To test my theory that the high voltage linear regulator chip in the LED filament lamp was likely to be running quite hot, I reconnected the lamp to its original PSU and monitored its temperature rise with a thermal imaging camera. I also monitored the power taken by the lamp as indicated by a power-monitor plug.
The chip gets very hot, even in open air. It started regulating the current down automatically at an exterior temperature of about 70 degrees Celsius and the current/dissipation finally balanced at about 92 degrees. This was in open air in a cold workshop though, so it would probably have regulated the current a bit lower if it was in the enclosed base. As it was, the power reduced from 4W down to the lowest value of 3.5W.
What this means is that the lamp is doing the exact opposite of a compact fluorescent lamp. Instead of starting dull and getting brighter, it actually starts at full intensity, then gradually lowers its output slightly. The adjustment is imperceivable due to the smoothness and time taken.
The chip gets very hot, even in open air. It started regulating the current down automatically at an exterior temperature of about 70 degrees Celsius and the current/dissipation finally balanced at about 92 degrees. This was in open air in a cold workshop though, so it would probably have regulated the current a bit lower if it was in the enclosed base. As it was, the power reduced from 4W down to the lowest value of 3.5W.
What this means is that the lamp is doing the exact opposite of a compact fluorescent lamp. Instead of starting dull and getting brighter, it actually starts at full intensity, then gradually lowers its output slightly. The adjustment is imperceivable due to the smoothness and time taken.
I really do think that the glue which bonded the socket to the lamp did cool this chip on the PCB just enough so that this current drop does not occur. I think it does only occur on hot summer days, where you certainly want cool white illumination.
Nice wee FLIR camera 🙂
I've one LED bulb (Integral 4.6W) that seems to use a thermal management chip also, where the lux level on a light meter started dropping after a minute of the light being on. With the lux meter on the floor and the light switched on, it read 22.8 lux. This dropped to 19.8 after 2m, 18.2 after 5m and was settled at 17.0 after 30m. So based on the lux level, this bulb got 25% dimmer after 30 minutes, i.e. 350 lumen instead of its rated 470 lumen.