STM32WLE5CCU6 LoRaWAN IN865 Deployment Showing 30–40% Packet Loss in Field but Only 2–5% in Office – Looking for Root Cause

Question

High Packet Loss in IN865 Field Deployment (30–40%) with STM32WLE5CCU6 – Looking for Root Cause Analysis and Scalability Guidance

Hello Community,

We are currently facing a packet loss issue in our LoRaWAN deployment and would appreciate guidance from members who have experience with large-scale deployments, particularly in the IN865 region.

System Configuration

End Device

MCU: STM32WLE5CCU6
LoRaWAN Version: 1.0.4
Device Class: Class A
Message Type: Unconfirmed Uplinks
ADR: Enabled
Initial Data Rate: SF10 (ADR manages the final data rate)
TX Power: TX_POWER_0 (ADR controlled)
Payload Size: 1 Byte
Uplink Interval: Every 30 Minutes

Network Infrastructure

Gateway: Milesight UG67
Network Server: ChirpStack
Region: IN865

Channel Configuration

The following 8 uplink channels are configured:

865.2 MHz
865.6 MHz
865.8 MHz
866.1 MHz
866.3 MHz
866.5 MHz
866.7 MHz
866.9 MHz

Deployment Scenario

We currently have approximately 150 end devices deployed.

Office Testing

When all 150 devices are powered on and operating in our office environment:

Packet loss is typically around 2–5%
Communication is generally stable
No major reliability concerns are observed

Field Deployment

When the same devices are deployed in the actual field environment:

Packet loss increases significantly to approximately 30–40%
Packet loss is observed across all nodes, not just a subset
Successfully received packets show good RSSI and SNR values
The deployment area is an open outdoor parking lot with no roof structures
Only trees are present in the surrounding area
Gateway installation height is approximately 15–20 feet above ground

Key Observation

The most interesting observation is that this issue appears to be region-specific.

When we test the same application and deployment using EU868 settings:

Packet loss is minimal
Overall performance is significantly better

When operating with IN865:

Packet loss increases to approximately 30–40% in the field

Because the hardware, firmware, network server, and gateway remain unchanged, we are trying to understand what could cause such a significant difference between EU868 and IN865 deployments.

Questions

Has anyone experienced similar behavior in IN865 deployments?
Could external interference in the 865–867 MHz ISM band be responsible for the observed packet loss, even when RSSI and SNR appear healthy for received packets?
Would performing an RF spectrum survey be the recommended next step to identify potential interference sources?
Are there known challenges or best practices specific to IN865 that should be considered?
Could gateway channel configuration, regional parameter configuration, or ADR behavior contribute to this issue?
What metrics should we investigate first to identify the root cause?
- RSSI/SNR trends
- Channel utilization
- Gateway logs
- Packet collision statistics
- ADR behavior
- Duty cycle limitations
- Other diagnostics
Is a gateway installation height of 15–20 feet potentially insufficient for a large outdoor parking deployment?
Could the issue be related to packet collisions, hidden node problems, or the near-far effect, even though traffic volume is relatively low?

Future Deployment Concern

Our eventual deployment target is approximately 4,000–5,000 end devices.

Given the packet loss we are already observing with 150 devices in the field, we would appreciate recommendations on:

Network scalability planning
Gateway density requirements
Capacity estimation methods
ADR optimization
RF survey methodology
Best practices for large-scale IN865 deployments

We are trying to determine whether the root cause is related to RF interference, gateway placement, network design, regional configuration, or some other factor.

Any guidance, troubleshooting suggestions, or experiences from similar deployments would be greatly appreciated.

Thank you.