more details on homemade LED retrofit

How did you measure that ?Are you sure that you measured real (103 W) instead of apparent (103 VA) power ?

At least in Europe the thick 1200 mm (40 W) consume about 55 W and the thin 36 W variant about 46 W total.

Since there is an inductor is in series, the power factor cos(phi) is about 0.5 so 103 VA x 0.5 = 51 W, which sounds believable.

Put a 5 uF capacitor rated for mains voltage across the luminary terminals and measure the power again, I guess it shows something less than 60 W even on an unreliable meter.

Since your ability to make reliable real power measurements on simple reactive loads, I am quite sceptic about your measurements regarding the power consumption of your LED replacement, since the current waveform is far more complex due to the rectifiers, causing high peak to average ratios, fooling many meters.

The eye is a very unreliable device for absolute illumination level measurements and will be fooled even when doing relative measurements.

The traditional method of making relative illumination comparison is taking a cardboard, making a round hole into it and put grease proof paper over it. Move this device between light sources A and B and adjust the distance so that the illumination appears uniform regardless if it comes through the hole or not. Measure the distance to light sources A and B and using the inverse square law to calculate the light output of B if the light output of A is known.

If the panel is equally illuminated half way between A and B, then the light output of A and B is the same.

This assumes identical spectral distribution, which is not the case in when comparing fluorescent vs. LED lamps, but at least you get some idea about the relative output, when moving the cardboard around the middle position between A and B.

Yes, it is a Simpson WATTmeter. It has buttons to select volts, Amps and Watts, and the coils in the memter movement perform a realtime multiplication of V * A.

A true magentic Wattmeter should be able to handle fairly nonlinear loads and still give a valid reading. Another point that helps make the reading believable is the heat. The electronic LED current regulator pod become detectably warm with no heat sinking after

I DID search for hours through many web sites to find the most efficent LEDs (in L/W) and a very efficient LED power supply (rated at 86% efficiency).

Jon

Jon,

Thank you for posting your LED retrofit experience. It gave me the idea to experiment ( successfully ! :-) ) with my garage lighting.

Ed

I was going to ask the same sort of questions as I find it hard to believe one of these fixtures draws so much power because I don't believe the get this warm in my experience. That is the ultimate measure of the power they consume, the heat they produce. That won't lie.

I checked a fixture with two 40W bulbs and after 10 minutes I could not feel any source of heat other than the bulbs. Although the ballast is inside the metal enclosure, a 100 Watt relatively point heat source would be felt I expect.

Perhaps your fixture is defective?

I don't know how much LED efficiency varies in the lighting market, but of the available replacement bulbs I have not seen a large difference between CFLs and LEDs. LEDs are a bit more efficient, but CFLs do pretty well. I don't think a properly working florescent fixture is much worse than a CFL and so in the same ballpark with LEDs.

The reason I want to replace the durn florescent fixtures is because they all hum! I have a 20 W fixture over my sink and it is very annoying. LEDs to the rescue... maybe. I just don't like having to hide the power supply. :(

Well, these are 35 year old magnetic ballasts, and probably nearing the end of their life. They DO get hot, REAL HOT, after a couple hours run. So hot, in fact, that I had trouble holding them when removing them from the fixture. I could easily believe they were burning

Modern electronic ballasts will draw a lot less, and the large mass of the old magnetic ballasts have a long time constant.

Jon