Interesting domestic LED design circuitry

Yes, it is, because a lot of voltage is dropped across the LEDs. E-caps have some surge capability too. Best regards, Spehro Pefhany

Which refers to

I don't believe in their 200 lm/W efficiency claims.

First of all, the lamps are available in two different color temperature with quite different spectrum, but still claiming the same efficiency to 3 digits.

The lumen scale depends of the human visual system, which is very sensitive for yellowish green (555 nm) , but less for red and much less sensitive for blue. Thus measuring different light source spectra in lumens will give quite different results while the two measurements in absolute power units (watts) is the same .

Is 200 lm/W claim given with three or just one significant decimal digit ? If it is just one significant digit, the actual value could be anything between 151 and 250 lm/W. Thus at 151 lm/W you could still claim 200 lm/W.

Modern LEDs claim about 100-120 lm/W, so achieving at least 151 lm/W would still be significant.

Due to the Lumen scale, the 100 % absolute efficient for true white light (6000 K true black body) is about 250 lm/W. A monochromatic 555 nm radiation the maximum is 683 lm/W by definition.

Achieving 200 lm/W usually would require in practice some badly skewed "white" light spectrum.

The LPW part marking for the fet indicates NXP 2N7002CK, which has internal gate protection.

I didn't watch the whole video originally, but reviewing it a second time, I see the presenter does measure the lamp string voltage and the mosfet drain voltage. The sim's string can be modified to reflect those values, to see what line voltage, temperature and part tolerances will do.

The advertizing sheet predicts 20K switch-on cycles of endurance and 25K power-on hours of life (3yrs).

RL