Hi guys I have a bit of a problem. I mounted two super bright white LED's on either side of my fireplace for uplighting. It looked good not too bright not too soft but after a week they started to dim one at a time then they finally died 80,000hrs. MY ASS! So what did I do wrong? The electronics store that made them up for me sodered a resistor to each LED the resistors are light beige with a green band at one end looks like two red bands in the middle and one gold band at the other end. I am getting power from a 12 volt transformer that runs an alcove light in my home. Here are the specs from my transformer maybe you guys can figure it out the LED that I am using is suppose to be a 16,000 MCD that's what the guy on e-bay said anyway. Thanks Deano 0084
Height: 1.06"
Width: 1.38"
Extension: 2.63"
Weight: 0.3"
Details: INPUT: 120V
MAX. LOAD: 60W
MIN. LOAD: NONE
SECONDARY VOLT: 11.5VAC
INPUT CURRENT: 0.5 AMPS
OUTPUT CURRENT: 5 AMPS
OUTPUT: 12V CLASS 2 60W
OUTPUT VA: 60W
FUSE: 5 AMPS
WIRE LENGTH: 4 3/4"
USE OF TRANSFORMERS MUST COMPLY WITH THE NATIONAL ELECTRIC CODE AND ALL LOCAL BUILDING CODES
Sorry Litfuse for my foul language didn't mean to offend anyone. The LED's are mounted 3 feet on either side of the fireplace flush mounted into the brick. The fireplace has yet to be turned on so I am thinking heat might not be an issue but more of a resistor issue than anything but thanks again for your help. Deano
Your power supply is A/C out and I have been told by "experts" that this is not a good idea for LED's. I have also found that many of the inexpensive "12 volt" power supplies can put out well in excess of 12 volts until you put a reasonable load on them. Two properly driven LED's is not enough of a load. Get a good voltage regulated 12VDC power supply, a couple good quality LED's (Nichia) and a proper sized resistor (around 470 ohms if I recall correctly) and you should get some years service out of the LED's.
If this sounds like greek to you, go back to the electronics store who made these up for you and request that they do it again but this time addressing the correct needs of the intended application. [img]/ubbthreads/images/graemlins/icon3.gif[/img] Good luck.
Thanks McGizmo.My alcove light is only 10 watts so I thought that I could just tap into the 12 volt power supply to run the 2 LED's sounded like a good idea, sounded easy but I guess there's a little bit more to it. Thanks again for your advice.
deano,
You still could but put a rectifier bridge between the 12V A/C supply and the LED's to convert the supply to DC and you may need to increase the resistance over what you were using before. The Rectifier Bridge should cost less than $5 and your electronics store should have one for you. I use these on the Malibu power supplies which are a dirty 12 VAC output.
Hi raggie33 no my alcove light is a standard 20 watt halogean low voltage t4 bipin bulb but believe it or not 20 watts was too bright so I installed a 10 watt bulb. The LED'S are two 5mm..................Deano
[/ QUOTE ]
This is the problem right here. LEDs require a DC source. 11.5 VAC means that a reverse voltage of over 16V (the peak value of the sine wave) was impressed upon the LED. White LEDs generally cannot tolerate a reverse potential of more than 5V. If this value is exceeded, they will fail in short order, as you discovered. Anyway, you should never run LEDs straight off AC. A full wave recifier and filter capacitor will give you nearly pure DC of roughly 15V. You can then put the two LEDs in series along with a resistor of about 470 ohms, and the current will be about 20 mA. There will be no reverse voltages because the voltage source will be DC instead of AC.
McGizmo is correct. AC voltage is measured as Root Mean Squared (RMS). It is the 'equivalent' DC voltage. 100 volts AC into a 100 ohm load is 100 watts, and 100 volt DC into a 100 ohm resistor is 100 watts. If you think about this for a minute, you realize that for part of the cycle the AC voltage must be less than 100 volts, and to compensate, so the 'effective' voltage is still 100 volts, for part of the cycle, the AC voltage must be more than 100 volts. The way this is done is to measure AC voltage as RMS.
Essentially the mean voltage of the input voltage squared over time. For a sine wave, RMS is .707 of the peak voltage.For other wave forms, it can be more complex (which iis why 'TRUE RMS' Meters are more expensive).
In my youth I needed 250V for a project, so I used an isolation transformer (safety) and a voltage doubler circuit 2 x 120V=240, close enough?
Imagine my surprise when the open circuit voltage on the supply was 315 volts. With no load I wasn't getting twice the RMS voltage, I was getting twice the peak voltage, 2 x 1.414 x 120. Fortunatley in this particular application it made very little difference.
Now if you take your 11.5 VAC secondary, unless you regulate the output, or seriously load it up, the output voltage is not going to be 11.5 volts, it is going to be much closer to
11.5 x 1.414 or about 16.2 volts, figure about .6 volts drop from a half wave bridge, 1.2 volts from a full wave bridge, so until you get somewhere near the rated capacity on the supply, the actual voltage output will not be 12V, but something around 15 to 15.6 volts if the input is 120V. That's another fallacy, it is often higher.
Semi-Conductor Diodes (even light emitting ones) have very non-linear characteristics, so small increases in voltage over Vf can create big increase in current, and big increases in I^2 R losses, which can (and will) fry the diode. So you either need to be very careful the voltage or limit the current (either series resistor, or constant current supply).
I have some Nichias that have been running all night every night for a few years now and they may be putting out less than they were originally but they still put out plenty. When I first joined CPF and discovered what these LED's in flashlights were being driven at, I was blown away!!
Nichia has since reduced the half life expectancy to around a 10th of what originally was. A couple 5 mm LED's driven at 15 or 20 mA can provide all the light some applications might require. I think they have a very viable place in low level fixed lighting. Even a Luxeon driven at 50 mA is way too much light in some cases! [img]/ubbthreads/images/graemlins/icon15.gif[/img]
I'd guess that the AC-output transformer was intended for use with low-voltage halogen lamps. I have a fiber optic Christmas tree that uses one of these. With an incandescent lamp, a rectifier is unnecessary and would hurt the overall efficiency of the circuit.
That's exactly the problem. AC is fine for halogen lamps although it still produces a slight, barely perceptible flicker. With LEDs the flicker will be more pronounced, and more importantly the reverse voltage will burn them out in short order.
The technical stuff has been well covered here already [img]/ubbthreads/images/graemlins/smile.gif[/img] But the solution is easy, a diode connected across the LED with it's open direction opposite of the LED. It will pass the current when the AC swings the opposite direction and not ask the LED to do so. LED's can be used as diodes but their ability to survive the reverse current is much less than a regular diode. Or better yet, if you've got an AC source then use 2 led's wired parallel but backwards to each other so in one direction 1 led conducts and when it swings the other one lights. I have done this for 120v night light circuits and it increases the brightness without having to up the current through the LED and significantly reduces the flicker, but doesn't eliminate it.
I've got 2 dual led night lights that I built that run at slightly less than 20ma through the LED's that are running strong still after 3 years 24/7. I cannot perceive that they are dimming yet at all.
Hi jtr1962.I was just about to get a rectifire bridge and some 470 ohms resistors but now reading your post seems like my AC transformer still might be a problem. So what I want to do is start from scratch.I want to do this project right so if you or anyone else on the forum could tell me the proper type of hard wire DC transformer that I could use that is small (fit in gang box) that could easily drive one low voltage alcove light (20 watts max) and 2 nichia LED's. That would be greatly appreciated.
The transformer you have should be OK. Just use the rectifier bridge to drive the LEDs only. Put a filter capacitor on the output of the rectifier bridge (watch for polarity). A few hundred uF should be enough since we're only dealing with a current of about 20 mA. You only need a one-amp rectifier bridge (those are about the size of an 8-pin DIP package). In fact, even an IN4001 diode should be OK if you have a filter cap of around 470 uF or more (make sure to use one rated for 25VDC or better). You should have about 15VDC across the capacitor so a 470 ohm resistor will give you a current of roughly 18 mA if you connect the two LEDs in series. The low-voltage alcove light remains connected straight to the AC output of the transformer.
The heirarchy concurs!!! [img]/ubbthreads/images/graemlins/grin.gif[/img]
The AC from the transformer is what's killing the LEDs.
Use a bridge rectifier to change that to pulsating DC, and place a good size (1000µF or more) electrolytic capacitor (they're polarity sensitive too, be sure you have + and - on there right) across the DC output to help smooth out those pulsations; this should greatly reduce or even elimate flicker.
Buy a "wall wart". Plug-in DC power supplies come in a variety of voltages and current capabilities...relatively cheap. Hell I might even have one I could send you in the old junk box if you figure out what voltage and current you want.
btw i have an early nichia white that i installed in a 12v fluorescent desklamp as a nightlight.i converted this fixture in 97 and it has been on ever since.
only goes off when the main light is on as it is across the switch.running 15-20 ma and still looks as good as the day i installed it.i swapped it with another from the same batch and saw no difference.estimate over 50k hours on it.
80,000hrs. MY ASS!
Reply With Quote
