| Well, there is one huge difference between the CPU vs LED, you have even mentioned it:
It is the complexity the given technology allows.
With the LED, the technology allows to form just a single diode, in the most advanced form I've seen so far (the first generations of the ACRiche from SSC) it mean just multiple interconnected PN junctions, so just diodes, nothing else. From that you can not create any functionality, not speaking about temperature protection.
But the processor technology is way more complex (but by far not that complex as the processes for a mains voltage integrated SMPS or the specialized HV capable mixed signal processes), so on top of the PN junctions it alows to design some amplifiers, so base building blocks for e.g. the thermal protection feature throttling down the clock to prevent overheating.
Building something like that into the LED chips alone just isn't possible.
But that does not mean it won't be possible to protect the LED modules from overheating. The LED chips alone are dumb and can not be made any smarter, but their power dissipation, so temperature rise above the substrate they are sitting on should be well known to the module designer. And adding a sensor onto the same board and wire it so, it reduces the LED power (or shut it completely down) on the board temperature corresponding to the maximum safe temperature for the LED junctions isn't as difficult.
Or it isn't as difficult to design the system so, it simply won't overheat even on a direct hot sun.
By the way I doubt it would be the LED module, what failed there. I even strongly doubt any electronic actually died there, but I would rather guess it is the relay, what gave up there.
The reason is the rather high power dissipation in the relay coil when energized, yielding really high temperatures inside, so the plastics start to deform so, the contacts are unable to work properly, although it is still clicking. When that is placed into an area exposed to the heat from the LED, it becomes beyond what the relays are able to tolerate.
If you go through some datasheets, ~80degC is about the maximum the relay makers do allow as the relay ambient and only with limited operation time, but with the 35degC plus sun plus operating LED's the heatsink (so the fixture body) would be above that.
My guess the primary fault was the non working, or improperly set ambient light detection (the PIR processing IC's generally used in the cheepeese PIR sensors I've seen have that feature, it is just a matter of connecting the CdS or a photodiode to that pin)
In such fixture what I don't understand is: There is a supply unit generating about 12..48VDC for the LED's. There is as well an electronic module processing the PIR sensor signal, with power requirement of just few mA (normally it is in the range of 1..5mA for the complete sensor circuit, include the ambient light detection and a timer). So it would mean just a simple dropping resistor from the 12..48V, controlling a FET switch operating the LED's. Or the two units could be integrated together so, the PIR sensor circuit takes over the power regulation feedback signal of the supply unit so, the LED's would light only when required. No relay whatsoever, the PIR sensor on the secondary side,...
|
