Starling 2 Max board runs 6 to 10 C hotter on older drone

[PQL]

Hi ModalAI,

I have the following two variations of the Starling 2 Max Dev Drone:

Starling 2 Max GPS-denied Development Drone (Purchased Sept 2025)
with Battery Pack and Controller and Case / Dual 4k & Dual Tracking & ToF (C29) / Microhard pMDDL2450 WiFi and Ghost Atto 2.4GHz
ECCN: EAR99, NLR

Starling 2 Max GPS-denied Development Drone (Purchased May 2026 and July 2026)
with Battery Pack and Controller and Case / Dual 4k & Dual Tracking & ToF (C29) / Microhard pMDDL2450 and ELRS 915MHz
ECCN: EAR99, NLR

The drone purchased in Sept 2025 runs approximately 6 to 10 degrees Celsius hotter than the ones purchased in May and July of 2026. Bottom line, the board as a whole runs hotter of which causes the entire drone to behave differently because of the IMU sensors being heat sensitive. Can someone at ModalAI help explain why this would be the case and how we can fix this issue?

[Alex Kushleyev]

Hi @PQL,

This is most likely due to the updated version of camera server and vio applications, as performance enhances may have impact on the cpu / gpu usage. We can help double check this. A few suggestions:

• if you are able, can you please use voxl-inspect-cpu to measure the top CPU consumers for both drones and post here. This way we can check if something looks abnormal. Both drones should be in the same state (on bench, everything running.. or whatever environment you want to test in).
• you can downgrade your new drone to earlier SDK and see if the CPU usage / temps reduce to original value

Alex