Following on from my thread here, I'm starting to get somewhere on pinpointing the issue.
TL;DR: Using this packet forwarder on a Multitech Conduit AP is suboptimal and results in severe packet loss, likely due to incorrect values for the antenna.
I have since switched to the inbuilt LoRa packet fowarder, and it's working perfectly. Because of the upcoming v3 stack, I intend to stay on the inbuilt forwarder, since this packet forwarder won't be needed... Therefore someone else would need to test this, but here's my theory:
From what I hear from MT support (I don't have much radio knowledge), the tx_lut values are very device-specific. While the MTCAP and MTCDT are very similar internally, they both have very different antenna designs.
When I switched to the default packet forwarder, I also erased the global_conf file with multitech_overrides, and replaced it with the default gateway-conf file from TTN (no modifications). Could it be that by adding a new multitech_overrides with values tailored for the MTCAP, that this issue could be fixed?
I can provide the stock conf files for both MTCDT and MTCAP if needed.
Following on from my thread here, I'm starting to get somewhere on pinpointing the issue.
TL;DR: Using this packet forwarder on a Multitech Conduit AP is suboptimal and results in severe packet loss, likely due to incorrect values for the antenna.
I have since switched to the inbuilt LoRa packet fowarder, and it's working perfectly. Because of the upcoming v3 stack, I intend to stay on the inbuilt forwarder, since this packet forwarder won't be needed... Therefore someone else would need to test this, but here's my theory:
From what I hear from MT support (I don't have much radio knowledge), the
tx_lutvalues are very device-specific. While the MTCAP and MTCDT are very similar internally, they both have very different antenna designs.When I switched to the default packet forwarder, I also erased the
global_conffile withmultitech_overrides, and replaced it with the default gateway-conf file from TTN (no modifications). Could it be that by adding a newmultitech_overrideswith values tailored for the MTCAP, that this issue could be fixed?I can provide the stock conf files for both MTCDT and MTCAP if needed.