I have Wisblock with RAK11300 that is unreliable (now failed). I operate it on as a temperature sensor on the I2C pins of Wisblock, using Chirpstack with many other nodes. It worked perfectly for about 2 months (1 report per minute) then began to randomly fail. A restart used-to fix the problem but now it no longer does. The node always connects to my RAK Lora GW (even in a failure condition) but fails to send sensor messages when in a failure state. When I connect it to a serial port I see that it appears to be happily sending messages with the correct temperature reading but my hub does not see any incoming messaging from this node.
I’ll be happy to hear any comments or suggestions that may lead me to a solution to this.
Hello @PeterL
Can you check the antenna connection on the RAK11310 and the antenna itself if there is any visible damage or if the iPEX connector is loose.
What code are you using? Is it one of our example codes or your own sources?
Yes … the antenna connection certainly looks perfect … it gives a proper “snap” when connected and there is no appearance of any physical damage. It appears to work properly and when the occasional frame gets through, the reported SNR is normal. Actually, as an engineer, I observe that this hardware is beautifully made. It was working as expected for several months at 1 report per minute to my Chirpstack through a RAK gateway forwarder.
It has been operated indoors at all times and I tried different power supplies.
The code I am using is a modified version of the weather monitoring example sketch. I changed it to read a SHT30 sensor and report 4 bytes on Port 2. The Arduino framework is RAKwireless RP boards ver 0.0.7 and the SX126x library is version 2.0.14. The problem persists even if I bypass the I2C routine to collect SHT30 data.
While it seems like a hardware problem, I’m guessing its not. Likely something I don’t know yet 
I see this now: While the failures were random I see now that the node no longer actually sends periodic data despite the fact that the “send_lora_frame” function reports that it does. It does however send one unacknowledged data frame immediately after it joins. If I look at the serial data for the node I get the “lmh_send OK count message 1” AND successive similar messages , one by one as they appear to be sent (but only the 1st one is actually sent).
LORAWAN_DATERATE DR_0
LORAWAN_TX_POWER TX_POWER_5
TX_POWER_0 to TX_POWER_15*/
JOINREQ_NBTRIALS 3
uint8_t gAppPort = 2; // LORAWAN_APP_PORT;
DeviceClass_t g_CurrentClass = CLASS_A;
LoRaMacRegion_t g_CurrentRegion = LORAMAC_REGION_US915;
lmh_confirm gCurrentConfirm = LMH_UNCONFIRMED_MSG;