What's in LED Bulbs?

4.0K195

Intro: What's in LED Bulbs?

I'm a maintenance man and as such I change a lot of lightbulbs. I've been picking up LED bulbs lately and I was curious what kind of components were in them. I picked up two of the less expensive bulbs that you can get at your local big box home improvement store because I didn't want to be taking apart any $10-20 bulbs right off the bat.

STEP 1: Bulbus Bulb

Starting with the bulb I buy most often, because they're inexpensive, I set about disassembling the unit. There's an obvious seam between the translucent part of the bulb and the base and I started there.

It didn't twist off so I tried jamming a knife in it and that surprisingly worked well. Usually it just breaks things. By working my way around the seam I found the translucent globe slowly pried away from the base. The globe is glued in place.

STEP 2: LEDs Revealed

Inside the base is an array of 11 LEDs. They're firmly mounted to an aluminum heat sink. There are two small screws in the face of the heat sink but removing them did not help to separate the heat sink from the rest of the unit.

Again, resorting to jamming a knife blade into the seam and prying removed the heat sink. Do this on the side furthest away from the wire leads that are soldered to the face of the LED array as they are quite short. I broke one of them in prying off the heat sink.

STEP 3: The Innards

Once the heat sink is pried out I was able to pull the circuit board free of the hull.

The inside of the hull is also aluminum, presumably acting as part of the heat sink.

As expected, there's a step down transformer. I plugged it in and found the DC voltage choppy. I was either getting 62v at the array or 38v. 62v means there's a 5.6v drop across each led. 38 means there's a 3.4v drop across each, which seems more reasonable but I'm not too sure about.

STEP 4: Stick Bulb

The stick bulb proved easier to disassemble, mainly because of its easy to grip shape. It was easy to pry the bulb apart with my hand. In this bulb, there is an array of six LEDs.

It was also easier to remove the heat sink. For one, there are openings that are left from the globe's removal. For another, the leads that power the LED array are sliding rails, so the heat sink is easily removed without any damage.

STEP 5: Friendlyier Circuit

This bulb uses a simpler circuit and seems much more maker friendly. The basic system follows the same logic but there are much fewer surface mount components and the circuit is much easier to trace.

This bulb tested 38.5 volts at the LED array, which breaks out to 6.4v per LED.

STEP 6: Conclusion

Both bulbs have some promise for using their LED arrays. I was hoping they would light when connected to a 20v power supply (because most 18v lithium ion drill battery packs are actually 20v) but unfortunately they didn't.

It'll take another step to get these arrays to be useful in the voltage range I'm interested in. That will be a different Instructable if I can pull it off. I'd love to swap out the crummy flashlight bulb that came with my battery powered tools for one of these.

STEP 7: Addendum

I had hoped that there would be a circuit on the back side of the led arrays. Now that I think of it, that wouldn't be possible. Indeed, I pried one off the aluminum heat sink and all I found was more aluminum. I found that it's difficult to see, but there is a circuit under a coating that the leds are surface mounted to.

The bulb that seemed the friendliest to hack, now became the more difficult and vice versa. The bulb that had the more complicated circuitry has convenient test points that can be used to power some of the LEDs.

I probed the test points with 20v and found that I could light three or four LEDs but only very dimly. I could light a single LED to the brightness that is generated when the full circuit is connected to mains. That's a bit disappointing and doesn't make a lot of sense to me. I'll continue testing and update as I go.

5 Comments

I've taken apart these LED light-bulbs before, The LED's are usually pretty strong, And most of them have a pretty big heat-sink too

The irregular voltage that you're experiencing is pulse width modulation (PWM) of the LED's. To provide the brightest possible light without burning out the LED's, the engineers designed a circuit that rapidly increases then decreases the voltage applied to the LED's in a square wave function (many actually just have the voltage low state as "off" making this a true rectified square wave). This permits the "high" voltage to be higher than the LED's rated maximum voltage and provides a bright light (when on....or in the "high" state). When in the "off" or "low" state, human persistence of vision allows us to disregard the "flashing" nature of the light over about 50 Hz. Typical PWM circuits operate at much higher frequencies. The ratio of "high" (on) time to that off the total cycle ("high+ low" time or on+ off time....The reciprocal of this sum in seconds, incidentally, is the frequency of the circuit in Hz) is known as the Duty Cycle of the circuit and will dictate how bright it gets and how hot the LED's will get. The designers then selected a combination of low and high voltages and then the period associated with each that optimized brightness while keeping the LED's from burning out and maximizing their life. This approach also saves energy. So, the three outcomes from the selection of these four variables are: LED life (capital expense to the consumer), LED brightness (basically, first cost to the manufacturer because the same can be achieved with more LED's per bulb), and energy cost (operating expense to the user). Guess which two wind up always being the second and third considerations....especially with cheaper bulbs.
Excellent information! Good to know, thanks! That makes sense, I didn't think of it though.
Could you use a DC-DC Adjustable Voltage Step Up Boost Converter to get your results you looking for to take it from 18/20v to your 36 v ?

I'm going to try separating the arrays from the heat sink and then see if I can split the arrays in half. That may make it possible to power both in parallel off 20v it may need a resistor to limit current, but these are pretty beefy, they may be ok without one.