanks a lot for your time!

2

What’s your device? Helps to be able to look into it ourselves.

My first thought is maybe a LoRaWAN revision version issue (device profile setting).

3

For the gateway it’s a dragino LPS8, for the device it’s an NKE ModBus Master RS485 Class C

I use the same LoRawan version as mentioned in their technical documentation 1.0.2

4

To give more details on the device profile am using :
Region : EU868
MAC version : LoRaWAN 1.0.2
Regional Parameter : B
Default ADR algorithm
Flush queue on activate (on)
Expected uplink interval : 3600s
Device status request frq: 1
RX1 : 0

Class C : on
Confirmed downlink timeout : 5s

and then there is the codec javascript.

I don’t know if this would help but I thought it might give an insight