Visible in bright sunlight

[ranzbak]

Given the ambient light sensor, the brightness can be changed based on the light outside.
This does mean we will have more power being consumed when there is a lot of light, so heat from sun + heat from electronics.

LED's need to be able to produce enough light during day time.
We want a fan in the design ?

[Stoneshop]

Theoretical maximum current draw for the LED panel would be 405 * 3 * 12mA = 14.5A @ 5V, so 72W power draw. This is in the same ballpark as a conventional traffic light with a filament bulb, which has no additional cooling facilities.

The maximum current draw will be restricted to what the chosen step-down converter can deliver, which is likely 10A or less. Because of the power source being a pack of LiPo batteries having a limited capacity, displaying a full panel at maximum available brightness should be discouraged.

Solar irradiation can be limited by painting the exterior silver (it's light grey), and additional cooling if required can be realised through mounting heatsinks on the outside of the enclosure. Using a fan to provide additional cooling would require protection against dust ingress, and preferably a fan speed monitor to detect failure.

An internal temperature sensor might be useful.

Any heatsinks mounted on the outside of the enclosure should be positioned so that they do not interfere with the storage requirements.

[ranzbak]

Let's keep heat in mind with component choices, so at least heat sensitive components, like cheap electrolytic capacitors are avoided.
Also add heat sinks to components that get warm but usually don't need them, in order to give them better heat dissipation.

[Stoneshop]

The stepdown converter requires a heatsink for the controller IC. As the converter PCB will be mounted on an L-shaped aluminium bracket (size appx 7x7 cm, lip appx 3cm, not finalised) which will be bolted to the bottom of the lamp enclosure it is feasible that this provides sufficient cooling capacity even in an enclosed environment. This will be tested; the controller has thermal shutdown.

A temperature sensor could be provided via I2C, but this is currently not part of the design.